[FFmpeg-user] How to extract core DTS / DCA stream from a DTS 96/24 stream without re-encoding

Onetel tghewett2 at onetel.com
Sat May 19 18:23:23 EEST 2018



> On 19 May 2018, at 15:32, Moritz Barsnick <barsnick at gmx.net> wrote:
> 
> On Sat, May 19, 2018 at 14:21:47 +0100, Onetel wrote:
> 
>> I tried that as well but the output is just a duplicate of the DTS
>> 96/24 stream, and has the same low volume problem as the original
>> stream:
> 
> That's interesting. I found a different sample, and ffmpeg indeed isn't
> capable of "reducing" it.
> 
> I was in doubt whether 96/24 was an extension, but at least Wikipedia
> confirms:
> "DTS 96/24 is implemented as a core DTS stream plus an extension
> containing the deltas to enable 96/24 sound reproduction." I'm not sure
> that means that the extensions need to be there, or whether they are
> optional.
> 
> If there are extensions in there, there seems to be a shortcoming in
> ffmpeg's dca_core bitstream filter.
> 
> BTW, if you use ffmpeg's decoder (and thereby re-encode, of course), it
> does have the option:
>  -core_only         <boolean>    .D..A.... Decode core only without extensions (default false)
> 
> I haven't checked whether that works though...

Re: the -core_only option, the output seems to be the same as this when the -c:a copy option is removed, i.e. a re-encoding but only of the core stream, hence in my case the volume in the output sample being normal. So it seems the dca_core filter is working at least in part.

I guess there is a problem - naturally I am more interested in the bit-perfect core stream so hope that a solution will happen. To whomever is relevant I am happy to try out patches…

Tim.


More information about the ffmpeg-user mailing list