[FFmpeg-devel] Ticket 165, grayscale Cinepak in mov

[Michael Niedermayer]

On Sat, Sep 17, 2011 at 04:24:59AM +0200, Michael Niedermayer wrote:
> On Wed, Sep 07, 2011 at 12:49:30PM +0200, Michael Niedermayer wrote:
> > On Wed, Sep 07, 2011 at 06:01:09AM +0200, Michael Niedermayer wrote:
> > > On Tue, Sep 06, 2011 at 08:20:34PM -0700, Baptiste Coudurier wrote:
> > > > On 09/06/2011 07:51 PM, Michael Niedermayer wrote:
> > > > > On Mon, Aug 29, 2011 at 11:22:42PM +0200, Carl Eugen Hoyos wrote:
> > > > >> Hi!
> > > > >>
> > > > >> I had attached this patch to ticket 165, it seems to fix all samples.
> > > > >>
> > > > >> Please comment, Carl Eugen
> > > > > 
> > > > >>  mov.c |    3 ++-
> > > > >>  1 file changed, 2 insertions(+), 1 deletion(-)
> > > > >> 1f3316d440d4702ea0f6743d5a3513c8cb502dec  patchmovcinepak.diff
> > > > > 
> > > > > in absence of a better solution i think this LGTM
> > > > > baptiste, whats your oppinon ?
> > > > 
> > > > What's wrong with hacking the decoder ?
> > > 
> > > that looks like this:
> > > 
> > > diff --git a/libavcodec/cinepak.c b/libavcodec/cinepak.c
> > > index d2e0c26..7f408b9 100644
> > > --- a/libavcodec/cinepak.c
> > > +++ b/libavcodec/cinepak.c
> > > @@ -394,9 +394,9 @@ static av_cold int cinepak_decode_init(AVCodecContext *avctx)
> > >      s->width = (avctx->width + 3) & ~3;
> > >      s->height = (avctx->height + 3) & ~3;
> > >      s->sega_film_skip_bytes = -1;  /* uninitialized state */
> > > -
> > > +av_log(0,0, "bits %d\n", avctx->bits_per_coded_sample);
> > >      // check for paletted data
> > > -    if (avctx->bits_per_coded_sample != 8) {
> > > +    if (avctx->bits_per_coded_sample != 8 || 1) {
> > >          s->palette_video = 0;
> > >          avctx->pix_fmt = PIX_FMT_YUV420P;
> > >      } else {
> > > 
> > > so the else would never be reached, which is likely breaking something
> > > better suggestions are welcome
> > 
> > heres a better fix:
> >  diff --git a/libavformat/mov.c b/libavformat/mov.c
> > index 5464b78..70cd776 100644
> > --- a/libavformat/mov.c
> > +++ b/libavformat/mov.c
> > @@ -1052,6 +1052,9 @@ int ff_mov_read_stsd_entries(MOVContext *c, AVIOContext *pb, int entries)
> >                      color_index = 255;
> >                      color_dec = 256 / (color_count - 1);
> >                      for (j = 0; j < color_count; j++) {
> > +                        if (id == CODEC_ID_CINEPAK){
> > +                            r = g = b = color_count - 1 - color_index;
> > +                        }else
> >                          r = g = b = color_index;
> >                          sc->palette[j] =
> >                              (r << 16) | (g << 8) | (b);
> 
> i will commit this in a few days if there are no objections

locally commited will push tomorrow

[...]

-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

I hate to see young programmers poisoned by the kind of thinking
Ulrich Drepper puts forward since it is simply too narrow -- Roman Shaposhnik
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL: <http://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20110920/98fe4454/attachment.asc>