I have a MP4 file generated with ffmpeg that has an mp4a-encoded audio track that beeps every second. It contains an edit list with a specified media_time of 1024, which makes Mediabunny (correctly) begin emitting samples offset from the first priming packet.
https://github.com/user-attachments/assets/e37d5e3c-7b7d-4f66-b220-b834a49604d6
I have another which is basically the same file, just with no edit list (media_time of 0). This would cause Mediabunny to emit the priming silence in the first packet when decoding.
https://github.com/user-attachments/assets/14ba3588-d227-4032-8e65-34da818b8d78
Audacity seems to do the same thing and include the silence:
However, Audition ignores the first packet, honouring the intent of the original file:
If the edit list isn’t present and we know the first packet doesn’t produce samples, should we assume an implicit offset of the first packet’s duration?
I have a MP4 file generated with ffmpeg that has an
mp4a-encoded audio track that beeps every second. It contains an edit list with a specifiedmedia_timeof 1024, which makes Mediabunny (correctly) begin emitting samples offset from the first priming packet.https://github.com/user-attachments/assets/e37d5e3c-7b7d-4f66-b220-b834a49604d6
I have another which is basically the same file, just with no edit list (
media_timeof 0). This would cause Mediabunny to emit the priming silence in the first packet when decoding.https://github.com/user-attachments/assets/14ba3588-d227-4032-8e65-34da818b8d78
Audacity seems to do the same thing and include the silence:
However, Audition ignores the first packet, honouring the intent of the original file:
If the edit list isn’t present and we know the first packet doesn’t produce samples, should we assume an implicit offset of the first packet’s duration?