[FFmpeg-devel] Copy user data in H264 for captions

[Michael Niedermayer]

On Wed, Oct 30, 2013 at 09:43:57PM +0100, Michael Niedermayer wrote:
> On Wed, Oct 30, 2013 at 06:20:14PM +0530, Radha Krishna Ramachandruni wrote:
> > Patch for copying ITU_T registered user data to AVFrame for extracting
> > closed captions.
> > 
> > 
> > rkrishna
> 
> >  libavcodec/h264.c     |   17 ++-
> >  libavcodec/h264.h     |    9 +
> >  libavcodec/h264_sei.c |  228 ++++++++++++++++++++++++++++++++++++++++++++------
> >  libavutil/common.h    |    5 +
> >  4 files changed, 228 insertions(+), 31 deletions(-)
> > 6669e36e8a72f265bc701f5263151652aa42c22f  ffmpeg-h264cc.patch
> > diff --git a/mythtv/external/FFmpeg/libavcodec/h264.c b/mythtv/external/FFmpeg/libavcodec/h264.c
> > index 3a83b4b..e8c0735 100644
> > --- a/mythtv/external/FFmpeg/libavcodec/h264.c
> > +++ b/mythtv/external/FFmpeg/libavcodec/h264.c
> > @@ -1742,10 +1742,19 @@
> >  
> >      pic->f.reference            = h->droppable ? 0 : h->picture_structure;
> >      pic->f.coded_picture_number = h->coded_picture_number++;
> >      pic->field_picture          = h->picture_structure != PICT_FRAME;
> >  
> > +
> > +    /* Put ATSC captions cached from parse_user_data into the correct frame */
> > +    memcpy(pic->f.atsc_cc_buf, h->tmp_atsc_cc_buf, h->tmp_atsc_cc_len);
> > +    pic->f.atsc_cc_len = h->tmp_atsc_cc_len;
> > +    h->tmp_atsc_cc_len = 0;
> > +    memcpy(pic->f.scte_cc_buf, h->tmp_scte_cc_buf, h->tmp_scte_cc_len);
> > +    pic->f.scte_cc_len = h->tmp_scte_cc_len;
> > +    h->tmp_scte_cc_len = 0;
> 
> there are no such fields on AVFrame
> 
> 
> > +
> >      /*
> >       * Zero key_frame here; IDR markings per slice in frame or fields are ORed
> >       * in later.
> >       * See decode_nal_units().
> >       */
> 
> > @@ -2775,12 +2784,11 @@
> >       * and a bad error table. Further, the error count goes to
> >       * INT_MAX when called for bottom field, because mb_y is
> >       * past end by one (callers fault) and resync_mb_y != 0
> >       * causes problems for the first MB line, too.
> >       */
> > -    if (CONFIG_ERROR_RESILIENCE &&
> > -        !FIELD_PICTURE && h->current_slice && !h->sps.new) {
> > +    if (CONFIG_ERROR_RESILIENCE && !FIELD_PICTURE && h->current_slice && !h->sps.new && h->avctx->skip_frame < AVDISCARD_ALL) {
> >          h->er.cur_pic  = h->cur_pic_ptr;
> >          ff_er_frame_end(&h->er);
> >      }
> >      emms_c();
> >  
> 
> looks unrelated
> 
> 
> > @@ -4519,12 +4527,13 @@
> >                  case NAL_SLICE:
> >                      first_slice = hx->nal_unit_type;
> >                  }
> >  
> >              // FIXME do not discard SEI id
> > -            if (avctx->skip_frame >= AVDISCARD_NONREF && h->nal_ref_idc == 0)
> > -                continue;
> > +            if (avctx->skip_frame >= AVDISCARD_NONREF && h->nal_ref_idc == 0
> > +            		&& hx->nal_unit_type != NAL_SLICE && hx->nal_unit_type != NAL_SEI)
> > +            	continue;
> >  
> 
> looks unrelated
> also contains tabs, which arent allowed in c files in ffmpeg git
> 
> also the patch doesnt apply
> 
> patching file libavcodec/h264.c
> Hunk #1 succeeded at 1951 with fuzz 2 (offset 209 lines).
> Hunk #2 FAILED at 2784.
> Hunk #3 FAILED at 4528.
> 2 out of 3 hunks FAILED -- saving rejects to file libavcodec/h264.c.rej
> patching file libavcodec/h264.h
> Hunk #1 succeeded at 668 (offset 30 lines).
> patching file libavcodec/h264_sei.c
> Hunk #1 FAILED at 92.
> Hunk #2 FAILED at 221.
> 2 out of 2 hunks FAILED -- saving rejects to file libavcodec/h264_sei.c.rej
> patching file libavutil/common.h

also, what i forgot, its very important code is tested too
(cosidering that it doesnt apply and doesnt look complete, i suspect
 it wasnt tested with ffmpeg HEAD, and thats very bad)


[...]
-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

Complexity theory is the science of finding the exact solution to an
approximation. Benchmarking OTOH is finding an approximation of the exact
-------------- 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/20131030/07327f21/attachment.asc>