[FFmpeg-devel] VQA v3 GSoC questions, need help

Kostya kostya.shishkov
Mon Mar 30 09:17:47 CEST 2009


On Mon, Mar 30, 2009 at 02:56:13AM -0400, The Deep Explorer wrote:
> Thanks Mike a bunch for all the inputs.
> 
> Please bear with me for I have some more questions :
> 
> Regarding the frequency of the codebooks being replaced.
> The HC-VQA.TXT states the following :
> "There's a major difference between the old (8 bit, 256 color) and the new,
>  HiColor VQAs. Lookup tables are no longer split up into several CBP?
>  chunks, instead they come in one piece (a CBFZ chunk). Two lookup tables
>  can now be many frames apart, not just 8 (as usual). This is indicated
>  by the CBParts entry of the header (see VQA_INFO.TXT), which is set to 0.
>  Subsequent frames use the last lookup table loaded, of course."
> 
> So it means keep using the same codebook for all frames till another
> CBFZ block comes
> and then we create a new codebook which will be used by subsequent frames ?
> 
> 
> > A VPTR chunk has a table of codebook indicies. A VPRZ chunk contains the
> > same data as a VPTR chunk, only it's compressed. You will not see 'VPTR'
> > inside the data after decompressing a VPRZ chunk. You will just see a
> > vector map.
> 
> So a VPRZ chunk should be decompressed like the following :
> 
> decode_format80(&s->buf[vprz_chunk], chunk_size,
>         s->decode_buffer, s->decode_buffer_size, 1);
> 
> and that the decoded buffer should be good to be used for rendering ?
> 
> Also, specifically for version 3, what is the rendering algorithm or
> the vector index mechanism
> specifically for v3 vqas. I did not find it , in the docs. Please give
> me some pointers.

You have http://multimedia.cx/VQA_INFO.TXT for that, it should be quite
similar. Also see VPTR description in HC-QA.TXT again.

It looks to me that codebook contains full 16-bit colour blocks and when
you get some information from VPTR you just draw that block from codebook
(look at the sizes of uncompressed VPTR and CBFZ, should be clear from that).
 
> I am just trying to learn, pardon me if these are too naive a question.
> Look forward to all the replies.
> 
> Thanks,
> -tde



More information about the ffmpeg-devel mailing list