[FFmpeg-devel] Codec IDs for AC3/DTS variants (Was: Re: [RFC] additinaldesc_type for dtshd mpeg-ts demuxer)

Benjamin Larsson banan
Tue Jun 10 17:06:35 CEST 2008


Michael Niedermayer wrote:
> On Tue, Jun 10, 2008 at 09:36:33AM +0200, Benjamin Larsson wrote:
>   
>> Andreas ?man wrote:
>>     
>>> Benjamin Larsson wrote:
>>>   
>>>       
>>>> elupus wrote:
>>>>     
>>>>         
>>>>>>> Yes, we need to/should add support for spdif  output/pass through.
>>>>>>>       
>>>>>>>           
>>>>>>>               
>>>>>> That would be entirely up to the calling application to handle.  All I'm
>>>>>> asking that such all application be given enough information to make the
>>>>>> right choice.
>>>>>>
>>>>>>     
>>>>>>         
>>>>>>             
>>>>> It's actually more the job of an encoder / bitstream filter. You need to add a 
>>>>> S/PDIF header to the data, and padd to a certain length. Demuxer should never 
>>>>> output S/PDIF'ed data, but sure a AC3_SPDIF, DTS_SPDIF, AAC_SPDIF, WMA_SPDIF 
>>>>> codec id's could be nice to have.
>>>>>   
>>>>>       
>>>>>           
>>>> Don't forget mpeg audio over spdif. Btw we have figured out how to send 
>>>> wmapro over spdif if you are interested.
>>>>
>>>>     
>>>>         
>>> Btw, does anyone know how to send AAC over SPDIF? I think i've seen
>>> AV receivers that claim to support it at least.
>>>
>>> Perhaps we should add a output muxer format "IEC958" that generates
>>> SPDIF output, I've been thinking about that previously..
>>>   
>>>       
>> I'm not sure that will work. Wmapro needs/can use parameters from the 
>> asf container and the packets need to be split into frames. So in that 
>> case it would be better if the decoder could get a spdif generation 
>> function.
>>     
>
> a solution with a muxer or bitstream filter would be much cleaner ...
>   

Hmm, yeah when I rethought the requirements it would actually fit better 
with a muxer. I'll try to draft how it should work for wmapro over spdif.

MvH
Benjamin Larsson






More information about the ffmpeg-devel mailing list