I record the audio in Audacity, where I change the speed to increase the pitch of my voice to make it cartoonish, then I export an MP3.
I bring the MP3 into the timeline in Lightwave 3D and coordinate the lip movements with the audio. Basically, when the character is speaking, their lips move. The audio file is used ONLY as a reference and is not rendered out from Lightwave.
From there, I export a .MOV video from Lightwave at 30 frames per second - 30FPS - using the Quicktime Animation codex. So it’s basically a string of animated frames that match the audio when in Lightwave.
However, when I bring both the MP3 and the .MOV video into Filmora, the MP3 is 40:12 and the .MOV is 44:22 (40 seconds, 12 frames versus 44 seconds 22 frames) in time length. The default Filmora frame rate is 25FPS, but I change that to 30FPS which has no effect on the MP3.
The number of frames in the animation matched the audio exactly in Lightwave. How do I get them to match up again in Filmora? That is, is there a conversion I need to do to the MP3 file that’s used in Filmora?
I figured out how to do this manually by re-importing the MP3 that I used as a reference in Lightwave back into Audacity and adjusting it with the Tempo effect so the length matches the MOV video in Filmora. It gets within a frame or two in the overall length, so it’s close enough for animation. This works, but it seems like there may be something more automagical that I may be missing.
Yes, it’s set to exactly 30FPS in Lightwave, but see next below.
Never do production in MP3. MP3’s full family name is MPEG-1, Layer 3. It’s part of a very early video format and it’s notoriously unstable when used by itself in critical applications.
Try it in WAV once, or one of the newer compressed sound formats such as M4A.
That did it! Using a WAV file in Lightwave indicates the shorter timeline, so I can sync to that. IOW, Lightwave was mis-reading the MP3 when using it as a reference. I just didn’t notice where the change was occurring.
When you convert to mp3 unavoidably about 100ms is added to the start, that’s enough knock-out lip-sync.
Either don’t use mp3, (use wav instead), or shift mp3 the audio back ~100ms in the video editor.
Use change speed effect rather than change tempo, as the latter adds artefact noises.
The solution is to use a WAV file in Lightwave as a reference audio file for setting the animation. For whatever reason, Lightwave misinterpreted the MP3 file in terms of total length. I didn’t notice it in Lightwave because that timeline is set to frames instead of minutes and seconds. I’ve now set the timeline in Audacity to NTSC frames, which, at 29.97 is close enough to 30 that I’ll at least notice it if there’s a big difference in the number of frames. Lightwave doesn’t actually incorporate the WAV file, it’s just there at the bottom of the screen so I can scrub through it for setting the animation. The WAV file gets combined with a Lightwave-generated MOV file of the animation when I import them both into Filmora and render out a final MP4.