[FFmpeg-devel] [PATCH] Add support for Audible AA files

[Vesselin Bontchev]

27.07.2015, 16:16, "Nicolas George" <george at nsup.org>:
> Le nonidi 9 thermidor, an CCXXIII, Vesselin Bontchev a écrit :
>>  This patch adds support for Audible AA files.
>>
>>  Audible samples can be obtained from,
>>
>>  https://gitlab.com/vesselin.bontchev/audible-samples/tree/master
>>  https://samples.ffmpeg.org/audible/
>
>>  Currently, this code generates corrupt audio output (in some places, and deterministically).
>
> Does this happen with the Go and Python implementations too?

There are no existing Go, or Python implementations actually.

>>  From ba73543efc3fdc4b5c61e9cb56f998d748716e00 Mon Sep 17 00:00:00 2001
>>  Subject: [PATCH] Add support for Audible AA files

>>  + avio_read(s->pb, src, trailing_bytes);
>>  + memcpy(buf + written, dst, trailing_bytes);
>
> Looks strange: you read to src but copy dst. May this be the cause of the
> glitches?

Good catch! This fixed the glitches entirely.

...

I am attaching a new revision (v2) of the patch. I still need to take care of memory leaks, and also harden input processing. This said, the code should already be looking much better (thanks for all the help!).

Vesselin
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0001-Add-support-for-Audible-AA-files.patch
Type: text/x-diff
Size: 14220 bytes
Desc: not available
URL: <http://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20150728/55342874/attachment.patch>