On Thu, Apr 02, 2015 at 11:51:29AM +, Rainer Hochecker wrote:
> Michael Niedermayer gmx.at> writes:
>
>
> >
> > have you tried this with both annex b and .mp4 style h264 ?
> >
> > [...]
>
> I have updated the patch and tested both styles:
> https://github.com/FFmpeg/FFmpeg/pull/129
appli
Michael Niedermayer gmx.at> writes:
>
> have you tried this with both annex b and .mp4 style h264 ?
>
> [...]
I have updated the patch and tested both styles:
https://github.com/FFmpeg/FFmpeg/pull/129
___
ffmpeg-devel mailing list
ffmpeg-devel@ffm
Michael Niedermayer gmx.at> writes:
>
> could it be that this should be using nal_length_size ?
>
> [...]
I tried nal_length_size and it does not fix the issue. Instead it breaks some
other files. Seems silly VDA does only support 4 byte avcC flavour. The
patch brakes annexb. Should we add a
Rainer Hochecker online.de> writes:
> Carl Eugen Hoyos ag.or.at> writes:
>
> > (guessing, needs --enable-libx264)
> > $ ffmpeg -f lavfi -i testsrc -pix_fmt yuv420p -t 10 out.mp4
> > $ ffmpeg -f lavfi -i testsrc -pix_fmt yuv420p -t 10 out.h264
>
> my vda decoder does not even open with those
>
Carl Eugen Hoyos ag.or.at> writes:
> (guessing, needs --enable-libx264)
> $ ffmpeg -f lavfi -i testsrc -pix_fmt yuv420p -t 10 out.mp4
> $ ffmpeg -f lavfi -i testsrc -pix_fmt yuv420p -t 10 out.h264
>
> Carl Eugen
>
my vda decoder does not even open with those generated samples. it fails in
VDA
Rainer Hochecker online.de> writes:
> > have you tried this with both annex b and .mp4 style h264 ?
>
> could you please point me the samples you want me to test?
(guessing, needs --enable-libx264)
$ ffmpeg -f lavfi -i testsrc -pix_fmt yuv420p -t 10 out.mp4
$ ffmpeg -f lavfi -i testsrc -pix_fmt
Michael Niedermayer gmx.at> writes:
>
> have you tried this with both annex b and .mp4 style h264 ?
>
> [...]
could you please point me the samples you want me to test?
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailm
Michael Niedermayer gmx.at> writes:
>
> could it be that this should be using nal_length_size ?
>
> [...]
I am quite sure that the fix is correct. Our VDA decoder has always been
working like this.
Further I could not find any info on developer@apple that the current
code is valid. No idea wh
On Wed, Apr 01, 2015 at 12:40:54PM +0200, Rainer Hochecker wrote:
[...]
> @@ -330,22 +337,6 @@ static int vda_h264_decode_slice(AVCodecContext *avctx,
> const uint8_t *buffer,
> uint32_t size)
> {
> -VDAContext *vda = av
On Wed, Apr 01, 2015 at 12:40:54PM +0200, Rainer Hochecker wrote:
> ---
> libavcodec/h264.c | 2 +-
> libavcodec/vda_h264.c | 25 -
> 2 files changed, 9 insertions(+), 18 deletions(-)
Why does this makes a difference ?
data is collected during the decode_slice() call
---
libavcodec/h264.c | 2 +-
libavcodec/vda_h264.c | 25 -
2 files changed, 9 insertions(+), 18 deletions(-)
diff --git a/libavcodec/h264.c b/libavcodec/h264.c
index e8d5120..60f8aa1 100644
--- a/libavcodec/h264.c
+++ b/libavcodec/h264.c
@@ -1589,7 +1589,7 @@ again:
11 matches
Mail list logo
