Downsampling Expression Clips

Due to its inherent algorithm, the Smooth filter can decimate animation details within an expression clip. Downsample, on the other hand, will preserve subtle movements for a more life-like performance.

Downsampling works by extracting the initial keys on a specified frame interval, which can remove animation juddering.

1. Deploy an expression clip by using facial motion capture or facial puppeteering.
   
2. Playback to view the result. In this case, the head juddering is quite obvious.
   
3. Open timeline and find the expression clips on the Expression track of the character.
   
4. Right click on the clip and choose the Downsample command from the context menu.
   
5. You will see the Downsample panel show. You can drag the sliders to determine the downsampling per frames for individual facial parts
   
6. Click the Apply button.
7. Playback to view the downsampled result.
   
8. Compare the expressions before and after being downsampled.
   
9. The following video shows the results by Smoothing and Downsampling.