[MPlayer-dev-eng] [PATCH] Automatic downmix

[Nicolas George]

Le jour du Labour, an CCXVIII, Clément Bœsch a écrit :
> So yes, there is a final downmix filter added (and this is the same
> without -channels 2 since it's the default).
> 
> However, if you set -channels 6, then only the pan filter for upmix
> (specified in command line) is added.

Yes, but this is not the only consequence of -channels 6, it also affects
the way some codec do their work.

For example, you could imagine a strange AC3 file that crashes the decoders
with -channels 6 or just plays horribly, but works fine with -channels 2. Of
course, you would file a bug report, but in the meantime, you may want to
play the file on a 5.1 speaker system, thus the af pan=6.

I concede that this scenario is quite unlikely, but it is not strictly
impossible, and the principle it illustrates stands: once in a while, some
user will have a good reason to combine -channels 2 and -af pan=6.

The cornerstone is, I believe, the meaning of the -channels option.
Currently, it is "have the codec try to decode to n channels if it can". The
current version of your patch changes it to "have the codec try to decode to
n channels if it can, and force the audio to n channels before outout", and
I do not think this is right, since those are two very different features.

The change should IMHO rather be: "have the codec decode to n channels by
any means necessary". That would probably mean add a pan filter just after
the codec, at the very start of the filter chain. And probably doing it just
after the codec is initialized.

Regards,

-- 
  Nicolas George
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL: <http://lists.mplayerhq.hu/pipermail/mplayer-dev-eng/attachments/20100919/b010fcd5/attachment.pgp>