Shortest video an AI dub or face swap accepts: 5 s vs 4 s

Synthesia's AI dubbing page says "Video must be at least 5s long". Sume does not offer a dubbing product in the docs read for this post; its nearest video-in tool, Beta face swap, plans for source videos of about 4-15 seconds with usable audio, and avatar videos take a 4-60 second script.

What does Synthesia's dubbing page list?

Besides the 5-second minimum, the page lists "140+ languages and dialects", accepts "MP4, MOV, and WebM files", and says the "first minute of video is completely free". It adds that "Subtitles are auto-generated for every dubbed video". These are Synthesia's statements about its own product.

What length does each Sume video tool take?

The face swap docs say Beta worker validation targets source videos "currently planned for about 4-15 seconds" with usable audio, and the request takes only avatar_handle, video_url and quality. The avatar video docs accept an estimated 4-60 seconds.

Is face swap the same as dubbing?

No. Face swap puts a Sume avatar's face onto your source video; it does not translate speech. The docs list transcripts and prompts among unsupported fields on that endpoint. If you need translated audio, see lip sync vs dubbing for how the pieces differ.

What if my clip is shorter than the minimum?

For face swap, the docs give a planned range of about 4-15 seconds and do not say how a shorter clip is handled. Test with a real sample before batching.