RPlusD Manual Contents RPlusD Software version 1.2.xx Acoustisoft Inc rev June Douglas H. Plumb

Size: px
Start display at page:

Download "RPlusD Manual Contents RPlusD Software version 1.2.xx Acoustisoft Inc rev June Douglas H. Plumb"

Transcription

1 RPlusD Manual Contents RPlusD Software version 1.2.xx Acoustisoft Inc rev June Douglas H. Plumb Chapter 1: Taking Data Required Hardware Equipment...3 Software Settings...4 Sound Card Loop Through Test...5 USB Calibrated Mic & Preamp...10 Set Up For Room Tests...11 Test Signal Types & Noise...16 Display And Related Functions...18 Special Measurement Functions...21 Chapter 2: Time Measurements Impulse Response...24 Noise In Measurements...25 Time Response Options...26 Subwoofer Measurements...28 Energy Time Curves...30 Impulse Pseudo RTA...34 Reverberation Time Measurements...35 Chapter 3: Frequency Response Introducing Gating...44 Types Of Frequency Response...46 Gating & Frequency Response...46 Practical Frequency Response...49 Averaging Measurements...54 Low Frequency Room Response...56 PsychoAcoustic Response...57 Pseudo RTA Response...59 Chapter 4: Practical Measurements Chapter 5: Equalization Reflection Measurement...61 Room Resonant Measurements...66 Subwoofer Placement...71 Loudspeaker Direct Sound...74 Loudspeaker Reflected Sound...78 Loudspeaker Power Response...80 Further Reading...84 EQ Types...85 Manual Equalization...86 EQ Emulation...89 Resampling To Match Hardware...91 Setting EQ Controls...92 Getting EQ Settings...93

2 Chapter 6: Digital Crossovers Setting Hardware EQ's and the EQ RTA...98 Curve Fitting...99 Measurement & EQ Solution Reliability Advantages Of Digital Xovers Physical Experimental Setup Software Settings Required Running Emulator On Measurements Setting Driver Correction Filters Emulating Crossover Filters Additional Notes Chapter 7: Measuring Distortion Conventional Distortion Practical Audio Components & Distortion Two Tone Measurements Multi Tone Measurements IMD & HD Measurement Accuracy Testing Microphones Dynamic Compression Dynamic Compression Example Chapter 8: Wireless Data Gathering This chapter should be read by anyone that wishes to use the wireless data gathering method. Appendix A: Printing Appendix B: User Licenses Appendix C: Functionality Licenses Appendix D: Software Requirements & Limitations...169

3 Chapter 1: Taking Data Introduction RPlusD was designed for measurements taken indoors or outdoors in very low wind conditions. Wind will cause noisy inaccurate results. The size of room that can be analyzed with the software is that of small theaters seating less than about 250 persons and smaller. RPlusD is not a noise analyzer. It provides the test signal for system excitation and this excitation cannot be substituted for anything else such as a noise test CD. The software does permit measurements taken from a test CD but this test CD signal must be made from the software. The software takes measurements for energy - time reflection analysis, frequency response, distortion and EQ emulation. EQ emulation allows the user to see the affect of equalization applied to measurements through a hardward emulation of an equalizer rather than just a simple "delta curve" method. The EQ emulation provided in RPlusD runs the measurement data through an actual software emulated equalizer. This and the fact that RPlusD can store up to 32 measurements makes this software the ideal tool from which to set an EQ and correct a system response. The affect of the equalizer can be checked for both the time and frequency responses. RPlusD also has an electronic crossover emulator as well. This can be used to set an electronic crossover and view the settings affect on many measurements at one time. The software was designed to be better than the earlier ETF is every possible way as well as provide this new EQ emulation and electronic crossover emulation capability. Equipment Needed The software runs on all Microsoft operating systems including Vista. In runs on MACs as well but a PC emulator may be required. The demo software download can be used to verify correct operation on an Apple computer. Any common sound card can be used to take the measurements. A sound card that may otherwise sound terrible will take measurements as accurately as the very best available. RPlusD uses a loop back connection by connecting (usually the left) channel output directly to the left channel input. The software uses this to measure the sound card response and remove its error from the measurement channel response. The right channel in this case uses its output to drive the audio system with the test signal and the right input takes the microphone signal. Built in sound cards in laptop computers almost never work correctly. An outboard USB sound card is recommended if a laptop computer is used to take measurements. Desktop computer sound cards almost always work. Any sound card that can both play and record in stereo (stereo full duplex) will suffice for RPlusD software. This is a common feature for even the lowest priced sound cards. Users of this software should obtain a microphone boom stand so that the microphone can be held stationary in any position in space while a measurement is being taken. It is critical that the microphone remain absolutely stationary during tests. Boom stands can be obtained from musical intrument stores at very low cost. The microphone needed will depend on the measurements desired. For most applications that do not include equalization the Radio Shack SPL meter will be all that is necessary. It can be taped to a boom stand using common black electrical tape. The tripod stands that work for this unit are often much more expensive that the mic boom stands that can be obtained at musical instrument stores. This unit behaves in an ideal fashion within the critical band of human hearing, 100 Hz Hz. It is useful for analyzing reflections as well as low frequency modal activity to

4 frequencies as low as 30 Hz. The output of this unit is not depenedent on the weighting settings on the unit. The weighting settings only apply to the actual meter movement and do not affect the electrical output of the unit. The electrical connection can be found on the side of the unit and provides a standard line level output with an RCA jack type connector. More expensive microphones such as ones available from Behringer can provide a better response due to the smaller size of the microphone. These measurements can be useful up to 20,000 Hz and will provide better performance in measureing reverberation time. These units are not individually calibrated and can be used for taking measurements for equalization provided the limitations of an approximate calibration curve is taken into consideration. The best possible measurements for equalization require a custom calibrated microphone. These can be provided by AcoustiSoft at the prices given on the RPlusD order page. Two types are offered, one requires the user to have a sound card and the second more expensive USB unit permits the microphone preamp to be connected directly to the computer without a sound card. This unit is recommended for professionals because it requires fewer electrical connections. Both preamps use the same individually calibrated microphone. Each microphone ships with its own calibration file that can be loaded from the software. The software also has a tool available from the Tools ->Microphone Calibration menu selection. This allows the RPlusD user to import calibration data from any microphone and convert it to the file format required by the software. The software uses a linear extrapolation between points so approximately 30 evenly spaced points (logarithmic spacing) would be required to create a useful calibration file for a third party microphone. Mic calibration is done in the frequency domain only. Software Settings The main data gathering control window is shown below. It is accessible from File->New->Normal Measurement from the top menu. Main Data Gathering Window Clicking on the button circled in red leads to the Measurement Settings window below. RPlusD

5 offers 3 sample rate settings. For most users the 48 KHz or 44 KHz settings should be chosen. Both should be tested to determine which one works better during the initial tests of a sound card. The sample frequency that works best will be the one that does not require hardware resampling on the part of the sound card. Higher sample rates are not provided because they provide no advantages and slow the calculations down considerably. If higher sample frequencies are required RPlusD has internal software resampling. This is necessary for emulating DSP equalizers inside the software and will be explained in a later chapter on DSP equalization. Measurement Settings The Measurement Settings window provides a place to load a microphone calibration file, select a sound card and select which channel will be the mic input channel. The left channel in this case will be the loop back channel. These connections will be explained later in this chapter. Sound Card Connections And The Loop-Through Test The sound card, software and computer must first be connected to ensure that the mixer is properly set and that the hardware works correctly when used with the software. This first test is called a "loop through" test and is done as explained below. This test is necessary for people that are using a sound card and not the USB mic/preamp that is supplied by AcoustiSoft. Users of the USB mic/preamp can skip the rest of this section and proceed to the next section that describes its use. The sound card is connected in the loop through configuration as illustrated below.

6 Loop Through Sound Card Test Wiring Configuration This test will verify that the sound card is working and that low cost cables do not have the left and right channel connectors confused. Many people use a low cost 1/8 inch stereo connector for this and problems arise when these cables have left and right confused in their internal connections. With most types of software and for most purposes this does not matter but RPlusD uses a different test signal for the right channel than the left and getting these connections correct is very important. Usually the internal wiring of the cables used is the problem when we get support calls asking why the loop through test will not work. It happens quite often when low cost connectors from overseas are used. The next step is to ensure that the RPlusD automatic mixer is selected. Select this from the main menu from Options->Mixer. The automatic mixer will remember all the settings for your sound card and mixer each time you take a new measurement. For this to work correctly the same sound card must always be used to take measurements. The mixer is set as illustrated below. The windows mixer may not be available on all sound cards. In cases when the windows mixer is not available for a given card this can be downloaded as a third party WDM driver by searching "<sound card make and model> WDM driver". RPlusD requires the WDM driver and this does not always ship with specialty sound cards. The windows mixer can be accessed from a little speaker icon in the right hand side of the bottom task bar on the desktop display. From the Options -> Properties main menu option on the Windows mixer the sound card can be selected.

7 Window Mixer Sound Card Selection The sound card device that will be controlled is selected at the top. All controls should be selected to be shown. The mixer is really two mixers, one for playback and one for recording. The playback mixer is selected and appears as below.

8 Play Control Mixer The test signal should contain no other sources mixed with its output therefore all other sources should be muted, turned off or have their levels reduced to minimum. Accessing the Properties window from the above Options menu permits the record mixer to be shown. Record Control Mixer The record mixer should only record from the line input source. The microphone input as well as all other inputs should be muted, shut off or reduced to zero. RPlusD never works correctly from the mic input because the loop back connection for the left (or right) channel requires a line level

9 standard input. The microphone must therefore be connected to a line level input. Once the above connections are made and the mixer is set the setup can be tested. Level Test The levels circled in red should be lower than 0 db after taking the level check. The S/N ratios should be better than 25 db for this sound card only test. In most cases they are 50 db or better with low cost sound cards. An S/N ratio of 25 db or better is required for accurate measurements.

10 Good Result The USB Microphone/Preamp Users of the USB sound card that is included in the USB mic/preamp can plug the unit directly into a USB port, wait for Windows to find the driver then set the mixer as shown below and proceed to take room tests. The above test procedure is not necessary. The mixer for the USB sound card / microphone preamp should be set as shown below. USB mic/preamp mixer

11 This unit does not include input controls and they will work as preset in the unit. The automatic mixer should be enabled from Options->Mixer-> Mixer On/Off. Select mixer on. This setting will be saved when the software is closed and the next time the software is used it will load these settings. Taking Room Tests Once the loop through test has been verified to show that the system is in working order the wiring arrangement on the sound card may be changed. It is important to only test one loudspeaker at a time. Many users make the mistake of testing the system with the left and right speaker playing at the same time. This leads to destructive interferance and a meaningless result. The audio system line may be split using a "Y" type connector so that the left and right channels can be connected. The selection of left or right can be made with the balance control when doing actual tests. The USB mic/preamp uses the right channel for signal input and the connections below are not part of this system. Connections required for this unit connect the microphone to the preamp input and the USB output to the sound system. Room Test Wiring Configuration

12 Red Circled "Start Test" button The Start Test button is then used to conduct a test with the sound card connected to the microphone and computer. Full Range speaker measurements are normally conducted with the Mic/Spkr Distance set to Auto. One way to verify the correctness of your measurements is to verify that the software measured the mic - speaker distance correctly as shown in the right lower corner of the window below. The manual section on experimental room acoustics should be read before conclusions are drawn from any room tests taken.

13 Good Room Test The mic-speaker distance should be very close to the actual mic-speaker distance. Propogation delays through DSP processors may make this figure larger than the actual physical distance. The impulse peak should be seen to start at very close to t = 0 ms on the bottom axis. The signal to noise ratio should normally be above 15 db. The estimate of signal to noise ratio is a very crude estimate shown here. The input signal level should be above approximately - 15 db and below 0 db. The actual level is not critical. In the above measurement the auto-range feature was used to automatically measure micspeaker distance. In some cases the software may not be able to align the impulse correctly. In this case select "manual" and type in the distance between the mic & speaker manually. If only a subwoofer is measured the "manual" selection for mic-spkr distance should be selected and the results should look something like the example below.

14 Set of 32 Subwoofer Measurements Clicking the button circled in red in the above window makes the options box shown below. Click the option for "Apply Impulse Filter" in the window below. Options Box For Time Display The full low frequency impulse can be seen when the single curve arrows are selected providing a full time span view of the low frequency impulse response.

15 Single Low Frequency Impulse The arrows circled in the window above show the single curve impulse response over a long period of about 650 ms when clicked. In this way each measurement can be viewed clearly. Short cut Level Check, Stop and Start Test buttons are provided on the data gathering form so that the Audio Data Gathering window does not have to be opened every time a new measurement is taken. These appear in color after a first measurement is taken as shown below circled in red. The Audio Data Gathering window can be allowed to fall behind this main window and use these shortcuts after the system is verified to be working.

16 Measurement Short Cut Buttons Test Signal Types In the Audio Data Gathering window there are three types of test signals from which to choose if the basic version and the add on enhancement package is licensed. The basic version has two types: Hybrid & Sweep. The third type of test signal should only be used when taking measurements for the electronic crossover network emulator.

17 Audio Data Gathering The hybrid test signal is the test signal that people will most often use. The hybrid test signal provides sweep test signals for low frequencies and pink noise filtered MLS test signals for high frequencies. This test signal provides signal to noise ratio at low frequencies necessary for accurate measurements at the lowest frequencies and a pleasant sounding test signal at mid and high frequencies allowing many measurements to be taken before operator fatigue sets in due to a harsh sounding test signal. Some users wish to have the very high signal to noise ratio at mid and high frequencies necessary to measure reverberation time. The full range sweep is provided for only this purpose. Users are advised to wear hearing protection and be mindful of the power they are sending to their high frequency speakers when using this signal. The reverberation time measurements require specific use of the software outlined in the Time Measurements chapter. The signal should never be required to be played louder than a normal conversation SPL level. The test signal chosen has no direct effect on actual accuracy in any direct way, it impacts signal to noise ratio which in turn affects accuracy. To verify accuracy of measurements with a given test signal, a test can be run twice, one immediately after the other. After this is complete switch to the frequency response type that is desired and see if there is any difference between the two measurements in the frequency range of interest. If there is an appreciable difference then noise is effecting the results. The solution is to first increase the output of the hybrid test signal through the sound system and re-try the experiment. In an extremely noisy environment sweeps should be used. In normal high performance audio environments the hybrid test signal can be played at a level somewhere between a whisper and a normal conversation SPL level to obtain good results. The MLS (XOver Sim) test signal should never be used. It contains very little test signal energy below the specified frequency. It is used to test individual drivers for electronic crossover simulations where low frequencies are neither necessary or desireable - particularly when testing tweeters. This setting is only visible for users that have the enhancement add on license.

18 Alternative to comparing measurements RPlusD provides the signal to noise approximation for each measurement as shown below. The best method is to take two succesive measements and compare them. This alternative is provided for use after measurements have been taken and this cannot be done. Signal To Noise Ratio Estimate When the button circled in red is clicked all signal to noise ratio estimate curves appear. The arrows circled in blue can be used to select any one in the set for clearer viewing. The button circled in green is the average on/off button. It is in the off position and this provides visual access to each individual measurement as opposed to an average signal to noise ratio of requency response measurement. The signal to noise ratio is estimated by comparing the signal level of the impulse with that of the trailing edge of the impulse. The gating determines how much of the impulse is used to compare with the same time period of signal at the trailing edge of the impulse. The gating used for this measurement should be approximately the same as the gating applied for the frequency response measurement. The concept of gating will be fully explained in the chapter on frequency response. The above measurement shows a single measurement that has the mid/high signal to noise ratio at approximately 14 db. This is more than adequate for mid / high frequencies. Low frequencies require a greater signal to noise ratio for people that wish to look at the waterfall plots. Display & Miscellaneous Functions The display limits for RPlusD are set with the mouse. The default limits for the display can be set by clicking the display with the mouse. The default display settings are avaliable from the button

19 circled in red in the figure below. Default Display Clicking the above button circled in red leads to the window below appearing. Default Display Limits The display defaults to these limits whenever the graphical display part of the main window is clicked. It must be clicked after setting these limits and closing this window for the display to change. RPlusD also provides a notepad where notes regarding measurements can be made. It is

20 accessible from the main menu from View->Details. File Details This window provides critical file information as well as textual notes to be entered regarding the data taken. Information may include a customer name, monitor type or listing on each curve number described. Custom microphone calibration files can be created in RPlusD. This is usually done from manufactered microphone data and converts data to the required *.cal file format for RPlusD. Data can be manually entered from graphical data supplied by calibrated microphone manufacterers. This window is available from Tools->Microphone Calibration. Mic Cal Window Detailed instructions and format information are provided in the help file accessible from this

21 window. Lin/Log Selection and Color Selection Graphical display in linear or log format as well as curve color is selectable from the buttons circled in red in the left hand side of the lower portion of the main window. Special Measurement Functions RPlusD also provides some other measurement functions that are used for specific functions. The ETC curve can be updated continuously with the ImpulseRTA. This is avalailable from File- New->ImpulseRTA. This is a continuous pseudo RTA measurement that continuously updates the ETC or impulse response. It is convenient for optimizing absorber placement while watching the display update. ImpulseRTA

22 The Time Measurements chapter in this manual explains the filtering function on the ETC measurement. Generally it should be set as shown. The "Period [ms]" setting should be set to a time much larger than the room RT/60. This sets the update period. A detailed discussion of this is beyond the scope of this manual. A shorter period will have a faster update time but will contain more noise. An update period of a little over one second as above is not conveniently slow for the intended application. Connection details are available in the help file that is available from the above blue help button ("?"). The help file shows the connections required if the mic/system channel connections are selected as the right channel. Users should observe this display with the system in the loop through configuration first to observe the affect the filter has on the measurements. EQ RTA Tuning This function allows the user to connect a hardware EQ and compare its response to the ideal EQ curve found inside RPlusD. The hardware EQ is adjusted as the frequency response display updates in real time until the hardware EQ matches the internally determined RPlusD EQ. The hardware EQ is connected in the channel that is normally used for the microphone & sound system connections.connection details are available in the help file that is available from the above blue help button ("?"). The help file shows the connections required if the mic/system channel connections are selected as the right channel. The SPL measurement tool is useful for measuring looses that occur though doorways and walls as well as other similar functions. SPL Measurement This tool requires a separate noise CD. It is explained in the help file available from the above window.

23 Test CD Data Gathering RPlusD may be used in the wireless mode if the enhancement add on package is purchased. This function is explained in its respective help file and is only recommended for cases where a hard connection between the audio system and computer sound card is not possible. The system is not ideal but does provide ideal measurements when using fractional octave frequency response displays. This function was designed for measuring the response of automobile sound systems.

24 Chapter 2: Time Measurements Time based acoustic measurements are used to measure times and levels of reflections. Many users mistakenly use the impulse response for this purpose but the impulse response can be read incorrectly and is therefore not an ideal measurement from which to examine reflections. The impulse response is only good for verifying that a measurement has been taken correctly and the the computer, sound card, microphone and audio system are all in working order and connected in a functioning arrangement. There are many aspects of this measurement that allow the verification of a correct measurement. Impulse Response Measurement The first characteristic of a good measurement can be seen in the lower right hand corner. Here we see a level at or below 0 db "Input Signal = db". This shows that there is no overload. The software zero level is set at 6 db below the normal setting in most types of software. This is done for reasons beyond the scope of this manual. The input level for this application should be kept at or below 0 db to prevent overload. The input signal to noise ratio is crudely estimated as "Input Signal S/N = 17.7 db". This should normally be above 15 db unless band limited speakers such as subwoofers are being measured in which case it will be much lower.

25 The mic-speaker distance shown here as "Mic-Spkr Dist = m" will be the actual mic speaker distance. If there is a signal processing component in the signal path this distance may be larger than actual. If the impulse response is aligned with t=0 ms then this distance will be very accurate if there are no signal processing components that delay the signal in the path. Signal processors in the path may add more than a meter to this distance. In some cases the impulse response may not actually align at t = 0 and the mic-speaker distance will be in error. The impulse can be manually shifted in the software. When the impulse is shifted the mic/speaker distance changes as well and will reflect the correct mic speaker distance when the impulse is moved to align with t=0 ms. Shifting of the impulse will be explained in a later section in this chapter. The signal to noise ratio of the measurement can also be checked with the signal to noise display. The impulse must be aligned approximately to t = 0 for this display to function properly. Signal To Noise Ratio Display Click on the Signal To Noise button shown in the Results window and circled in red (above) to see an estimated signal to noise ratio of all measurements. Click the arrows circled in red to view any one signal to noise ratio. The average button is circled in blue and it should be in the shown position for the individual curves to be displayed with the arrow buttons.

26 Impulse Response: Options Button Clicking the Options button (circled in red in the above window) makes the following window appear. Impulse Response Options: Impulse Shift The impulse response can be shifted left or right using the arrows that are circled in red in the above form. Notice that the impulse response is selected in the Single Curve Impulse Display frame in the above window. The arrows are used to shift the impulse to align with t=0 in the display.

27 The impulse response may be filtered as with the settings circled in red as shown below. Impulse Response Options: Impulse Filtering Closing this window and clicking on the arrow key in the window below causes the filtered impulse to be displayed. In this case the impulse is filtered between 160 Hz and 2000 Hz. Impulse Response: Filtered between 200 Hz and 2000 Hz This filtered display can be useful when the response at mid and high frequencies is desired and

28 there is too much low frequency noise. With the filter applied to the displayed data you can verify that the impulse starts at t=0 or shift it accordingly. When measuring only subwoofers it is best to invoke the impulse filter to eliminate out of band noise from the single curve display. The signal to noise graph can be used to verify that the data in the frequency range of interest is not corrupted with noise when measuring subwoofers. Subwoofer only measurements often show some "grassiness" (high frequency noise) in the display that can be filtered out. This is out of range noise and does not affect results in the frequency band of interest. A typical subwoofer measurement may look like the curve shown below. Subwoofer Measurement With "Grassy" Noise The impulse filter can be used to select a more suitable frequency range from which to view the "low frequency impulse". If the frequency range only includes typical subwoofer frequencies the time range on the graph is automatically expanded to a time more suitable for looking at low frequency information.

29 Time Options: Subwoofer settings The above settings for a subwoofer measurement in the options window set the low frequency band filter in place and expand the time interval for better observation.

30 Filtered Impulse For Subwoofer Measurement The time range now covers almost 700 ms for better display of the low frequency response of the subwoofer. Measuring Energy Time Curves The Energy Time curve is the appropriate time display to look at reflections in the measurements. The selections shown circled in red in the window below provide the Energy Time Curve display.

31 Selecting The Energy Time Curves Clicking "Apply" and closing the above Options window and then clicking the arrows circled in red in the window below causes the Energy Time Curves to display.

32 Energy Time Curves The Options box shows two Energy Time Curves selected, one filtered between 1000 Hz & 2000 Hz and the second filtered between 2000 Hz and 3000 Hz. A bandwidth of 1000 Hz was selected to provide the necessary resolution of 1/1000 seconds or 1 ms. If the two drivers (HF & LF) were identical units then these two curves would be identical. In this measurement one driver is an HF unit and one driver is an LF unit and the crossover frequency is 2000 Hz. This display permits comparison of how the two drivers radiate sound into the room. In the above case the reflection levels are higher from the smaller HF speaker because of its increased radiation dispersion. To compare the radiation patterns of the two drivers correctly its important that the same bandwidth be chosen, in this case 1000 Hz. Bandwidth is the difference between the highest and lowest frequency in a chosen band. The band filtered ETC curves also show a direct to reverberant energy ratio in the lower right hand corner text box when the band filtered ETC curves are shown. This ratio is the direct energy divided by the reverberant energy for each of the filtered bands. In the above settings the direct energy is taken as the energy over the first 5 ms of the impulse and the reverberant energy taken over the later 45 ms of the impulse (50 ms - 5 ms). The above values were found to be incorrect and the software has since been corrected in versions and later of the software. Many screen shots contained in this manual will show these figures in error. The above display has its limits fixed and is good for only determining the differences in radiation between different drivers in a two or three way speaker as well as the D/R ratio. To get a more flexible display of energy time behaviour the Large Room Tools can be used. These can be accessed from the top menu Tools -> Large Room

33 Large Room Tools Energy Time Curves This display provides band filtering as well. The chosen frequency band provides much finer time resolution of 1/(5000 Hz -500 Hz) = 1/4500 Hz = seconds = 0.22 ms. This provides a finer display of the Energy Time Curve than the previously shown ETC's and any part of the graph can be selected for close examination with the mouse. De-selecting the BP checkbox provides much more resolution than is necessary. The critical band of hearing is between 100 Hz and 5000 Hz but we know drivers will be omni directional below 500 Hz. The 500 Hz Hz band is therefore usually the most useful.

34 Large Room Tools: High Resolution 0.22 ms display The large room tool will often dissapear behind the main measurement window. It is helpful to minimize the main measurement when looking at data in the separate large room tools window. In addition top these two ETC curve displays RPlusD provides a real time ETC measurement tool. This is ideal for placing absorbers while watching the ETC change in real time. Before this tool is used correct operation of the setup should be verified using the Audio Data Gathering window available from File->New->Normal Room/Speaker Measurement. The reason for this is that this pseudo real time tool has the same test signal for both the left and right channels and can be fooled when the only signal is leakage across the input and output of the soundcard. It can be fooled into showing results that are OK when there is no acual sound card connections. This was permitted in the design to improve the speed of this display. The pseudo real time ETC tool is avaliable from File->New->ImpulseRTA menu slection. It also measures the impulse response in real time. ImpulseRTA Window

35 The filter range can be user selected and its usage is identical to the filter described above. The existing settings have the impulse or ETC curve update approximately every seconds. Slower computers may see this update slower. Shorter times can be chosen but the above settings are recommended for users that do not have the signal processing background to understand the disadvantage of shorter times. Measuring RT/60 The RT/60 measurement is calculated from the schroeder plot, labeled "Sch" in the above form. RPlusD uses the two measurement microphone to measure RT/60 and to calculate the schroeder plot. Each schroeder plot is the result of two successive measurements taken from a single microphone location. The measurement curve 1 and 2 combine to make the first schroeder plot, the measurements 2 and 3 combine to make the next one, etc. The window below shows the schroeder plot formed from curve number 5 and 6. Schroeder Plot (Curves 5 & 6) The schoder plot default frequency range is 500 Hz Hz but it can be changed by enabling the band pass filter with the checkbox shown above.

36 Schroeder Plot With Noise after 257 ms (Curves 7 & 8) Another Schroeder Plot With Noise after 300 ms (Curves 13 & 14)

37 The schroeder plot goes "crazy" when it becomes infected with excessive noise after a certain period of time. The display may show the plot going downward to "negative infinity" or becomming jagged. Selecting the reverberation time button directly from here will provide an RT/60 graph with the time limits set to the default 50 ms ms. The RT/60 is the slope of all the schroeder plots in the curve set and averaged. The RT/60 calculation recalculates the schroeder plot for each band spaced one octave apart across the full human hearing frequency range then calculates the best linear approximation to a straight line that the data provides using linear regression between the time limits for each band. The RT/60 value for each frequency is the average slope of all the schroeder plots at each frequency band. RT/60 Calculated With Default Time Range Switching back to the schroeder plot we can select a linear region of the schroeder plot with the mouse and scan through each schroeder plot in the set using the arrow keys to check that the displayed data is approximately linear for all curves over the time range of interest. Selecting the most linear range with the RIGHT mouse key held down causes the RT/60 to be recalculated with the new time limits.

38 Linear Portion Of Schroeder Plot This linear portion can be re grabbed while holding the right mouse key down to produce the new RT/60 result.

39 RT/60 With New Time Limits Notice that the upper frequency range has a band that shows a result of zero. The software determined that there was too much noise in this range of frequencies to get an accurate result. Another calculation should be attempted with a smaller upper time limit such as 200 ms rather than the above 250 ms selected from the schroeder plot set.

40 Data Gathering: RT/60 Settings The two measurement method for measuring reverberation time used to two curves for each schroeder plot calculation. This provides a method to remove background noise. The method used compares the similarity of the two impulses when calculating the schroeder plot. The method depends on the two measurements being from the same measurement mic location for each pair and the autoranging option for data gathering should be turned off. Microphone positions for the reverberation time should be selected from points around the boundary of the room. It is helpful to re orient the monitor position so that it radiates directly to the microphone. For each measurement two succesive tests are taken. It may be necessary to delete certain curve pairs because their schroeder plots are infected with noise. This can be done from the main measurement window using the delete button circled in red in the figure below. The delete button will have to be clicked twice to remove both curves in the curve pair and preserve the arrangement so that curve 1 & 2 form the first set, curve 3 & 4 form the second set, etc.

41 Curve Deletion The curve to be deleted with the "X" button is selected using the arrow buttons circled in blue. The signal to noise ratio above is only an estimate. It is usually best to let the software calculate all schroeder plots as above then select the curves to be deleted from odd looking schroeder plots. Various noisy schroeder plots are shown below.

42 Noisy Schroeder Plot In this case curve 13 and curve 14 would be deleted. This can be done by selecting curve 13 on the main form and pressing the delete button twice.

43 Noisy Schroeder Plot It is impossible to characterise what a noisy schroeder plot will look like. It is a random occurance.

44 Chapter 3: Frequency Response Impulse Response and Gating Frequency response measurements are the most important measurement in audio. They are calculated directly from the impulse response measurement. The impulse response of the system is calculated from the test signal recorded at the input terminals of the sound card. The beginning topic of this chapter may be a bit difficult to understand on the first read. Our advice is to read the section on impulse response a few times before continuing to read the rest of this chapter until the concept of gating is firmly understood. The impulse response of a system tells us how energy is dissipated in the system that is being measured. Knowledge of the impulse response provides the information necessary to predict how the system will respond to any given input. This measurement has a real world example - in the case of checking a cars suspension you can give it a quick bump and see how the suspension dissipates the added energy that was supplied from the kick. The kick that you give the car is the energy input in the form of an impulse. If the car oscillates then the suspension is underdamped and new shock absorbers (read energy absorbers) are required. In this case the system takes too long to dissipate energy. The times involved with this behaviour allow us to see it and easily observe it. The car may oscillate once per second but this is a simple vibration. In the case of more complex systems that have large numbers of energy storage devices, complex natural vibrations occur and its too difficult to get insight to relative system behaviour using the impulse response. The impulse response contains undesirable components such as room reflections which are hear as a separate sound from the direct sound or filter out in the ear-brain system. Room reflections typically arrive at the ear from a different direction than the direct sound allowing this filtering to take place. We wish to remove these reflections from the impulse response before converting the impulse response to a frequency response. Room reflections are seen at a later time in the impulse response than the direct sound because the path length of reflections is always greater. The fact that reflections always arrive after the direct sound permits them from being removed from the impulse response before the frequency response calculation is performed.

45 Loudspeaker Direct & Reflected Energy Room Response With Reflections Gating is sometimes refered to as windowing. It is something that is always done to an impulse response before a frequency response can be calculated when measuring audio systems in practical environments. In the beginning examples we will be using the Bode or "unsmoothed" frequency response to illustrate the principle of gating and how it is applied. Some examples illustrate the principle of gating below.

46 Impulse Response The above impulse response shows the impulse and 30 ms of room reflections. The frequency response below is calculated using the full 30 ms of data from the impulse shown above. The frequency response options window is selected from the button circled in blue when the button circled in red is depressed. Frequency Response Options After the frequency response type is chosen the "Apply" button can be clicked and the Bode or FFT magnitude response is shown. (The Fract Oct and PsychoAcoustic options will be discussed in the next section of this chapter.)

47 Frequency Response: Gating = 33 ms The above graph contains multiple dips in the response that make it almost useless. These dips are considered artifacts because they do not represent any real frequency response anomality as heard by a listener. The reflections arrive at the human ear from a different direction then the direct sound and this causes them to be perceived differently by our human hearing system. Each individual perceives them differently and the stereo human hearing and brain filters them out. What you see is not what you get in this measurement. We know that a sound system that provides a flat frequency response will be perceived as the most accurate sounding from research done at various institutions such as that done by Floyde Toole at the National Research Council in Canada as well as work done at the British BBC many years ago. To remove these reflections and their associated affect of the frequency response we must gate most of the impulse response so that the reflections do not get included in the calculation.

48 Gated Impulse The above impulse response is gated to only include the first 1.9 ms. This new gated impulse is used to calculate the frequency response shown below. This frequency response includes only a first few very early reflections as well as the direct sound from the loudspeaker.

49 Frequency Response:Gating = 1.9 ms The shorter gate time removes all the impulse response after 1.9 ms and produces a much smoother response than the 33 ms response shown above. The disadvantage of the shorter gate is that it limits the lowest frequency that can be included in the frequency response measurement. The above response starts at approximately 500 Hz because 1/(1.9ms) = 1/ (0.0019) = approximately 500Hz. The 33 ms gate gives us a frequency response that that has a minimum low frequency of 1/ (33ms)=1/(0.033)= approximately 30 Hz. Long gate times are required if a frequency response that includes low frequencies is required. Long gate times require a very detailed full range Bode response. It would require as many as ,000 points to graph a full range frequency response with a long enough gate time to include the lowest of frequencies. These thousands of points would show far too much detail at high frequencies to be useful. Most of this detail would just be due to artifacts caused by room reflections. The time required to print this great number of points is also slows down the calculation therefore RPlusD only plots a maximum of 2000 points. When long gate times are used RPlusD truncates useless high frequency data that is mostly comprised of reflection artifacts. The Bode response is not often used in audio because it has this fundamental disadvantage. A variety of processing and measurement techniques is employed to overcome the practical difficulties in measuring frequency response in real rooms. Practical Frequency Response Measurements The chapter on practical acoustic measurements will outline more detailed experiments for correctly measuring frequency response as well as the rest of the acoustic quantitites that people often wish to measure in small to medium sized rooms. This chapter will outline the various types of frequency response measurements and explain why many individual response measurements must be averaged to form a meaningful result across the full range of human hearing for an audio system in a typical setting. One way of removing the artifacts from a standard Bode or "unsmoothed FFT" response is to apply fractional octave smoothing. RPlusD provides fractional octave soothing from 1 octave to 1/20 of an octave. An octave is defined as the halving or doubling of any frequency. A frequency one octave above 100 Hz is 2^1 times 100 Hz = 200 Hz. An octave below 100 Hz is 100 Hz / (2^1) = 50 Hz, where "2^x" is read "2 to the power of x". A frequency 1/10 of an octave above 100 Hz is: 100 * 2^(1/10) =107 Hz. A 1/3 octave smoothed frequency response uses a weighting window that is 1/3 of an octave wide to average data at each frequency point. A 1/10 octave smoothed frequency response has a much smaller weighting window and is not as effective as the 1/3 octave window in smoothing data. The 1/10 octave resolution response shows more detail. The concept of an octave is used frequently in the discussion of frequency response and in the subject of electronic equalization.

50 Frequency Response Options: Fractional Octave Smoothing Making the above selection and clicking the apply button causes all of the fractional octave measured results to appear. After the Apply button is clicked we make the selections shown below. Frequency Response Options: Bode Response Click the apply button again and both measurement types are loaded into the buffers. (The "Filter" checkboxes must also be checked but this will be addressed in the section on EQing). Closing the above form and clicking the arrow buttons in the Results frame makes the graph below appear.

51 Bode and 1/3 Oct Fractional Octave Response The fractional octave response is the red curve in the above window. It smooths the response curve and eliminates many of the artifacts due to room reflections. The fractional octave response resolution can be increased as shown below.

52 Bode And 1/20 Oct Fractional Octave Response In both cases above the fractional octave response was calculated with a 50 ms gate time. If the FFT response is calculated with a 50 ms gate time the minimum frequency on the curve would be 1/(50 ms) = 1/(0.05 s)=20 Hz. In the fractional octave response the minimum frequency is much higher. Many types of software will plot the minimum frequency as 20 Hz in this case but RPlusD does not make this error. The fractional octave response starts to become smoother than the specified smoothness at a higher frequency than that determined by the simple formula, Fmin = 1/GateTime. The minimum frequency shown in RPlusD fractional octave responses is the minimum frequency that can be shown while still maintaining the specified degree of smoothing, in these two cases this being 1/3 oct and 1/20 oct respectively. The "cure" for this is to increase the gate in the fractional octave response until the desired minimum frequency is shown. This is the purpose of the gate setting in the fractional octave response curves in RPlusD. This method ensures that the smallest amount of impulse is used to make the calculations leading to a minimum amount of room response included in the response graph. Another way of dealing with the artifacts in the Bode response caused by room reflections is to take measurements with the measurement microphone placed at many points around the listening area. The reflections arrive at a different time for each measurement microphone location and the dips in each response can be averaged out with a "power response average". When many curves are loaded into the software they can be averaged.

53 32 Bode Response Measurements Little sense can be made of the above set of 32 measurements. RPlusD provides an averaging utility that averages the above set of curves. This removes the artifacts associated with the measurements and provides a single average response over a loudspeaker listening window for which these measurements were taken. More detail on this is illustrated in the chapter of practical measurements.

54 Averaged Bode Response Note the position of the average button (circled in red) in the above window. Notice the "G" button in the Results frame in the above figure. The G button means that the average includes the gain of each response measured. Clicking on the "G" button converts it to show a label of "N", in which case the responses are normalized before an average is taken. For most applications the setting of this button really doesn't matter and it should be left at the "G" position to avoid confusion that occurs when band limited speakers are measured and this button position is on "N". The above average still shows quite a bit of jagged behaviour due to room reflection artifacts but a general shape can be seen.

55 Averaged Fractional Octave Response The above is an averaged 1/3 octave fractional octave response in a listening window. A set of low frequency measurements is shown below.

56 Low Frequency Response: Gate Time = 614 ms Notice how the measurements look similar up to and below approximately 80 Hz. The region below 80 Hz is known as the modal region. The region above is called the diffusion region.

57 Averaged Low Frequency Response: Gate Time = 614 ms Averaging the Bode response works best for low frequencies because the fractional octave response smoothing tends to dull resonances when low frequency responses contain sharp peaks (resonances). Averaging the fractional octave response works best for mid and high frequencies because the averaged Bode response does not filter out the room reflection artifacts as well as the fractional octave response. Resonances will be discussed in greater detail in the chapters on electronic equalization and the chapter on practical measurements for room acoustics. The fractional octave response uses an SPL average and the Bode response uses a power response average. The SPL response average simply adds the SPL's of the curve and then divides this by the number of curves to generate each point in the average. The power response puts each point into its original (non db) linear form, sums them, determines the average and then re converts this average value back to a logarithmic (db) value. The power response method ensures that deep dips in the response pull the averaged value down by only a minimal amount for each frequency value in the response. At mid & high frequencies the averaged fractional octave response removes artifacts and shows the underlying response of the system most clearly. At low frequencies the averaged FFT response gives the clearest indication of the system pathologies (resonances) that are of most concern. PsychoAcoustic Response There is a way of looking at a full range frequency response in RPlusD. The phychological model uses a "time equivalent bandwidth" model to calculate frequency response. There some presets available for this response but experimentation in many different listening environments has lead to the options preset in the software.

58 To explain the time equivalent bandwidth model we first look at the fractional response again. In the above fractional octave responses, gate time needed to be increased to reduce the lowest frequency plotted. In the time equivalent bandwidth model each point in the curve is recalculated with the mimimum gate time required for the required preset resolution. At very low frequencies a longer gate time is required to make the curve have a 0.75 octave resolution (a preset). As the points are gathered for higher frequencies a reduced gate time is used until that gate time reaches 5 ms (a preset) at approximately 200 Hz. At frequencies above 200 Hz the necessary smoothing to maintain a 1/3 octave resolution is used. The presets for the PsychoAcoustic Response are shown below. PsychAcoustic Presets These preset values are explained in detail in the help topic available from the help button above. This window is available from the "Param." button in the frequency response options box. This provides a much better indication of overall frequency balance as heard by a listener. It can be used to adjust coarse tone controls such as Bass and Treble as found on many systems.

59 Bode/FFT (Blue), Fract Oct (Light Blue) & PsycoAcoustic (Red) The psychoacoustic response is the best overall indicator of full range frequency balance and can often be quite different than the fractional octave response. The accuracy of the fractional octave response at high frequencies can be improved with a reduced gate time. There is still more to discuss with regard to taking meaningful and accurate measurements. This chapter was written to explain the concept of frequency response so that the methods illustrated in the chapter on practical acoustic measurements and the chapter on equalization can be understood. Pseudo Real Time Frequency Response RPlusD also provides a pseudo real time frequency response tool available from the File->New-> EQRTA tuning menu selection. RTA EQ

60 This tool is not used for room response measurements. Its sole purpose is to be used to measure electronic components. This tool provides a sampled Bode response in the real time display that is used to set equalizers to an RPlusD solution for equalization that is determined from many averaged measurements. The RPlusD method of equalization is far better than the typical method of equalization employed by setting an EQ to correct a real time display of frequency response. This older method can lead to significant error. RPlusD was developed partly because this newer method could be done on computers with large reserve memory to store many curves in the buffers from which to apply equalization in emulated form. Equalization is a separate topic and has a manual chapter devoted to it.

61 Chapter 4: Practical Measurements This section will outline practical measurements used to identify early reflections, room resonances, loudspeaker / room response as well as the direct sound that gets radiated from the loudspeaker as a first arrival at the listener location. The subject of relative liveness and critical distance will be discussed at the end of this chapter. Readers are advised to read the chapters on time and frequency response before reading this chapter. Reflection Measurement An often necessary task in the measurement of reflections is to determine the location of the offending surface. Fortunately sound waves behave like light waves when they reflect from a surface. A trick known as "the mirror trick" can be used to locate surfaces from which the direct sound from loudspeakers reflects. Ideal absorber placement positions on rear walls, side walls and ceilings can be identified using this method. The mirror trick is best illustrated as below. Locating Reflection Points As a mirror is moved along a reflecting surface an observer seated at a listening position watches the mirror until the reflection of the loudspeaker is seen in the mirror. At this point the mirror is in a spot where sound is being reflected. Actual sound absorbers should be used minimally and on surfaces that are the source of only early reflections as identified using the method above. Ceilings are a good place to place absorbers since these reflections come from above and the human hearing system is not good at processing sounds that originate from above. We were never hunted from above. Absorbers will be effective to frequencies as low as ones with wavelengths as 4 times the thickness of absorbers. For an absorber to absorb a 500 Hz signal its thickness must be: Speed Of Sound = 344 m/s Wavelength = 344/500 = 0.68 meters 0.68 meters / 4 = 17 cm = 6.5 inches thick

62 A two way monitor may start to become directional at 1 KHz and therefore remains directional up to its 2 KHz crossover frequency. In this case an absorber 1/4 as thick as the thickness required at 500 Hz will be required to work at 2 KHz where the HF unit takes over radiation and becomes omni-directional. In this case the thickness required is under 2 inches. The length and width of the absorber should be much greater than the wavelength of the lowest frequency to be absorbed and it must allow for changes in the actual listening position. The mirror trick can be used with the listener placed at the limits of the listening position and observing the position of the mirror required for the loudspeaker to be visible in the mirror. The RPlusD energy time curves can be used to test the reflection levels at various bands of interest. In this case the reflection levels between 500 Hz and 1500 Hz can be used to measure the reflective energy in this band. A filtered band for the energy time curve may also be set to 1 KHz - 2 KHz to look at this band. In both bands the ETC curves will look identical for ideal situations because both bands have a bandwidth of 1000 Hz. Keeping the bandwidth constant allows the direct comparison of reflection levels between the different bands of frequencies Hz Hz (yellow), 1000 Hz Hz (red), 500 Hz Hz (blue) The bands in the above ETC bands were selected from the options box circled in red. Selections are shown below.

63 ETC Curve Filter Bands The earliest reflections have the yellow band as the most significant therefore some substantial lowering of early reflection levels may be accomplished by some relatively thin absorbers because this band covers the upper frequency region starting at 2 KHz. An absorber to deal with frequencies of 2 KHz and above isn't very thick. The above ETC graphs are designed to compare ETC's over different bands. This is convenient for comparing how a two way speaker radiates sound into the room from its HF driver to the sound radiation of its LF driver. To get a more detailed look at the ETC results the large room tool can be used. To access this go to Tools->Large Room from the main menu. The affect of a single Micro Trap on a single early reflection is illustrated below. The graphic limits are 0 db to -20 db for vertical limits and the horizontal limits are from -1 ms to 2 ms on these two results shown in the following examples. Wide Band ETC Early Reflection (No Absorber)

64 Wide Band ETC: Early Reflection (Microtrap Trap in place) The reflection at 1.2 ms is completely removed with this absorber. It is particularly important to control early reflections before 5 ms. Using a smaller bandwidth gives less resolution in the time domain as a penalty for the greater resolution in frequency. The band was reduced from 500 Hz - 10 KHz as above to 500 Hz KHz below. The reduced resolution is a mathematical limitation arising from the uncertainty principle of obervation rather than a weakness of the software. The narrower frequency band chosen causes a smearing of the time domain data. The graphical limits on the results shown below are from - 24 db to 0 db for the vertical and -3 ms to 9 ms on the horizontal.

65 Narrow Band ETC: Early Reflection (No absorber) Narrow Band ETC: Early Reflection (Micro Trap in place) These tests should be run with the microphone at various physical limits of the listening region to ensure that all seated listeners do not hear the early reflection. The best frequency range to choose when using this tool is about 500 Hz - 5 KHz or 1 KHz to 5 KHz. In the case of multi-way speakers each speaker supplies energy for the reflection therefore two peaks may be seen for

66 each reflection, each peak representing a driver operating in the band of the ETC filter. It is not always a good idea to use side wall absorbers and the only time they absolutely should be used is when one speaker has a side wall nearby and the other speaker has an open space. Stereo loudspeakers should be used in a room that has acoustic symmetry. In cases where absorbers are not used for the purpose of increasing symmetry, listening tests with and without their placement is the only way to know if absorbers improve sound quality. Systems will almost always benefit from the use of ceiling absorbers. In testing absorbers and their affectiveness for early reflection control around the loudspeaker it may be necessary to move the listening seats out of the way to prevent reflections from them from interfering with what is being measured at the microphone. Seats and chairs often provide reflections that interfere with what is being observed. Its important not to use more absorption than necessary to control early reflections for reasons that will be explained in the last section of this chapter. Room Resonance Measurements Many small rooms have a boomy sound to them at frequencies that are un-naturally emphasised by the dimensions and shape of the room. This is particularly true of rooms with few openings and constructed with hard surfaces such as concrete or thick drywall. There are many ways of reducing sharp room resonances. The first step is to take a measurement and find out the frequencies that are problematic. The chapter on frequency response explains why many measurements are necessary to get a clear view of the actual response in a room. The artifacts caused by room reflections get filtered out from an averaged measurement but the room resonances, themselves created by room reflections do not get filtered out in the averaging process used in the un-smoothed FFT (Bode) response. Of course we know that room resonances are caused by room reflections because if all the room surfaces were kept absorptive at very low frequencies no room modes (resonances) would occur. The set of room measurements taken around the listening area of a mostly concrete room is shown below.

67 Room Resonances / Listening Area Measurements Notice that the curves remain approximately consistent in the modal region below approximately 80 Hz in the above measurement set. Above 80 Hz the curves are no longer uniform. This is the diffusion region of the room and measurements must be averaged to identify overly excited modes. The average button is circled in red in the above window. In the region above 80 Hz the results become highly sensitive to microphone position. The above result shows that we can move a subwoofer(s) around a room and optimize its placement for the response below 80 Hz with a single measurement microphone location at the listening area used for all measurements. Above the 80 Hz frequency the response is too sensitive to the actual microphone position to compare various subwoofer locations. A few curves similar to the above above measurement set should be taken to determine the lower frequency of the diffusion region. After this is complete the single mic location measurement can be used for each subwoofer location test and the results compared only below this frequency (of 80 Hz in this case). There would be no point in optimizing a subwoofer location for the frequency range above 80 Hz in this case because if the microphone was in a slightly different position measurement results would be highly divergent. It is helpful to set the default display to the low frequency range of interest and to set the display to average. The display limits button is circled in red in the window below.

68 Averaged Room Response Over Listening Area (resonances circled in red) Display Lock The display is shown locked for the band between 25 Hz and 200 Hz. Clicking the display with the mouse causes the graph to be redrawn with these limits after the OK button is clicked and the Display Lock window is closed. The same room as above was measured by placing the microphone at various spots around the perimeter of the room. This result is shown below.

69 Averaged Room Response Around Room Perimeter All of the resonances in the listening area appear in this perimeter result along with a few that do not appear in the listening area measurements. A strong 150 Hz resonance appears in both sets. It is circled in red in the above figure. Its is likely a floor-ceiling mode but will be corrected by equalization rather than absorbers for aesthetic reasons. Ultimately using the traps would be a better way of controlling this sharp peak than equalization but aesthetic considerations lead to an EQ solution. This example will be discussed in the chapter on equalization. To measure the affect of passive bass traps, a good microphone position from which to take a measurement and observe the modes of interest should first be identified by taking a few measurements at positions around the listening area. Trap placement can be experimented with until the smoothest frequency response is found as measurements are repeated with the microphone in this single stationary position. The effect of Mondo Traps being placed in a different 11' X 16' room is illustrated below.

70 Empty Room 4 Mondo Traps

71 Notice the peak at 70 Hz is controlled with only a few well placed absorbers. 17 Mondo Traps The traps have the effect of smoothing the low frequency response and dulling the sharp peaks associated with resonant behaviour. The advantage of using the Mondo traps is that they are reflective at high frequencies and therefore only affect the low frequency range of interest. The best places to put traps are usually in the corners rooms but experimentation in absorber placement is the key to getting the best performance. A poorly placed set of absorbers may not work as well as one that is well placed. Measurements should be taken with any doors or windows closed. Room openings add absorption and may cause a resonance to dissapear when it would otherwise be present when the sound system is being used and the opening is closed. Its most important to clean up the region above 100 Hz. Sharp resonances below 100 Hz are not nearly as damaging as the ones that exist within the critical band for human hearing of 100 Hz Hz. If a room sounds "boomy" or unnatural then it likely has a sharp room mode in the range above 100 Hz. Subwoofer Placement Measurements There are usually several places in a room that can accept subwoofers. In this case multiple measurements for each subwoofer location should be taken over the listening area. The example below illustrates subwoofer placements and measurements

72 The Two X's mark the locations for which the tests below were done. In this room the frequency range of interest was below about 80 Hz. The subwoofer should be placed to maximize the output in this region. At frequencies above 80 Hz the response cannot be measured reliably because this range is the diffusion range so optimization of placement is useful at frequencies below about 80 Hz. Multiple measurements are taken over the listening area for each placement to average out dips that would be caused at any one particular location. Individual curves for one measurement are shown below.

73 Position 1: Six Measurements 16 Hz Hz Position 1 Averaged 16 Hz Hz

74 Position 2 Averaged 16 Hz Hz Clearly position 2 is the more favourable position. Position 1 has much less outputput in the region of 40 Hz to 70 Hz. After the best position is chosen various methods to smooth out the response can be applied. In cases where more than one subwoofer is used, phase can be switched on one or more subs and measurements for comparison should be taken as shown in the above example. Loudspeaker Direct Sound There are two aspects to loudspeaker response that affect the sound heard. One aspect is the direct sound that will be discussed in the section. The other aspect is the power response that determines the amount of total reflected energy that is present in the room. Good loudspeakers have (most importantly) a flat axial response as well as a linear and smooth non axial response so that reflected energy is well balanced. In this section a method for measuring the direct sound of loudspeakers will be illustrated. The following section will deal with power response and suggest a possible reason why live rooms sound better. In the chapter on measuring frequency response the fact that many microphone positions were required to take accurate frequency response measurements was illustrated. The large number of measurements were necessary to filter out artifacts in the measurements caused by room reflection interferance. In characterizing a loudspeaker direct sound it would be more practical to measure the response over a narrow "listening window" where the listener may be seated relative to the loudspeaker than it would be to characterize the response using a single measurement. This creates a need to take many measurements and average them to form a single response.

75 The narrow listening axis should be defined by the positions in the listening area relative to the loudspeaker. Its important to use truly random microphone locations rather than a carefully arranged grid. This will help to randomize the artifacts caused by room reflections in the results and permit the affect of these artifacts to be removed in the averaging process. It is important to set up the speaker in a position as far away from any room reflections as possible when doing this experiment. For this experiment the speaker was set up on its stand on top of the coffee table in the middle of the room. The loudspeaker should be shifted in position by a few inches for each measurement to further randomize reflections. Loudspeaker Direct "listening window" Response Measurement This experiment should be repeated twice and the two averages of the experiments compared to check consistency in the experimental result. Some loudspeakers will provide virtually no consistency when they are used in this experiment. This is often true of horn type loudspeakers. The standard direct radiator loudspeakers that is often used will almost always yield consist results when this experiment is performed.

76 Set Of 32 Measurements Thirty two measurements may sound like a lot but after the microphone is set up and the speaker in the room center on a high stand the work required to take 32 measurements will be done in short order - it takes about 30 seconds to move the speaker and take a single measurement. The setup and execution time for this entire experiment is less than 1/2 hour once working familiarity with the software is obtained. The 32 measurements will used to form two averages, one of the first half of the curves and the second using the last half of the curves. This will comprise two experiments and give us a check for the validity (repeatability) of the experiment. After a set of curves is gathered from this experiment the curves may be checked to ensure that no two are overly similar. The defining characteristic of each individual curve will be the artifacts from the room reflections. We wish to delete one of any two curves that show a large similarity. RPlusD has a similarity indicator for each curve that shows its relative similarity to the rest over the displayed frequency range with the applied gate time. In checking for curve similarities it is best to choose a wide frequency range and employ a wide band that has a higher minimum frequency than 500 Hz. The band 500 Hz to 5 KHz is a good approximate choice. The Bode /FFT response should be chosen for the similarity test because it shows the dips caused by room reflection artifacts most clearly and best differentiates between mic positions. A gate time should be chosen so that a wide frequency band can be shown in the Bode response. In choosing random positions for curve generation it is likely that a few mic positions that are too close will be used. The similarity function will be used to identify measurements that are suspect of near duplication in the set.

77 Curve Similarities 600 Hz - 6 KHz In this set, four curves, one being curve #8 showed a greater than 0.8 similarity rating between it and any other curve. These four curves were deleted to increase the randomness of the individual measurements. The value 0.8 was chosen because only a few curves had a similarity above 0.8 in this set. The actual similarity ratings will be highly dependent on the frequency range chosen as well as the gate time used. The best thing to do is to select the four or five curves that are more similar than the rest for deletion. The set contained a remaining 28 curves. The first 14 curves were deleted and the remaining saved as one file. The original file was opened and the second 14 curves were deleted and the remaining average compared to the set with the first 14 curves deleted. The two experiments were found to be highly consistent. The total curve set of 28 curves is then used to give a result similar to the one below.

78 Loudspeaker Frequency Response Over Listening Window The above curve was smoothed with a 1/6 octave smoothing. The vertical limits are -2dB to + 3dB, the frequency band is 500 Hz to 10 KHz. Room effects interfer to a large degree for measurements at frequencies below 500 Hz. A smoother version of it could be found by using the 1/3 octave response. The best approach is to smooth the curve to a degree that provides maximum smoothing but does not dull the major response peaks and dips. This response is remarkably good and is the response of a PSB model M2 loudspeaker. The response could be better if a single axial response was used but no one ever sits directly at the axis of the loudspeaker. Many manufacterers of loudspeakers will optimize the design of their products to have the single ideal axial response and then fudge the data for their product specification and carefully select single measurements that look the best. PSB is my favorite loudspeaker manufacterer because they don't fudge data and their products are designed to provide good sound and good value rather than a specific good looking specification. Few loudspeakers at any price point would match the above performance. This is an accurate indicator of the direct loudspeaker response over the listening area and can be used for further equalization. The topic of equalization will be discussed in a separate chapter devoted to this complex, confusing and most miss-understood and misused subject area of audio. Loudspeaker Reflected Sound This topic area generates about as much confusion in audio as equalization. There many ways in which the reflected sound field in a small room may be measured and quantified. The most popular method of quantifying the reflected sound in small rooms is the use of the reverberation

79 time measurement. Reverberation time is convenient, easily understood and widely accepted. These three things, however, do not make it correct - particularly in a field such as acoustics where everyone is an "expert". Part of the goal of this chapter is to impart the necessary understanding to the reader so that the relative unimportance of reverberation time is understood and the truly relevant aspects of reflections in small rooms can receive the required attention. The reverberation time measurement is included in the software because it is demanded by customers not because the measurement is recommended by AcoustiSoft for small room application. The guiding principle of RPlusD is to make all aspects of the software far better than that in ETF5. The measurement of reverberation time is no exception to this principle and a substantial amount of time was used to develop the RPlusD reverberation time measurement for customers that demand the higher standard of performance associated with this software. Reverberation is the sound that has time to reflect around the room to become diffuse and random. It must reflect around for a long enough period of time to be heard as separate from the direct sound and this takes approximately 50 ms. The issue with reverberation in small rooms is that sound is rapidly absorbed before it exists in the room for a long enough period of time to be heard as distinctively separate from the direct sound. This contrasts greatly with environments where most of the energy heard is actually reverberation energy such as in concert halls and auditoriums when they are used for musical venues. Reverberation time is calculated as the slope of the schroeder plot which is shown below. The scroeder plot shows us exactly how much reverberation energy is in the room after an initial sound. From the plot below, the amount of energy left over after 50 ms is 10 db below the initial sound. This room has an RT/60 of 0.6 seconds, for a typical room that has an RT/60 of 0.3 seconds this amount of energy would actually be less than 20 db below the direct sound. Schroeder Plot The early reflections are at a level of between - 3 db and -6 db below the direct sound. The early reflections have a far greater amount of energy than that of the reverberant energy and they play a much greater role in determining the relative acoustic liveliness or deadness in the room. The reverberation energy is almost completely masked by the direct sound and its early reflections. In the case of a concert hall having a reverberation time of 3 seconds, the reverberant energy

80 that exists at a time period of 50 ms after the direct sound is heard is only 3 db down. This means that the reverberation in these larger rooms can be heard quite clearly and is therefore an important large room parameter. Small room acoustics demands that we consider the direct sound and its early reflections because that is where the energy is. The Direct to Reflected energy ratio determines how "live" or "dead" a room actually sounds- the actual reverberation is hardly heard at all! The D/R ratio is a computation of the direct sound energy divided by the reflected sound energy at the microphone. This ratio varies with the microphone position in the room where this measurement is taken. Close in monitoring is a technique from which a listener uses a small set of loudspeakers placed in close proximity to the listener to monitor recordings. The closeness of the loudspeaker to the listener reduces the effect of room acoustics on the sound heard. It is a technique that brings the listener half way between in room monitoring and monitoring on headphones. It is similar to listening to a recording in a completely acoustically dead room. The sound is dry and unnatural but room independent. As the listener moves away from the loudspeakers the reflected sound in the room plays a larger role in the total sound energy heard by the listener. The required distance from the speaker that the listener must be placed to hear an equal amount of direct and reflected energy is known as the critical distance. The critical distance in any room can be found using RPlusD and a set of trial and error measurements in successively taking measurements with the microphone being moved to positions further away from the source loudspeaker as measurements are taken. The D/R ratio will be 0 db when the microphone is at the critical distance. For most listening environments and loudspeakers this distance will be approximately 1-2 meters away from the loudspeaker. At distances closer than the critical distance the D/R ratio is positive. For measurements taken at distances greater than the critical distance, the D/R ratio will be a negative quantity because the reverberant energy will be greater than the direct energy. The ideal D/R ratio is a subject of some debate, much of it can be found on the internet. The D/R ratio should not concern small room acousticians directly because there are other forces at work. Ultimately the control of very early reflections, particularly ones that come from the vertical direction such as ceiling reflections should be controlled and the reflections that arise from the non symmetrical nature of the room should be controlled. If the right speaker has a wall beside it and the left speaker has empty space then placing an absorber on the right wall is something that should be done. If the room is symmetrical and reflections from both the left wall and right wall are present then it is not known by this author if these reflections should be controlled or not. The solution here is to experiment with removable absorbers. Other than these considerations the room should be kept as live as as possible. The next section of this chapter will illustrate a possible reason for this. Loudspeaker Power Response The reflected energy that is heard in a room will ultimately depend on how the loudspeaker radiates sound in all directions. High performance loudspeaker manufacterers go to great lengths to measure the total acoustic power that gets radiated from their products. It is important that this radiated power response be as smooth as possible in nature. The most common direct radiator types of loudspeakers have a power response that will decline with frequency due to the physical nature of the drivers and their sound dispersion characteristics. Fortunately the measurement of total radiated power of the loudspeaker is not something that needs to be measured nor can easily be measured using the resources available to the home theater or studio owner. The measurement that can be done is the power response over the listening area of the in place loudspeakers. The reason for measuring this response to be use it

81 to observe the affect equalization has on it. RPlusD uses a non destructive approach to equalization that requires this measurement to be taken. Equalization will be explained in the following chapter in great detail. The reason the room should be kept as live as possible except for the control of possibly damaging early reflections is that it helps to improve the power response as heard over the listening area. The power response is tricky to measure and the method of measurement of power response is explained in this remaing section of this chapter. The measurement of power response requires a set of independent measurements to be taken and averaged over the listening area. The various microphone position spacing will determine the lowest frequency of validity for this measurement. In a typical 6 foot by 2 foot listening region six microphone positions can be used as illustrated below. The lowest frequency for which this would be valid is: Mic Spacing = 2 ft Lowest Wavelength = 4 X 2 = 8 ft. Minimum Frequency = 1100 ft / 8 = 140 Hz Below 140 Hz the measurement data should be considered random and useless. In this case two measurement sets can be taken, one grid spaced 2 ft above the other grid. This would cover the listening space adequately. A spacing of one foot between various mic locations will provide a greater number of measurements over the area and improve accuracy but be valid for frequencies as low as 280 Hz. Four times the number of microphone locations can be included in a grid that has all mic locations spaced 1 foot apart. The greater number of mic locations will increase the resolution that can be obtained in the measurement. A valid experiment is a repeatable experiment and this experiment is repeatable provided the resulting average frequency response is suitably smoothed using the fractional octave response. To test the degree of repeatability of this experiment it can be repeated by shifting the listening location half the mic spacing and repeating the measurement. The fractional octave response can be adjusted starting at 1 octave to various fractions such as 1/2 octave until the consistency between the two measurements is lost. The two experiments will look completely dissimilar if high resolution responses such as Bode/FFT or 0.1 octave measurements are used. Consistency in experiment will normally be obtained with the fractional octave resolution set between 1 octave and 1/2 octave. It is not possible to take enough measurements to provide a repeatable experiment with 1/3 octave or higher resolution. These measurements will almost always show a decreasing energy level with increasing frequency and this is why the room should be kept as live as possible. Some experimental results are illustrated below with six microphone locations used for each averaged result. The resolution is 0.75 octave in these results.

82 Six Measurements: Experiment #1

83 Six Measurements: Experiment #2 Notice the increase in power that is present at frequencies just above the crossover frequency. This is due to the increased radiated power of the HF driver because of its greater dispersion in this region than that of the larger LF driver. This may seem like large number of measurements but each measurement takes about 30 seconds to complete and a good power response measurement involving 24 measurements would take a little longer than 10 minutes. Our recommendation is that the user perform these experiments once for practise and to understand the software while making note of mistakes. The experiments can be repeated after this to provide the useful results sought. To perform all of the experiments outlined in this chapter will take a few hours but the results of these experiments will be both repeatable and correct. One more measurement can be used to characterize the sound of loudspeakers. The dynamic response shows how the loudspeaker frequency response varies with level. Lower cost loudspeakers tend to provide a very good response at between 80 db and 90 db but when the level is increased to 90 db to 100 db the response can change quite dramatically. A loudspeaker that has a response that changes with level is heard as one that has dynamic compression. This test can be quickly performed with the speaker and mic in stationary positions and successive frequency response measurements can be taken with a gradual increase in signal level being fed to the loudspeakers. The PSB M2 measurements for this were very good but were lost and cannot be shown here. A

84 high SPL calibrated microphone is required to carry out these tests. The rest of the tests only require that the test signal be played at an SPL level somewhere between a whisper and a normal conversation. The hybrid test signal should be used because it will lead to minimal fatigue when performing many measurements. Chapter Summary The theme of this chapter has been experiment repeatability. A set of measurements taken in a room can be considered an experiment, for that experiment to be useful it must have two essential characteristics: (1) A basis in reality (2) repeatability. If a loudspeaker is measured for frequency response at only one point then that experiment does not have a basis in reality because listeners are positioned within an area when listening to speakers. No one puts their head in a vice to hold a precise postion when listening to an audio system. A loudspeaker or room measurement from a single point in space does have some usefulness, but only when the experimental result can be repeated with approximately the same result with the microphone in a slightly different but still realistic position. Any set of measurements taken or any single measurement taken must have the principle of repeatability in mind when its resulting data is being examined from which to draw conclusions. Clearly, a single mic measurement of a room response shown in this chapter above 80 Hz has no basis in reality. If any experiment is to have a basis in reality then the same results must be obtained with a shift in the measurement microphone position(s) and a repeat of the experiment. Further Reading & Recommended Web Sites The list of resources below is an incomplete list of reliable sources of information on small room acoustics and loudspeakers. (1) Harmon International web site. This site has some very good tutorials on the basics of small room acoustics and subwoofer placement as well as many aspects of loudspeaker performance. (2) Gedlee LLC web site (Earl Geddes) (3) The Book "Premium Home Theater" by Earl Geddes. Earl Geddes is one of the worlds top audio and small room acoustics experts. (4) Real Traps web site (source for Mondo and Micro traps decribed in this chapter). Ethan makes very good passive acoustic products that get professionally tested in laboratories. (5) The book "Home Recording Studio: Build It like the Pros" by Rod Gervais (6) PSB Speakers web site. I have recommended PSB speakers to everyone I know for over 20 years.

85 Chapter 5: Full Range Equalization Equalization generates a lot of confusion in audio. Many people swear against it, others swear by it. Equalization is something that is done according to measurement data, whether or not it is beneficial has more to do with the data collected and the conclusions drawn from the data than the actual process of equalization. Equalization is most often done incorrectly and the hardware equalizer itself is often named at fault. Even if you believe that electronics can add distortion to a typical audio processing chain it is difficult to argue that the added electronic distortion is greater than the amount of distortion associated with a loudspeaker response anomality. Electronic components that have a characteristic sound usually have that characteristic traced to a direct fault in the circuitry and its operation such as in the case of older low cost D/A and A/D converters or an undesirable effect observable in its frequency response. A gradual downward or upward slope in a frequency response that has a magnitude no greater than 0.5 db can make a component sound overly detailed or "fast" or as "slow" as some people describe a component with a gently decreasing level with frequency in the response. For the purposes of this chapter on equalization the affect of distortions associated with electronic components will be ignored. Types Of Equalizers There are many types of equalizers, the most well known is that of a graphic equalizer. Graphic equalizers have limited usefulness but can be used to improve sound when properly understood. A professional graphic equalizer has bands spaced 1/3 of an octave apart. The 1/3 octave equalizer can be used to make corrections to anomalities that are much less than 1/3 octave in bandwidth. An example of this would be an incorrectly biased tape recording where a gentle but precise boost or cut to the high frequency response is necessary. For sharper anomalies, unless the anomality has a precise center frequency that is one of the EQ center frequencies, the equalizer is virtually useless. This is true of all graphic type equalizers. Most problems that demand the use of an equalizer require a unit that is much more flexible that a graphic equalizer. The parametric equalizer provides custom controls from which the frequency and the effective bandwidth of the control can be set by the user. This provides far greater degree of control and requires a greater degree of care when being used. Unless the equalizer is very expensive, an analogue parametric equalizer will have its controls behave in an ideal way only over a small range of Q or bandwidth settings. (Q is approximately defined as 1/Bandwidth but there has never been a universally accepted standard definition of Q for equalizers, consequently a Q of 2 on one unit may mean a Q of 3 or even 1 on another) The non standard Q on parametric equalizers means that when parametric equalizers are set according to a measurement result the measurement must be taken again to verify that the equalizer was set correctly. This problem is very serious when more than one measurement is required to define the anomality being corrected. It is a nearly impossible situation in reality. The parametric equalizer falls into the category of IIR filters. IIR means "infinite impulse response" - that is theoretically an impulse that "rings" forever. A filter when adjusted in a parametric equalizer makes the resulting impulse response have a trailing edge that never fully decays to nothing. There is a certain advantage to this because it is the way resonances actually behave. In this case when an IIR filter is adjusted to exactly correct for a typical resonance the ringing is completely canceled out. The biquad IIR filter provides an exact solution to the problem. Analogue IIR filter equalizers have their limitations in ranges over which ideal behaviour can be attained. Limitations in the setting of bandwidth associated with lower cost analogue equalizers

86 prevent them from being useful in ideally cancelling high Q resonances. An accurate analogue equalizer is very expensive (accuracy implies that the EQ is doing what its control settings say it is doing). Digital equalizers on the other hand can acheive near ideal performance in terms of accuracy - that is they are doing exactly what the settings dictate they do at very low cost. Digital equalizers come in two forms, one being the IIR type described above and the second being implemented in the form of FIR filters. FIR filters are normally used in equalizers where automatic algorithms determine the ideal settings from a given measurement or set of measurements. FIR filters are much easier to design and the ideal settings can easily be determined by simple algorithm implemented inside the software of the unit. An FIR filter can be designed to exactly match a given response anomality, the only question is how accurately can such an anomality be measured. Sharp resonances require large FIR filters and the FIR filter never perfectly matches the resonance because FIR filters do not have an impulse response that continues for an infinite time period. IIR filter implementations are complex and require a certain amount of human input to find their ideal settings when more than one measurement is used to characterize a system and provide a reference curve from which to set the equalizer. Methods Of Equalization FIR filter type equalizers are automatic filter generators that are found in some sophisticated home theater equipment. The operation of these will not specifically be discussed here but the discussion on certainty of measurement later in this chapter certainly applies. The remaining part of this chapter will be devoted to the standard parametric control or "biquad section" as in engineering literature. The biquad section ideally matches resonance behaviour and is the ideal filter for audio in this authors opinion. The most widely known method of equalization is to take a measurement, estimate the filter needed, apply it then verify the filter with another measurement. This is slow & tedious and can only be applied to one measurement at a time. This method can work well for people that have a good understanding of the mathematical behaviour of loudspeakers and rooms. This requires an engineering background or lots of experience. A better method would involve taking many measurements, averaging them, then guessing the filter needed from the measurements, applying the filter then retaking the measurement to verify the filter. This is a huge task - impractical for even the most determined user of equalization. RPlusD allows the user to apply the above step with only one set of measurements. The software contains an internal biquad filter emulator that precisely emulates biquad filters and this saves the user from having to take the second set of verification measurements. Users that have the Behringer DSP 1124P, FBQ 2496, or DSP 2496 electronic equalizers benefit from this unit being emulated inside the software. Users of other equalizers can use a built in utility to set their equalizer to the ideal filters found in RPlusD. Although RPlusD was written to be much better than the earlier ETF software, it was renamed to reflect the new capabilities for equalization. The new software allows the user to take and store up to 32 measurements from which to set an equalizer from. Manual Equalization The problem with equalization is the certainty of which the measurements can be taken. Some averaged sets of measurements of a loudspeaker axial response averaged about an approximate ten degree window are illustrated below.

87 Average Ten Degree Window Set #1 Average Ten Degree Window Set #2

88 Average Ten Degree Window Set #3 The use of any measurement below 700 Hz would result in gross errors added because the response below 700 Hz has not been consitently characterized from this experiment. Single measurements show a far greater degree of variance.

89 Group Of Single Measurements If EQ is set to correct to one of the above averaged responses to be flat then measurements taken from most measurement locations would result in a smoother & flatter response. The above measurement contains a big dip centered at approximately 2000 Hz. This is due to the fact that the LF driver in this product works up to 2 KHz and has non ideal dispersion up to 2 KHz. Response is restored above 2 KHz by the HF unit that has nearly ideal dispersive characteristics at this relatively low frequency in its operating bandwidth. Equalization cannot correct for this and its important to understand this important point. An equalizer cannot change the sound radiating characteristics of this speaker. Before setting filters in RPlusD the user must decide which emulation to use for the internal RPlusD equalizer. In this case the FBQ 2496 was selected but if one of the Behringers is not being used the Bristow - Johnson or Classic Audio can be selected (it doesn't matter which). In the case where the EQ used is not emulated, RPlusD utility can be used to set any equalizer to the ideal RPlusD solution.

90 EQ Emulation Selection Generating a filter inside RPlusD to deal with these response anomalities is first done by going to the main menu ->Filters -> New. The following window appears. Adding A Filter Normally the 1/3 octave band EQ's will never be used because they do not provide the kind of resolution required for precise equalization. The parametric equalizer is selected with an estimated frequency of 3500 Hz. Clicking the load button creates the following filter in the main measurement window.

91 Main Window With Filter Loaded Notice that the filter created is at 3515 Hz. The FBQ 2496 equalizer does not have a 3500 Hz filter, the closest filter is 3515 Hz in the FBQ 2496 equalizer. To see the affect of the filter on any of the measurement data click the arrows shown in the Results window boxed in blue. Clicking on the arrow button in this case causes a problem because the measurement data is sampled at 48 KHz but the FBQ 2496 samples data at 96 KHz. RPlusD cannot emulate this equalizer internally with 48 KHz data and the data must be resampled. Resampling To Match Emulated EQ Once the measurement data is resampled the above arrows can be used to see the affect of the filter on the measurement data. Always save data before resampling - the software prompts this when resampling is attempted on data that is not saved. RPlusD will not save resampled data. In cases where the hardware equalizer is not being emulated in the software the data should be resampled to the hardware sampling frequency of the EQ being used so that it can be properly emulated by the internal generic biquad filter. Resampling is done from the main menu Tools-> Resampling menu choice.

92 Filter Adjustment With FBQ 2496 Emulation The task bar shown boxed in blue must first complete before the filter window opens. Each measured response is calculated with the filter then averaged to show an equalized curve beside the measured data. In the above curve they overlap because the level is set to 0 db. For the response anamality above we wish to correct it by 2 db, so the level is set to -2 db and the frequency and bandwidth fine tuned until the desired result is acheived. Each time an adjustment is made to the filter the software emulates the filter through the measurement data while the above task bar appears.

93 Fine Tuned Filter Once the filter is fine tuned the filter window can be closed. The affect of the filter can be seen on any of the measured results with the arrows in the Results window. This applies to averaged or single curve data.

94 Filter Applied To Curve 7 If the filter is applied to averaged data the software is slower because the filter must be emulated through each measurement before the EQ measurements are averaged. This is still faster than observing 32 real time analyzers react to the filter setting change but is the equivalent. Many types of EQ software do not provide actual filter emulation, they simply calculate the response of the EQ and superimpose the result onto the measured curve. This is not the same as emulation when gating is applied to a measurement impulse to provide a frequency response measurement and can lead to incorrect results. This is why RPlusD uses a full time domain emulation of the equalizer.

95 Filter Applied To Average Fractional Octave Measurement The affect of the filter on any measurement can be viewed. Sometimes the task bar may have to complete a few times to gather the necessary data.

96 Filter Applied To Averaged Psychological & Fractional Octave Data Filter Applied To Curve #2

97 It may be necessary to add another filter but the existing settings waste too much time in calculating both the PsychoAcoustic and Fractional Octave curves. The PsychoAcoustic curve computation is more time consumming so it gets deselected. Changing the window below, Filtered Curves For Both Types Shown To the settings shown below speeds up the calculations. Unchecking any of the boxes saves the software the trouble of recalculating curves during filter adjustment. Filtered Curves for Only Fractional Octave Shown Another filter was added to flatten the average response further.

98 Two Filters Applied To Average This equalizer can be saved as a file and its affect on any other measurements can be evaluated. This EQ file can also be exported as a text file that shows the exact settings applied to the physical FBQ 2496 equalizer. EQ Text File Output: *******************************************,"AcoustiSoft R+D EQ Settings","","Graphic EQ Settings","","Controls set to zero db are not listed","","","parametric EQ Settings","Freq = Hz, Level = 3 db BW:1","Freq = Hz, Level = -2 db BW:0.75","","Controls set to zero db are not listed","","","modal EQ Parametric Settings","","Controls set to zero db are not listed" ******************************************* In cases where an emulated equalizer will not be used a utility is included to match and hardware EQ settings to the ideal curves found in RPlusD. This utility is from File->New->EQ RTA Tuning and is shown below.

99 EQ RTA The EQ controls that comprise the solution should be opened in the software and appear in the right hand control list that has been illustrated in this section before this utility is accessed. This utility requires one channel of the hardware EQ to be connected across the sound card inputs and outputs. If sharp low frequency filters form part of the solution the MLS period should be increased to its maximum. The above setting would apply for mid frequency and high frequency equalization where the trailing part of the filter impulse is less than 85 ms. These settings provide a graphic refresh rate of 85 ms. The system works by displaying a frequency response in real time of the filters. The hardware EQ connected to the system is adjusted until a perfectly smooth frequency response is displayed graphically. This will normally be gently downward with frequency in shape until about 20 KHz at which point it will be down by 0.5 db - 1 db. This gentle decrease in high frequency response is due to anti - aliase filters used in hardware equalizers. RPlusD also has a curve fitting option for those who purchase the enhancement add on package. This speeds up the fitting of filters and will be illustrated in the following part of this chapter Curve Fitting The curve fitting algorithm in RPlusD is available as part of the add on enhancement package. It allows faster and more complex filter fitting. The functions of the curve fitting system are illustrated below. These functions appear beside the Results frame when the software has the add on package registered, otherwise the Filters frame does not appear. This tool is somewhat tricky to use, the explanation below supports the explanation illustrated by video that is also posted on the web site for RPlusD. This tool does require some practise for effective use but it has the advantage of being very powerful in finding duplicate room modes. Multiple room modes that appear at near the same frequency often cause large sharp peaks like the one in this example making them very difficult or sometimes impossible to fit filters to using manual methods. The curve fitting tool will be illustrated in the context of low frequency correction but could equally be used for mid & high frequency speaker equalization.

100 Averaged Low Frequency Response The curve fitting system is envoked when a part of any of the frequency response graph is grabbed with the mouse while holding the right hand mouse key. The curve fitting algorithm is best used with the frequency range of interest set by the (co - ordinate lock) on the lower left corner of the above window. In this case limits were locked at 25 Hz to 200 Hz. The above response shows an average of the curve set below

101 Low Frequency response Curve Set The above set of measurements was taken over a 8 ft X 2 ft X 3 ft listening volume above the listening area in a small room. Small room acoustics is often spoken of in terms of the modal region, diffusion region and reverberant region. In the above measurement set we see consistency in measurements up to approximately 80 Hz. Above 80 Hz measurement response diverges and can almost be considered random, this is diffusion. The demarkation point between the modal region and the diffusion region is therefore approximately 80 Hz in this particular room. The response below 80 Hz can therefore be equalized over the listening region. Human hearing has a critical band of between 100 Hz and 5000 Hz. This is the voice range and the frequency band over which hearing is most acute. The resonance at approximately 150 Hz is far more audible than any of the deviations below 100 Hz and is therefore far more important to correct. This 150 Hz resonance has a very colorful affect on the sound heard from this sytem, after a period of time one becomes accustomed to it but it would be a dangerous thing to be accustomed to when mixing music for other people to listen to on their systems. Recordists agree that the sound they use to make reference recordings should be as uncolored as possible. In fitting filters to this anomoly the graphic limits can be changed by grabbing the region of the response around this region for a close up view.

102 Close Up 150 Hz View -Ready For Correction Filters can then be generated to fit this resonance using the button. The button can be set to generate filters for each individual curve or for the single curve shown in the Results frame, in this case curve #1. Its current setting is to generate filters for each curve in the set. If the average response is shown then only one filter set will be generated, if all curves are shown then a filter set for each curve will be generated as in this example. In the above setup filters will be generated for each curve. In some cases this process may take too long and the (stop) key may be used to stop the process. There are two algorithms used, called the fast algorithm and the slow algorith. The button is set to the fast algorithm. Clicking the button causes the software to create the following filter list to appear after a short period of time in which the software executes a curve fitter on each curve.

103 Filter List Shown The P's indicate "poles" which are response peaks. The "Z" indicates a dip or a "zero". Zeros are the response nulls that occur from comb filtering and they can be considered as artifacts for the purpose of this discussion. The frequency of these artifacts is highly dependent on the measurement microphone position so they cannot be effectively equalized. Their effects must be removed from each curve with filters to allow the peaks to be found more accurately. The Q number is given as "Q: 26.1/0.47" at the top of the above list. The second 0.47 number is the estimated 60 db decay time of the resonance. The "L" indicates the amount of correction applied in db. This list provides an estimate of the exact resonance and Q for each curve. Each result is an estimate and the list can be sorted in frequency or Q using the button. The two buttons (sort by frequency) or (sort by Q) can be used to sort the list and pick out the most likely value from the range of estimates. Pressing the button provides the list shown below.

104 List Showing Curves #17,15,5,18 estimates for the mode. It is not important to pick the perfect value because this value can be fine tuned by hand using the control illustrated in the first part of the section. Before going further these values could all be polished yielding hopefully closer estimates. This is accomplished using the button and the software goes through each curve and fine tunes the filter values used to smooth the curve. The smoothing operation can sometimes cause some filters to be thrown out for each curve. In addition to this it may be that some possible filters were missed. The curve fitter can be re - executed and if the (accumulate) button is in the position, more filters can be added to each of the curves. After this operation is complete the best fit filter to the 150 Hz resonance can be chosen from the above list by pressing the (list all modes) button or the best fit filter may be chosen visually by looking through each of the curves using the arrow buttons boxed in blue in the figure below. Curve Scan Arrows (boxed in blue) Scanning through the curves using the above arrow buttons provides us with this display for each curve gives the display below for curve # 27

105 Accurate Filters Applied to 150 Hz Resonance (boxed in blue) The above result provides a smooth correction to the above resonance. A bad estimate is shown below.

106 Bad Estimate of 150 Hz Resonance The bad estimate does not provide a well smoothed area around the correction. It is chopped off with sharp edges. Based on the physical appearance of the correction curves, the filters generated for curve #27 give the best estimate. Its filters can be smoothed further. The (all curves) button is in the all curves position. It can be changed to this position (single curve) so that only the filters generated for curve 27 will be polished when the (polish) button is clicked. The filter values can be polished out and in this case the new set for further polishing. button can be used to generate a The above filter estimates included three filters for the above resonance. Polishing this result cut the number of filters required to smooth out the resonance to two as shown below.

107 Accurate Filters Applied to 150 Hz Resonance The filters for Hz and Hz will be selected for application. Clicking on one filter from the list causes the following window to appear. Clicking on "Apply" adds it to the applied filters list in the upper window. Note that any filter can be deleted from the list for before further polishing is applied as well. The software does generate duplicate values, almost all curve fitters suffer from this problem. RPlusD does an exceptional job at eliminating double values but it is not perfect. Selecting the two filters provides the display indicated below:

108 Filters Applied To Curve #5 These filters fit the problem very well because the resulting filtered curve is both smoother and flatter than the uncorrected response. It would be extremely difficult to fit filters to this complex mode by hand or by less sophisticated forms of estimation such as done from waterfall plots. This resonant peak is the result of two or possibly three near coincident room modes.

109 Damped Resonance The appearance of the peak after these filters were applied suggest a third resonance operating in this narrow region. It may be that the smoothing filters estimated for curve 27 containing 3 filters may have provided a better fit. The example discussed in the video of the curve fitting section of the software illsutrates this example with three filters applied. This video is posted on the RPlusD web site. The emulation was set for the Behringer DSP 1124 P equalizer from the (emulation) button. If the FBQ 2496 EQ was used, the emulation could be changed and the nearest fit filters that would work in the FBQ 2496 EQ would appear in the above list. The filter set can be checked to see that it both flattens and smooths every curve in the measurement set. It may be that this sharp peak could be notched out forming a narrow dip on the response in the vicinity of this narrow band region. This example was chosen to illustrate the power and flexibility of the curve fitting algorithm in RPlusD. The buttons above have function summaries below: Start Curve Fit Polish Values (fine tune values) Stop Finding Filters The option buttons below apply to the above functions

110 Slow curve fitting, Fast curve fitting New filter list, Accumulate (add) filters to existing list found Fit filter to all curves, Fit filters to the single curve shown in the Results window Measurement & Equalization Certainty In setting an EQ one is tempted to take measured results and set filters until the response is perfectly flat. The problem with this approach is that measurements can never be taken with absolute certainty even when sets of measurements are averaged. If there is only one measurement microphone location used to take the measurement then the equalizer would only work for that one microphone location. If a new measurement microphone location was chosen to take a new measurement and the same filters applied then many of the filters would be observed to be in error and effectively subtract from the filters that actually improve the response in the new microphone location. Averaged measurements can show trends from which simple filters can be applied for correction. Filters for mid and high frequency equalization can only be used to equalize loudspeakers. These can be generated from the axial measurements described in the frequency response chapter. These measurements are taken with the microphone inside a very narrow cone that resides about the axis of the speaker. Normally seated listeners hear this response as their direct sound. The experiment to measure the axial response of the loudspeaker as outlined in the frequency response chapter should be performed twice to ensure consistency between experiments. Both files can be saved and added togather to form one file from File->Add->Measurements To Existing. The experiment to measure power response over the listening area should be performed twice with the array of microphone measurement locations shifted by 1/2 of the mic distance for the second experiment. Both of these files should be checked against the EQ file generated from the axial measurements. An example of low frequency measurements is shown below with three distinct regions. These measurements were taken over an 8 ft X 3 ft X 2 ft listening volume that encloses normally placed listeners.

111 Noise, Modal & Diffusion Region The region outlined in blue shows consistent response measurements over the area. Parametric filters can be applied to flatten this response but this region is outside the critical band for human hearing and the affects of flattening this response would not be highly beneficial. The region outlined by the green square is noise. This is observable from the fact that these curves should share the consistency of the curves outlined in the blue region because the frequency is lower. The fact that these measurements are corrupted with noise would make any filter obtained from these results mostly in error. The red region is the diffusion region and each curve is different. Only very sharp resonances that appear with consistency through all measurements should be equalized. With the exception of sharp resonances, the response over this region is almost impossible to determine. Each of the experiments shown in this manual take only minutes longer to perform than single measurements. Too many measurements is never a problem, always take as many measurements as possible because it takes only a few extra minutes after the system is set up and the first measurement is taken.

112

113 Chapter 6: Digital Crossovers The crossover network is a critical part of any loudspeaker system. It prevents the low frequency signals from going to the tweeter and the high frequency signals from going to the woofer. It must ensure a proper vector addition of sound over its optimal listening area to ensure accurate sound reproduction through the critical band of human hearing of 100 Hz - 5 KHz. Often crossover points lie well within this critical band making the affect of the crossover a very audible and important aspect of speaker design. Passive crossover networks are often limited in complexity because of signal loss and expense. Complex passive filter networks that perform as well as their digital counterparts can be very complex and very expensive. The complexity often requires higher tolerance components than would be required in a simpler network. Filter complexity that can be practically implemented in passive networks can be a limiting factor in their design. Digital filtering provides exact filters of any complexity to be implemented much more easily and sometimes less expensively. High order crossover filters are easily possible. Passive Crossover: Large expensive components The filtering is done when the signal power is very high and therefore circuit board components tend to be expensive.

114 DSP Crossover: Small low cost low power components The only disadvantage of the DSP crossover is that two power amplifiers are needed. This is partially offset by the increase in power delivery to the drivers that occurs as a consequence of not having a passive filter in the signal path. The increased efficiency of this arrangement may in some cases offset the additional cost of the extra power amplifier required. Advantages of DSP Xovers (Introduction) This software was created to allow its user to optimize a crossover filter for a loudspeaker that is already built with the drivers in their final configuration on the baffle. This tool therefore requires that the user already have the loudspeaker constructed or is using a manufactered model without its factory crossover network in the signal path. The remainder of this document illustrates how to use RPlusD to optimize the settings of almost any DSP equalizer including the low cost high performance Behringer DCX The price, widespread use and availability of this processor makes it the ideal choice for users that would like to take adavantage of DSP processing in a crossover network. This is a complex task to perform, however, this method is also very powerful and its methods from which to derive the best crossover settings will far exceeds any convetional methods normally used to set up an electronic crossover network. The method is cumbersome and the reader is advised to read through this document a few times until it is fully understood before doing any actual measurement experiments. The user of this method should already be familiar with RPlusD operation and EQ emulation before this is attempted due to its complexity. The sophisticated user of this program will recognise the performance advantages in the following methods. Physical Setup & Taking Measurements Carefully placing a microphone at various locations in a three dimensional space can be a time consumming task. A large number of randomly spaced microphone positions within a narrow physical region was used instead to make the process of taking these measurements much easier and faster. Measurements required can all be taken in less time than 1 hour. This entire experiment can be performed in only a few hours. A large number of measurements ensures that the loudspeaker response is adequately characterised over its listening window for the optimal setting of the DSP filter to be performed. The multiple measurements help filter out room reflections and provide greater resolution in the measurements than would be possible when only using one or a few microphone locations for measurement.

115 There are additional measurements that can be taken to more carefully optimize the filter. These are explained at the end of this document. The following example illustrates the use of the DCX 2496 for the filtering used in a small two way high fidelity speaker system. Figure A Microphone Locations (16) For each microphone location a measurement of the woofer is taken followed by a measurement of the tweeter, this generates two measured results for each microphone location. This generates a single file containing a set of measured curves. The odd curves will be the low frequency speaker measurements and the even numbered curves will represent the HF driver response. It is important that the mic position remain stationary during each set of two measurements to avoid timing and summation errors between measurements of the LF and HF driver for each mocrophone location. The procedure required the electrical connection from the power amplifier be switched between the successive measurements. Curves 0,2,4,...etc were taken with the power amplifier connected to the woofer. Curves 1,3,5,,,,etc contain the response of the HF driver. (It is better to switch the speakers at the speaker terminal levels than at the preamp/power amp levels to avoid possible ground loops in changing connections) A two channel system can also be set up to quickly measure a system containing two drivers by using one channel from the power amp for each driver and the balance control on the preamp to change signal path from left (HF driver) to right (LF driver) and visa versa. A preamp with a mono switch or a set of "Y" jacks that enable to single output from the sound card to drive both of the two channels in a two channel system would be required when using the balance control to switch between drivers.

116 Audio Data Gathering In setting up these measurements care must be taken. The MLS (Xover Sim) test signal type must be chosen. This test signal is band limited above 300 Hz and this prevents low frequency energy from getting to lower power HF units during tests. This cutoff frequency is optional and should be made to be approximately two octaves below the crossover frequency. This will provide a result that has a very poor low frequency signal to noise in the lower frequency range which does not interest us for the purpose of setting a crossover at mid & high frequencies. This noise affects the impulse response appearance and makes it appear very noisy. That noise does not affect the measurements in the frequency range above this critical frequency but makes results below this frequency useless. This affect is illustrated and explained further in the measurements below. The noise tends to interfere with the autorange feature and means that the mic - speaker distance must be manually entered. This number does not need to be exact, here all measurements were taken with this distance set to 1 meter. This setting should not be changed although the physical mic-speaker distance may change. This method enables RPlusD to have perfect timing so that timing errors between the LF unit measurement and the HF unit measurement do not occur. Once the audio data gathering is configured and a level test is performed short cut buttons appear on the left side of the measurement result window. Use these buttons to take further measurements to prevent accidental selection of a different test signal type while clicking around the Audio Data Gathering form with the mouse. (These buttons are greyed out in the window below-see lower left) The recommendation is to use the two second test signal to keep power dissipation requirements of the speaker low and to speed up measurement taking. The short signals provide much more than adequate signal to noise in even noisy environments for accurate mid and high frequency measurements. Once this set of measurements is complete a measured result similar to what is seen in the figure below will be ready for filtering and the addition of frequency dividing filters.

117 Running The Emulator on Measurements Initial Measurement File Containing Both LF & HF response This file gets saved as mainxover.cfg from File -> Save.Parse Files to Get Separate HF & LF response files from the Crossover Window as explained in the following. These individual files will be used to set up separate EQing for the LF and HF drive units. To do this, open Xover window but do not resample when the software promts you to do so because the orginal measurements should be preserved.

118 Note Load/View Option at bottom of Window Selection of Set1 from Load/ View at the bottom of the Xover form provides a display of only the woofer response on the main display when "Load" is clicked. This file gets saved as the "LFDriver.cfg" response file. Returning to the XOver window we select Set2 and "Load" and only the HF response gets shown. The HF data file gets saved as the "HFDriver.cfg" file. There are now three files saved: The original measurement and the separate files for each of the driver responses. Its important that the sample rate never changes from the original measured rate during this process. The software will always prompt the user to resample if required - do not ever resample untill all these files are saved. Resampling raw measured data prevents the user from being able to save it. The next step will be to open the HFDriver.cfa file for the purpose of setting an EQ to flatten its response. Most users will be using the DCX 2496 equalizer/ digital speaker management system. The software emulates the filters in the DCX If this unit is to be used then the next step is to set the emulation to the DCX This can be done from the Filters->EQEmulation selection from the main menu. When the first filter is selected the software will prompt you to resample this already saved data as required for correct filter emulation. Users that do not have the DCX 2496 should select Classic or Bristow Johnson for the emulationit doesn't really matter which. These filters will be set as required and a separate utility is provided to match any EQ filters to the required software filters called the EQ RTA. It can be enabled from the File->New->EQ RTA tuning menu selection. The help file contains the instructions for this procedure to set a hardware EQ to match the internal software filters.(all biquad filters are essentially the same, one parametric filter may have a Q setting of for example Q = 3 and another parametric filter provided by another manufacterer may have the same response with a Q setting of 1.5) Other than the different ways of labeling the values of Q, all parametric equalizers do exactly the same thing. Any one equalizer can be set to duplicate another equalizer but its labeled Q may be different and this is the reason why this procedure is

119 necessary for users of something other than the behringer DCX To set the parametric corrections for the HF driver the file "HFDriver.cfa" is opened. The EQ emulation is set to DCX 2496 emulation. HF Driver Response The driver response shows a large increase in outputbelow about 300 Hz. The reason for this is that the original test signal was filtered with parts of the signal having frequencies lower than 300 removed. This rise is due to background noise in the room - there is no test signal to mask this noise below 300 Hz, but we don't care because the range around the crossover frequency is our only concern.

120 HF Driver signal to noise estimate The signal to noise ratio is affected by both the physical response of the driver and the energy in the test signal. The lesser amount of lower frequency signal energy at the microphone is lower making room noise a more dominant factor in frequencies far below the Xover frequency. Notice the single yellow curve that shows a poor signal to noise ratio. This measurement will have to be removed from the set. The corresponding woofer measurement for this result will also have to be removed. Scanning each curve gives the curve number and it can be deleted using the arrow and delete button from the above Results frame shown in the HF Driver response above. Its paired woofer measurement will have to be deleted from the "mainxover.cfg" file. The display was switched back to frequency response and the resolution of this measurement was changed to 1/20 of an octave providing a little more detail.

121 HF Driver Response: 1/20 octave. The above 1/20 oct measurement is also compared to the bode (unsmoothed) response shown below.

122 HF Driver Bode Response The short gate time in the Bode response was used to restrict room reflections from affecting measurements. The gate time was changed to evaluate the effect of room reflections on the basic identifiable shape of the result. The displayed results became grassier but the overall response did not change. This and the similarity between this response and the 1/20 octave response shows that this is a good representation of the response of this driver and baffle configuration. The curve does not show any peaks due to responance and could be flattened with some broad band filters manually. The expected crossver frequency is around 2 KHz so the behaviour of the driver between about 1 and 4 KHz is of particular concern. There are no resonances to dampen and no correction needs to be applied to the band of 1000 Hz Hz. In the interest of keeping this document short corrections will only be applied to the LF driver measurements. The response of the LF unit is then opened and shown below.

123 LF Driver 1/10 Octave Response It is quite clear that the LF driver has some odd behaviour within the crossover band of 1 KHz to 4 KHz. Some parametric filters can be applied to correct this response. All the curves can be viewed at once to see how consistent these response anomalies is between the various measurements.

124 LF Driver 1/10 Octave Response - all curves shown Consistent response anomalies above 2.5 KHz show that these response anomalies are due to the driver response and not addition of random data. Parametric EQing will now be used to correct the response. These filters can be found using the curve fitting feature of RPlusD or can be found by manually setting a parametric filter until a response peak is smoothed out.

125 LF Driver: Applied Parametric Controls The two resonances were first estimated with the curve fitting routine. This is invoked when the right hand mouse key is held down while a portion of the graph is selected. The resulting filters are generated and shown in the bottom right hand data box. Clicking on each of these allows the filters to be placed in the upper right hand data box where they are actually applied to measured data that is shown when the left or right arrow key is clicked in the Results frame in the main window shown above. Left clicking on a filter listed in the upper right data display causes the parametric control window to appear. When the filter was clicked the software noticed that the Behringer DCX 2496 emulation was selected and that this unit samples at 96 KHz and that the measured data is sampled at 48 Khz. An option box appears so that the data can be resampled to the required hardware sampling frequency of the physical equalizer being emulated. This is necessary so that the applied filter shown here behaves exactly as the DCX 2496 would and that the settings match those available on the Behringer DCX This EQ is now saved as the "LFEQ.EQ" file from File-->Save->EQ File. The EQ filters for the individual drivers have now been created - in this case correction was applied to only the LF driver. It is now time to open the original measurement file containing both HF and LF driver responses (mainxover.cfg) and apply an electronic crossover to these drivers and evaluate the resulting response set. The currently measured results and current EQ can be easily deleted from the software from File->Clear on the main menu before loading this new data. Clear the existing data and EQ file before loading the new data. After the original measurement set file is loaded the crossover modeling window is opened from Filters->XOver modelling and the process of modelling a crossover can begin.

126 The Crossover window appears when the resampling routine is finished and all data is resampled to 96 KHz as shown below. Be sure that this sample frequency shown in this window is the same as the hardware unit being emulated. (Here Fs = Hz) Crossover Window: Set 1 Set 1 applies filters to curves 0,2,4,..etc and the set 1 filter will be the low frequency driver filter. The parametric controls for this driver were loaded as the LFEQ.EQ file as shown in the text box in the above window. This EQ file should not appear in the main measurement window. Applying a check to the checkbox labeled "EQ" will enable these filters to be applied to the measured LF data- this is the only visual indicator of this filter set being used. The crossover filter type will be a low pass crossover at initially 2000 Hz.

127 Set 1 Crossover (LF Driver) Clicking the "load" button updates the Set 1 curves to show the affect of the applied filters and the selected 4th order low pass crossover filter. Notice the resonant peaks at 3603 and 4576 Hz have been filtered. Also notice that no EQ controls appear on the right side of this measurement result window. They could be added but they would be pre crossover filters- as if another EQ was in the signal path before the signal reaches the frequency divider.

128 LF Response with the Set 1 crossover filter applied Moving to the Set 2 controls we can set the HF Driver Crossover. The HF driver required no parametric filters for correction. The EQ box remains unselected.

129 Set 2 Crossover (HF Driver) The HF driver response with the crossover filter is shown below and obtained by clicking on the Set 1 button from "Load/View" in the crossover modelling window. A second order Bessel filter was chosen to combine with the driver response and obtain the forth order response. The driver itself forms a second order high pass filter.

130 HF Response with the Set 2 crossover filter applied The final response can be viewed from clicking "Sum" from the "Load / View" window. This sums curve 0 with curve 1, curve 2 with curve 3, etc, generating 16 curves within the software. These curves are then averaged to produce one listening window response or the curves can be viewed individually depending on how the window is set below. This loaded measurement set behaves as if it was just 16 curves that were taken as a result from a full range speaker. All RPlusD functions are available.

131 Driver Sum in Crossover Region It can be seen here that the response is held to within 3 db across the band. An equivalent passive crossover would never achieve this response without using many high grade passive elements. We can view the summed response from this window with the same degree of flexibility that we can look at any other standard file and all of RPlusD standard features are available to further work with this listening window averaged response. Additional full range equalization can be added. The DCX 2496 provides pre filter parametric controls and these can be emulated here by selecting an additional parametric control from the Filters menu (Filters-> New Filter).

132 Signal To Noise Ratio The summed impulses are shown below. Even the signal to noise ratio that would be obtained if the speaker was measured with the DCX 2496 in place can be shown. It is of course limited to above 300 Hz because the test signal used had very low signal energy below 300 Hz. The software contains 16 curves as if the DCX 2496 was already in place and the speaker was measured with 16 different mic locations within its listening window using a test signal that only had energy above 300 Hz.

133 Impulse Response of summed output This noise does not affect the measured frequency response at the range of interest between about 1 Khz and 4 KHz. Also the impulses are not lined up with t = 0. The difference is the actual physical mic - speaker distance and the user input of 1 meter. These kinds of misalignments are not consequential when using the fractional octave display. If the Bode response is viewed gate times should be viewed in the impulse response display to ensure that the gate time captures enough of all the trailing response of all the impulses, ie, if one impulse was delayed by 2ms then one would be careful as to only use gate times greater than 2 ms. Individual curves can also be viewed as with normal room measurements taken independently of this feature. The 1/10 octave and bode response of the summed loudspeaker is shown for mic location 11 below.

134 1/10 Octave and Bode Summed Response at a Single Microphone Location The process of averaging many curves filters out parasitic affects of the room.

135 Averaged Full Range Response These results are very preliminary results and more experimentation with the crossover frequencies and types would be appropriate. This existing response is already exceptional, holding the entire response within a 3 db window across the range. Additional Notes To extend the accuracy and ultimate quality of this filter another set of measurements could be taken with the imaginary cone that borders the mic locations could be shifted by 30 degrees. This would provide a set of off axis measurements from which to judge the crossover filter after it is initially optimized from the listening window measurements. This set of measurements would be loaded and the same XVR file used above can be loaded to see this XOver affect on the off axis response of the speaker. This is useful to see the off axis response of an LF unit that may have some beaming at higher frequencies in its range. The Behringer FBQ 2496 and the DSP 1124 equalizers are perfectly emulated in RPlusD. The emulation of the DCX 2496 should be considered accurate for filter settings and Q's of less than +/- 6 db and Q < 5. The reason for this is that the exact frequency list in the DCX 2496 is not yet known precisely. The discrete frequency values of the filters are also spaced to far apart for exact correction using high Q's unless the frequency of correction is close enough to one of the DCX 2496 discrete frequency values - this is a non ideal design feature of the actual unit but does not cause real world problems. For out of passband resonances the frequency error will normally be inconsequential. The discrete frequency values for the actual crossover frequencies is not included in the actual Xover model. The low Q filters are unaffected by the slight differences in frequency that the simulation is carried out using and the array of discrete frequencies avaliable of the hardware

136 unit. A crossover frequency change from 2000 Hz to 2100 Hz is not going to effect performance in any significant way. The exact frequency list for the Behringer DCX 2496 will also be included in this soon to be released upgrade.

137 Chapter 7: Measuring Distortion Distortion measurements can be very misleading and they are only recommended for experienced users that understand its causes, effects and and audibility. They are also complex. This section of the manual was written as a "textbook refresher" on distortion and instruction for RPlusD for people who already understand it. This part of the manual may be complicated for people unfamiliar with the topic. Distortion takes many forms, radio interferance is often called distortion, someone yelling so loud that they can be heard above the music at a concert could be called distortion. The distortion we refer to in audio systems is caused by something called non linear behaviour. This non linear behaviour is caused when the sytem no longer behaves ideally, that is to follow the principle of superposition as outlined in the next topic. Distortion itself is difficult to characterize but standard types of measurements have been used to benchmark non linear systems for comparison. Conventional measurements have little use other than for this benchmark. Non linear distortion itself as it occurs in loudspeakers is practically inaudible until the component is overloaded. Overloaded components generate very audible distortion and measuring distortion can be used to check this overload level. Generally distortion will rise slowly with output until overload is reached at which point distortion increases rapidly with increasing SPL. Dynamic Compression is another method of measuring the output capabilities of a loudspeaker and will be covered near the end of this manual and has a more practical application. The first part begins with conventional distortion measurements in RPlusD. Conventional Non Linear Distortion Ideal audio components and systems that follow the principle of superposition are perfectly linear systems. An example of a perfectly linear system would be a woofer with a cone excursion that is always directly proportional to the constant frequency voltage applied. If we apply 2 volts a peak cone excursion is observed, if the voltage is doubled, the observed cone excursion is precisely doubled. The cone excursion remains exactly proportional to voltage, the system is perfectly linear and operates by the principle of superposition. No distortion frequencies are generated. If the cone displacement is not directly proportional to voltage applied then the system is non ideal. It does not operate by the principle of superposition, is not linear and therefore does generate distortion product frequencies not present at the input at the output. A perfectly linear system behaves according to the principle of superposition and only input frequencies are present at the output.

138 The kind of distortion that gets generated when a system is non linear is called modulation distortion. Harmonic distortion (HD) is really just a subset of modulation distortion. Modulation distortion includes both IM (Inter-Modulation) and HD(Harmonic Distortion) distortion products. Lets first consider modulation distortion with one single tone of excitation. If a constant 60 Hz tone is fed into a non linear system then distortion components at the output will result from the 60 Hz tone modulating the 60 Hz tone (itself). Distortion components are as follows: 1st product = 60 Hz + 1* 60 Hz = 120 Hz 2nd product = 60 Hz + 2*60 Hz = 180 Hz 3rd product = 60 Hz + 3* 60 Hz = 240 Hz In the case of a single tone, only multiples of the applied frequency are added as distortion products. These distortion products are also called harmonics because they are an integer multiplier of the fundamental frequency applied. A musical instrument sound is largely determined by its harmonic nature. You can strike the same fundamental note on two different instruments and hear a distinctively different sound. The different instruments have a different harmonic nature. Musical instruments produce harmonics and therefore harmonic distortion is not very audible in music. It is mostly co-incident with the natural overtones generated with the musical instruments. The more audible form of distortion occurs when two different frequencies are intermodulated through a non linear system. Music consists of many different frequencies. If a signal with frequency 60 Hz is applied at the same time as an 80 Hz signal, then an ideal system will exibit behaviour according to the principle of superposition and only have 60 Hz and 80 Hz signals at the output. A non ideal system will have distortion products that are a result of the tones modulating themselves, harmonic distortion frequency components are as follows: 60 Hz (fundamental) 120 Hz (2nd harmonic)

139 180 Hz (3rd harmonic) 80 Hz (fundamental) 160 Hz (2nd harmonic) 240 Hz (3rd harmonic) These above distortion frequencies are called harmonics and are listed up to third order harmonic distortion here. These are the distortion products that you don't typically hear in typical music, because they are masked by the natural harmonically related overtones of the instruments. The above distortion set is not included in what is conventionally called intermodulation distortion. Intermodulation distortion products are the distortion products that are not harmonically related to the excitation tones (not multiples of the excitation tones). The intermodulation distortion measurements are much more important in the context of high fidelity. Intermodulation products with two tones of excitation can be calculated as shown below. First order sideband products from a 60 Hz and 80 Hz signal are as follows: 60 Hz + 80 Hz = 140 Hz 60 Hz - 80 Hz = - 20 Hz 80 Hz + 60 Hz = 140 hz 80 Hz - 60 Hz = 20 Hz. Second order sidebands are: 60 Hz + 2*80 Hz = 220 Hz 60 Hz - 2* 80 Hz = 100 Hz 80 Hz + 2*60 Hz = 200 Hz 80 Hz - 2*60 Hz = -40 Hz. (a negative frequency is equal of the positive frequency with a phase inversion) Higher order sidebands follow in frequency as N, sideband (N) = tone1 + N*tone2 sideband (N) = tone1 - N*tone2 sideband (N) = tone2 + N*tone1 sideband(n) = tone2 - N*tone1 These IM sidebands are measured separately from the harmonically related products (HD) because they affect system distortion audibility to a greater extent. The two tone measurement allows higher order distortion products such as 3rd, 4th... 8th and 9th to be evaluated because there are fewer actual distortion products to evaluate with only two tones. The measurement can be completed within a short period of time because of the low density of distortion products. Practical Audio Components Real loudspeakers are not perfectly suspended and tend to have less than proportional cone excursion at high voltages than the cone excursion at low voltages would indicate for an ideal

140 system. It is this non ideal or non proportional behaviour gives rise to distortion. Loudspeakers have a suspension that slightly increases its tension as the cone is moved from its mechanical position of rest. In addition to this, small variations in magnetic flux density in the voice coil as the coil moves from its centered position in the gap contribute to non linear behaviour. The mechanical characteristics of the speaker are slowly change with cone movement and therefore result in mainly low order distortion products. Loudspeakers are usually tested with the order set to 2 when the two tone measurement is used. Amplifiers tend to be quite linear through most of their operating range. Amplifiers produce changes in linearity close to their operating limits. Electric inductors behave similarly to amplifiers near their operation limits when the flux limit of the iron core is reached. The rapid change in linearity near operating limits gives rise to high order distortion products and are the reason for RPlusD being capable of measuring up to the 9th sideband and 10th harmonic using the two tone measurement. Two Tone Measurements To illustrate the prodedure for measuring the distortion of a loudspeaker, the following example is provided. Loudspeaker distortion measurements are made accurately when room acoustics cannot interfere with the measurement setup. This can be set up using close mic measurements in non lab conditions. In a close mic measurement, the microphone is placed very close to the loudspeaker ( a few inches/cm) under test before any testing takes place. This increases the direct sound SPL without increasing the room sound SPL and reduces the effect of the room on the measurement. Step 1. Open the distortion window and choose the measurement type as two tone. Enter two frequencies, 30 Hz and 60 Hz for tone 1 and tone 2. Leave the NR checkbox unchecked.

141 RPlusD Distortion Module Set the remaining values as shown above. The results will be calculated for harmonics between 20 Hz and 20,000 Hz. Step 2 Press start. An error appears telling you that one of the frequencies must be changed slightly. Change the 60 Hz to 59 Hz and press start. If 60 Hz and 30 Hz were the two excitation frequencies, then harmonic distortion products and intermodulated products would be identical in frequency at one or more distortion product frequencies. The 2nd order harmonic of 30 Hz = 60Hz, which is also an excitation frequency. The 60 Hz excitation frequency and 2nd harmonic of 30 Hz are the same. This results in an incorrect measurement. An error is generated in RPlusD if any two calculated distortion products are closer than 0.4 Hz of each other. Reduce the Order of Products or change Tone 1 or Tone 2 by a small amount to fix this error. Step 3 This step is done to ensure that the measurement does not cause the microphone upper SPL limit to be exceeded. Start the measurement so that you hear sound coming from the loudspeaker to verify connections. Place the SPL meter at the location of the test microphone which should be close to the loudspeaker driver ( few inches), but far enough away so that SPL is less than the upper distortion limit for the microphone (<90 db for most electret condensers). Select C weighting for this measurement. Verify that this measured SPL is far below your measurement microphone upper limit. Step 4

142 Place the measurement test microphone at the location where the SPL microphone was to measure the SPL close to the driver. This will provide the close mic measurement required for distortion measurement. Step 5 Click "Start Test". The distortion test tone will be heard. Take note of the SPL at 1 meter or at the listener position and type it in as User Input SPL. This will provide standard distortion measurements with the mike at standard positioning and the speaker playing at standard SPL. This measurement will take a few moments to complete after the test signal has stopped playing. Step 6 Check the NR checkbox and click Start Test again. If the distortion measurements are the same for an unchecked NR as a checked NR, the measurements are not likely affected by background noise. Each measurement must be performed twice to check for the affect of background noise on the measurement. This can be repeated with increases in speaker SPL to estimate an overload level. This method is useful for comparing drivers to be put in a loudspeaker system. Example Measurements The following measurements were made with a high quality bookshelf speaker. The distortion figures, while high, remain in the domain of lower order distortion products showing a well designed low frequency driver. The lower and upper distortion product frequencies were set to 20 Hz and 1500 Hz. This 6.5 inch low frequency driver becomes directional above 1500 Hz and listeners are not seated at its axis. Distortion products above this frequency would not likely be heard for this reason. Measurements were taken with NR checked and unchecked to verify that background noise was not affecting measurements. Measured distortion must be the same if the NR is checked or unchecked to verify that background noise is not affecting measurements. This measurement was taken to illustrate performance properties of a well designed low frequency driver in terms of distortion. Measurement 1 SPL= 84 db Tone 1 = 40 Hz Tone 2 = 63 Hz Level 1 = Level 2 Order = 2 %HD = 0.2 % %IMD = 0.5 % The low distortion figures are due to the fact that the woofer cone does not have to move in and out very far to produce these frequencies at the stated SPL. These frequencies lie within the

143 stated frequency operating limits of the loudspeaker. Lets see how the speaker behaves at frequencies lower than its stated low frequency cutoff point. Measurement 2 SPL = 80 db ***Tone 1 = 31 Hz ***Tone 2 = 20 Hz Level 1 = Level 2 order = 2 %IMD = 1.61 % %HD = 0.6 % Distortion dramatically increases at low frequencies as a general rule. This bookshelf loudspeaker has a very low distortion for its size, output frequency and SPL output. Lets see how it behaves with an increased SPL. Measurement 3 ***SPL = 86 db Tone 1 = 31 Hz Tone 2 = 20 Hz Level 1 = Level 2 Order = 2 %IMD = % %HD = % The slight increase of SPL caused a dramatic increase in distortion. A well designed woofer will have the same distortion rating with higher order distortion products evaluated showing that the cone reaches its limit gradually rather than suddenly, such as when the voice coil hits the back of the magnet structure during high excursions with lesser products. This is illustrated in the measurements below. Measurement 4 SPL= 86 db Tone 1 = 31 Hz Tone 2 = 20 Hz Level 1 = Level 2 ***Order = 7 %IMD = % %HD = % The total distortion for higher order products included in the result are not much different from those of only low order, showing that this drive unit is of high quality construction and reaches its excursion limits slowly with excursion and with a minimum high order distortion products. We may of course isolate a distortion product and re- calculate the distortion without doing repeating the test. For example, these two tones give a first order IM product at 20 Hz + 31 Hz = 51 Hz. If the Result Parameter window frequency limits are changed to 50 Hz and 52 Hz, and the analyze button is clicked, the percentage of distortion reflects the single IM product contribution at 51 Hz. The % HD value shows 0 % because no harmonic components exist in the new

144 frequency range. The distortion calculation in RPlusD generates results in percent or in db and compares the recorded distortion energy with the excitation energy to generate a distortion spec. The distortion products analyzed include all distortion products between the frequency ranges Fmin & F max in the result parameters window. Multi-Tone Measurements RPlusD uses a multitone test consisting of 16 tones that are computer optimized in frequency so that none of the more than 600 resulting IM distortion components are identical in frequency. This prevents one distortion component from being counted twice in the measurement and it keeps HD frequency components separated from IM frequency components. The multitone measurement measures the 1st and 2nd sidebands for IMD and 2nd and 3rd harmonics for HD. Multi-Tone Selection The idea behind multi tone distortion measurements is to measure the approximate distortion that is produced by music playing though the system under test. For this, a test tone that simulates music in terms of spectral content and distribution is needed. Another advantage of the multitone measurement is that it can overcome the affects of room acoustics because of the large number of excitation frequencies and products error due to room acoustics averages out. Microphone placement can be at the listening area or anywhere in various positions to verify that the measurement is accurate and independent of room acoustics and mic placement. The disadvantage of multi-tone measurements is that they only give wide band distortion readings and should only be used on full range devices. Reducing the excitation bandwidth or the products bandwidth reduces the advantages of multi-tone measurements.

145 RPlusD uses a B weighted set of 16 excitation tones to approximate real music. The measured result includes all 1st and 2nd order intermodulation products as well as all 2nd and 3rd order harmonic products between the user input frequency range Fmin - Fmax. The 16 member set of excitation frequencies is spaced approximately 2/3 octave apart (with a small random component). This sequence was optimized for spectral spacing of the distortion products so that a 10 second test signal could be used to gather the data when using sound cards with significant actual hardware sample rate from specified, ie a 48 KHz card may actually be sampling at Hz. This possible error must be accounted for when using low cost sound cards for high quality measurements. We believe that this is the practical limit of complexity for analyzing distortion components of a 16 tone signal. Attempting to test for a larger density of distortion products such as what would be required for the 3rd or 4th modulation products would require a test signal time of an impractically long time period due to the frequency resolution and spacing required. Loudspeakers are the primary source of distortion in audio systems. Fortunately the distortion characteristics of loudspeakers can be represented well with only 2nd and 3rd order harmonics and 1st and 2nd modulation products. Loudspeaker non linearities can be characterized as low order products when the loudspeaker is operating in its normal SPL range. The formula used to calculate distortion is given by: Distortion = sqrt ( s2^2 + s3^2 +s4^+...) / sqrt ( e1^2+e2^2+e3^2+...) sn = distortion product frequency level in SPL en = excitation frequency level in SPL each distortion product is calculated as Product (f) = sqrt (level(f)^2) / sqrt ( e1^2+e2^2+e3^2+...) level(f) = SPL level at specified distortion component. One the multitone system is selected in the distortion module and the system is connected the Start Test button is clicked on twice, once with the NR check box checked and once without the check box checked. Many systems will show 10 % distortion at normal listening levels. IMD And HD Measurement Accuracy NR Checkbox RPlusD includes an NR checkbox. All distortion measurements (both multi-tone and two tone) should be repeated with this checked and unchecked. Similar results between measurements indicate measurements uncorrupted by background noise. Testing Microphones for Distortion To test a microphone for both multi - tone and two - tone measurements, repeat the procedure below. Begin this experiment by selecting Two Tone type distortion measurements and setting the following values in the distortion analyzer. Tone 1 = 2000 Hz

146 Level 1 = 0 db Tone 2 = 3100 Hz Level 2 = -10 db Order = 2 Fmin = 20 Hz Fmax = Hz Step 1. Place the microphone approximately 2 foot / 60 cm away from a test loudspeaker. Adjust the SPL of the speaker so that the distortion is 0.1 % - 1 %. The microphone is too far away for an accurate distortion measurement, we just want to be sure that the louspeaker is operating well within its limits. Step 2. Move the microphone very close to the loudspeaker driver (2 inch/ 6 cm) and re measure distortion. The distortion should be very high due to the extremely high SPL's with the test mic in this close proximity to the actual speaker cone. This large amount distortion will be that produced by only the microphone. Step 3. Decrease SPL of the speaker being tested and repeat distortion measurements until the distortion measurement becomes very low. Try and find a level where the distortion measurements are about 0.1 % - 1% IMD. The loudspeaker distortion will be very low and this measurement will approximate the distortion produced by the only the microphone. Carefully experiment by changing SPL levels to find a level for which distortion begins to rise dramatically. Repeat the test with these levels but place the SPL meter at the new close mic location of the test microphone. Read the SPL during another distortion test. The microphone can safely handle SPL's up to this value without adding accessive amounts of its own distortion. Repeat this measurement with the multi tone measurement type selected. Notice the increased amount of distortion shown with the multi tone test. The two tone test only measures a few distortion components. Many distortion components are generated with music and the multitone measurement more accurately reflects the amount of distortion generated by real music. Benchmark the low distortion value for the multitone test as 1% - 3% percent rather than the 0.1 % - 1 % for the Two Tone test. Find levels where the multi tone measurement is below a few percent. Increase SPL until it rises to 5 % - 15 %. Keep the SPL limit of the microphone in mind for both types of distortion tests. Dynamic Compression The experiment below was performed with the microphone 1 meter from the speaker. The measured SPL from the measurement was 84 db at the microphone. It helps to use the 6.2 second Hybrid signal to run the SPL test to give the meter time to respond. Set the meter on slow C weighted response. Take the SPL as the peak read by the meter as the signal plays through the speaker. It is not important to get a flat axial response or to use calbration with this step. Its important that the test microphone can handle the high SPL's without distortion.

147 The hybrid signal best approximates the power distribution of music. This experiment should not be performed using sweeps. Level Setting The first step is to adjust the loudspeaker volume so that it is at a known level using an SPL meter. This example has the speaker set to 84 db SPL. Note Input Levels on Right (mic) Channel The low input for the first curve gives lots of range as level is increased. Input levels on the mic and sound card should not be changed as the SPL output of the speaker is increased from its amplifier.

148 Average Level - Determined by guessswork on the 1.0 oct Smoothed Response The measured response averages to about -32 db on the RPlusD graph. This is a fractional octave curve with 1.0 octave resolution for maximum smoothness. The reference for this measurement can be calculated because the actual SPL level of the loudspeaker is known to be 84 db. -32 db + Ref = 84 db Ref (0 db) = 84 db + 32 db = 116 db The 0 db level on the RPlusD graph corresponds to 116 db SPL on the frequency response display. The SPL levels for the various measurements can now be read from the RPlusD graph. Input levels on the microphone side should be kept quite low because volume will increased as more curves are generated for the measurement. The curves will illustrate the differences in frequency response that occur as level is increased. Taking Measurements The experiment is restarted with the test signal length set to 2 seconds- The long 6 second tests is not required to actually measure these curves. A series of measurements were performed while increasing the volume output of the loudspeaker.

149 0.1 Oct Smoothed Frequency Response (84 db db)

150 1.2 ms Bode Response for SPL (84 db db) The lowest level curve (navy blue) is at approximately -33 db. 112 db is 0 db as determined from the level experiment. This curve level is therefore 80 db SPL. The highest level curve is about -15 db which is 116 db - 15 db = 101 db. Signal To Noise Ratios The signal to noise ratios of the measurements should be verified to ensure that the measurements for the lowest SPL measurements are not corrupted by noise. The conclusion is that this loudspeaker has a virtually identical frequency response at all levels between 84 db and 101 db. Lesser products would see changes in the frequency response at higher SPL levels due to dynamic compression. The PSB model M2 exibits very good dynamic compression characteristics and can be played loudly without audible distortion. Knowledgeable readers may notice that I didn't track the input power of the loudspeaker along with the SPL levels and track this data. The reason for this is that this is a two way loudspeaker and one driver is unlikely to have both smooth and evenly distributed dynamic compression effects and be identical in this regard to the other driver. Dynamic compression in either driver appear as changes in the wide band frequency response.

151

152 Chapter 8: Measuring Distortion RPlusD has a feature that allows data to be taken with the sound card excitation source replaced by a test signal CD. This means that a direct connection between the computer and the system under test is not required. RPlusD can quickly measure impulse response with this method. This wireless method is redesigned from the system used in ETF software. This method is not perfect and results must be checked carefully each time. The method works for full range speakers. To measure a sub woofer a full range speaker must also be operated so that the software can properly align the impulse. Reverberation time cannot be measured using this technique. This utility was developed to measure car audio systems where fractional octave smoothed results suffice for resolution and a wired connection to the system under test is not possible. The feature is only practical when a wire cannot be connected from the computer to the sound card. Many users will insist on using it on home audio systems when a cable could easily be connected, it will be against our advice to do this. Our recommendation is to use the wired configuration whenever possible. See the Appendix for information regarding RPlusD Licensing. Wireless Operation The wireless operation feature enables a test CD to be used to generate the excitation signal in the sound system, making an electrical connection between the sound system and the computer unnecessary. This is a great feature for professionals because it saves time in messing around with wires to get the computer connected to the sound system for quick measurements. The wireless system also enables the software to be used to test car audio systems. This system is not ideal and will include error in the measurements and for this reason it is not ideal for beginners. Users should familiarize themselves with the pathologies operating with this system by connecting a CD player directly to the line input of the sound card and take a few measurements. Another way to determine the affect of these pathologies on actual measurements is to run multiple tests with the microphone in the same physical location and compare results between measurements to determine the affect of error on the measurements of interest. The error is practically nothing for frequency response measurements that are smoothed as in the fractional octave or psychological response display but can be relevant when looking at unsmoothed measurements. Instructions (1) Create a test signal CD from Tools -> Make Test CD (main menu). This feature is available on the demo version so that a test CD can be produced on any computer without a licensed version of the software installed. The file created will be a stereo wave file. CD burning software can then independently be used to burn the test wave to a CD.

153 Test Signal Creation (2) Go to Options-> Mixer (main menu) and select "mixer on". This will enable the software to memorize your mixer settings that you set in step (3). If the auto mixer has been used with the wired configuration it will be set correctly for this and no mixer settings need to be changed. CD Data Gathering (3) Connect the microphone to the appropriate input on the sound card and adjust the mixer of the sound card to receive input from the mic input or line input from the sound card. Press record and check the "Level [db]" box. This level should be below 0 db and above approximately - 15 db. The microphone can be connected to the left or right channel of the sound card input. The remaining channel requires no connection. Measurements will appear after the recording (5 sec) takes place. It takes the software 5 sec - 10 sec to complete the post processing calculations after the test signal recording has completed. A progress bar appears when the computer begins the calculations. On older/slower computers this calculation can take as long as 30 seconds. **** The CD player / sound card combination can be calibrated so that frequency response errors associated with these components can be compensated for. This calibration operates independently of mic calibration. To do this connect the CD player directly to left or right input and take a test as in (3). Save this file as a measurement and open it as the calibration file from this window to calibrate future measurements. This calibration measurement should be repeated a few times to verify consistency of the result. The software detects automatically which input (L or R) contains the input signal in the sound card. The software automatically determines if the signal being received is left, right or center channel from the system. Only one channel can be tested at one time. Typical Results

154 Figures 1A & 1B shows the signal to noise ratio in a typical measurement with the CD player connected directly to the input of the sound card. S/N Ratio CD Player connected directly to Sound Card. (no mike)

155 Unsmoothed FFT, CD Player connected directly to Sound Card Typical error in frequency response between two measurements is shown below.

156 1/20 Octave Response (Room Response) Unsmoothed FFT Response (Room Response) The artifacts caused by the non ideal behaviour of the CD player and sound card are shown below. These artifacts are not seen in standard wired measurements.

157 Typical Frequency Domain Artifacts Some artifacts can be visible about 250 Hz when using long gate times such as the maximum 720 ms, but this long gate time is not used when frequencies above 250 Hz are being measured. The fractional octave response and the phsychological response both smooth these artifacts so that they dissapear.

158 Another way of making the above artifacts dissapear is to reduce the gate time. All pathologies can be best viewed in the waterfall plot The wirelss method of data gathering can be more than adequate in terms of accuracy for most measurements. In the case of measuring speaker response to very high accuracy, the standard method of test using the sound card may be preferable.

159 Verfying Results Test should be repeated twice for each microphone location to verify that measurements are consistent. The test below illustrates results with the CD connected directly to the sound card. The results must be verified for the impulse starting at 0 ms. Typical Impulse Measurements The impulse response should be verified using the large room tools (Tools -> Large Room) to verify that the impulse does start at t=0

160 Large Room Tools -50 ms to 1436 ms Time View A correct measurement will show the impulse starting at very close to 0 ms in the above Large Room Tools utility. The impulse can be corrected in position using the time options tools box from the main window but this is not recommended. If the impulse does not start at t= 0 ms the measurement should be considered as in error and repeated.

161 Typical Room Result: Large Room Tools -50 ms - 56 ms

162 Typical Room Result ETC (0 to -90 db // -50 ms to 1436 ms) The above shows an ETC result in the large room tools window. The verticle scale is from - 90 db to 0 db. The horizontal time scale is from - 50 ms to 1430 ms. The noise at the tail end of the impulse is seen to be about 30 db below the peak signal. The test CD measurements include noise that is artifacts of the technique. This noise can be checked in the main window signal to noise ratio graph or from the ETC results in the Large Room tools.

AcoustiSoft RPlusD ver

AcoustiSoft RPlusD ver AcoustiSoft RPlusD ver 1.2.03 Feb 20 2007 Doug Plumb doug@etfacoustic.com http://www.etfacoustic.com/rplusdsite/index.html Software Overview RPlusD is designed to provide all necessary function to both

More information

SREV1 Sampling Guide. An Introduction to Impulse-response Sampling with the SREV1 Sampling Reverberator

SREV1 Sampling Guide. An Introduction to Impulse-response Sampling with the SREV1 Sampling Reverberator An Introduction to Impulse-response Sampling with the SREV Sampling Reverberator Contents Introduction.............................. 2 What is Sound Field Sampling?.....................................

More information

Spectrum Analyser Basics

Spectrum Analyser Basics Hands-On Learning Spectrum Analyser Basics Peter D. Hiscocks Syscomp Electronic Design Limited Email: phiscock@ee.ryerson.ca June 28, 2014 Introduction Figure 1: GUI Startup Screen In a previous exercise,

More information

!! 1 of! 21. Magico Subwoofer Setup and DSP Control Manual. Password: Fact_ory

!! 1 of! 21. Magico Subwoofer Setup and DSP Control Manual. Password: Fact_ory Magico Subwoofer Setup and DSP Control Manual Password: Fact_ory Thank you for the purchase of your new Magico S-Sub or Q-Sub. The goal of this setup manual is to first integrate the subwoofer into your

More information

Features/Specifications

Features/Specifications Introduction Thank you for purchasing the DD Audio DSI-1(Digital Signal Integrator). The DSI-1 is a feature rich audio signal processor that will allow you to precisely tune the acoustics of your car audio

More information

Dirac Live for Emotiva Automatic Room Correction

Dirac Live for Emotiva Automatic Room Correction Dirac Live for Emotiva Dirac Live for Emotiva Automatic Room Correction ENTER Differential Reference Media Controller TUNER 1 2 3 4 5 6 7 8 VOLUME ------------------------------------------ INPUTS ------------------------------------------

More information

Getting Started with the LabVIEW Sound and Vibration Toolkit

Getting Started with the LabVIEW Sound and Vibration Toolkit 1 Getting Started with the LabVIEW Sound and Vibration Toolkit This tutorial is designed to introduce you to some of the sound and vibration analysis capabilities in the industry-leading software tool

More information

NOTICE: This document is for use only at UNSW. No copies can be made of this document without the permission of the authors.

NOTICE: This document is for use only at UNSW. No copies can be made of this document without the permission of the authors. Brüel & Kjær Pulse Primer University of New South Wales School of Mechanical and Manufacturing Engineering September 2005 Prepared by Michael Skeen and Geoff Lucas NOTICE: This document is for use only

More information

Performing a Sound Level Measurement

Performing a Sound Level Measurement APPENDIX 9 Performing a Sound Level Measurement Due to the many features of the System 824 and the variety of measurements it is capable of performing, there is a great deal of instructive material in

More information

A few quick notes about the use of Spectran V2

A few quick notes about the use of Spectran V2 A few quick notes about the use of Spectran V2 The full fledged help file of Spectran is not ready yet, but many have asked for some sort of help. This document tries to explain in a quick-and-dirty way

More information

NOTICE. The information contained in this document is subject to change without notice.

NOTICE. The information contained in this document is subject to change without notice. NOTICE The information contained in this document is subject to change without notice. Toontrack Music AB makes no warranty of any kind with regard to this material, including, but not limited to, the

More information

Quick Start for TrueRTA (v3.5) on Windows XP (and earlier)

Quick Start for TrueRTA (v3.5) on Windows XP (and earlier) Skip directly to the section that covers your version of Windows (XP and earlier, Vista or Windows 7) Quick Start for TrueRTA (v3.5) on Windows XP (and earlier) Here are step-by-step instructions to get

More information

Natural Radio. News, Comments and Letters About Natural Radio January 2003 Copyright 2003 by Mark S. Karney

Natural Radio. News, Comments and Letters About Natural Radio January 2003 Copyright 2003 by Mark S. Karney Natural Radio News, Comments and Letters About Natural Radio January 2003 Copyright 2003 by Mark S. Karney Recorders for Natural Radio Signals There has been considerable discussion on the VLF_Group of

More information

Please feel free to download the Demo application software from analogarts.com to help you follow this seminar.

Please feel free to download the Demo application software from analogarts.com to help you follow this seminar. Hello, welcome to Analog Arts spectrum analyzer tutorial. Please feel free to download the Demo application software from analogarts.com to help you follow this seminar. For this presentation, we use a

More information

Diamond Cut Productions / Application Notes AN-2

Diamond Cut Productions / Application Notes AN-2 Diamond Cut Productions / Application Notes AN-2 Using DC5 or Live5 Forensics to Measure Sound Card Performance without External Test Equipment Diamond Cuts DC5 and Live5 Forensics offers a broad suite

More information

D-901 PC SOFTWARE Version 3

D-901 PC SOFTWARE Version 3 INSTRUCTION MANUAL D-901 PC SOFTWARE Version 3 Please follow the instructions in this manual to obtain the optimum results from this unit. We also recommend that you keep this manual handy for future reference.

More information

System Requirements SA0314 Spectrum analyzer:

System Requirements SA0314 Spectrum analyzer: System Requirements SA0314 Spectrum analyzer: System requirements Windows XP, 7, Vista or 8: 1 GHz or faster 32-bit or 64-bit processor 1 GB RAM 10 MB hard disk space \ 1. Getting Started Insert DVD into

More information

The BAT WAVE ANALYZER project

The BAT WAVE ANALYZER project The BAT WAVE ANALYZER project Conditions of Use The Bat Wave Analyzer program is free for personal use and can be redistributed provided it is not changed in any way, and no fee is requested. The Bat Wave

More information

Liquid Mix Plug-in. User Guide FA

Liquid Mix Plug-in. User Guide FA Liquid Mix Plug-in User Guide FA0000-01 1 1. COMPRESSOR SECTION... 3 INPUT LEVEL...3 COMPRESSOR EMULATION SELECT...3 COMPRESSOR ON...3 THRESHOLD...3 RATIO...4 COMPRESSOR GRAPH...4 GAIN REDUCTION METER...5

More information

SigPlay User s Guide

SigPlay User s Guide SigPlay User s Guide . . SigPlay32 User's Guide? Version 3.4 Copyright? 2001 TDT. All rights reserved. No part of this manual may be reproduced or transmitted in any form or by any means, electronic or

More information

PS User Guide Series Seismic-Data Display

PS User Guide Series Seismic-Data Display PS User Guide Series 2015 Seismic-Data Display Prepared By Choon B. Park, Ph.D. January 2015 Table of Contents Page 1. File 2 2. Data 2 2.1 Resample 3 3. Edit 4 3.1 Export Data 4 3.2 Cut/Append Records

More information

Linrad On-Screen Controls K1JT

Linrad On-Screen Controls K1JT Linrad On-Screen Controls K1JT Main (Startup) Menu A = Weak signal CW B = Normal CW C = Meteor scatter CW D = SSB E = FM F = AM G = QRSS CW H = TX test I = Soundcard test mode J = Analog hardware tune

More information

USER S GUIDE DSR-1 DE-ESSER. Plug-in for Mackie Digital Mixers

USER S GUIDE DSR-1 DE-ESSER. Plug-in for Mackie Digital Mixers USER S GUIDE DSR-1 DE-ESSER Plug-in for Mackie Digital Mixers Iconography This icon identifies a description of how to perform an action with the mouse. This icon identifies a description of how to perform

More information

USB AUDIO INTERFACE I T

USB AUDIO INTERFACE I T USB AUDIO INTERFACE EN DE FR ES IT JA Contents Introduction...3 Contents in this Operation Manual... 3 Features... 3 Panel Controls and Terminals (Details)...4 Rear Panel... 4 Front Panel... 6 Panel Controls

More information

USB Mini Spectrum Analyzer User Manual PC program TSA For TSA5G35 TSA4G1 TSA6G1 TSA12G5

USB Mini Spectrum Analyzer User Manual PC program TSA For TSA5G35 TSA4G1 TSA6G1 TSA12G5 USB Mini Spectrum Analyzer User Manual PC program TSA For TSA5G35 TSA4G1 TSA6G1 TSA12G5 Triarchy Technologies, Corp. Page 1 of 17 USB Mini Spectrum Analyzer User Manual Copyright Notice Copyright 2013

More information

MTL Software. Overview

MTL Software. Overview MTL Software Overview MTL Windows Control software requires a 2350 controller and together - offer a highly integrated solution to the needs of mechanical tensile, compression and fatigue testing. MTL

More information

SNG-2150C User s Guide

SNG-2150C User s Guide SNG-2150C User s Guide Avcom of Virginia SNG-2150C User s Guide 7730 Whitepine Road Revision 001 Richmond, VA 23237 USA GENERAL SAFETY If one or more components of your earth station are connected to 120

More information

White Paper Measuring and Optimizing Sound Systems: An introduction to JBL Smaart

White Paper Measuring and Optimizing Sound Systems: An introduction to JBL Smaart White Paper Measuring and Optimizing Sound Systems: An introduction to JBL Smaart by Sam Berkow & Alexander Yuill-Thornton II JBL Smaart is a general purpose acoustic measurement and sound system optimization

More information

USB Mini Spectrum Analyzer User Manual TSA Program for PC TSA4G1 TSA6G1 TSA8G1

USB Mini Spectrum Analyzer User Manual TSA Program for PC TSA4G1 TSA6G1 TSA8G1 USB Mini Spectrum Analyzer User Manual TSA Program for PC TSA4G1 TSA6G1 TSA8G1 Triarchy Technologies Corp. Page 1 of 17 USB Mini Spectrum Analyzer User Manual Copyright Notice Copyright 2013 Triarchy Technologies,

More information

Lab 1 Introduction to the Software Development Environment and Signal Sampling

Lab 1 Introduction to the Software Development Environment and Signal Sampling ECEn 487 Digital Signal Processing Laboratory Lab 1 Introduction to the Software Development Environment and Signal Sampling Due Dates This is a three week lab. All TA check off must be completed before

More information

456 SOLID STATE ANALOGUE TAPE + A80 RECORDER MODELS

456 SOLID STATE ANALOGUE TAPE + A80 RECORDER MODELS 456 SOLID STATE ANALOGUE TAPE + A80 RECORDER MODELS 456 STEREO HALF RACK 456 MONO The 456 range in essence is an All Analogue Solid State Tape Recorder the Output of which can be recorded by conventional

More information

NJU26125 Application Note PEQ Adjustment Procedure Manual New Japan Radio Co., Ltd

NJU26125 Application Note PEQ Adjustment Procedure Manual New Japan Radio Co., Ltd NJU26125 Application Note PEQ Adjustment Procedure Manual New Japan Radio Co., Ltd Version 1.00 CONTENTS 1.ABSTRACT...2 2.NJU26125 FIRMWARE BLOCK DIAGRAM...2 3.EQUIPMENT...2 4.ATTENTION...2 5.GENERAL FLOW

More information

Signal Stability Analyser

Signal Stability Analyser Signal Stability Analyser o Real Time Phase or Frequency Display o Real Time Data, Allan Variance and Phase Noise Plots o 1MHz to 65MHz medium resolution (12.5ps) o 5MHz and 10MHz high resolution (50fs)

More information

White Paper JBL s LSR Principle, RMC (Room Mode Correction) and the Monitoring Environment by John Eargle. Introduction and Background:

White Paper JBL s LSR Principle, RMC (Room Mode Correction) and the Monitoring Environment by John Eargle. Introduction and Background: White Paper JBL s LSR Principle, RMC (Room Mode Correction) and the Monitoring Environment by John Eargle Introduction and Background: Although a loudspeaker may measure flat on-axis under anechoic conditions,

More information

USB Mini Spectrum Analyzer User s Guide TSA5G35

USB Mini Spectrum Analyzer User s Guide TSA5G35 USB Mini Spectrum Analyzer User s Guide TSA5G35 Triarchy Technologies, Corp. Page 1 of 21 USB Mini Spectrum Analyzer User s Guide Copyright Notice Copyright 2011 Triarchy Technologies, Corp. All rights

More information

Lab experience 1: Introduction to LabView

Lab experience 1: Introduction to LabView Lab experience 1: Introduction to LabView LabView is software for the real-time acquisition, processing and visualization of measured data. A LabView program is called a Virtual Instrument (VI) because

More information

OPERATIONS MANUAL FOR EDISON PROFESSIONAL Professional ABS Molded Loudspeaker M4000

OPERATIONS MANUAL FOR EDISON PROFESSIONAL Professional ABS Molded Loudspeaker M4000 M4000 Introduction: Congratulations on your purchase of an M-4000 powered loudspeaker, engineered and manufactured by BriteLite Enterprises. The M-4000 includes a high-output compression driver, and 15

More information

Scan Converter Installation Guide

Scan Converter Installation Guide Scan Converter Installation Guide Software on supplied disks Please note: The software included with your scan converter is OPTIONAL. It is not needed to make the scan converter work properly. This software

More information

Topic: Instructional David G. Thomas December 23, 2015

Topic: Instructional David G. Thomas December 23, 2015 Procedure to Setup a 3ɸ Linear Motor This is a guide to configure a 3ɸ linear motor using either analog or digital encoder feedback with an Elmo Gold Line drive. Topic: Instructional David G. Thomas December

More information

Multiband Noise Reduction Component for PurePath Studio Portable Audio Devices

Multiband Noise Reduction Component for PurePath Studio Portable Audio Devices Multiband Noise Reduction Component for PurePath Studio Portable Audio Devices Audio Converters ABSTRACT This application note describes the features, operating procedures and control capabilities of a

More information

Manual of Operation for WaveNode Model WN-2m. Revision 1.0

Manual of Operation for WaveNode Model WN-2m. Revision 1.0 Manual of Operation for WaveNode Model WN-2m. Revision 1.0 TABLE OF CONTENTS 1. Description of Operation 2. Features 3. Installation and Checkout 4. Graphical Menus 5. Information for Software Expansion

More information

XB-14 Quick Operation Manual V1 23/10/2013

XB-14 Quick Operation Manual V1 23/10/2013 XB-14 Quick Operation Manual V1 23/10/2013 14. MIXER ON/OFF SWITCH 19. USB GAIN CONTROL 17. ST1 18. ST16 SELECTOR SELECTOR 7. GAIN CONTROL 6. 100Hz HIGH PASS FILTER 13. MAIN 16. GAIN 5. EQ METERS 12. PHANTOM

More information

Calibrating Measuring Microphones and Sound Sources for Acoustic Measurements with Audio Analyzer R&S UPV

Calibrating Measuring Microphones and Sound Sources for Acoustic Measurements with Audio Analyzer R&S UPV Product: R&S UPV Calibrating Measuring Microphones and Sound Sources for Acoustic Measurements with Audio Analyzer R&S UPV Application Note 1GA47_0E This application note explains how to use acoustic calibrations

More information

Panaray 802 Series III TECHNICAL DATA SHEET. loudspeaker. Key Features. Product Overview. Technical Specifications

Panaray 802 Series III TECHNICAL DATA SHEET. loudspeaker. Key Features. Product Overview. Technical Specifications Panaray 82 Series III Key Features Articulated Array design provides 12 x 1 coverage to deliver wide-range reproduction over a broad dispersion area Eight Bose 4.5" (114 mm) full-range drivers for unsurpassed

More information

THE DIGITAL DELAY ADVANTAGE A guide to using Digital Delays. Synchronize loudspeakers Eliminate comb filter distortion Align acoustic image.

THE DIGITAL DELAY ADVANTAGE A guide to using Digital Delays. Synchronize loudspeakers Eliminate comb filter distortion Align acoustic image. THE DIGITAL DELAY ADVANTAGE A guide to using Digital Delays Synchronize loudspeakers Eliminate comb filter distortion Align acoustic image Contents THE DIGITAL DELAY ADVANTAGE...1 - Why Digital Delays?...

More information

TF5 / TF3 / TF1 DIGITAL MIXING CONSOLE. TF Editor User Guide

TF5 / TF3 / TF1 DIGITAL MIXING CONSOLE. TF Editor User Guide TF5 / TF3 / TF1 DIGITAL MIXING CONSOLE EN Special notices Copyrights of the software and this document are the exclusive property of Yamaha Corporation. Copying or modifying the software or reproduction

More information

spiff manual version 1.0 oeksound spiff adaptive transient processor User Manual

spiff manual version 1.0 oeksound spiff adaptive transient processor User Manual oeksound spiff adaptive transient processor User Manual 1 of 9 Thank you for using spiff! spiff is an adaptive transient tool that cuts or boosts only the frequencies that make up the transient material,

More information

E X P E R I M E N T 1

E X P E R I M E N T 1 E X P E R I M E N T 1 Getting to Know Data Studio Produced by the Physics Staff at Collin College Copyright Collin College Physics Department. All Rights Reserved. University Physics, Exp 1: Getting to

More information

Using the BHM binaural head microphone

Using the BHM binaural head microphone 11/17 Using the binaural head microphone Introduction 1 Recording with a binaural head microphone 2 Equalization of a recording 2 Individual equalization curves 5 Using the equalization curves 5 Post-processing

More information

WAVES H-EQ HYBRID EQUALIZER USER GUIDE

WAVES H-EQ HYBRID EQUALIZER USER GUIDE WAVES H-EQ HYBRID EQUALIZER USER GUIDE TABLE OF CONTENTS CHAPTER 1 INTRODUCTION...3 1.1 WELCOME...3 1.2 PRODUCT OVERVIEW...3 1.3 CONCEPTS AND TERMINOLOGY...4 1.4 COMPONENTS...7 CHAPTER 2 QUICK START GUIDE...8

More information

MIE 402: WORKSHOP ON DATA ACQUISITION AND SIGNAL PROCESSING Spring 2003

MIE 402: WORKSHOP ON DATA ACQUISITION AND SIGNAL PROCESSING Spring 2003 MIE 402: WORKSHOP ON DATA ACQUISITION AND SIGNAL PROCESSING Spring 2003 OBJECTIVE To become familiar with state-of-the-art digital data acquisition hardware and software. To explore common data acquisition

More information

MDynamicsMB. Overview. Easy screen vs. Edit screen

MDynamicsMB. Overview. Easy screen vs. Edit screen MDynamicsMB Overview MDynamicsMB is an advanced multiband dynamic processor with clear sound designed for mastering, however its high performance and zero latency, makes it ideal for any task. It features

More information

MaxView Cinema Kit Quick Install Guide

MaxView Cinema Kit Quick Install Guide SYSTEM SETUP The MaxView will work at any of the following display settings: INSTALLATION MaxView Cinema Kit Quick Install Guide Step 1 - Turn off your computer. Disconnect your monitor s VGA cable from

More information

BER MEASUREMENT IN THE NOISY CHANNEL

BER MEASUREMENT IN THE NOISY CHANNEL BER MEASUREMENT IN THE NOISY CHANNEL PREPARATION... 2 overview... 2 the basic system... 3 a more detailed description... 4 theoretical predictions... 5 EXPERIMENT... 6 the ERROR COUNTING UTILITIES module...

More information

USER GUIDE FOR NETmc MARINE X-Ops

USER GUIDE FOR NETmc MARINE X-Ops USER GUIDE FOR NETmc MARINE X-Ops Rev.2.2 Firmware v.1.2.8 X-Ops, 1.2.9 Pipeline September 2013 NETmc Marine X-Ops Manual-Rev.2.2 September 2013 1 of 17 Contents 1. Introduction... 3 1.1 Multi-channel

More information

4 MHz Lock-In Amplifier

4 MHz Lock-In Amplifier 4 MHz Lock-In Amplifier SR865A 4 MHz dual phase lock-in amplifier SR865A 4 MHz Lock-In Amplifier 1 mhz to 4 MHz frequency range Low-noise current and voltage inputs Touchscreen data display - large numeric

More information

Introduction To LabVIEW and the DSP Board

Introduction To LabVIEW and the DSP Board EE-289, DIGITAL SIGNAL PROCESSING LAB November 2005 Introduction To LabVIEW and the DSP Board 1 Overview The purpose of this lab is to familiarize you with the DSP development system by looking at sampling,

More information

J R Sky, Inc. tel: fax:

J R Sky, Inc.  tel: fax: STEREO OPTICAL RECORDING SYSTEM N UOPTIX STEREO OPTICAL RECORDING MONITOR LEFT SYSTEM MODE PREVIEW RECORD BIAS RECORD REV SETUP TEST RIGHT INPUT SETUP INPUT BIAS SETUP BIAS INPUT STEREO AUX MONO DIRECT

More information

Configuring the Stack ST8961 VS Module when used in conjunction with a Stack ST81xx series display.

Configuring the Stack ST8961 VS Module when used in conjunction with a Stack ST81xx series display. Configuring the Stack ST8961 VS Module when used in conjunction with a Stack ST81xx series display. Your Stack ST8961 VS module allows you to synchronize, overlay, and record data available on your Stack

More information

MAutoPitch. Presets button. Left arrow button. Right arrow button. Randomize button. Save button. Panic button. Settings button

MAutoPitch. Presets button. Left arrow button. Right arrow button. Randomize button. Save button. Panic button. Settings button MAutoPitch Presets button Presets button shows a window with all available presets. A preset can be loaded from the preset window by double-clicking on it, using the arrow buttons or by using a combination

More information

Fluke 190-Series II Firmware Upgrade V11.44

Fluke 190-Series II Firmware Upgrade V11.44 Fluke 190-Series II Firmware Upgrade V11.44 Requirements 1. Fluke 190- Series II ScopeMeter with firmware prior to V11.44 2. Supported models are: 190-102, 190-104, 190-062, 190-202, 190-204, 190-502,

More information

V4.7 Software Quick Start Guide

V4.7 Software Quick Start Guide V4.7 Software Quick Start Guide INTRODUCTION TO V4.7 The 4.7 software update for the Vi Series includes a major update to the functionality of the Vi4 console in particular, bringing a new level of power

More information

Version 1.10 CRANE SONG LTD East 5th Street Superior, WI USA tel: fax:

Version 1.10 CRANE SONG LTD East 5th Street Superior, WI USA tel: fax: -192 HARMONICALLY ENHANCED DIGITAL DEVICE OPERATOR'S MANUAL Version 1.10 CRANE SONG LTD. 2117 East 5th Street Superior, WI 54880 USA tel: 715-398-3627 fax: 715-398-3279 www.cranesong.com 2000 Crane Song,LTD.

More information

WAVES Cobalt Saphira. User Guide

WAVES Cobalt Saphira. User Guide WAVES Cobalt Saphira TABLE OF CONTENTS Chapter 1 Introduction... 3 1.1 Welcome... 3 1.2 Product Overview... 3 1.3 Components... 5 Chapter 2 Quick Start Guide... 6 Chapter 3 Interface and Controls... 7

More information

USER GUIDE MULTI-CHANNEL STEREO MIXER S MIX 4M2S UD

USER GUIDE MULTI-CHANNEL STEREO MIXER S MIX 4M2S UD USER GUIDE MULTI-CHANNEL STEREO MIXER S MIX 4M2S UD FEATURES 4 mono input channels and 2 stereo input channels. Frequency EQ on each input channel. Low-noise mic pre-amp on microphone inputs. Low cut filter

More information

UB22z Specifications. 2-WAY COMPACT FULL-RANGE See NOTES TABULAR DATA for details CONFIGURATION Subsystem DESCRIPTION

UB22z Specifications. 2-WAY COMPACT FULL-RANGE See NOTES TABULAR DATA for details CONFIGURATION Subsystem DESCRIPTION DESCRIPTION Ultra-compact 2-way system Wide projection pattern LF on angled baffles to maintain a wide upper/midrange beamwidth High output, high definition sound DESCRIPTION The UB22z is engineered for

More information

Audyssey MultEQ Pro User Guide

Audyssey MultEQ Pro User Guide Audyssey MultEQ Pro User Guide for the NAD T 775, T 785, and T 175 Audyssey Laboratories, Inc, 350 South Figueroa Street, Suite 233, Los Angeles, CA 90071 www.audyssey.com Table of Contents Introduction...

More information

USER MANUAL FOR DDT 2D. Introduction. Installation. Getting Started. Danley Design Tool 2d. Welcome to the Danley Design Tool 2D program.

USER MANUAL FOR DDT 2D. Introduction. Installation. Getting Started. Danley Design Tool 2d. Welcome to the Danley Design Tool 2D program. USER MANUAL FOR DDT 2D ( VERSION 1.8) Welcome to the Danley Design Tool 2D program. Introduction DDT2D is a very powerful tool that lets the user visualize how sound propagates from loudspeakers, including

More information

Manual for the sound card oscilloscope V1.41 C. Zeitnitz english translation by P. van Gemmeren, K. Grady and C. Zeitnitz

Manual for the sound card oscilloscope V1.41 C. Zeitnitz english translation by P. van Gemmeren, K. Grady and C. Zeitnitz Manual for the sound card oscilloscope V1.41 C. Zeitnitz english translation by P. van Gemmeren, K. Grady and C. Zeitnitz C. Zeitnitz 12/2012 This Software and all previous versions are NO Freeware! The

More information

VTAPE. The Analog Tape Suite. Operation manual. VirSyn Software Synthesizer Harry Gohs

VTAPE. The Analog Tape Suite. Operation manual. VirSyn Software Synthesizer Harry Gohs VTAPE The Analog Tape Suite Operation manual VirSyn Software Synthesizer Harry Gohs Copyright 2007 VirSyn Software Synthesizer. All rights reserved. The information in this document is subject to change

More information

Using Extra Loudspeakers and Sound Reinforcement

Using Extra Loudspeakers and Sound Reinforcement 1 SX80, Codec Pro A guide to providing a better auditory experience Produced: December 2018 for CE9.6 2 Contents What s in this guide Contents Introduction...3 Codec SX80: Use with Extra Loudspeakers (I)...4

More information

CDM10: Channel USB Mixer. Item ref: UK User Manual

CDM10: Channel USB Mixer. Item ref: UK User Manual CDM10:4 19 4 Channel USB Mixer Item ref: 171.135UK User Manual Caution: Please read this manual carefully before operating Damage caused by misuse is not covered by the warranty Introduction Thank you

More information

Installer s Manual. DEQX- Cal Version 2.90

Installer s Manual. DEQX- Cal Version 2.90 Installer s Manual DEQX- Cal Version 2.90 Table of Contents Copyright... 4 License... 4 Welcome to DEQX Calibration... 6 Features:... 7 How to use this manual... 8 Introduction... 9 Quick Start... 10 Installing

More information

Audyssey MultEQ Pro User Guide

Audyssey MultEQ Pro User Guide Audyssey MultEQ Pro User Guide for the Denon AVR-5805CI Audyssey Laboratories, Inc, 350 South Figueroa Street, Suite 233, Los Angeles, CA 90071 www.audyssey.com Table of Contents Introduction... 1 MultEQ

More information

Log-detector. Sweeper setup using oscilloscope as XY display

Log-detector. Sweeper setup using oscilloscope as XY display 2002/9/4 Version 1.2 XYdisp user manual. 1. Introduction. The XYdisp program is a tool for using an old DOS PC or laptop as XY display to show response curves measured by a sweeper log-detector combination.

More information

REPORT DOCUMENTATION PAGE

REPORT DOCUMENTATION PAGE REPORT DOCUMENTATION PAGE Form Approved OMB No. 0704-0188 Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instructions,

More information

M R X

M R X MRX 710 / 510 / 310 AWARD-WINNING A/V RECEIVERS ANTHEM ROOM CORRECTION SYSTEM A fantastic-sounding product well worthy of a serious audition don t delay. ~ What Hi Fi on the MRX 700 MRX 710 shown Among

More information

Using the new psychoacoustic tonality analyses Tonality (Hearing Model) 1

Using the new psychoacoustic tonality analyses Tonality (Hearing Model) 1 02/18 Using the new psychoacoustic tonality analyses 1 As of ArtemiS SUITE 9.2, a very important new fully psychoacoustic approach to the measurement of tonalities is now available., based on the Hearing

More information

RoomMatch RM and RM TECHNICAL DATA SHEET. asymmetrical array modules. Key Features. Product Overview. Technical Specifications

RoomMatch RM and RM TECHNICAL DATA SHEET. asymmetrical array modules. Key Features. Product Overview. Technical Specifications RoomMatch RM281 and RM281 Key Features All the benefits of the original 2 RoomMatch array module loudspeakers - Concert-quality sound quality for live music and outstanding spoken-word clarity with industry-leading

More information

RoomMatch Utility RMU208 TECHNICAL DATA SHEET. small-format foreground/fill loudspeaker. Key Features. Technical Specifications

RoomMatch Utility RMU208 TECHNICAL DATA SHEET. small-format foreground/fill loudspeaker. Key Features. Technical Specifications RoomMatch Utility RMU28 Key Features Award-winning RoomMatch sound now in smaller 2-way point-source designs for high-level foreground music, under-balcony, zone-fill and vocalrange floor monitor applications

More information

Field Test 2. Installation and operation manual OPDAQ Installation and operation manual

Field Test 2. Installation and operation manual OPDAQ Installation and operation manual Field Test 2 Installation and operation manual OPDAQ 17.08.25 Installation and operation manual January 2016 How to get copies of OpDAQ technical publications: 53, St-Germain Ouest Rimouski, Québec Canada

More information

Measuring Delays. A Step by Step instruction. (c) 2017 by Thomas Neumann

Measuring Delays. A Step by Step instruction. (c) 2017 by Thomas Neumann Measuring Delays A Step by Step instruction A few things to mention This document shows a simple step by step way to determine the distance in time between to audio sources (reproducing Mid High frequency

More information

RT-DRIVE DLM808 DIGITAL PROCESSOR AUDIO MATRIX PROCESSOR

RT-DRIVE DLM808 DIGITAL PROCESSOR AUDIO MATRIX PROCESSOR RT-DRIVE DLM808 DIGITAL PROCESSOR AUDIO MATRIX PROCESSOR 2 1. Introduction 2. Features 3. Usefull Data 4. Function Buttons and LED Indicators 5. Rear Panel 6. DSP Control 1. Configuration of IP Address

More information

The 01X Configuration Guide

The 01X Configuration Guide The 01X Configuration Guide A Very Brief Introduction Welcome to the world of learning! Like many of you, I have spent countless hours reading and re-reading the 01x and the Cubase SX owner's manuals,

More information

Analyzing and Saving a Signal

Analyzing and Saving a Signal Analyzing and Saving a Signal Approximate Time You can complete this exercise in approximately 45 minutes. Background LabVIEW includes a set of Express VIs that help you analyze signals. This chapter teaches

More information

LUCAS NANO 600 Series

LUCAS NANO 600 Series LUCAS NANO 600 Series PROFESSIONAL SOUND TO GO > LUCAS NANO 602/600/608i 1 LUCAS NANO 600 Series PROFESSIONAL SOUND TO GO > The LUCAS NANO 600 series packs professional audio performance into the most

More information

soothe audio processor Manual and FAQ

soothe audio processor Manual and FAQ soothe audio processor Manual and FAQ Thank you for using soothe! soothe is a spectral processor for suppressing resonances in the mid and high frequencies. It works by automatically detecting the resonances

More information

CHAPTER 3 AUDIO MIXER DIGITAL AUDIO PRODUCTION [IP3038PA]

CHAPTER 3 AUDIO MIXER DIGITAL AUDIO PRODUCTION [IP3038PA] CHAPTER 3 AUDIO MIXER DIGITAL AUDIO PRODUCTION [IP3038PA] Learning Objectives By the end of this chapter, students should be able to: 1 State the function of the audio mixer in the sound studio. 2 Explain

More information

2 MHz Lock-In Amplifier

2 MHz Lock-In Amplifier 2 MHz Lock-In Amplifier SR865 2 MHz dual phase lock-in amplifier SR865 2 MHz Lock-In Amplifier 1 mhz to 2 MHz frequency range Dual reference mode Low-noise current and voltage inputs Touchscreen data display

More information

Transmitter Installation and Operation

Transmitter Installation and Operation Transmitter Installation and Operation Easy-to-follow instructions on how to program and use your Talking House / i A.M. Radio Transmitter Questions? Just call (616) 772-2300. Contents Quick Start... 3

More information

Using Extra Loudspeakers and Sound Reinforcement

Using Extra Loudspeakers and Sound Reinforcement 1 SX80, Codec Pro A guide to providing a better auditory experience Produced: October 2018 for CE9.5 2 Contents What s in this guide Contents Introduction...3 Codec SX80: Use with Extra Loudspeakers (I)...4

More information

MDistortionMB. Easy screen vs. Edit screen

MDistortionMB. Easy screen vs. Edit screen MDistortionMB Easy screen vs. Edit screen The plugin provides 2 user interfaces - an easy screen and an edit screen. Use the Edit button to switch between the two. By default most plugins open on the easy

More information

Scan Converter Quick Installation Guide

Scan Converter Quick Installation Guide Scan Converter Quick Installation Guide Software Note: No software is required to use your scan converter. Please complete the hardware installation and system setup before you determine the need to the

More information

Meetinghouse Webcast Setup Guide for the Europe Area Sending Site

Meetinghouse Webcast Setup Guide for the Europe Area Sending Site Meetinghouse Webcast Setup Guide for the Europe Area Sending Site Rev 03 15 Nov 2009 Introduction This guide provides helpful information and details for successfully setting up a webcast in the Europe

More information

Welch Allyn CardioPerfect Workstation Tango+ Interface Notes

Welch Allyn CardioPerfect Workstation Tango+ Interface Notes Welch Allyn CardioPerfect Workstation Tango+ Interface Notes To setup Tango+ with the CardioPerfect stress system, simply follow the directions below. 1. Verify Correct RS-232 and ECG Trigger Cables RS-232

More information

Eventide Inc. One Alsan Way Little Ferry, NJ

Eventide Inc. One Alsan Way Little Ferry, NJ Copyright 2017, Eventide Inc. P/N: 141263, Rev 5 Eventide is a registered trademark of Eventide Inc. AAX and Pro Tools are trademarks of Avid Technology. Names and logos are used with permission. Audio

More information

User s Guide Contents

User s Guide Contents User s Guide Contents Chapter 1 Introduction Video Conferencing on your PC Image and Video Capture Chapter 2 Setting Up your PC for Video Conferencing Overview How to Setup AVerMedia AVerTV Studio for

More information

UltraPioneerAVR3 HSPI User s Guide A HomeSeer HS3 plug-in to enable control of Pioneer A/V Receivers that support the IP & RS232 control system.

UltraPioneerAVR3 HSPI User s Guide A HomeSeer HS3 plug-in to enable control of Pioneer A/V Receivers that support the IP & RS232 control system. UltraPioneerAVR3 HSPI User s Guide A HomeSeer HS3 plug-in to enable control of Pioneer A/V Receivers that support the IP & RS232 control system. Copyright 2014 ultrajones@hotmail.com Revised 01/20/2014

More information

TL AUDIO M4 TUBE CONSOLE

TL AUDIO M4 TUBE CONSOLE TL AUDIO M4 TUBE CONSOLE USER MANUAL TL AUDIO M4 TUBE CONSOLE M4 INTRODUCTION... 3 M4 MIXER TECHNICAL SPECIFICATION... 4 Mic Input:... 4 Line Input:... 4 Phase Rev:... 4 High Pass Filter:... 4 Frequency

More information

Interface Practices Subcommittee SCTE STANDARD SCTE Measurement Procedure for Noise Power Ratio

Interface Practices Subcommittee SCTE STANDARD SCTE Measurement Procedure for Noise Power Ratio Interface Practices Subcommittee SCTE STANDARD SCTE 119 2018 Measurement Procedure for Noise Power Ratio NOTICE The Society of Cable Telecommunications Engineers (SCTE) / International Society of Broadband

More information