[FFmpeg-user] amix with atempo: Inconsistent behaviour creating m4a with aac

[Jonathan Girven]

> I haven't tested such use case, but I can tell you atempo also changes
> output pts (computes it based on the number of samples output so far).  So
> shouldn't mix_audio2 asetpts expression reference the output of atempo for
> STARTPTS?

I was under the impression that the concat filter that I use to join
the final sections together (com_a_slowmo) just put the sections one
after the other. Therefore I need each section to "start from time
zero" for lack of a better expression. If I were using amix to merge
the three sections, I would agree with you that my asetpts would need
to be something more complex. Am I wrong?