[FFmpeg-devel] CBP question

Germán Milano g.milano87 at gmail.com
Wed Nov 2 14:12:05 CET 2011


Thanks for the quick reply!, is good to know there's a dedicated team
behind FFmpeg.

I've got one more question, hope you don't mind:

   - In the function decode_cabac_luma_residual, there's a parameter (int)
   called *p*, which is later used to pass *index = 16*p + i4x4* to the
   function decode_cabac_residual_nondc; it is also used to infer the qmul
   param. What do the params *p* and *index* refer to?

Thank you very much for your help,

Jeremy

On Tue, Nov 1, 2011 at 11:08 AM, Michael Niedermayer <michaelni at gmx.at>wrote:

> On Tue, Nov 01, 2011 at 10:11:19AM -0200, Germán Milano wrote:
> > Hi everyone,
> >
> > I'm trying to print out the DCT coeffs AFTER dequantization. In order to
> do
> > that, i had to take a look at the decode_cabac_residual_internal function
> > (in h264_cabac.c). Now, i know that the cbp param represents whether a
> luma
> > block (and corresponding colour diferences blocks) has any DCT coeffs
> > distinct from zero. If i understand correctly, that is taken into account
> > in that function, in the following line:
> >
> > h->cbp_table[h->mb_xy]
> >
> > I started with FFmpeg a couple of months ago, so i'm trying to understand
> > the terminology in order to fully understand the functions. So, i have
> the
> > following questions:
> >
> >    - h -> mb_x and h -> mb_y are the coordinates of the mb in the frame?,
> >    is h -> mb_xy the dimension?
>
> mb_xy is the mb index into 2d mb based arrays
>
>
> >    - The DCTELEMENT passed to the function decode_cabac_residual_internal
> >    function is the one that gets 'filled up' with the DCT coeffs after
> >    dequant, isn't it?, i.e., h -> mb contains the DCT coeffs after
> dequant?
>
> yes, thats except for DC
>
>
> >
> > I hope i'm not bothering anyone with my questions (sorry in advance if i
> > am).
>
> no, questions are very welcome. The more people ask, the more they know
> and the more they know, the more they can contribute
>
> [...]
> --
> Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB
>
> I know you won't believe me, but the highest form of Human Excellence is
> to question oneself and others. -- Socrates
>
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel at ffmpeg.org
> http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>
>


More information about the ffmpeg-devel mailing list