[Ffmpeg-devel] [PATCH] theora huffamn table

[castet.matthieu at free.fr]

Hi,

On Thu, May 19, 2005 at 06:02:10PM -0600, Mike Melanson wrote:
> matthieu castet wrote:
> >Hi,
> >
> >this patch allow to read the huffman table in theora header.
> 
> 	Got it, thanks. And thanks for providing the sample. So we now read 
> intra-Y coeffs, intra-C coeffs, inter coeffs, and Huffman tables. I 
We also already read coded_{ac,dc}_scale_factor tables.
> thought there were 6 items to read in all? I recall that loop filter 
> limit values were another item. What others are we missing?
There are 3 matrix [1] to read in order to compute a Quantization Matrix.

So I think init_dequantizer should be redone for theora in order to use
those tables.

The number of base matrix (coded_int*_dequant) could be also different of 3.



Matthieu


[1]
                        A 2 ? 3 array containing the
NQRS    2D Integer 6 No
        array           number of quant ranges for a
                        given qti and pli , respectively.
                        This is at most 63.
                        A 2 ? 3 ? 63 array of the
QRSIZES 3D Integer 6 No
        array           sizes of each quant range for a
                        given qti and pli , respectively.
                        Only the first NQRS[qti ][pli ]
                        values are used.
                        A 2 ? 3 ? 64 array of the
QRBMIS  3D Integer 9 No
        array           bmi 's used for each quant
                        range for a given qti and pli ,
                        respectively. Only the first
                        (NQRS[qti ][pli ] + 1) values
                        are used.