[FFmpeg-devel] [PATCH] PSMF audio support, trac ticket #3233

[Maxim Polijakowski]

>> +        avctx->sample_rate    = ctx->sample_rate;
>> +        avctx->block_align    = ctx->frame_size;
>> +        avctx->bit_rate       = ctx->sample_rate * ctx->frame_size * 8 / 2048;
>> +        avctx->channels       = ff_oma_chid_to_num_channels[ctx->channel_id - 1];
>> +        avctx->channel_layout = ff_oma_chid_to_native_layout[ctx->channel_id - 1];
> if you set these from the parser instead of the decoder then you
> must be a bit careful as the decoder can run in a seperate thread as
> the parser so the parser cannot change anything midstream it could
> only set the parameters once before the decoder can start
> (that is when the parameters where not set before)

Ok, I see. The problem is that ATRAC3+ streams will be added dynamically 
in the demuxer. Stream parameters like frame_size or sample_rate seem to 
be sent along with packets. I don't know whether they are available in 
headers or not.

For the case I won't find any stream configuration before dealing with 
packets, what's the right way to supply stream parameters to the decoder?
The whole parser thingy seems to be poorly documented so I don't know 
where to proceed...

Best regards
Maxim