Internal Mixing of Dance Music

Transcription

1 First bachelor thesis Internal Mixing of Dance Music Completed with the aim of graduating with a Bakkalaureat (FH) in Telecommunications and Media from the St. Pölten University of Applied Sciences Telecommunications and Media degree course under the supervision of Hannes Raffaseder Completed by Sascha Galley tm St. Pölten, Signature:

2 Declaration I declare that the attached research paper is my own, original work undertaken in partial fulfilment of my degree. I have made no use of sources, materials or assistance other than those which have been openly and fully acknowledged in the text. If any part of another person s work has been quoted, this either appears in inverted commas or (if beyond a few lines) is indented. Any direct quotation or source of ideas has been identified in the text by author, date, and page number(s) immediately after such an item, and full details are provided in a reference list at the end of the text. I understand that any breach of the fair practice regulations may result in a mark of zero for this research paper and that it could also involve other repercussions. I understand also that too great a reliance on the work of others may lead to a low mark. Undersigned & Date - I -

3 Abstract This bachelor thesis provides a general overview about mixing dance music primarily on a computer with the focus on digital audio workstations and plugins. To begin with, it describes the variety of tools that are available on the market and how to use and set up the studio equipment. Furthermore, the processing of audio files is explained and the target for the final product is set. The next chapter focuses on the theory of mixing. It introduces the three dimensions of sound and how those are defined and used for the purpose of mixing. Moreover, it describes the most commonly applied filters and effects like equalizers, reverbs, and delays. Dynamics and its alteration with tools like compressors and limiters are also covered in this part of the thesis. In the next chapter the practice and workflow of mixing from planning a soundstage to the final mix is described. It provides information about how to set up the digital audio workstation and how to adjust the mixer. Then it continues by explaining how to set the volume and panning position of each track correctly. After that it shows the procedure of editing single instruments and fitting them into the mix. Furthermore, an overview about automations is provided. The last chapter deals with professional mix characteristics, final steps for the mixdown, and the preparation for the mastering process. - II -

4 Table of Content 1. PREPARING THE MIX STUDIO EQUIPMENT Digital Audio Workstation (DAW) Monitoring Headphones PROCESSING AUDIO FILES Technical Conditions DC Offset Normalization KNOW YOUR GOALS 6 2. MIXING THEORY THE THREE DIMENSIONS Layering Front-to-Back Perspective Panorama Horizontal Perspective Frequency Vertical Perspective FILTERS AND EFFECTS Equalizer Reverb Delay Other Effects 19 - III -

5 2.3 DYNAMICS Compressor Side-Chaining and Ducking Limiter MIXING PRACTICE AND WORKFLOW CREATING A SOUNDSTAGE INITIATE THE WORK SPACE Arrangement Sub Groups Mixer and Signal Routing SETTING VOLUMES PANNING SINGLE INSTRUMENT EDITING Kick / Bass Drum Snare Hi-hats and Cymbals Bass Vocals Synthesizers / Pianos / Guitars Atmos, Pads & Sound FX AUTOMATIONS THE FINAL MIX WHERE TO GO 42 - IV -

6 4.2 FINAL STEPS OUTLOOK ON MASTERING CONCLUSIO 45 LIST OF FIGURES 46 LIST OF LITERATURE 47 - V -

7 1. Preparing the Mix Before starting to mix it is essential to prepare a solid and well arranged setup. High quality audio tracks, excellent studio components, and a proper starting point are crucial for professional mixes and will simplify matters. Nevertheless, the art of mixing is not only an issue of technical know-how, but also a very creative and inventive process. Mixing engineers can only reach their goals with spatial imagination, musically trained ears, and other artistic skills. Furthermore, an ample knowledge of ones software and hardware is needed to develop an appropriate and individual workflow. 1.1 Studio Equipment Due to the rapidly growing number of home studios and the vast spread of personal computers, high end components have become rather affordable. The big amount of available audio devices offers a huge range of possibilities for your personal setup, depending on style, genre, affectations, and independent means. However, a well- planned acquisition of software and hardware is indispensable for a reliable working environment, especially when it comes to mixing Digital Audio Workstation (DAW) To hear real time changes in your mix accurately you need a system with low latency. Only fast computers having the right hardware configuration offer a failure-free workflow. Hence you have to know your system like the back of your hand, especially the characteristics of your soundcard and the functioning of your software. There are endless solutions for good setups and it is your decision on 1

8 how much you will invest in it. 1 Keep in mind that electronic dance music owes a big part of its personality to the creative use of low-tech, home studio gear. That dirty sound is part of its heritage. 2 Beneath a proper DAW you also need plug-ins or external devices to deliver professional mix-downs. Most common audio programs are provided with an interface called Virtual Studio Technology (VST), invented by Steinberg, which communicates with other software. There are a lot of plug-ins available on the market, either as VST instruments or as VST effects. 3 Since these plug-ins are certainly sufficient in most cases you don t have to worry about expensive and huge external audio hardware Monitoring To get a precise image of the current work progress it is very important to hear veridical what you are doing. All hi-fi speakers deliberately employ frequencyspecific boosts and cuts throughout the sonic range to make the sound produced by them appear full and rounded. As a result, if you rely entirely on these while mixing you ll produce mixes that will not translate well on any other hi-fi system. 4 Thus it is far better to use flat and natural loudspeakers, so-called studio monitors, with an excellent frequency response. When you have chosen the right monitors for your requirements you have to consider some setup guidelines, because inaccurate positioning can affect the 1 Cf. Hawkins (2004), p See Hawkins (2004), p Cf. Wikipedia (2008), 23th July See Snowman (2004), p

9 perception of frequencies and, moreover, the stereophonic sound. First of all, the distance between your monitors shall neither be too short nor too long for a clear spatial definition. A rule of thumb is that the speakers should be as far apart as the distance from the listening position. 5 Furthermore, they have to be angled properly at your mixing position. At best you get an equilateral triangle, meaning that both monitors need to have an angle of about 60, like shown in figure 1. Some mixing engineers, though, prefer to set the intersection point of their speakers about three feet behind the mixing position to dispose of the mostly hyped sounds of modern speakers. 6 Figure 1: Equilateral triangle of monitors (Source: Owsinski (2006), p.73.) Headphones Mixing with headphones often sounds better and more transparent than with normal boxes. This is due to the fact that the signal attains directly to the ear without any reflections and reverberation. So instruments are perceived louder and more perspicuous than they actually are and thus it gets impossible to set 5 See Owsinski (2006), p Cf. Owsinski (2006), p.72ff. 3

10 volumes correctly. Moreover, there are still no headphones available which have a linear frequency response. While medium notes are sensed better, deep basses are neither heard nor felt with headphones, because the body does not experience any vibrations in the air. On the other hand headphones can turn out to be a very useful tool, though, as signals are perceived more intensive and clearer and therefore headphones can be used as a magnifying glass. Unwanted noises and clicks can be located and eliminated much better. But since the overall picture is pretty distorted, you should change again to your monitors immediately Processing Audio Files A good mix starts with good material. Recording errors, poor microphones, generally lousy quality, or badly arranged MIDI programming will lead to a miserable mix. It is vital at this stage that you don t settle for anything less than the very best it may be an overused analogy but you wouldn t expect to make a great tasting cake if some of the ingredients were out of date. 8 So it is necessary to listen to each track once more to correct errors early enough. Nonapplicable instruments can be rejected or replaced with others that are more suitable now. 9 7 Cf. Eisner (2006), p.57f. 8 See Snowman (2004), p Cf. Snowman (2004), p

11 1.2.1 Technical Conditions To archive the current state of production all MIDI files are exported to WAV files with a minimum depth of 24-bit to obtain high accuracy. Yet it is far better to use 32-bit floating-point audio tracks to increase dynamics and to prevent clipping and rounding errors. While you may not recognize a big difference with one single file, it becomes impossible to obtain a transparent mix in more or less comprehensive projects. Furthermore, the co-operation between data, software and hardware is facilitated since most popular digital audio workstations and CPUs work internally with 32-bit DC Offset The DC offset can be imagined visually as a shift from zero within the waveform and can occur in all recorded audio signals. It can be heard as a muddy effect in the bass range, as the subwoofer uses a lot of unnecessary energy for those events, so the whole mix will sound scruffy. Moreover, zero crossings are harder to find and it will get impossible to accomplish clean cuts. Each good DAW has a built-in feature to remove the DC offset very quickly to avoid problems from the beginning on Normalization Normalization is the reinforcement of a signal so that the highest occurring amplitude in the signal is lifted to 0 dbfs (decibel full scale is the maximum available level in digital workstations) and thus the entire signal and at the same time undesirable noise gets louder. Normalization does not reduce the dynamic 10 Cf. Tischmeyer (2006), p.37ff. 11 Cf. Tischmeyer (2006), 289ff. 5

12 but leaves little or no headroom and furthermore it can lead to inter sample peaks. Of course, it always depends on the material you are working with, but in most cases normalization does not make mixing easier Know Your Goals To begin with imagine how the final mix should sound like and think of how the instruments can be positioned. If it is not your own song you have to get familiar with it. Try to figure out the direction of the song while finding the most important constituent of the track so that you can highlight it. 13 Your goal should be a clear, warm and powerful mix in which every element can be heard perspicuously. Therefore you need to find sufficient space for every event and instrument except for pads and ambience elements which only create background warmth. When there isn t enough place for a certain element, think about if you really need it or if it can be renounced. Keep in mind, that sometimes less is more! 14 Above all, dance music is ultimately about groove, vibe and feel, and without these fundamental elements it isn t dance music. A track may be beautifully programmed and arranged, but if it has no feel then the whole endeavour is pointless. 15 To find out how a final mix should sound like it is further advisable to train your hearing by analyzing titles of similar genres. Look for details in 12 Cf. Hometracked (2008), 15 th August Cf. Owsinski (2006), p.7f. 14 Cf. Tischmeyer (2006), p See Snowman (2004), p

13 commercial mixes by listening to them closely and learn how professional mixers did their jobs. 7

14 2. Mixing Theory Now that you have the sound of the final mix in mind you need to employ your spatial sense to imagine where and how you can arrange the instruments. Because, as with our eyes, we experience the surrounded environment with our ears in three dimensions, too. Furthermore, you will learn how to alter sound and dynamics of audio signals. 2.1 The Three Dimensions The most important thing about mixing is the arrangement of audio signals in the given spatial sonic range. To do so you have to think in three dimensions: tall, deep, and wide. A well made mix must imply all frequencies from low to high in a balanced form (tall or vertical), and, furthermore, it requires enough depth (deep or front-to-back) and a wide spread stereo dimension (wide or horizontal) to sound consummate and clear. 16 This will be described closer in the following three chapters. But to obtain a fully balanced track you also need to worry about a proper arrangement, which can avoid problems right from the start. If instruments with a similar frequency and volume are played at the same time, the human ear can t tell them apart. Instead of trying desperately to find space for both instruments in the three dimensions it is often better to just mute one of them Cf. Owsinski (2006), p.8f. 17 Cf. Owsinski (2006), p.11f. 8

15 2.1.1 Layering Front-to-Back Perspective The distance of a sound source is generally specified through the intensity of air pressure our ears perceive, because the inverse law says that sound pressure decreases proportionally to the square of the distance from the source 18. So the more the sound has to travel, the quieter it gets, thus the louder an instrument is, the more in your face it will appear to be 19. This effect is definitely essential in dance music for the drums and bass to impart rhythm we can dance to. To avoid a fight for attention at the front side of the mix you should not use equal volumes for all instruments. To get the appearance of three dimensions you need to put some instruments further back. It is obvious that the easiest method is to reduce gain to create the impression of depth, but, in addition to that, another characteristic of distance is the amount of high frequencies we perceive. Since those frequencies have a shorter wavelength and, hence, have to travel a longer way than their lower companions, they lose more energy. By reducing the higher frequencies with an equalizer we can reproduce this effect to obtain a more realistic distant sound. Moreover, every sound is reflected by walls which cause echoes and reverberation. The further away a sound source is, the more often the sound waves will be rejected by obstacles crossing their way. Hence, while the amount of reverberation will increase, the stereo width will be reduced. By applying this natural effect to the desired instruments they will be pushed to the rear of the mix. 20 This procedure will be described in chapter Filters and Effects. 18 See Snowman (2004), p See Snowman (2004), p Cf. Snowman (2004), p.312ff. 9

16 2.1.2 Panorama Horizontal Perspective One of the most overlooked or taken-for-granted elements in mixing is panorama, or the placement of a sound element in the sound field. 21 Panning is not only useful to make a track seem more spacious but can also cause more overall action through movement and spreading. 22 The placement of the signal within the left and right border is done by the panorama potentiometer (pan), which is included in almost any audio software. Principally stereophony only simulates a spatial sound. We can hear a voice centrally although we only have a left and a right speaker. So it depends on how the brain processes the information it receives from the ears. If the signal arrives simultaneously and at the same volume we sense the sound coming from the middle. But if we hear a signal louder and earlier on one side of our head, we think that the sound source is not central. Stereophony takes advantage of this effect and uses different volumes on each speaker to control the position of a signal. 23 Nowadays nearly every sound-making software and hardware produces a stereo signal. But if you only use stereo files, you will not be able to make a transparent mix. While an instrument that is played alone may sound fantastic due to its stereo behaviour, it will occupy space for another one. It will get impossible to solve this problem by only using the pan pot or trying to fix it with unnecessary EQ, particularly in dance music, where many various elements are played at the same time. Furthermore, comb filtering effects, phase coherency, or other mono 21 See Owsinski (2006), p Cf. Owsinski (2006), p Cf. Eisner (2006), p

17 compatibility problems can appear by using stereo files or too many stereo effects. Therefore it is always better to use mono files throughout the mix Frequency Vertical Perspective The human ear can only perceive a certain range of frequencies and thus the mixing engineer has to arrange the instruments within these borders. With the usage of an equalizer it will be possible to assign an instrument to its place in the vertical layer. The sense of hearing can be divided into the following frequency bands: - Sub-Bass (0-50 Hz): The sub-bass, which is the lowest part of kick and bass, is mostly felt subconsciously. To avoid a reduction of volume and to prevent the subwoofers of a club from getting broken it is generally advisable to cut off these low frequencies. - Bass ( Hz): As the name suggests this range is used for instruments that primarily contain bass. EQ cuts (or boosts) around this area can add definition and presence to the bass and kick drum Mid-Range ( Hz): If too many instruments are represented here, the mix will sound muddy and unclear. - True Mid Range ( Hz): Due to the reason that this frequency range stands for the human voice it is perceived as the loudest. Therefore you have to pay utter attention when boosting instruments at these frequencies. 24 Cf. Snowman (2004), p See Snowman (2004), p

18 - High Range ( Hz): In this range the body of hi-hats and cymbals is located. To get a brighter sound you can try to boost dull instruments in this area. - Hi-high Range ( Hz): This area is occupied by some higher frequencies of hi-hats and cymbals. While some boosts at around 12 khz may produce music with a higher fidelity, too much gaining will cause unwanted noise and hiss. 26 The perception of loudness is not only depending on the acoustic intensity but is also controlled by the frequencies. Our ears perceive various frequencies in a different manner by their very nature. Due to the complex anatomy of the ear canal we can hear some frequencies better than others. The human voice, the most essential sound of all, resides between two and four khz and hence our sense of hearing is the best in this area. Quite contrary to that the sensitivity is decreasing below 500 Hz and above 10 khz. 27 This circumstance was found out experimentally and was henceforward called Fletcher-Munson contour control curve. For a mixing engineer it is very important to bear this curve in mind as the amount of bass is totally depending on the volume, as you can see in figure 2. Try to make your mixes sound great at loud volumes and passable on lower ones. 28 That is, of course, assuming that the clubs are your market they don t play music at low or medium levels Cf. Snowman (2004), p.311f. 27 Cf. Raffaseder (2002), p Cf. Snowman (2004), p.310f. 29 See Snowman (2004), p.310f. 12

19 Figure 2: Fletcher-Munson curve (Source: Snowman (2004), p.311.) 2.2 Filters and Effects Filters and effects are algorithms to alter signals to make them more interesting or to fit them into the surroundings. While filters are usually used to exclude certain parts of the signal, effects will add something to it. Both are available as software and hardware. Effects are further subdivided into real-time and offline effects. Latter are altering the audio file in a destructive way meaning that they change the file directly. Real-time effects are applied as plug-ins by internal routing of the signal path. Due to that the sound can be changed on the fly without the risk of losing the original file Equalizer The equalizer is the most important filter and is used as a frequency volume control. This will let you increase or decrease the gain of specific frequency 30 Cf. Raffaseder (2002), p.194f. 13

20 ranges to make space for other instruments. While an instrument needs all its frequencies when played solo, we will perceive it as a whole in a mix in which only the main frequencies of the timbres are left, because our hearing will pretend that the hidden ones are masked behind other instruments. 31 The simplest type is a Shelving EQ that only lets you cut or boost above or below a certain frequency. While low frequencies are passed by a low-pass shelving EQ and everything above the cut-off frequency will be diminished, a high-pass shelving EQ works the other way round. Those two filters are commonly known as bass and treble, which appeared quite often as knobs on older radios or hi-fi systems. On a typical mixer you will also find a third knob to adjust the mid range. This filter allows you to cut or boost all frequencies within a specific range and is called Band-pass filter. With a parametric EQ you can also alter the associated frequencies and the width of the filter response (Q). A high Q value creates a narrow filter that handles small frequency ranges while a low Q value produces a wider bandwidth and smoother crossovers. 32 In nature all frequencies of several sounds are reduced by walls and other objects. Consequently, while some boosts may be required for creative reasons, you should look towards mostly cutting to prevent the mix from sounding too artificial. Keep in mind, that you can effectively boost some frequencies of a 31 Cf. Snowman (2004), p Cf. Sound on Sound (2008), 28 th August

21 sound by cutting others, as the volume relationship between them will change. This will produce a mix that has clarity and detail Reverb As sound waves travel through the air they are reflected by every obstacle getting in the way. Depending on the surface the energy of the sound will decrease more or less while the frequency is modified. Therefore our brain notices some different echoes, from which it can recognize the surroundings, after the direct sound reaches our ears. Due to the reason that we are used to these natural reflections we also expect them to appear in everything we hear. But most digital samples and synthesizers will not give you a feeling of a real life surrounding, because their sound is mostly dry and without natural reverberation. 34 Usually the reverb effect unit is connected with the mixer through aux-send and is routed post fader to control the amount of reverberation with the volume fader of the channel. For this reason the ratio of the reverb effect is mostly set to 100%. 35 On most reverb effect units or plug-ins you can alter the following settings: - Ratio (or Mix): This setting lets you adjust the amount of direct sound and reverberation. When it is set to 0% there will only be direct sound, while at 100% only reverberation is let through. 33 See Snowman (2004), p Cf. Snowman (2004), p.109f. 35 Cf. Tischmeyer (2006), p

22 - Pre-delay Time: After a sound occurs, the time separation between the direct sound and the first reflection to reach your ears is referred to as the pre-delay. 36 When applying reverberation to important instruments that should stick out it is important to use a longer predelay to avoid moving them backwards in the mix due to the overlap of reflections. - Diffusion: To control the displacement of a signal you also have to define the stereo width of early reflections. A signal further away will tend to have more reverberation and less stereo width and the other way round. To obtain a correct stereo image it is vital to bear this in mind. - Decay Time: While the decay of reflections holds on longer in bigger buildings, the reverberation will shrink away very fast in smaller ones. To avoid overlapping of reverbs you have to pay attention to successional reflections from prior notes. This could cause a continuous increasing of feedback and, furthermore, desultory and washed-out frequencies. - High-Frequency and Low-Frequency Damping: Another vital thing to remember is the equalization of frequencies to get the impression of distance. 37 The further reflections have to travel, the less highfrequency content they will have, since the surrounding air will absorb them. 38 The following picture shows a typical reverb effect plug-in. 36 See Snowman (2004), p.110f. 37 Cf. Snowman (2004), p.110f. 38 See Snowman (2004), p.110f. 16

23 Figure 3: Reverb effect plug-in (Waves Rverb) (Source: author s own.) Every instrument should have its own ambient environment created by reverbs and other layering effects. To obtain a sonic layering these environments should not stay in conflict with each other. By returning the reverb in mono and panning it anywhere but hard left or right you can avoid clashing of those environments. Furthermore, a long reverb should be brighter than a shorter one. To weld the ambient environments together you may apply just a little bit of the longest reverb to all main components Delay In nature we experience echoes on mountains where an apparent recurrence of sound occurs. First reflections that arrive 50 ms after the direct sound are heard 39 Cf. Owsinski (2006), p.41f. 17

24 as a delay. 40 A digital delay line can emulate this effect for all incoming audio signals. On these effect units you can either set the delay time in milliseconds or in note values. To produce more echoes in a row the unit sends some of the signal back to itself. A feedback knob will allow you to determine the amount of the signal that comes back, like shown in figure Figure 4: Delay effect plug-in of Steinberg Cubase (Source: author s own.) If you have to set the delay time manually, because the effect unit cannot do the job, you need to use some mathematics to get the time of one quarter note: 60,000 (ms in a minute) / bpm (beats per minute). You need to set the time accurately, because otherwise the delay will stick out. 42 To get a realistic and nice delay effect the unit needs to add ambience to the signal and, furthermore, the sound must be duller. To get this effect done manually you need to add a reverb after the delay and equalize the signal before 40 Cf. Raffaseder (2002), p.218f. 41 Cf. Snowman (2004), p Cf. Owsinski (2006), p.44f. 18

25 it goes back as a feedback. If the unit or plug-in doesn t have a damping parameter, you will need to turn off the feedback completely and send the returning signal back to the delay unit again by using AUX-send. On the channel you are using for the AUX-return you can now equalize the signal and take away the high frequencies. As a result, not only the whole delay is duller, but also every repeat Other Effects Beneath reverb and delay, which are the two most important effects, there are a lot of other effects available. The most common and known are: - Chorus: By duplicating a signal, adding a little delay, and altering the sound height slightly it seems as if two instruments were playing together. Though most chorus effects now come as stereo, you can fatten up a mono chorus by panning the original, untreated sound to one side and the modulated delay to the other. The result is a moving wide sound source that seems to hang between the speakers Phaser and flanger: These two effects work similar to the chorus and are slightly different from each other. Both use an LFO to either modulate the phase shifting of the phaser or the time delay of the flanger. This creates a series of phase cancellations, since the original and delayed signals are out of phase with one another. 45 This will 43 Cf. Sound on Sound (2008), 10 th September See Sound on Sound (2008), 15 th September See Snowman (2004), p.112f. 19

26 result in a concave, strange sound and is often applied to various dance music elements Distortion: Normally distortion is not eligible but you can use this effect to get a dirtier, uglier, and more interesting sound. In most cases this is used for electric guitars to appear fatter and more aggressive Exciter: This effect is applied to get [ ] a brighter, airier sound without the stridency that can sometimes occur by simply boosting the treble. It is often accomplished with subtle amounts of high-frequency distortion, and sometimes by playing around with phase shifting. 48 Exciters add presence and clarity to instruments and, hence, are very useful for recordings with a lack in high frequencies Enhancer: When a signal crosses a certain threshold the enhancer emphasizes dynamically its heights. While an equalizer also gains the unwanted noise, the enhancer changes the signal to a more brilliant and transparent sound. 50 The usage of those effects and how to adjust their settings depends on the material and your purpose. Try whatever you want to find weird sounds that possibly have not been discovered yet. But be aware of using effects too heavily, because this may cause a very unnatural and intransparent global sound. 46 Cf. Snowman (2004), p.112f. 47 Cf. Wikipedia (2008), 13 th August See Zölzer (2002), p.128ff. 49 Cf. Zölzer (2002), p.128ff. 50 Cf. Eisner (2006), p

27 2.3 Dynamics The human perception of loudness depends on the average level of a signal. Sound events with a great dynamic possess large differences between minimum and maximum levels and are thus perceived quieter as a whole. The dynamic is an important factor for the sound design and composing of classical music, in contrary to dance music where especially kick and bass are extremely compressed to obtain as much pressure as possible. Furthermore, generally in pop music, vocals are very much restricted in their dynamics. In the mastering process the final songs are heavily compressed to compete with other titles in all kinds of media Compressor Recorded instruments and especially voices with large dynamics are a challenging task to be fit into an appropriate mix, because the volumes of those are difficult to adjust. If the fader is set to the loudest parts, the quieter ones will vanish into thin air. In the opposite case the louder sections would be too dominant to fit well. To prevent this compressors are used, whose task it is to reduce the dynamics and to prepare a signal so that it can be amplified without the risk of distortion. Thus quieter passages are lifted and the song gets generally louder and more powerful. 52 To achieve this you have to set a threshold in db on the compressor. This indicates when the compressor begins to work. Once a signal exceeds this threshold it will be diminished with a certain factor, named ratio. The signal is now restricted in its dynamics and can be boosted with the gain or output 51 Cf. Raffaseder (2002), p.196f. 52 Cf. Snowman (2004), p

28 controls. Two other important parameters of every compressor are the attack and release time. These indicate how quickly the compressor steps in after the threshold is crossed and how much time will have to pass until it stops when the amplitude falls below the value. In addition, there are some compressors that have the opportunity to choose between soft- and hard-knee. This means that you can decide if either a soft or a hard transition should be used. 53 In the following figure you will find the design of a typical compressor. Figure 5: Waves RComp compressor plug-in (Source: author s own.) A special form is a multiband compressor that splits the spectrum of the signal into three to ten bands. This allows you to compress the individual areas harder without mutual influence of the others. It is advisable to use this tool rather less 53 Cf. Raffaseder (2002), p.196f. 22

29 when mixing to ensure that possible corrections can still be done during the mastering process. 54 One very important thing to bear in mind when compressing is that, if you use a fast attack, the transient will be restricted and so high frequencies are curbed. This will lead the instrument to appear further back in the mix than you probably want it to. So you either use a compressor with a longer attack or apply a multiband compressor to only affect the lower frequencies Side-Chaining and Ducking Another very useful and frequently built-in compressor technique can make your mix sound much better, a lot punchier and even more sophisticated. The actual procedure is based on combined signal routes where the sound s envelope of another track is used to control the compressor of the actual signal, on what score this method is called side-chaining. When the current signal gets lower because the other one is triggering the threshold of its compressor we are speaking of ducking. This effect can be heard when a DJ speaks into his microphone and the currently played music gets hushed. For the purpose of mixing we can use the same effect when two signals use the same frequency but one should be more present. So you can use ducking when the lead sound and human voice are played together to let the vocals pull through the mix. This method can also be applied to bass lines to make room for the kick leading to a very pumping bass drum Cf. Tischmeyer (2006), p.157f. 55 Cf. Snowman (2004), p Cf. Snowman (2004), p

30 2.3.3 Limiter A limiter can be best described as a compressor with an infinite ratio. Because of its fast impact and heavy dynamic modification it should be used quite cautiously. A limiter is applied during recording instruments or vocals to prevent transient signals crossing the user-defined threshold. In the mastering process it will also avoid clipping errors and can boost the overall volume of the final mix by adding 3-6 db Cf. Snowman (2004), p.105f. 24

31 3. Mixing Practice and Workflow Now that you know the theory you can start your mixing software of choice. But before you touch a single fader or knob you should think about the final product once again. You really have to know what you want and where to place the instruments in the three dimensions. 3.1 Creating a Soundstage It is always advisable to start with a draft of a soundstage on which you position the instruments from left to right and front to back, like shown in figure 6. Your goal is to find a unique space in the mix for every instrument by keeping the three dimensions in mind. In the mixing process this is realized by using panorama, volume, equalizer, and reverberation. It is vital that every sound is heard well and fits in the mix perfectly, so this draft is an eligible concept to start with. Moreover, this will help you to keep an overview about positioning and arrangement of the various instruments Cf. Snowman (2004), p

32 Figure 6: Arrangement of instruments (Source: Snowman (2004), p.313.) 3.2 Initiate the Work Space If you want to mix your own songs, you have to make sure that you separate production from mixing. A vantage point to start with is a completely new project with dry audio tracks. This means that no automations, send, or insert effects are implemented yet, except for those that form the sound in a creative way. Finally, the pan and volume controllers have to be set to default Cf. Tischmeyer (2006), p.97f. 26

33 3.2.1 Arrangement Be sure to arrange similar audio files among each other. I prefer to place drums and percussions at the top of the window, followed by basses, synths, and vocals. Those are the main components of every dance track and, hence, require enough space in the sonic range to unfold powerful and transparent mixes. Pads, effects, and other harmonical instruments are positioned beneath. A typical arrangement is shown in figure 7. Figure 7: Arrangement of tracks and sub groups (Source: author s own.) Furthermore, it is also possible to define colours for audio tracks and sub groups in most popular digital audio workstations. However, it is your decision on how you want to configure your project, as long as you keep track of it. 27

34 3.2.2 Sub Groups Finally divide the audio files into sub-groups to disburden quick muting or common editing of tribal instruments. Sub-groups will also save valuable resources when you put compressor and equalizer into its inserts. A compressor can merge pieces together to build up consistent sounds while a common equalizer is great for getting required space in the frequency spectrum easily and fast. But panning and automations are not advisable in groups because this can lead to a lack of clarity Mixer and Signal Routing Instead of starting aimless mixing you should initialize your mixer with all necessary presets. Route all audio tracks into their dedicated sub-groups and arrange them accurately for a proper overview. If you use the same effects for a certain type of audio track you can add it by now to enhance your workflow. Moreover, some sequencers also offer the opportunity to save mixer presets as templates to use them in every similar project. Though early preparations will facilitate your working progress keep your hands off effect parameters. When you add them for later usage turn them off again to get an uncommitted start. 3.3 Setting Volumes Now let s begin with the first adjustments by muting all tracks, except of one: starting with the drum group, mainly focusing on the bass drum, all relative volumes are set. After all tracks are fit to one another in their regarding subgroups, the volumes of each sub-group are adjusted. Be sure to bring the 60 Cf. Tischmeyer (2006), p.103ff. 28

35 defining elements of dance music, typically the drums and bass, to the front of the mix. 61 There are also some things that have to be regarded conscientiously to obtain a sophisticated achievement. First of all it is obvious that you should never gain the volume of any audio track above 0dBFS to stem distortions and generally unwanted noise. Another important rule is that channel faders should never be set higher than their respective sub-group and, furthermore, all together should remain below the master volume. In earlier times every bit was significant for recording engineers but nowadays, since we deal with 24-bits and beyond, the more important issue is headroom. Every instrument, especially drums and percussions, exhibits extremely fast and impulsive signals called transients. Our ears recognize these peaks as a very natural sound but unfortunately they are hard to detect with standard peak meters. If you allocate enough headroom by adjusting your average volume to about -10dB, those important transients will not be cut and a realistic sound will remain. In fact you rather aim the volume of your final result a little lower and the overall dynamic higher to get better results during the mastering process Panning After setting volumes it is important to place the instruments from left to right. Timbres with a low frequency, especially kick and bass, should always remain in the centre of the soundstage to achieve mono compatibility. Particularly dance 61 Cf. Snowman (2004), p Cf. Owsinski (2006), p.105f. 29

36 music, which is most frequently played in the clubs, has to have a central fundament, so that the energy spreading is equal at all points. 63 The snare, which mostly supports the kick, should stay relatively central to exploit the energy, but don t shrink away from varying its position a little bit. Cymbals and percussions can be spread all over the sound stage to increase dynamic behaviour. Another good trick is to position Hi Hats further left or right with a delayed version on the other site. Another part of the mix which should stay central is the lead vocal because everyone would expect the vocalist in the centre of the stage. The lead synth should be a stereo file which is panned hard left and hard right to obtain space for the vocals. In parts of the song where there are no vocals you can also place a mono synth into the centre of the soundstage. In general it is recommended to set higher signals to the outside and deeper ones central. 64 In many cases it may be important to exaggerate the positions of each instrument to make a mix appear clearer and more defined. Nevertheless, you shouldn t position a sound in a different area of the soundstage just to avoid any small frequency clashes with other instruments. In theses circumstances you should try to EQ the sound beforehand, then if it still doesn t fit consider panning it out of the way Cf. Tischmeyer (2006), p Cf. Snowman (2004), p See Snowman (2004), p

37 3.5 Single Instrument Editing Now that you have an adequate global setup the next steps will focus on single instruments. Of course it always depends on the actual project, but the following lines can help you to find a good entry point and can be seen as a general guideline Kick / Bass Drum The bass drum is the most important part of all dance tracks and so you should give your best attention to it. The kick drum must assert itself with power and pressure in the low frequencies to lay the foundations together with the bass. To emphasize the pressure of a kick drum you can try to boost the gain of the low-frequency impact, residing around 40 to 120 Hz, a little bit. But that won t solve the problem if the sound of the bass drum has no punch at all, so you rather replace the kick with a more significant timbre. 66 The attack of a bass drum resides between 2 and 8 khz and controls the ability of locating the rhythm. In most cases you can increase this component a little bit but this always depends on the specific kick and the arrangement as a whole. Furthermore, the bass sector has to be kept clean so a low-cut with a narrow bandwidth should be set around 30 Hz, like shown in figure 8. In the muddy mid range between 120 and 350 khz you should decrease the gain with a bell filter to offer space for other elements of the mix, because this is the foundation for many other instruments. 66 Cf. Snowman (2004), p

38 Figure 8: A sample of a kick equalizer (Source: author s own.) Another important part for editing the kick drum is of course the compressor. It is used to limit its dynamics to intercept single peaks and, moreover, to make it punchier by adding artificial transients by using long attack times. If you want to use a reverb for your bass drum it may gain the higher frequencies and due to that the transients seem fickle and not precise. For this reason a reverb should only be inserted without pre-delay to avoid this unwanted fluttering. The reverb, on the other side, is vital for the spatial connection between the bass drum and other percussions. The reverb should be used unobtrusively, so that it is missed when turned off but not heard to strong and obvious when applied Cf. Tischmeyer (2006), p.196ff. 32

39 3.5.2 Snare Snares are very powerful at the bottom of the frequency range, but this energy is not needed, because it is mostly produced by the kick. Therefore you have to set a high-pass filter at around 150 Hz to get rid of the disturbing frequencies below this limit. The main body of a snare is located between 400 Hz and 1 khz. Some cuts or reductions in this area may help to fit better into the whole mix. To emphasize the snap of the snare it is generally recommended to give it a little boost at around 8 to 10 khz to brighten it up and to get a more striking snare drum at the high frequencies. 68 This is shown in figure 9. Figure 9: Snare equalizer (Source: author s own) 68 Cf. Snowman (2004), p

40 When you want to compress your snare drum to decrease dynamics it is very important to bear attention that you do not affect the transients of it, because otherwise it will sound as if someone has deflated the drum. To avoid that and to support its crispness you need to use long attack times Hi-hats and Cymbals Due to the reason that hi-hats and cymbals are only represented at higher frequencies you can erase the lower ones by using a high-pass filter at around 300 Hz or above. Figure 10 shows how narrow a filter can be set to fit an instrument into a mix. To increase the brightness of these instruments a shelving filter is set for all frequencies between 8 and 15 khz. If you took higher ones, the overall mix would get an ugly hiss. If you miss a little presence, try to boost the instrument at about 600 Hz Cf. Tischmeyer (2006), p Cf. Snowman (2004), p

41 Figure 10: Hihat equalizer (Source: author s own.) Bass Perhaps the most difficult task of a mixing engineer is balancing the bass and drums (especially the bass and kick). Nothing can make or break a mix faster than how these instruments work together. It s not uncommon for a mixer to spend hours on this balance (both level and frequency), because if the relationship isn t correct, the song will never sound big and punchy. 71 Once more a foresighted production is vital to avoid the kick and bass, which have similar frequencies, from clashing together. This is done by off-setting the bass so that it is not played at the same time as the kick does. If, however, your arrangement disallows that, you will have to equalize the bass radically to let both instruments fit together well See Owsinski (2006), p Cf. Snowman (2004), p

42 For getting more clarity you can try to gain the bass at around 800 Hz a little bit. Try boosting at 100 Hz and cutting at 140 Hz, if the bass is too warm, to make it more distinct without getting rid of the fundamental. To avoid muddiness it is always better to cut other instruments in the 250 Hz area than removing it from the bass. With a low-pass and high-pass filter you can further fit the bass into the mix. 73 A suitable ratio for the compression of the bass is somewhere around 8:1, like shown in figure 11. Connect the threshold of the compressor with the kick to avoid simultaneous occurrence. Generally it is recommended to use a fast attack and release, but as with all settings it always depends on the overall project situation and the timbre being used Cf. Owsinski (2006), p Cf. Snowman (2004), p

43 Figure 11: Typical offbeat bass compressor settings (Source: author s own.) Reverbs for basses are only applied discreet and virtually inaudibly to consolidate the interaction between kick and bass by combining their ambient environment. In other cases a stronger reverberation may be applied as a stylistic element, but this can lead to intransparent mixes. If you want to use chorus, flangers or phasers you will have to duplicate the bass and define the bottom on the one and the tonal perspicuity on the other track. This is essential to obtain a clear mono bass without phase problems and a well sounding stereo effect Cf. Tischmeyer (2006), p.209f. 37

44 3.5.5 Vocals Firstly, it should go without saying that the vocals should be compressed so that they maintain a constant level throughout the mix without disappearing behind instruments. Generally, a good starting point is to set the threshold so that most of the vocal range is compressed with a ratio of 9:1 and an attack to allow the initial transient to pull through unmolested. 76 If you have recorded the vocals correctly, you will not have to make big changes during the mixing process, because any boost will let the voice sound unnatural. To remove muddiness you may reduce it a little bit around 400 Hz. Many dance artists evade a lack of energy by increasing the speed and pitch of the vocal a little bit. Though this may not sound completely correct to some perfect trained ears it will produce the intended energy. Adding effects to vocals is a very sensible issue and should be handled with care. In most cases a little reverb is added to the vocals to obtain a more realistic sound. The effects have to be applied correctly from beginning on to avoid later frequency clashes. In this case you would have to equalize the effect of the vocal unnecessarily leading to an unnatural sound. 77 To get more presence you can also duplicate the vocal track and apply different EQ and compression settings to make them sound dissimilar. After that one track is placed left and the other one is panned to the right, but it should still sound as if the vocalist was in the centre of the stage See Snowman (2004), p Cf. Snowman (2004), p.332f. 78 Cf. Tischmeyer (2006), p

45 3.5.6 Synthesizers / Pianos / Guitars Synthesizers and all other melodic instruments have their basic frequencies in the mid range. Therefore it is important to edit them in the order of their significance. If vocals occur simultaneously, you will have to leave them untouched. You can either decrease the mid range of the instrument or apply a side chain of not more than 1 db. All frequencies that are not vital to appear in the mix have to be removed by using shelving filters. Try to find the lowest and highest frequency that change the sound apparently when eliminated and use them as the cut-off frequency Atmos, Pads & Sound FX Atmos, pads and sound effects will increase the depth of your production and will improve the dynamic behaviour and dramatization. These elements should not be too present and hence they should be edited in the end when every important instrument has its own space assigned. Try to find a place where they may fit in. 3.6 Automations Most common digital audio workstations that can handle MIDI input are capable of automating any knob or fader value. By using your MIDI controller along with your favourite sequencer software you can automate any volume, panning, EQ, and usually all parameters on VST instruments and effects, too. The software memorizes all of your movements and mostly lets you edit any recorded automation data with an editor afterwards Cf. Snowman (2004), p Cf. Snowman (2004), p

46 At first you need a static mix in which all volume and panning settings are already done consummately. This is a vantage point for your automations, because you should bear in mind that early usage of automations will lead to a very lousy workflow and so it ends up in extra work. As soon as you change the parameters of any track all of its automations regarding this value have to be modified, too. Thus, if the lead vocal is too quiet just within one part, try to solve this little difference with on-board techniques, so called offline methods, like simply gaining the volume only in this section. 81 Figure 12 shows the automation editor in FL Studios Cf. Tischmeyer (2006), p

47 Figure 12: Automation curves in FL Studios (Source: author s own.) 41