I don’t know how many people in Japan use ffmpeg to create singing cover videos, but I think ffmpeg is the fastest so I use ffmpeg.
It’s recommended to standardize file names before batch processing with ffmpeg.
mp4 to wav
Convert the downloaded original video and off vocal video to wav.
ffmpeg -i input.mp4 -vn -acodec pcm_s16le -ar 44100 -ac 2 output.wav
# e.g.
ffmpeg -i offv_*.mp4 -vn -acodec pcm_s16le -ar 44100 -ac 2 offv.wav
ffmpeg -i org_*.mp4 -vn -acodec pcm_s16le -ar 44100 -ac 2 org.wav
Seems highly versatile.
Remove audio from video
Remove audio from the original video.
ffmpeg -i org*.mp4 -vcodec copy -an org_noSound.mp4
Combine audio with video
Combine the singing cover audio with the original video. (Assumes the singing cover audio is encoded to match the original video’s audio. Otherwise it will be out of sync.)
# note
ffmpeg -i org*.mp4 -i comp_*.wav -vcodec libx264 -acodec libmp3lame 歌ってみた.mp4
# e.g
ffmpeg -i org*_noSound.mp4 -i comp_oke-3_normalize_calc.wav -vcodec libx264 -acodec libmp3lame comp_oke-3_normalize_calc.wav.mp4
- There seems to be an option to delay audio file audio (not verified)