[FFmpeg-devel] [RFC] Channel layouts

[Måns Rullgård]

Peter Ross <pross at xvid.org> writes:

> On Fri, Aug 29, 2008 at 10:21:28PM -0400, Justin Ruggles wrote:
>> Peter Ross wrote:
>> > On Fri, Aug 29, 2008 at 04:28:00PM +1000, Peter Ross wrote:
>> >> Hi.
>> >>
>> >> This patch adds the notion of channel layouts to libavcodec.
>> > [...]
>> 
>> I think your solution works well for now.  I am not 100% clear about how
>> the encoders/decoders should handle unsupported or undefined layouts.
>> Firstly, is the channel layout is guessed by default?  If so, that would
>
> At this stage, the guess function is used only for display purposes.
>
>> eliminate it matching for some codecs.  Also, if the layout is
>> unspecified (and not guessed) I'm assuming an encoder or decoder can
>> decide whether to accept or reject it?  If it can accept it, then I
>> suppose it would be up to the encoder/decoder to make its own guess?
>
> If the decoder doesnt know the layout, then it should set layout to 0
> (i.e. "no statement").
>
> Each encoder will publish a list of supported input layouts, and libavcodec
> will provide transformation functions to convert(remix) between different
> layouts. All encoders will be capable of accepting a layout of 0.
>
>> For the ffmpeg program, I think it should revert to undefined if
>> rejected once, then if the codec doesn't support undefined layout return
>> an error.
>
> Yes, that's doable. Eventually lavc will provide a layout transformation
> (down/upmixing) function. My vision to have FFmpeg negociate the layout
> in much the sample way it currently negociates different pixel and audio
> formats.

Don't forget that the channel layout must specify not only which
components are present, but also their order.  Currently lavc tells
you neither, making it practically useless for multi-channel audio.

>> > +int64_t avcodec_guess_channel_layout(int nb_channels)
>> > +{
>> > +    switch(nb_channels) {
>> > +    case 1: return CHANNEL_LAYOUT_MONO;
>> > +    case 2: return CHANNEL_LAYOUT_STEREO;
>> > +    case 4: return CHANNEL_LAYOUT_QUAD;
>> > +    case 6: return CHANNEL_LAYOUT_5POINT1;
>> > +    case 8: return CHANNEL_LAYOUT_7POINT1;
>> > +    default: return 0;
>> > +    }
>> > +}
>> 
>> Why not add 3 and 5 to the list?  My vote would be for left/right/center
>> and 5.0.

2.1 and 4.1 are equally, if not more, likely.  I have such DVDs, but
none with x.0 where x > 2.

-- 
M?ns Rullg?rd
mans at mansr.com