When the format change, the VAAPI context cannot be destroyed.
Otherwise, the reference frame surface will lost.
Signed-off-by: Yan Wang
---
libavcodec/decode.c | 6 ++
1 file changed, 6 insertions(+)
diff --git a/libavcodec/decode.c b/libavcodec/decode.c
index 6c31166ec2..3eda1dc42c 100644
On Mon, May 27, 2019 at 7:52 AM Nicolas George wrote:
> Jun Li (12019-05-27):
> > I compared the perf between vflip and this patch before, they take almost
> > the same time, around 9ms for a frame.
> > Hflip is a little special, but surely I can update the patch to avoid
> frame
> > copy, and it
On Mon, May 27, 2019 at 7:56 AM Nicolas George wrote:
> Jun Li (12019-05-25):
> > I believe it has been enabled by default, the 'autorotate' value is true
> by
> > default.
>
> Yes, but it is currently enabled for constant rotation. Variable
> rotation is more tricky, so the question of enabling
Fix #6945
Rotate or/and flip frame according to frame's metadata orientation
---
fftools/ffmpeg.c| 16 +++-
fftools/ffmpeg.h| 3 ++-
fftools/ffmpeg_filter.c | 28 +++-
3 files changed, 40 insertions(+), 7 deletions(-)
diff --git a/fftools/ffmpe
Add exif orientation support and expose an option.
---
libavfilter/hflip.h| 3 +
libavfilter/vf_hflip.c | 43 ++---
libavfilter/vf_transpose.c | 173 +
3 files changed, 171 insertions(+), 48 deletions(-)
diff --git a/libavfilter/hflip.h b/lib
On Thu, 2019-03-28 at 13:28 -0400, Shaofei Wang wrote:
> Fix the issue: https://github.com/intel/media-driver/issues/317
>
> the root cause is update_dimensions will be called multple times
> when decoder thread number is not only 1, but update_dimensions
> call get_pixel_format in each decode thr
From: Limin Wang
The testing command for the HDR10 output with nvenc:
$ ./ffmpeg_g -y -i 4K.mp4 -c:v hevc_nvenc -color_primaries bt2020 -colorspace
bt2020_ncl -color_trc smpte2084 -sei hdr10 \
-master_display
"G(13250,34500)B(7500,3000)R(34000,16000)WP(15635,16450)L(1000,50)"
-max_
From: Limin Wang
The testing command for the atc HLG output with nvenc:
./ffmpeg_g -y -i 4K.mp4 -c:v hevc_nvenc -g 7 -color_primaries bt2020
-colorspace bt2020_ncl -color_trc arib-std-b67 -sei hlg test.ts
Signed-off-by: Limin Wang
---
libavcodec/nvenc.c | 22 ++
liba
From: Limin Wang
Signed-off-by: Limin Wang
---
libavcodec/nvenc.c | 36 ++--
1 file changed, 22 insertions(+), 14 deletions(-)
diff --git a/libavcodec/nvenc.c b/libavcodec/nvenc.c
index 3fe64bba8b..75dda6d689 100644
--- a/libavcodec/nvenc.c
+++ b/libavcodec/nven
On 5/27/2019 7:13 PM, Timo Rothenpieler wrote:
> On 27.05.2019 23:49, Lance Wang wrote:
>> On Saturday, May 25, 2019, James Almer wrote:
>>
>>> On 5/22/2019 3:59 AM, lance.lmw...@gmail.com wrote:
From: Limin Wang
The testing command for the HDR10 output with nvenc:
$ ./ffmpeg_
On 27.05.2019 23:49, Lance Wang wrote:
On Saturday, May 25, 2019, James Almer wrote:
On 5/22/2019 3:59 AM, lance.lmw...@gmail.com wrote:
From: Limin Wang
The testing command for the HDR10 output with nvenc:
$ ./ffmpeg_g -y -i 4K.mp4 -c:v hevc_nvenc -g 7 -color_primaries bt2020
-colorspace
This is based on target_dec_fuzzer
Signed-off-by: Michael Niedermayer
---
Makefile | 3 +
tools/Makefile| 3 +
tools/target_dem_fuzzer.c | 163 ++
3 files changed, 169 insertions(+)
create mode 100644 tools/target_dem_fuzzer.c
On Saturday, May 25, 2019, James Almer wrote:
> On 5/22/2019 3:59 AM, lance.lmw...@gmail.com wrote:
> > From: Limin Wang
> >
> > The testing command for the HDR10 output with nvenc:
> > $ ./ffmpeg_g -y -i 4K.mp4 -c:v hevc_nvenc -g 7 -color_primaries bt2020
> -colorspace bt2020_ncl -color_trc smp
Michael Niedermayer wrote on 27.05.2019 at 00:22:
> ideally some seperation between frames would be kept
> for example each frame in its own file might be the easiest
> or maybe just add teh codec id to nut and try to store it in that
> with stream copy
Added the codec id but getting invalid `cur
On Mon, May 20, 2019 at 14:40:24 +0200, Romane Lafon wrote:
> +float tolerance = 1.0 - s->tolerance;
Strictly speaking 1.0f (or 1.f).
> +float angle = cos(s->angle/180.0*PI);
If you want a float as result, use cosf(s->angle / 180.f * M_PI). (I'm
aware PI and M_PI aren't explicitly marked
On 5/20/19, Romane Lafon wrote:
> I have updated the patch that extends aphasemeter filter.
> It allows to get metadata for out-of-phase or mono sequences of stereo
> streams.
> It displays start, end and duration as for silencedetect filter.
>
Please use M_PI, and do not invent own constants.
__
On Thu, May 23, 2019 at 10:36:10PM +0200, Michael Niedermayer wrote:
> On Thu, May 23, 2019 at 10:57:14AM +0200, Carl Eugen Hoyos wrote:
> > Am Do., 23. Mai 2019 um 01:47 Uhr schrieb Michael Niedermayer
> > :
> > >
> > > Fixes: Timeout (14sec -> 9sec)
> > > Fixes:
> > > 13398/clusterfuzz-testcase-
On 26-05-2019 10:32 PM, Gyan wrote:
On 26-05-2019 10:08 PM, Werner Robitza wrote:
On Sun, May 26, 2019 at 6:09 PM Gyan wrote:
Do you want to update it for all algorithms with missing parameter
details?
Not at this stage – would be a lot of guesswork since I didn't
implement the filters. A
On Sun, May 05, 2019 at 12:57:50AM +0200, Michael Niedermayer wrote:
> Fixes: signed integer overflow: -1877966852 + -469491713 cannot be
> represented in type 'int'
> Fixes:
> 14561/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_TRUEMOTION2_fuzzer-5167608359288832
>
> Found-by: continuous fu
On Sun, May 12, 2019 at 11:21:48PM +0200, Michael Niedermayer wrote:
> Fixes: Timeout (54sec -> 188ms)
> Fixes:
> 14585/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_DIRAC_fuzzer-5649933052411904
>
> Found-by: continuous fuzzing process
> https://github.com/google/oss-fuzz/tree/master/projec
On Tue, May 14, 2019 at 01:04:52PM +0200, Michael Niedermayer wrote:
> Signed-off-by: Michael Niedermayer
> ---
> libavformat/mp3enc.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
will apply
[...]
--
Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB
Let us c
On 5/27/2019 11:06 AM, Steve Lhomme wrote:
> It's better to do it before the buffers are actually created. At least in VLC
> we currently don't support changing some parameters dynamically easily so we
> don't use the information if it comes after the buffer are created.
>
> Co-authored-by: James
Jun Li (12019-05-25):
> I believe it has been enabled by default, the 'autorotate' value is true by
> default.
Yes, but it is currently enabled for constant rotation. Variable
rotation is more tricky, so the question of enabling it by default must
be asked again.
> Do you mean the orientation cas
Jun Li (12019-05-27):
> I compared the perf between vflip and this patch before, they take almost
> the same time, around 9ms for a frame.
> Hflip is a little special, but surely I can update the patch to avoid frame
> copy, and it will be as fast as hflip.
>
> But still this is duplicated feature
On 5/27/19, Jun Li wrote:
> On Mon, May 27, 2019 at 1:39 AM Paul B Mahol wrote:
>
>> On 5/27/19, Jun Li wrote:
>> > On Mon, May 27, 2019 at 12:32 AM Paul B Mahol wrote:
>> >
>> >> On 5/27/19, Jun Li wrote:
>> >> > On Sun, May 26, 2019 at 2:09 AM Paul B Mahol
>> wrote:
>> >> >
>> >> >> On 5/26
It's better to do it before the buffers are actually created. At least in VLC
we currently don't support changing some parameters dynamically easily so we
don't use the information if it comes after the buffer are created.
Co-authored-by: James Almer
---
The same problem may exist with H264 alter
On Mon, May 27, 2019 at 1:39 AM Paul B Mahol wrote:
> On 5/27/19, Jun Li wrote:
> > On Mon, May 27, 2019 at 12:32 AM Paul B Mahol wrote:
> >
> >> On 5/27/19, Jun Li wrote:
> >> > On Sun, May 26, 2019 at 2:09 AM Paul B Mahol
> wrote:
> >> >
> >> >> On 5/26/19, Jun Li wrote:
> >> >> > On Sun,
On Sat, May 18, 2019 at 10:32:42PM +0200, Michael Niedermayer wrote:
> On Sat, May 18, 2019 at 01:05:01PM +0200, Reimar Döffinger wrote:
> > On 18.05.2019, at 12:15, Michael Niedermayer wrote:
> >
> > > On Sat, May 18, 2019 at 12:02:55PM +0200, James Darnley wrote:
> > >> On 2019-05-18 09:39, Mic
On 5/27/19, Jun Li wrote:
> On Mon, May 27, 2019 at 12:32 AM Paul B Mahol wrote:
>
>> On 5/27/19, Jun Li wrote:
>> > On Sun, May 26, 2019 at 2:09 AM Paul B Mahol wrote:
>> >
>> >> On 5/26/19, Jun Li wrote:
>> >> > On Sun, May 26, 2019 at 1:16 AM Paul B Mahol
>> wrote:
>> >> >
>> >> >> On 5/25
On Mon, May 27, 2019 at 12:32 AM Paul B Mahol wrote:
> On 5/27/19, Jun Li wrote:
> > On Sun, May 26, 2019 at 2:09 AM Paul B Mahol wrote:
> >
> >> On 5/26/19, Jun Li wrote:
> >> > On Sun, May 26, 2019 at 1:16 AM Paul B Mahol
> wrote:
> >> >
> >> >> On 5/25/19, Jun Li wrote:
> >> >> > Add exif
On 5/27/19, Jun Li wrote:
> On Sun, May 26, 2019 at 2:09 AM Paul B Mahol wrote:
>
>> On 5/26/19, Jun Li wrote:
>> > On Sun, May 26, 2019 at 1:16 AM Paul B Mahol wrote:
>> >
>> >> On 5/25/19, Jun Li wrote:
>> >> > Add exif orientation support and expose an option.
>> >> > ---
>> >> > libavfilt
On Sun, May 26, 2019 at 09:27:55PM +0200, Reimar Döffinger wrote:
> On Sat, May 25, 2019 at 12:04:49PM +1000, Peter Ross wrote:
> > ---
> > This provides a small readability improvement.
> > I observe no performance change on x86_64 or arm6.
>
> Looks good to me, but probably the maintainer should
32 matches
Mail list logo
