Doesn't work.
Signed-off-by: Philip Langdale
---
libavcodec/libnvenc.c | 7 +++
libavcodec/libnvenc.h | 3 +++
libavcodec/nvencoder.c | 6 --
3 files changed, 14 insertions(+), 2 deletions(-)
diff --git a/libavcodec/libnvenc.c b/libavcodec/libnvenc.c
index 3c64234..83587a6 100644
---
Signed-off-by: Philip Langdale
---
libavcodec/libnvenc.c | 17 +
1 file changed, 17 insertions(+)
diff --git a/libavcodec/libnvenc.c b/libavcodec/libnvenc.c
index 6ba150e..3c64234 100644
--- a/libavcodec/libnvenc.c
+++ b/libavcodec/libnvenc.c
@@ -208,6 +208,23 @@ static av_cold i
To support b frames, we need to implement a queue of buffers, so that
frames can be held, pending their future reference frames. The nvenc
docs say that we need (num b frames) + 4 buffers, and the maximum
number of b frames is 16, so we need 20 buffers.
While we could allocate them dynamically, it
As promised, here are my current stack of improvements over the base
implementation. The interlacing one is probably pointless as the hardware
doesn't seem to support it, but you'll know better than me.
And hopefully you can explain the black magic in the DAR calculation.
Philip Langdale (4):
C
The encoder writes the display aspect ratio into the output frames, which a
good player will respect, so let's make sure it's correct.
The 1.02 scale factor is black magic, but produces correct results. I don't
know what nvenc is doing.
Signed-off-by: Philip Langdale
---
libavcodec/libnvenc.c
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA256
On 05/01/15 16:57, Michael Niedermayer wrote:
> On Mon, Jan 05, 2015 at 01:00:41PM +, tim nicholson wrote:
>> On 01/01/15 01:56, Michael Niedermayer wrote:
>>> This should allow us to insert idet before scale and let scale have
>>> interl=-1 as
On Mon, Jan 05, 2015 at 04:53:05PM -0800, Mark Reid wrote:
> Hi,
> MXF files generated by Media Composer or LibMXF can contain a
> project name property in the Preface. Lots of existing samples have them.
>
> http://samples.ffmpeg.org/MXF/issue2160/PW0805A0V01.4C5B5636.EFA330.mxf
> project_name
On Tue, Jan 6, 2015 at 5:47 AM, Michael Niedermayer
wrote:
> On Mon, Jan 05, 2015 at 06:20:15PM +0530, Anshul wrote:
> > On 01/04/2015 10:17 PM, Michael Niedermayer wrote:
> > > the table is constant and does not change, theres no need to have
> > > a copy of it in each context or to "make it eve
On Mon, 5 Jan 2015 21:10:31 +0100
Michael Niedermayer wrote:
> On Mon, Jan 05, 2015 at 06:56:20PM +0100, wm4 wrote:
> > I think this turned out pretty terrible. There's no good way to add new
> > custom tags that write to AVFormatContext->metadata.
> > ---
> > libavformat/id3v2.c | 50 ++
Signed-off-by: James Almer
---
I don't have MSVC or ICL 10.x, so i only tested this with gcc after forcing
HAVE_ALIGNED_STACK to 0 in config.asm
libswresample/x86/audio_convert.asm| 39 --
libswresample/x86/audio_convert_init.c | 4 ++--
2 files changed, 30
It's not an installed header. Use libavutil/avconfig.h instead.
Signed-off-by: James Almer
---
Untested as i don't have an OpenCL SDK installed.
configure | 2 +-
libavutil/opencl.h | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/configure b/configure
index 50458
2015-01-06 0:47 GMT+08:00 Titov, Alexey :
> Hi Wei,
> I will double check using yuv.
> What opencl platform/device you are running it on?
> Also, can you point me to /testfile/blueangles.m4v so I can reproduce
> correct and not correct outputs?
>
> Thanks!
>
Hi
Number of platforms:
Hi,
MXF files generated by Media Composer or LibMXF can contain a
project name property in the Preface. Lots of existing samples have them.
http://samples.ffmpeg.org/MXF/issue2160/PW0805A0V01.4C5B5636.EFA330.mxf
project_name: DNX145 PW Test
http://samples.ffmpeg.org/ffmpeg-bugs/trac/ticket345
On Thu, Jan 01, 2015 at 02:56:41AM +0100, Michael Niedermayer wrote:
> This should allow us to insert idet before scale and let scale have interl=-1
> as default in that case
>
> Signed-off-by: Michael Niedermayer
> ---
> doc/filters.texi |7 +++
> libavfilter/vf_idet.c | 41
On Mon, Jan 05, 2015 at 06:20:15PM +0530, Anshul wrote:
> On 01/04/2015 10:17 PM, Michael Niedermayer wrote:
> > the table is constant and does not change, theres no need to have
> > a copy of it in each context or to "make it every time decode is
> > called"
> > a simple static uint8_t parity_tabl
On Mon, Jan 05, 2015 at 03:17:32PM -0500, Vanista Herion wrote:
> Some MPEG 2 video will give bad interlacing information in AVFrame. If the
> parameters are used by a deinterlace filter such as yadif, frames will be
> processed needlessly and impair the quality.
>
> By reviewing the code in mpegv
On Mon, Jan 05, 2015 at 03:17:32PM -0500, Vanista Herion wrote:
> Some MPEG 2 video will give bad interlacing information in AVFrame. If the
> parameters are used by a deinterlace filter such as yadif, frames will be
> processed needlessly and impair the quality.
yes and i hope this will get resol
On Mon, Jan 5, 2015 at 9:17 PM, Vanista Herion wrote:
> Some MPEG 2 video will give bad interlacing information in AVFrame. If the
> parameters are used by a deinterlace filter such as yadif, frames will be
> processed needlessly and impair the quality.
>
> By reviewing the code in mpegvideo.c wh
Some MPEG 2 video will give bad interlacing information in AVFrame. If the
parameters are used by a deinterlace filter such as yadif, frames will be
processed needlessly and impair the quality.
By reviewing the code in mpegvideo.c which sets the interlaced_frame
and top_field_first, I noticed that
---
libavformat/mxfdec.c | 39 ---
1 file changed, 32 insertions(+), 7 deletions(-)
diff --git a/libavformat/mxfdec.c b/libavformat/mxfdec.c
index 4715169..5ade23a 100644
--- a/libavformat/mxfdec.c
+++ b/libavformat/mxfdec.c
@@ -200,6 +200,7 @@ typedef struct {
Hi,
This patch exports the full umid of packages as metadata. ffmpeg currently
only exports the material number portion of the umid.
The new format is
---
example:
060a2b340101010101010f00-13-00-53dc416b9a770251060e2b347f7f2a80
Some applications use the umids to link to mxf media instea
On Mon, Jan 05, 2015 at 06:56:20PM +0100, wm4 wrote:
> I think this turned out pretty terrible. There's no good way to add new
> custom tags that write to AVFormatContext->metadata.
> ---
> libavformat/id3v2.c | 50 ++
> 1 file changed, 50 insertions
On Mon, Jan 05, 2015 at 06:56:19PM +0100, wm4 wrote:
> ID3v1 fields have a fixed size, and they are padded either with zeros,
> or with spaces. Handle the latter case, instead of putting strings with
> trailing spaces into the AVDictionary.
> ---
> libavformat/id3v1.c | 11 ++-
> 1 file ch
On 04/01/15 8:49 PM, Michael Niedermayer wrote:
> On Sat, Jan 03, 2015 at 07:52:17PM -0300, James Almer wrote:
>> On 29/12/14 5:05 PM, James Almer wrote:
>>> x86inc can translate r*m into a register or stack on its own
>>>
>>> Signed-off-by: James Almer
>>> ---
>>> libavcodec/x86/flacdsp.asm | 18
On Mon, Jan 05, 2015 at 04:33:44PM +, Paul B Mahol wrote:
> Signed-off-by: Paul B Mahol
> ---
> libavfilter/Makefile | 2 +
> libavfilter/allfilters.c | 2 +
> libavfilter/f_deinterleave.c | 143
> +++
> 3 files changed, 147 insertions(
On Sat, Dec 27, 2014 at 7:35 AM, Tomas Härdin
wrote:
> On Sun, 2014-12-21 at 17:46 -0800, Mark Reid wrote:
> > hi,
> >
> > Media Composer generates audio OPAtom mxf files that don't have index
> segments.
> > All the uncompressed pcm audio essence data contained in a single KLV
> packet.
> >
> >
On Mon, Jan 05, 2015 at 04:33:44PM +, Paul B Mahol wrote:
> Signed-off-by: Paul B Mahol
> ---
> libavfilter/Makefile | 2 +
> libavfilter/allfilters.c | 2 +
> libavfilter/f_deinterleave.c | 143
> +++
> 3 files changed, 147 insertions(
I think this turned out pretty terrible. There's no good way to add new
custom tags that write to AVFormatContext->metadata.
---
libavformat/id3v2.c | 50 ++
1 file changed, 50 insertions(+)
diff --git a/libavformat/id3v2.c b/libavformat/id3v2.c
ind
ID3v1 fields have a fixed size, and they are padded either with zeros,
or with spaces. Handle the latter case, instead of putting strings with
trailing spaces into the AVDictionary.
---
libavformat/id3v1.c | 11 ++-
1 file changed, 10 insertions(+), 1 deletion(-)
diff --git a/libavformat/
On Mon, Jan 05, 2015 at 01:00:41PM +, tim nicholson wrote:
> On 01/01/15 01:56, Michael Niedermayer wrote:
> > This should allow us to insert idet before scale and let scale have
> > interl=-1 as default in that case
>
> Good thinking.
>
> How would interl=-1 then get set as default for auto
Hi Wei,
I will double check using yuv.
What opencl platform/device you are running it on?
Also, can you point me to /testfile/blueangles.m4v so I can reproduce correct
and not correct outputs?
Thanks!
2015-01-05 14:50 GMT+08:00 Titov, Alexey :
Hi Wei,
> This is the first part of the whole patch
Signed-off-by: Paul B Mahol
---
libavfilter/Makefile | 2 +
libavfilter/allfilters.c | 2 +
libavfilter/f_deinterleave.c | 143 +++
3 files changed, 147 insertions(+)
create mode 100644 libavfilter/f_deinterleave.c
diff --git a/libavfilter
On Mon, Jan 05, 2015 at 02:04:47AM +0100, Michael Niedermayer wrote:
> Hi all
>
> as you probably already noticed videolans git server is not working
> since yesterday and as they generously provide also our main git,
> ffmpeg git is neither.
> I do hope the server will be alive again soon/?tomorr
On date Monday 2014-12-29 10:20:26 +0100, Stefano Sabatini encoded:
> On date Sunday 2014-12-28 19:45:18 +0530, arwa arif encoded:
> [...]
> > From ebf5f36d9e9fc80264bb605dfee5c09cbf1f10ef Mon Sep 17 00:00:00 2001
> > From: Arwa Arif
> > Date: Thu, 25 Dec 2014 09:50:24 +0530
> > Subject: [PATCH] l
On 01/01/15 01:56, Michael Niedermayer wrote:
> This should allow us to insert idet before scale and let scale have interl=-1
> as default in that case
Good thinking.
How would interl=-1 then get set as default for auto inserted filters.
Or did I miss something?
>
> Signed-off-by: Michael Nied
On 01/04/2015 10:17 PM, Michael Niedermayer wrote:
> the table is constant and does not change, theres no need to have
> a copy of it in each context or to "make it every time decode is
> called"
> a simple static uint8_t parity_table[256];
> or even
> static const uint8_t parity_table[256] = {...}
On date Friday 2015-01-02 11:13:29 +0100, Werner Robitza encoded:
[...]
> (By the way, should I use in-reply-to or do you want separate threads
> for new patches? The developer doc doesn't say.)
I tend to prefer to have a single thread, so I can follow the
evolution and discussion related to a pat
On date Sunday 2015-01-04 12:31:30 +, Paul B Mahol encoded:
> Signed-off-by: Paul B Mahol
> ---
> libavfilter/vf_blend.c | 32
> 1 file changed, 20 insertions(+), 12 deletions(-)
>
> diff --git a/libavfilter/vf_blend.c b/libavfilter/vf_blend.c
> index 538774b
38 matches
Mail list logo
