24bit flac converted to alac, bit-depth atom in output file incorrectly set at 16bit

[pssturges]

Summary of the bug: 24bit flac converted to alac, bit-depth atom in output file incorrectly set at 16bit not 24bit
How to reproduce:

% ffmpeg -i 24bitflac.flac -y -map 0:0 -c:a:0 alac 24bitalac.m4a
ffmpeg version 2.8.14-0ubuntu0.16.04.1
mp4file --dump 24bitalac.m4a | grep sample
        "24bitalac.m4a": sampleSize = 16 (0x0010)
       "24bitalac.m4a": sampleSize = 0 (0x00000000)
       "24bitalac.m4a": sampleCount = 6096 (0x000017d0)

This causes the resultant file to play as white noise in some players as they attempt to play it as 16bit file. Logitech Media Server is one such player.

An extensive investigation of this issue can be found on the slim devices forum found here:
​https://forums.slimdevices.com/showthread.php?102059-24-bit-ALAC-files-being-truncated-to-16-bit

Forum member 'bpa' has suggested a patch in that thread which I have tested and found to correct the problem

[Carl Eugen Hoyos]

Replying to pssturges:

Forum member 'bpa' has suggested a patch in that thread which I have tested and found to correct the problem

Then why did you open this ticket?
Please send the patch - made with correct attribution and git format-patch - to the development mailing list.

[bpa]

My understanding of patches sent to developmentlist is that have to had run the regression test which I havne't done.
​https://www.ffmpeg.org/developer.html#Submitting-patches

[bpa]

The patch is to libavformat/movenc.c but it has been updated since I created the "patch" so line number may be wrong.
1081,1083c1085

<             
<             if (track->par->codec_id == AV_CODEC_ID_FLAC ||
<                 track->par->codec_id == AV_CODEC_ID_ALAC) {
---
>             if (track->par->codec_id == AV_CODEC_ID_FLAC) {

[Elon Musk]

Fixed in d9a91d58a17ff485a02cd63581ba6f528315c767.