[theora-dev] VP3 IDCT

Michael Niedermayer michaelni at gmx.at
Thu Mar 6 01:59:52 PST 2003


On Thursday 06 March 2003 03:47, Mike Melanson wrote:
> On Thu, 6 Mar 2003, Michael Niedermayer wrote:
> > btw, does theora/vp3 have a maximum allowed keyframe (or co-located
> > non-skiped non-intra macroblock) limit like mpeg* ? IMHO it should have
> > one, otherwise different decoders with different IDCTs will drift too
> > much like u mentioned
> >
> > another related issue is the dequantizer, mpeg1/2/4/h263 does various
> > weird oddifications to reduce the idct drift, i havnt found anything like
> > this in the vp3 source, but possibly i missed it
> > anyway, as soon as melansons implementation is finished we can try vp3
> > with a few different IDCTs (& DCTs) ...
> 	Note that I have strictly been a decoder guy thus far in my
> multimedia hacking career. I will need some help when it comes to writing
> the VP3 encoder. Fortunately, I am preparing a VP3 document, similar to
encoding isnt much more difficult then decoding here very likely / hopefully 
ffmpeg does the motion estimation, ratecontrol, ... u just have to do the 
bitstream encode mostly ...


> 	Again, all 3 Q matrices and both scaling tables are hardcoded in
> the VP31. In the upcoming Theora spec, they will be transported within the
> Ogg file header, along with the 80 32-element histograms for DCT token
> decoding.
yes, u allready mentioned that ...
btw, why does theora store the histograms, instead of just the code lengths
16bit *80*32 for the histogram     -> 5120byte
5bit*80*32   for the code lengthes -> 1699byte
and with some difference encoding for the code lengthes it will be another 
2-3x smaller

u very likely say now that it doesnt matter, as its just global headers which 
are transmitted just once, but what about (very) low bitrate cases, there it 
could add a noticeable startup delay, for example with 64kbit/sec its 0.64 
sec with 5120 byte


--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to 'theora-dev-request at xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is needed.
Unsubscribe messages sent to the list will be ignored/filtered.

More information about the Theora-dev mailing list