Did you cut first seconds of the FaceForensics dataset?

[universome]

Hi! FaceForensics contains "video starting" artifacts for its first ~0.5 seconds for many its videos (see the gif), which might produce the corresponding training artifacts. Did you remove them?

Here are random samples from FFS, cut to the first 0.5 seconds:

Also, did you account for them in any way when computing FVD?

[alanspike]

Hi, we follow the data processing code from TGAN v2 for a comparison w/ their methods.

[universome]

I see, thank you. I checked the code in TGANv2 and they do not cut the videos in any way