3 CHOPS - LIP SYNCHING

Transcription

1 3 CHOPS - LIP SYNCHING In this Lesson you will use CHOPs to match lip movements with an existing audio file. You will use blend operations in the SOP editor to create the different facial shapes when saying different words. This will be done using previously created target shapes which are the mouth positions when saying different vowels and consonants. This is one of many different ways to do lip synching. Using CHOPs to do this will greatly reduce the amount of time spent animating the mouth. Houdini Training 4.0 Side Effects Software

2 3.1 GETTING STARTED 1. Launch Houdini. 2. From the desktop selector button located at the top of the viewport, select the CHOPs_Training desktop. 3. From the File menu Open the locate and open the following.hip file: /lipsynch.hip This.hip file you have loaded is a simple curve that is modelled in different ways so that it form the different vowels as well as a rest position and the letter B and R. These letters have been chosen based on the letters that occur in the audio file. Each model has been named so that it is easy to understand the SOP network. Each of the model SOPs are being fed in to a Blend SOP called mouth_blend. This Blend SOP will have Channels exported to it form CHOPs, which will blend between the vowels and consonants to make quick work of lip synching. 4. Working in the CHOPs (left side) of the desktop, navigate up one level to rename the CHOP network to be lipsynch. Press the i key to return to the CHOP network editor. 5. From the string of generator CHOPs, place a File CHOP into the tile layout window. Rename it to be Audio_in. Click on the Channel File option button to select a new file. Navigate to the file called beezer.aiff. Click on the green audio flag of this CHOP. From the Dialogs menu located at the top of the interface, select Audio Panel. The hot key for this is Ai. This will bring up the audio panel. Working with the Audio Panel. Toggle on the Time Line option. This will play the audio back while you are playing back the time slider. Toggle on the Graph flag of the Audio_in CHOP. Home the Graph to display it in it s entirety. Home Button Hold down the information button of the Audio_in CHOP. 18 Side Effects Software 1999 Houdini Training 4.0

3 The audio file s sample rate is 8000 Hz This is roughly the quality of AM radio reception. (not the greatest quality, but it is very small in size, so the computer will have an easy time dealing with it.) Increase the volume on your speaker. Starting at frame 1. Press play to listen to the audio sample. You should hear You re a good better Bucky Beezer. Toggle on the Frame Indicator button. Frame Indicator Button This will allow you to scroll the playback interactively. Grab the blue timebar with the left mouse button and move it to the right along the graph. You should now hear the audio playback as you are scrolling. Move the timebar to the left. You should now hear the playback in reverse. This technique is called scrubbing. What you are going to do now is to enter in the location of the words as they are spoken along the audio channel. With the left mouse button, press the toggle options button located above the graph window as shown below. This allows you to make notes onto the graph window at any time location. 6. Change the option from Notes Off to Audio_in 7. Starting from the far left of the audio, interactively scroll the audio until you hear the first sound of the word Your. Once you hear the sound of the Y (this should be around frame 4) In the notes section located above the graph window, type in the following note which is words Your a spelled phonetically, then press enter: Yourraa Houdini Training 4.0 Side Effects Software

4 Your graph window should have the words Yourraaa spelled along the top of the graph as shown below. 8. Advance the time bar so that you hear the beginning of the word good. This should be around frame 16. Keep in mind that advancing the timebar will allow you to advance in very small increments, small.er than a per frame basis. This is different than scrubbing with the frame counter which only allows you to advance one frame at a time. In the case of the word good the exact frame mark is closer to the mark. In the notes section, enter the following: goooooooood Advance to the beginning of the word batter Because there are two syllables in the word batter you will notice a definite division in the wave where the word is spoken. For animating this we will treat this as two separate words. At the beginning of the word batter enter the following in the notes field: baaattt Advance the time bar to the second syllable of the word batter Enter the following in the notes field: terrrrr By now your graph should resemble the image shown below: 20 Side Effects Software 1999 Houdini Training 4.0

5 9. Continue with the word Buckey, splitting it up into two different sections 10. Finish with the word Beezer so that your Graph window resembles the one shown below. What this allows you to do is to visualize where the mouth movements will occur. By spelling out how the audio sounds and when it happens, it will make it easier and faster to animate. You are now going to prepare the channels to be exported out to the object editor 11. Place a Constant CHOP to the left of the Audio_in CHOP. Open the Channel folder and disable the single frame option. Houdini Training 4.0 Side Effects Software

6 Change the start from 0 to 1. Change the End value to be 90. Open the Common folder tab. Change the Units parameter from Seconds to Frames. 12. Append a Rename CHOP to the Constant CHOP. Working in the Objects/SOPs area of the Desktop, navigate to the Mouth_flap s SOP editor. The reason why you have returned to the Mouth_flap s SOP editor is to make note of the order in which the mouth target shapes are piped into the mouth_blend SOP. This is important because you are going to be exporting channels up to each of the blend channels, and you want to know whether the blend2 channel is going to make the mouth form an o sound, or an a sound. Below is an illustration from Preston Blair s book on Character Animation it illustrates the target shapes to form various consonants and vowels. The input order of the mouth_blend SOP is as follows: input 1 = rest position, input2 = A, input 3 = E, input 4 = O, input 5 = U, input 6 = B input 7 = R, input 8 = Y 22 Side Effects Software 1999 Houdini Training 4.0

7 To find the proper channel names for the Blend input channels, hold down the help button for the mouth_blend SOP. You should note that the channel names are very straight forward: blend1 blend2 blend3 blend Return to the CHOPs editor, Append a Rename CHOP to the Constant CHOP. In the To field enter mouth_blend/blend2. Open the Common folder. In the Scope field change it so that it reads: /obj/mouth_flap Toggle on the Export flag of this CHOP. Hold down the information button of the Rename CHOP. You should see in the lower right corner of the information pop-up window that the channel is now being exported up to the Blend SOP. You are now going to edit the animation curve in the channel editor. This will be done for each of the blend channels. For the quote You re a good batter Buckey Beezer, you will create channels that will reflect what is being spoken. In other words for the word good you will create a channel that will export to the blend4 channel in the Blend SOP. The channel will rise up to a value of 1 where the o sound is heard. By doing this for each of the consonant you will create an animated blend sequence that will blend between the different vowels and consonants matching the audio playback. 14. Ac copy the Rename CHOP. Av to paste it. Repeat this step so that you have seven Rename CHOPs. Working with the Rename1 CHOP in the Name field. Rename it from rename1 to letter_a. Now do the same with the rest of the existing Rename CHOPs Rename the Rename2 CHOP to be letter_e. In its Rename to field, change it to be mouth_blend/blend3 Rename the Rename3 CHOP to be letter_o. In its Rename to field, change it to be mouth_blend/blend4 Rename the Rename4 CHOP to be letter_u. In its Rename to field, change it to be mouth_blend/blend5 Rename the Rename5 CHOP to be letter_b. In its Rename to field, change it to be mouth_blend/blend6 Rename the Rename6 CHOP to be letter_r. In its Rename to field, change it to be mouth_blend/blend7 Rename the Rename7 CHOP to be letter_y. In its Rename to field, change it to be mouth_blend/blend8 You have now prepared the necessary rename CHOPs to control each of the blend channels. 15. Return back to the letter_a CHOP. Ensure that graph flag for it and the Audio_in CHOP are being displayed at the same time Houdini Training 4.0 Side Effects Software

8 Enter into it s Channel editor. (\ on the icon display and select Edit Data Channels...) From the beginning of the audio advance the timebar to the first sounding of the letter a (in this case it is around frame 8) In the channel editor insert a keyframe at frame 5. In the channel editor insert a keyframe at frame 8. Insert another keyframe at frame 11. Insert one more keyframe at frame 14. Make a bounding around the keyframe at frame 8 and frame 11. Increase their values up to 1. You should now have a curve which resembles the channel as shown below: The hump in the curve which rises up to a value of 1 and remains at that value for the duration of the pronunciation of the a sound. Advance the timebar to the next uttering of the a sound. This should be around frame 36. Insert four more keyframes at the following marks: Using the left mouse button, make a bounding box around the two keyframes at 37 and 41. Increase their values to 1. Because that is the only pronunciation of the a sound that is all we have to do for this particular channel. By leaving the interpolation type at cubic, it will allow for a smooth transition between the mouth movements. Your letter_a channel should now resemble the illustration below: You are now going to continue doing this for each of the remaining channels. 16. Make the letter_e CHOP the active CHOP. 24 Side Effects Software 1999 Houdini Training 4.0

9 Toggle on the Graph flag of the Audio_in CHOP as well as the Graph flag of the letter_e CHOP. Hold down it s pop_up menu, and select Edit Data Channels... You are now going to create a similar curve, but this one will match the pronunciation of the e sound. Advance the timebar to the first e sound. This is easy due to the fact that you spent the time typing out the words as they are spoken. The firs e sound is at frame 43. In the channel editor, insert three keyframes at frames and 47. Make a bounding box around the keyframe at 43. Increase it s value to 1. Advance the timebar to the next e sound. Insert 4 keyframes at the following frame values: and 62 Make a bounding box around the keyframes at 57 and 61. Increase their values to 1. Insert four more keyframes at frames and 69. Make a bounding box around the 64 and 68 keyframes. Increase their values to 1. Insert two more keyframes at frames 71 and 74. Make a bounding box around the keyframe at frame 71. Increase it s value to 1. If you compare the way the channel rises and falls to the text you wrote on the graph window, it will match the areas where the e sound is heard. Your letter_e channel; should resemble the illustration below: 17. Continue with the same strategy so that your channels match the ones shown below. Houdini Training 4.0 Side Effects Software

10 letter_o CHOP keyframes at all set to all set to 1 letter_u CHOP has keyframes at all set to set to 1 letter_u CHOP keyframes at all at set to 1 letter_r CHOP keyframes at all at all at 0 26 Side Effects Software 1999 Houdini Training 4.0

11 letter_y CHOP keyframes at set at 0 1 set at set to Place a Merge CHOP below the CHOP list. Starting form the left, pipe all of the letter_* CHOPs into the Merge CHOP. Open the Common folder tab of the Merge CHOP. In the Export Prefix field change it so that it reads /obj/mouth_flap Toggle on the Export flag of the Merge CHOP. Hold down the information button of the Merge CHOp to confirm that all of the channels are being exported up to the mouth_blend SOP. 19. Starting at frame 1 press play to preview the animation sequence. Provided that you have the curves which represent the lips in view, you should see it lip synched as it plays. Although there are many ways to do lip synching, this is just one of them. In the case of one quick 3 second quote, it is very easy. This however, does not take full advantage of CHOPs, but it does illustrate how you can integrate the audio into the work environment and with the use of the CHOPs interface it makes it quite easy to set up the blend channels. The other important lesson here is the integration of the Jive channel editor in CHOPs, although using the Jive editor will result in a CHOP being locked off, it does allow you to created custom channels within CHOPs, almost like a CHOPs modeller. Houdini Training 4.0 Side Effects Software

12 28 Side Effects Software 1999 Houdini Training 4.0