X video audio must be AAC-LC, not HE-AAC: check an AI video

X's media best practices require audio to be AAC with the Low Complexity profile; High-Efficiency AAC is not supported. Audio should be mono or stereo, not 5.1 or more. Sume's default exact trim re-encodes video to libx264 and yuv420p and remuxes kept audio as AAC, but the docs do not say which AAC profile or channel layout, so verify the file before upload.

X rules are from its best-practices page, read 2026-10-01. Sume behavior is from Video trim.

What does X require?

The audio and codec rows, as X lists them:

What does Sume's exact trim do?

precision: "exact" is the default: a frame-accurate re-encode with libx264 and yuv420p. With audio: "keep", the docs say exact remuxes kept audio as AAC. That is consistent with X's AAC, H264 and YUV 4:2:0 rows. It does not by itself prove H264 High Profile, the audio profile, or the channel count, and the docs make no such claim.

What about keyframe mode?

precision: "keyframe" is stream copy. Nothing is re-encoded, so the output carries whatever codecs the source had. If the source audio is HE-AAC or 5.1, keyframe mode keeps it, and X would not accept it. For an upload to X use the default exact mode; see trim without re-encoding for when stream copy is fine.

How do I get a mono audio track?

Audio detach can return a track with channels set to mono (default source), as wav or mp3. That is a separate audio file, not a re-muxed video. Check the final MP4 with a probe before posting.