Re: [FFmpeg-devel] [FFmpeg-devel, RFC] lavfi: add opencl tonemap filter.

Hello Niklas, Thanks so much for your valuable feedback. > -Original Message- > From: Niklas Haas [mailto:ffm...@haasn.xyz] > Sent: Saturday, May 5, 2018 2:00 AM > To: Song, Ruiling > Cc: ffmpeg-devel@ffmpeg.org; s...@jkqxz.net > Subject: Re: [FFmpeg-devel,RFC] lavfi: add opencl tonemap

Re: [FFmpeg-devel] [PATCH v2 04/10] lavfi/nlmeans: add AArch64 SIMD for compute_safe_ssd_integral_image

On Tue, May 08, 2018 at 04:06:26AM +0200, Michael Niedermayer wrote: > On Mon, May 07, 2018 at 07:24:16PM +0200, Clément Bœsch wrote: > > ssd_integral_image_c: 49204.6 > > ssd_integral_image_neon: 28346.8 > > --- > > libavfilter/aarch64/Makefile | 3 + > > libavfilter/aarch64/vf_nlmeans_

Re: [FFmpeg-devel] [PATCH] Added the possibility to pass an externally created CUDA context to libavutil/hwcontext.c/av_hwdevice_ctx_create() for decoding with NVDEC

Thank you so much! We will test this hopefully today, and verify the expected behavior with NSIGHT. By the way, I'm new to ffmpeg, so... I don't know if you use your fork to test things first, before adding the changes to the main ffmpeg project? Or may we consider to compile and use your fork?

Re: [FFmpeg-devel] Misc improvements in nlmeans filter [v2]

On Mon, May 07, 2018 at 07:32:55PM +0200, Paul B Mahol wrote: > On 5/7/18, Clement Boesch wrote: > > Changes since v1: > > > > - fixed float operation in double as pointed out by Moritz > > - fix broken commit split as pointed out by Michael > > - added patch 10: "use unsigned for the integral pat

Re: [FFmpeg-devel] [PATCH] Added the possibility to pass an externally created CUDA context to libavutil/hwcontext.c/av_hwdevice_ctx_create() for decoding with NVDEC

On 08.05.2018 10:11, Oscar Amoros Huguet wrote: > Thank you so much! > > We will test this hopefully today, and verify the expected behavior with > NSIGHT. > > By the way, I'm new to ffmpeg, so... I don't know if you use your fork to > test things first, before adding the changes to the main ff

Re: [FFmpeg-devel] fate/hap : add test for hap encoding

> > > Sorry, i missread the code before writing the test, and > > didn't see it use float for one step. > > I will send a patch to remove these tests. > > Or just commit. > > Tests removed. Martin ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org htt

Re: [FFmpeg-devel] avdevice/sdl output : fix window_size and add new option (WIP)

2018-04-20 22:24 GMT+02:00 Moritz Barsnick : > On Sun, Apr 08, 2018 at 17:28:24 +0200, Martin Vignali wrote: > > > - 001 : Fix -window_size option > > Before this patch, window_size is always set to the source size > > In other word, -window_size option have no effect. > > Makes sense, since the o

Re: [FFmpeg-devel] [PATCH 01/13] avformat/mxfenc: Correct KAG alignment of preface

mån 2018-05-07 klockan 12:38 +0200 skrev Michael Niedermayer: > Signed-off-by: Michael Niedermayer > --- >  libavformat/mxfenc.c|  1 + >  tests/ref/fate/copy-trac4914|  4 ++-- >  tests/ref/fate/mxf-reel_name|  2 +- >  tests/ref/fate/time_base|  2 +- >  tests/ref/lavf/mx

Re: [FFmpeg-devel] [PATCH 03/13] avformat/mxfenc: Add Product Version, Toolkit version and Platform

mån 2018-05-07 klockan 12:38 +0200 skrev Michael Niedermayer: > > Signed-off-by: Michael Niedermayer >  static void mxf_write_identification(AVFormatContext *s) >  { >  MXFContext *mxf = s->priv_data; > @@ -790,7 +809,7 @@ static void mxf_write_identification(AVFormatContext *s) >   >  ver

[FFmpeg-devel] [PATCH v2 1/3] lavf/network: fix doxygen comments.

Signed-off-by: Jun Zhao --- libavformat/network.h | 8 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/libavformat/network.h b/libavformat/network.h index e3fda4d..efaa789 100644 --- a/libavformat/network.h +++ b/libavformat/network.h @@ -87,9 +87,9 @@ int ff_network_wait_f

[FFmpeg-devel] [PATCH v2 2/3] lavf/tcp: add option to setting Maximum Segment Size

This can change the the MSS value announced to the other end in the initial TCP packet, it's can be used when failed Path MTU discovery. Signed-off-by: Jun Zhao --- libavformat/tcp.c | 11 +++ 1 file changed, 11 insertions(+) diff --git a/libavformat/tcp.c b/libavformat/tcp.c index b028

Re: [FFmpeg-devel] [PATCH 04/13] avformat/mxfenc: Add object model version

mån 2018-05-07 klockan 12:38 +0200 skrev Michael Niedermayer: > > Signed-off-by: Michael Niedermayer >  mxf_write_metadata_key(pb, 0x012f00); >  PRINT_KEY(s, "preface key", pb->buf_ptr - 16); > -klv_encode_ber_length(pb, 130 + 16LL * > DESCRIPTOR_COUNT(mxf->essence_container_count));

[FFmpeg-devel] [PATCH v2 3/3] doc/protocols: documents tcp_mss

Signed-off-by: Jun Zhao --- doc/protocols.texi | 3 +++ 1 file changed, 3 insertions(+) diff --git a/doc/protocols.texi b/doc/protocols.texi index e19504d..e9091e0 100644 --- a/doc/protocols.texi +++ b/doc/protocols.texi @@ -1397,6 +1397,9 @@ Set send buffer size, expressed bytes. @item tcp_n

Re: [FFmpeg-devel] [PATCH 05/13] avformat/mxfenc: Fix stored width

mån 2018-05-07 klockan 12:38 +0200 skrev Michael Niedermayer: > This fixes the width to have computations matching the height > > > Signed-off-by: Michael Niedermayer > --- >  libavformat/mxfenc.c  |   3 +- >  .../ref/fate/concat-demuxer-extended-lavf-mxf |   2 +- >  .../f

Re: [FFmpeg-devel] [PATCH 06/13] avformat/mxfenc: Add Sample width/height/x offset/y offset, Display x offset and F2 offset

mån 2018-05-07 klockan 12:38 +0200 skrev Michael Niedermayer: > > Signed-off-by: Michael Niedermayer > +if (sc->interlaced) { > +//Display F2 Offset > +mxf_write_local_tag(pb, 4, 0x3217); > +avio_wb32(pb, -((st->codecpar->height - display_height)&1)); Negative values f

Re: [FFmpeg-devel] [PATCH 07/13] avformat/mxfenc: Add vertical subsampling support

mån 2018-05-07 klockan 12:38 +0200 skrev Michael Niedermayer: > > Signed-off-by: Michael Niedermayer > > @@ -1149,6 +1151,8 @@ static void mxf_write_cdci_common(AVFormatContext *s, > AVStream *st, const UID ke >  desc_size += 5; >  if (sc->interlaced) >  desc_size += 8; > +  

Re: [FFmpeg-devel] [PATCH 08/13] avformat/mxfenc: add white/black ref /color range

mån 2018-05-07 klockan 12:38 +0200 skrev Michael Niedermayer: > > Signed-off-by: Michael Niedermayer > > +if (st->codecpar->color_range != AVCOL_RANGE_UNSPECIFIED) { > +int black = 0, > +white = (1 +color = (1 +  

[FFmpeg-devel] [PATCH] lavc/amfenc: moving amf common code (library and context) to

This patch moves AMF common parts from amfenc to hwcontext_amf. Now av_hwdevice_ctx API is used for AMF context creation/destroying This patch does not change component behaviour, it contains only restructurization for further patches with new amf components --- sending patch one more time in May

Re: [FFmpeg-devel] [PATCH 09/13] avformat/mxfenc: Add Padding Bits

mån 2018-05-07 klockan 12:38 +0200 skrev Michael Niedermayer: > > Signed-off-by: Michael Niedermayer > @@ -1228,6 +1229,10 @@ static void mxf_write_cdci_common(AVFormatContext *s, > AVStream *st, const UID ke >  mxf_write_local_tag(pb, 1, 0x3303); >  avio_w8(pb, sc->color_siting); >   > +

[FFmpeg-devel] [PATCH v2 0/3] Add TCP MSS setting.

v2: - fix build fail in win32/mingw32/mingw64 - change the tcp_mss docments as Gyan's comments Jun Zhao (3): lavf/network: fix doxygen comments. lavf/tcp: add option to setting Maximum Segment Size doc/protocols: documents tcp_mss doc/protocols.texi| 3 +++ libavformat/network.h |

Re: [FFmpeg-devel] [PATCH 10/13] avformat/mxfenc: Set color siting to 0 for D10-MXF

mån 2018-05-07 klockan 12:38 +0200 skrev Michael Niedermayer: > > Signed-off-by: Michael Niedermayer > --- >  libavformat/mxfenc.c  |   1 + >  .../ref/fate/concat-demuxer-extended-lavf-mxf |   2 +- >  .../fate/concat-demuxer-extended-lavf-mxf_d10 |   2 +- >  .../ref/fate/co

Re: [FFmpeg-devel] [PATCH 12/13] avformat/mxfenc: Add Stored F2 Offset / Image Start/End Offset for D10

mån 2018-05-07 klockan 12:38 +0200 skrev Michael Niedermayer: > > Signed-off-by: Michael Niedermayer >  desc_size += 8; >  if (st->codecpar->color_range != AVCOL_RANGE_UNSPECIFIED) >  desc_size += 8 * 3; > +if (s->oformat == &ff_mxf_d10_muxer) > +desc_size += 8 + 8

[FFmpeg-devel] [PATCH] lavc/amfenc: moving amf common code (library and context) to lavu/hwcontext_amf from amfenc to be reused in other amf components

This patch moves AMF common parts from amfenc to hwcontext_amf. Now av_hwdevice_ctx API is used for AMF context creation/destroying. This patch does not change component behaviour. it contains only restructurization for further patches with new amf components --- sending patch one more time in May

Re: [FFmpeg-devel] [PATCH 13/13] avformat/mxfenc: Write transfer characteristic

mån 2018-05-07 klockan 12:38 +0200 skrev Michael Niedermayer: > > Signed-off-by: Michael Niedermayer > > +static int get_trc(UID ul, enum AVColorTransferCharacteristic trc) > +{ > +switch (trc){ > +case AVCOL_TRC_GAMMA28   : > +case AVCOL_TRC_GAMMA22   : > +memcpy(ul, > (UID)

Re: [FFmpeg-devel] [PATCH 11/13] avformat/mxfenc: Write Audio Ref Level for D10

mån 2018-05-07 klockan 12:38 +0200 skrev Michael Niedermayer: > > Signed-off-by: Michael Niedermayer >  // Generic Sound Essence Descriptor >  { 0x3D02, > {0x06,0x0E,0x2B,0x34,0x01,0x01,0x01,0x04,0x04,0x02,0x03,0x01,0x04,0x00,0x00,0x00}}, > /* Locked/Unlocked */ >  { 0x3D03, > {0x06

[FFmpeg-devel] [PATCH] avfilter: add fftdnoiz filter

Signed-off-by: Paul B Mahol --- configure | 2 + doc/filters.texi | 32 +++ libavfilter/Makefile | 1 + libavfilter/allfilters.c | 1 + libavfilter/vf_fftdnoiz.c | 691 ++ 5 files changed, 727 insertions(+) create

[FFmpeg-devel] [PATCH 5/6] avutil/hwcontext_cuda: explicitly synchronize cuMemcpy calls

--- libavutil/hwcontext_cuda.c | 18 +++--- 1 file changed, 15 insertions(+), 3 deletions(-) diff --git a/libavutil/hwcontext_cuda.c b/libavutil/hwcontext_cuda.c index 8024eec79d..7f55881ba5 100644 --- a/libavutil/hwcontext_cuda.c +++ b/libavutil/hwcontext_cuda.c @@ -267,13 +267,19 @@

[FFmpeg-devel] [PATCH 4/6] avcodec/nvdec: pass CUstream in vpp parameters

--- libavcodec/nvdec.c | 7 ++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/libavcodec/nvdec.c b/libavcodec/nvdec.c index d98f9dd95e..c5c5c83dc7 100644 --- a/libavcodec/nvdec.c +++ b/libavcodec/nvdec.c @@ -39,6 +39,7 @@ typedef struct NVDECDecoder { AVBufferRef *hw_devi

[FFmpeg-devel] [PATCH 1/6] avutil/hwcontext_cuda: add AVCUDAFramesContext and AVCUDAFramesContext.flags

Frames can be mapped from nvdec/cuvid, not needing any actual memory allocation, but all other features of the hw_frames_ctx. Hence the dummy-mode, which does not allocate any (notable amounts of) memory but otherwise behaves the exact same. --- doc/APIchanges | 3 +++ libavutil/hwcon

[FFmpeg-devel] [PATCH 6/6] avcodec/cuviddec: explicitly synchronize cuMemcpy calls

--- libavcodec/cuviddec.c | 6 +- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/libavcodec/cuviddec.c b/libavcodec/cuviddec.c index 122c28f6e8..4d3caf924e 100644 --- a/libavcodec/cuviddec.c +++ b/libavcodec/cuviddec.c @@ -550,12 +550,16 @@ static int cuvid_output_frame(AVCodecCo

[FFmpeg-devel] [PATCH 2/6] avcodec/nvdec: avoid needless copy of output frame

Replaces the data pointers with the mapped cuvid ones. Adds buffer_refs to the frame to ensure the needed contexts stay alive and the cuvid idx stays allocated. Adds another buffer_ref to unmap the frame when it's unreferenced itself. --- libavcodec/nvdec.c | 83 +--

[FFmpeg-devel] [PATCH 3/6] avutil/hwcontext_cuda: add CUstream in cuda hwctx

--- configure | 6 -- doc/APIchanges | 3 ++- libavutil/hwcontext_cuda.c | 3 +++ libavutil/hwcontext_cuda.h | 1 + libavutil/version.h| 2 +- 5 files changed, 11 insertions(+), 4 deletions(-) diff --git a/configure b/configure index 7c143238a8..cae8a235a4

Re: [FFmpeg-devel] [FFmpeg-devel, RFC] lavfi: add opencl tonemap filter.

Hello Ruiling, > Thanks for sharing your idea with me. I basically also noticed some poor > quality tone mapping result for some hdr stream. > I will try your suggestion to see whether I can make it in good state so I > can include it in next version. > In fact I have not thought detecting scene

[FFmpeg-devel] [PATCH] Revert "configure: check that the required header for Linux Perf is available"

This reverts commit 234a5e08313c6b8b0774796dfa1f285a3f877d14. Linux perf is not enabled if you set --target=android, which you are supposed to when building for Android. --- configure | 3 --- 1 file changed, 3 deletions(-) diff --git a/configure b/configure index 7c143238a8..bdca93a9ee 100755 -

Re: [FFmpeg-devel] [PATCH] Revert "configure: check that the required header for Linux Perf is available"

On 5/8/2018 12:21 PM, Clément Bœsch wrote: > This reverts commit 234a5e08313c6b8b0774796dfa1f285a3f877d14. > > Linux perf is not enabled if you set --target=android, which you are Isn't it --target-os=android? > supposed to when building for Android. > --- > configure | 3 --- > 1 file changed,

Re: [FFmpeg-devel] [PATCH 2/6] avcodec/nvdec: avoid needless copy of output frame

On Tue, 8 May 2018 15:31:28 +0200 Timo Rothenpieler wrote: > Replaces the data pointers with the mapped cuvid ones. > Adds buffer_refs to the frame to ensure the needed contexts stay alive > and the cuvid idx stays allocated. > Adds another buffer_ref to unmap the frame when it's unreferenced it

Re: [FFmpeg-devel] [PATCH] avutil/hwcontext_cuda: add AVCUDAFramesContext and AVCUDAFramesContext.flags

On Mon, 7 May 2018 23:46:48 +0200 Timo Rothenpieler wrote: > Frames can be mapped from nvdec/cuvid, not needing any actual memory > allocation, but all other features of the hw_frames_ctx. > Hence the dummy-mode, which does not allocate any (notable amounts of) > memory but otherwise behaves the

Re: [FFmpeg-devel] [PATCH 3/6] avutil/hwcontext_cuda: add CUstream in cuda hwctx

On Tue, 8 May 2018 15:31:29 +0200 Timo Rothenpieler wrote: > --- > configure | 6 -- > doc/APIchanges | 3 ++- > libavutil/hwcontext_cuda.c | 3 +++ > libavutil/hwcontext_cuda.h | 1 + > libavutil/version.h| 2 +- > 5 files changed, 11 insertions(+), 4 d

[FFmpeg-devel] [PATCH] lavfi/vf_srcnn: use avio_check instead of access

The filter uses avio for file access already, and avio_check is portable. Fixes trac #7192. --- libavfilter/vf_srcnn.c | 5 + 1 file changed, 1 insertion(+), 4 deletions(-) diff --git a/libavfilter/vf_srcnn.c b/libavfilter/vf_srcnn.c index dace2b99b5..edffebb278 100644 --- a/libavfilter/vf_s

Re: [FFmpeg-devel] [PATCH] avutil/hwcontext_cuda: add AVCUDAFramesContext and AVCUDAFramesContext.flags

-frame->buf[0] = av_buffer_pool_get(ctx->pool); +if (frctx->flags & AV_CUDA_HWFRAMES_DUMMY_MODE) +frame->buf[0] = av_buffer_create(NULL, 0, NULL, NULL, 0); +else +frame->buf[0] = av_buffer_pool_get(ctx->pool); + Is this really needed? Because at least videotoolbox als

Re: [FFmpeg-devel] [PATCH] lavfi/vf_srcnn: use avio_check instead of access

On 5/8/2018 12:27 PM, Hendrik Leppkes wrote: > The filter uses avio for file access already, and avio_check is > portable. > > Fixes trac #7192. Remove the access dep in configure while at it. LGTM is confirmed working. > --- > libavfilter/vf_srcnn.c | 5 + > 1 file changed, 1 insertion(+)

Re: [FFmpeg-devel] [PATCH 3/6] avutil/hwcontext_cuda: add CUstream in cuda hwctx

Am 08.05.2018 um 17:26 schrieb wm4: On Tue, 8 May 2018 15:31:29 +0200 Timo Rothenpieler wrote: --- configure | 6 -- doc/APIchanges | 3 ++- libavutil/hwcontext_cuda.c | 3 +++ libavutil/hwcontext_cuda.h | 1 + libavutil/version.h| 2 +- 5 fil

Re: [FFmpeg-devel] [PATCH 02/13] avformat/mxfenc: Bump minor versions for S377-1-2009

On Mon, May 07, 2018 at 06:27:37PM +0200, Carl Eugen Hoyos wrote: > 2018-05-07 12:38 GMT+02:00, Michael Niedermayer : > > > diff --git a/libavformat/mxfenc.c b/libavformat/mxfenc.c > > index c0db10b3c2..00dfce977b 100644 > > --- a/libavformat/mxfenc.c > > +++ b/libavformat/mxfenc.c > > @@ -691,7 +

Re: [FFmpeg-devel] [PATCH] avutil/hwcontext_cuda: add AVCUDAFramesContext and AVCUDAFramesContext.flags

On Tue, 8 May 2018 17:43:49 +0200 Timo Rothenpieler wrote: > >> -frame->buf[0] = av_buffer_pool_get(ctx->pool); > >> +if (frctx->flags & AV_CUDA_HWFRAMES_DUMMY_MODE) > >> +frame->buf[0] = av_buffer_create(NULL, 0, NULL, NULL, 0); > >> +else > >> +frame->buf[0] = av_buf

Re: [FFmpeg-devel] [PATCH 2/6] avcodec/nvdec: avoid needless copy of output frame

Am 08.05.2018 um 17:25 schrieb wm4: On Tue, 8 May 2018 15:31:28 +0200 Timo Rothenpieler wrote: Replaces the data pointers with the mapped cuvid ones. Adds buffer_refs to the frame to ensure the needed contexts stay alive and the cuvid idx stays allocated. Adds another buffer_ref to unmap the

Re: [FFmpeg-devel] [PATCH 03/13] avformat/mxfenc: Add Product Version, Toolkit version and Platform

On Tue, May 08, 2018 at 12:28:48PM +0200, Tomas Härdin wrote: > mån 2018-05-07 klockan 12:38 +0200 skrev Michael Niedermayer: > > > Signed-off-by: Michael Niedermayer > >  static void mxf_write_identification(AVFormatContext *s) > >  { > >  MXFContext *mxf = s->priv_data; > > @@ -790,7 +809,7

Re: [FFmpeg-devel] [PATCH 3/6] avutil/hwcontext_cuda: add CUstream in cuda hwctx

On Tue, 8 May 2018 17:48:21 +0200 Timo Rothenpieler wrote: > Am 08.05.2018 um 17:26 schrieb wm4: > > On Tue, 8 May 2018 15:31:29 +0200 > > Timo Rothenpieler wrote: > > > >> --- > >> configure | 6 -- > >> doc/APIchanges | 3 ++- > >> libavutil/hwcontext_c

Re: [FFmpeg-devel] [PATCH 04/13] avformat/mxfenc: Add object model version

On Tue, May 08, 2018 at 12:32:17PM +0200, Tomas Härdin wrote: > mån 2018-05-07 klockan 12:38 +0200 skrev Michael Niedermayer: > > > Signed-off-by: Michael Niedermayer > > >  mxf_write_metadata_key(pb, 0x012f00); > >  PRINT_KEY(s, "preface key", pb->buf_ptr - 16); > > -klv_encode_ber_l

Re: [FFmpeg-devel] [PATCH 01/13] avformat/mxfenc: Correct KAG alignment of preface

On Tue, May 08, 2018 at 12:20:58PM +0200, Tomas Härdin wrote: > mån 2018-05-07 klockan 12:38 +0200 skrev Michael Niedermayer: > > Signed-off-by: Michael Niedermayer > > --- > >  libavformat/mxfenc.c|  1 + > >  tests/ref/fate/copy-trac4914|  4 ++-- > >  tests/ref/fate/mxf-reel_name 

Re: [FFmpeg-devel] [PATCH] lavfi/vf_srcnn: use avio_check instead of access

On Tue, May 8, 2018 at 5:47 PM, James Almer wrote: > On 5/8/2018 12:27 PM, Hendrik Leppkes wrote: >> The filter uses avio for file access already, and avio_check is >> portable. >> >> Fixes trac #7192. > > Remove the access dep in configure while at it. > > LGTM is confirmed working. > Amended, t

Re: [FFmpeg-devel] [PATCH 05/13] avformat/mxfenc: Fix stored width

On Tue, May 08, 2018 at 12:37:08PM +0200, Tomas Härdin wrote: > mån 2018-05-07 klockan 12:38 +0200 skrev Michael Niedermayer: > > This fixes the width to have computations matching the height > > > > > Signed-off-by: Michael Niedermayer > > --- > >  libavformat/mxfenc.c  |

Re: [FFmpeg-devel] [PATCH 06/13] avformat/mxfenc: Add Sample width/height/x offset/y offset, Display x offset and F2 offset

On Tue, May 08, 2018 at 12:40:49PM +0200, Tomas Härdin wrote: > mån 2018-05-07 klockan 12:38 +0200 skrev Michael Niedermayer: > > > Signed-off-by: Michael Niedermayer > > +if (sc->interlaced) { > > +//Display F2 Offset > > +mxf_write_local_tag(pb, 4, 0x3217); > > +avio_

Re: [FFmpeg-devel] [PATCH] avutil/hwcontext_cuda: add AVCUDAFramesContext and AVCUDAFramesContext.flags

Am 08.05.2018 um 17:49 schrieb wm4: On Tue, 8 May 2018 17:43:49 +0200 Timo Rothenpieler wrote: -frame->buf[0] = av_buffer_pool_get(ctx->pool); +if (frctx->flags & AV_CUDA_HWFRAMES_DUMMY_MODE) +frame->buf[0] = av_buffer_create(NULL, 0, NULL, NULL, 0); +else +frame->b

Re: [FFmpeg-devel] [PATCH 3/6] avutil/hwcontext_cuda: add CUstream in cuda hwctx

What is this? https://docs.nvidia.com/cuda/cuda-driver-api/stream-sync-behavior.html It allows asynchronous processing of CUDA workloads. The next couple patches make use of it. There's no change in behaviour if it remains unset/NULL, but if you set one, the workload won't block the main CUDA s

Re: [FFmpeg-devel] [PATCH v3] avformat/mxfenc: add h264 profiles

2018-05-07 10:40 GMT+02:00 Tomas Härdin : > sön 2018-05-06 klockan 21:31 +0200 skrev Thomas Mundt: > > 2018-05-06 13:32 GMT+02:00 Tomas Härdin : > > > > > fre 2018-05-04 klockan 01:52 +0200 skrev Thomas Mundt: > > > > Hi, > > > > > > > > this is a better version of the patch. > > > > 10 bit and TF

Re: [FFmpeg-devel] [PATCH 07/13] avformat/mxfenc: Add vertical subsampling support

On Tue, May 08, 2018 at 12:45:18PM +0200, Tomas Härdin wrote: > mån 2018-05-07 klockan 12:38 +0200 skrev Michael Niedermayer: > > > Signed-off-by: Michael Niedermayer > > > > @@ -1149,6 +1151,8 @@ static void mxf_write_cdci_common(AVFormatContext *s, > > AVStream *st, const UID ke > >  d

Re: [FFmpeg-devel] [PATCH 08/13] avformat/mxfenc: add white/black ref /color range

On Tue, May 08, 2018 at 12:47:35PM +0200, Tomas Härdin wrote: > mån 2018-05-07 klockan 12:38 +0200 skrev Michael Niedermayer: > > > Signed-off-by: Michael Niedermayer > > > > +if (st->codecpar->color_range != AVCOL_RANGE_UNSPECIFIED) { > > +int black = 0, > > +white = (1

Re: [FFmpeg-devel] [PATCH 09/13] avformat/mxfenc: Add Padding Bits

On Tue, May 08, 2018 at 12:50:16PM +0200, Tomas Härdin wrote: > mån 2018-05-07 klockan 12:38 +0200 skrev Michael Niedermayer: > > > Signed-off-by: Michael Niedermayer > > @@ -1228,6 +1229,10 @@ static void mxf_write_cdci_common(AVFormatContext > > *s, AVStream *st, const UID ke > >  mxf_write

Re: [FFmpeg-devel] [PATCH 10/13] avformat/mxfenc: Set color siting to 0 for D10-MXF

On Tue, May 08, 2018 at 12:55:18PM +0200, Tomas Härdin wrote: > mån 2018-05-07 klockan 12:38 +0200 skrev Michael Niedermayer: > > > Signed-off-by: Michael Niedermayer > > --- > >  libavformat/mxfenc.c  |   1 + > >  .../ref/fate/concat-demuxer-extended-lavf-mxf |   2 +- > >

Re: [FFmpeg-devel] [PATCH 11/13] avformat/mxfenc: Write Audio Ref Level for D10

On Tue, May 08, 2018 at 01:02:19PM +0200, Tomas Härdin wrote: > mån 2018-05-07 klockan 12:38 +0200 skrev Michael Niedermayer: > > > Signed-off-by: Michael Niedermayer > >  // Generic Sound Essence Descriptor > >  { 0x3D02, > > {0x06,0x0E,0x2B,0x34,0x01,0x01,0x01,0x04,0x04,0x02,0x03,0x01,0

[FFmpeg-devel] [PATCH] lavd/v4l2: Add ARGB and XRGB packed pixel formats

Formats ARGB32, XRGB32, ABGR32, and XBGR32 were added to V4L2 instead of ill-defined deprecated RGB32/BGR32 pixel formats. When pixel format is not specified explicitly FFmpeg tries formats in order in which they are stored in the table. Therefore formats are sorted as follows: BGR is preferred ov

Re: [FFmpeg-devel] [PATCH 12/13] avformat/mxfenc: Add Stored F2 Offset / Image Start/End Offset for D10

On Tue, May 08, 2018 at 12:58:09PM +0200, Tomas Härdin wrote: > mån 2018-05-07 klockan 12:38 +0200 skrev Michael Niedermayer: > > > Signed-off-by: Michael Niedermayer > >  desc_size += 8; > >  if (st->codecpar->color_range != AVCOL_RANGE_UNSPECIFIED) > >  desc_size += 8 * 3; >

[FFmpeg-devel] [PATCH 2/3] libavcodec: v4l2m2m: output AVDRMFrameDescriptor

This is V2 of my previous patch. I have worked together with Jorge to get this working properly. We have made it so that AV_PIX_FMT_DRM_PRIME output can be selected by setting avctx->pix_fmt. This allows v4l2 to export the buffer so we can use it for zero-copy. If AV_PIX_FMT_DRM_PRIME is not sel

[FFmpeg-devel] [PATCH 3/3] libavcodec: v4l2m2m: fix error handling during buffer init

From: Jorge Ramirez-Ortiz Signed-off-by: Jorge Ramirez-Ortiz --- libavcodec/v4l2_context.c | 19 --- libavcodec/v4l2_m2m_dec.c | 11 --- 2 files changed, 24 insertions(+), 6 deletions(-) diff --git a/libavcodec/v4l2_context.c b/libavcodec/v4l2_context.c index 9457fadb1e

[FFmpeg-devel] [PATCH 1/3] libavcodec: v4l2m2m: fix indentation and add M2MDEC_CLASS

This is just some formatting that is taken from the rkmpp decoder. I find that this make is more readable. --- libavcodec/v4l2_m2m_dec.c | 44 --- 1 file changed, 23 insertions(+), 21 deletions(-) diff --git a/libavcodec/v4l2_m2m_dec.c b/libavcodec/v4l2_m2m_de

[FFmpeg-devel] [PATCH 1/5] avcodec/nvdec: avoid needless copy of output frame

Replaces the data pointers with the mapped cuvid ones. Adds buffer_refs to the frame to ensure the needed contexts stay alive and the cuvid idx stays allocated. Adds another buffer_ref to unmap the frame when it's unreferenced itself. --- libavcodec/nvdec.c | 96 +++

[FFmpeg-devel] [PATCH 2/5] avutil/hwcontext_cuda: add CUstream in cuda hwctx

--- configure | 6 -- doc/APIchanges | 3 +++ libavutil/hwcontext_cuda.c | 3 +++ libavutil/hwcontext_cuda.h | 1 + libavutil/version.h| 2 +- 5 files changed, 12 insertions(+), 3 deletions(-) diff --git a/configure b/configure index 6626111ff2..9743de05d0

[FFmpeg-devel] [PATCH 4/5] avutil/hwcontext_cuda: explicitly synchronize cuMemcpy calls

--- libavutil/hwcontext_cuda.c | 18 +++--- 1 file changed, 15 insertions(+), 3 deletions(-) diff --git a/libavutil/hwcontext_cuda.c b/libavutil/hwcontext_cuda.c index f3e81680e9..668293bffc 100644 --- a/libavutil/hwcontext_cuda.c +++ b/libavutil/hwcontext_cuda.c @@ -258,13 +258,19 @@

[FFmpeg-devel] [PATCH 5/5] avcodec/cuviddec: explicitly synchronize cuMemcpy calls

--- libavcodec/cuviddec.c | 6 +- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/libavcodec/cuviddec.c b/libavcodec/cuviddec.c index 122c28f6e8..4d3caf924e 100644 --- a/libavcodec/cuviddec.c +++ b/libavcodec/cuviddec.c @@ -550,12 +550,16 @@ static int cuvid_output_frame(AVCodecCo

[FFmpeg-devel] [PATCH 3/5] avcodec/nvdec: pass CUstream in vpp parameters

--- libavcodec/nvdec.c | 7 ++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/libavcodec/nvdec.c b/libavcodec/nvdec.c index 502747adfd..e779be3a45 100644 --- a/libavcodec/nvdec.c +++ b/libavcodec/nvdec.c @@ -39,6 +39,7 @@ typedef struct NVDECDecoder { AVBufferRef *hw_devi

[FFmpeg-devel] [PATCH] avformat/mpegts: add skip_unknown_pids option

From: Aman Gupta Allows the user to skip streams that are not advertised in the PMT. --- libavformat/mpegts.c | 7 ++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/libavformat/mpegts.c b/libavformat/mpegts.c index 629631f60a..5464f48a8d 100644 --- a/libavformat/mpegts.c +++ b

[FFmpeg-devel] [PATCH] ffprobe: fix SEGV when new streams are added

From: Aman Gupta --- fftools/ffprobe.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/fftools/ffprobe.c b/fftools/ffprobe.c index 8b2a18b6b1..544786ec72 100644 --- a/fftools/ffprobe.c +++ b/fftools/ffprobe.c @@ -2371,11 +2371,11 @@ static int read_interval_packets(Writer

Re: [FFmpeg-devel] [PATCH 1/5] avcodec/nvdec: avoid needless copy of output frame

On 2018-05-08 11:36, Timo Rothenpieler wrote: Replaces the data pointers with the mapped cuvid ones. Adds buffer_refs to the frame to ensure the needed contexts stay alive and the cuvid idx stays allocated. Adds another buffer_ref to unmap the frame when it's unreferenced itself. --- libavcode

[FFmpeg-devel] [PATCH 1/2] avcodec/cbs_h2645: use simple data buffers for some parameter set extensions

There's no gain from using AVBufferRef for these, as no copies or references are ever made. Signed-off-by: James Almer --- There is however a raw copy of the struct storing these buffers, which is dangerous and fragile. This patch is in preparation to change how the above is handled. libavcodec

[FFmpeg-devel] [PATCH 2/2] avcodec/cbs_h2645: use AVBufferRef to store list of active parameter sets

This is more in line to how h264_ps handles it. Signed-off-by: James Almer --- Not really going to make much of a difference seeing parameter sets rarely show up in-band even in raw streams, but it's one less memcpy anyway. libavcodec/cbs_h264.h | 4 +-- libavcodec/cbs_h2645.c

Re: [FFmpeg-devel] [PATCH 2/5] avutil/hwcontext_cuda: add CUstream in cuda hwctx

On 5/8/2018 3:36 PM, Timo Rothenpieler wrote: > --- > configure | 6 -- > doc/APIchanges | 3 +++ > libavutil/hwcontext_cuda.c | 3 +++ > libavutil/hwcontext_cuda.h | 1 + > libavutil/version.h| 2 +- > 5 files changed, 12 insertions(+), 3 deletions(-) > >

Re: [FFmpeg-devel] [PATCH 13/13] avformat/mxfenc: Write transfer characteristic

On Tue, May 08, 2018 at 01:00:51PM +0200, Tomas Härdin wrote: > mån 2018-05-07 klockan 12:38 +0200 skrev Michael Niedermayer: > > > Signed-off-by: Michael Niedermayer > > > > +static int get_trc(UID ul, enum AVColorTransferCharacteristic trc) > > +{ > > +switch (trc){ > > +case AVCOL_TRC_

Re: [FFmpeg-devel] [PATCH 3/6] avutil/hwcontext_cuda: add CUstream in cuda hwctx

Hi! Responding the per device question (sorry I can't make it shorter, the topic is quite dense). A typical CUDA application uses a single cuda context, and multiple cuda streams to allow asynchronicity between cuda tasks (memory transfers, kernels, memsets) and make overlapping between those

Re: [FFmpeg-devel] [PATCH 1/3] vaapi_encode: Initialize the pointer

On 08/05/18 03:35, Xiang, Haihao wrote: > On Mon, 2018-05-07 at 21:46 +0100, Mark Thompson wrote: >> On 04/05/18 15:41, Haihao Xiang wrote: >>> Otherwise it might use unitialized last_pic in av_assert0(last_pic) >>> >>> Signed-off-by: Haihao Xiang >>> --- >>> libavcodec/vaapi_encode.c | 2 +- >>>

Re: [FFmpeg-devel] [PATCH 2/3] hwcontext_vaapi: Return error if can not find a VA RT format

On 08/05/18 03:53, Xiang, Haihao wrote: > On Mon, 2018-05-07 at 21:48 +0100, Mark Thompson wrote: >> On 04/05/18 15:41, Haihao Xiang wrote: >>> Otherwise va_rt_format might be unitialized >>> >>> Signed-off-by: Haihao Xiang >>> --- >>> libavutil/hwcontext_vaapi.c | 5 + >>> 1 file changed, 5

Re: [FFmpeg-devel] [PATCH 2/4] vaapi_encode_h265: Insert mastering display colour colume if needed

On 08/05/18 04:54, Xiang, Haihao wrote: > On Mon, 2018-05-07 at 22:03 +0100, Mark Thompson wrote: >> On 04/05/18 09:54, Xiang, Haihao wrote: >>> On Thu, 2018-05-03 at 22:43 +0100, Mark Thompson wrote: On 03/05/18 04:07, Haihao Xiang wrote: > '-sei xxx' is added to control SEI insertion, so

Re: [FFmpeg-devel] [PATCH] lavd/v4l2: Add ARGB and XRGB packed pixel formats

2018-05-08 15:11 GMT+02:00, Anton Leontiev : > Formats ARGB32, XRGB32, ABGR32, and XBGR32 were added to > V4L2 instead of ill-defined deprecated RGB32/BGR32 pixel formats. Just curious because I am quite sure I tested them: Why / how are they ill-defined? [...] > diff --git a/libavdevice/v4l2-co

Re: [FFmpeg-devel] fate/hap : add test for hap encoding

2018-05-08 11:45 GMT+02:00, Martin Vignali : >> >> > Sorry, i missread the code before writing the test, and >> > didn't see it use float for one step. >> > I will send a patch to remove these tests. >> >> Or just commit. >> >> > Tests removed. Thank you! Carl Eugen __

Re: [FFmpeg-devel] [PATCH 1/2] avcodec/cbs_h2645: use simple data buffers for some parameter set extensions

On 08/05/18 21:48, James Almer wrote: > There's no gain from using AVBufferRef for these, as no copies or > references are ever made. > > Signed-off-by: James Almer > --- > There is however a raw copy of the struct storing these buffers, > which is dangerous and fragile. > This patch is in prepar

[FFmpeg-devel] [PATCH] avformat/mpegts: skip non-PMT tids earlier

From: Aman Gupta This mimics the logic flow in all the other callbacks (pat_cb, sdt_cb, m4sl_cb), and avoids calling skip_identical() for non PMT_TID packets. Since skip_identical modifies internal state like MpegTSSectionFilter.last_ver, this change prevents unnecessary reprocessing on some str

Re: [FFmpeg-devel] [PATCH 1/2] avcodec/cbs_h2645: use simple data buffers for some parameter set extensions

On 5/8/2018 7:17 PM, Mark Thompson wrote: > On 08/05/18 21:48, James Almer wrote: >> There's no gain from using AVBufferRef for these, as no copies or >> references are ever made. >> >> Signed-off-by: James Almer >> --- >> There is however a raw copy of the struct storing these buffers, >> which i

Re: [FFmpeg-devel] [PATCH 2/2] avcodec/cbs_h2645: use AVBufferRef to store list of active parameter sets

On 08/05/18 21:48, James Almer wrote: > This is more in line to how h264_ps handles it. > > Signed-off-by: James Almer > --- > Not really going to make much of a difference seeing parameter sets > rarely show up in-band even in raw streams, but it's one less memcpy > anyway. > > libavcodec/cbs_

Re: [FFmpeg-devel] [PATCH 1/2] avcodec/cbs_h2645: use simple data buffers for some parameter set extensions

On 08/05/18 23:36, James Almer wrote: > On 5/8/2018 7:17 PM, Mark Thompson wrote: >> On 08/05/18 21:48, James Almer wrote: >>> There's no gain from using AVBufferRef for these, as no copies or >>> references are ever made. >>> >>> Signed-off-by: James Almer >>> --- >>> There is however a raw copy

Re: [FFmpeg-devel] libavutil/encryption_info: Allow multiple init info.

On Mon, May 07, 2018 at 04:59:33PM -0700, Jacob Trimble wrote: > On Mon, May 7, 2018 at 3:18 PM, Michael Niedermayer > wrote: > > On Mon, Apr 23, 2018 at 11:03:57AM -0700, Jacob Trimble wrote: > >> While integrating my encryption info changes, I noticed a problem with > >> the init info structs.

Re: [FFmpeg-devel] [PATCH 2/2] avcodec/cbs_h2645: use AVBufferRef to store list of active parameter sets

On 5/8/2018 7:37 PM, Mark Thompson wrote: > On 08/05/18 21:48, James Almer wrote: >> This is more in line to how h264_ps handles it. >> >> Signed-off-by: James Almer >> --- >> Not really going to make much of a difference seeing parameter sets >> rarely show up in-band even in raw streams, but it'

[FFmpeg-devel] [PATCH v2] avcodec/cbs_h2645: use AVBufferRef to store list of active parameter sets

Removes unnecessary data copies, and fixes potential issues with dangling references held in said lists. Signed-off-by: James Almer --- libavcodec/cbs_h264.h | 2 ++ libavcodec/cbs_h2645.c | 41 + libavcodec/cbs_h265.h | 3 +++ 3 files changed, 26 inse

Re: [FFmpeg-devel] [PATCH 2/3] libavcodec: v4l2m2m: output AVDRMFrameDescriptor

On 08/05/18 19:24, Lukas Rusak wrote: > This is V2 of my previous patch. I have worked together with Jorge to get > this working properly. > We have made it so that AV_PIX_FMT_DRM_PRIME output can be selected by > setting avctx->pix_fmt. This isn't how format selection in libavcodec works. Look

Re: [FFmpeg-devel] [PATCH v2] avcodec/cbs_h2645: use AVBufferRef to store list of active parameter sets

On 09/05/18 00:12, James Almer wrote: > Removes unnecessary data copies, and fixes potential issues with > dangling references held in said lists. > > Signed-off-by: James Almer > --- > libavcodec/cbs_h264.h | 2 ++ > libavcodec/cbs_h2645.c | 41 + > lib

Re: [FFmpeg-devel] [PATCH v2] avcodec/cbs_h2645: use AVBufferRef to store list of active parameter sets

On 5/8/2018 8:49 PM, Mark Thompson wrote: > Ah. I should have thought of this problem earlier - testing this patch, it > this breaks the VAAPI encoders because the parameter set structures there are > not reference counted (and shouldn't be - they are pointers to substructures > of the encoder

[FFmpeg-devel] [PATCH v3] avcodec/cbs_h2645: use AVBufferRef to store list of active parameter sets

Removes unnecessary data copies, and partially fixes potential issues with dangling references held in said lists. Signed-off-by: James Almer --- libavcodec/cbs_h264.h | 2 ++ libavcodec/cbs_h2645.c | 46 -- libavcodec/cbs_h265.h | 3 +++ 3 files chang

Re: [FFmpeg-devel] [PATCH] avformat/mpegts: add skip_unknown_pids option

On Tue, May 8, 2018 at 12:54 PM, Aman Gupta wrote: > From: Aman Gupta > > Allows the user to skip streams that are not advertised in the PMT. > Disregard this patch. It combines two different things (skip pmt, skip streams without pmt) into one badly named option. Aman > --- > libavformat/m

Re: [FFmpeg-devel] [PATCH 1/3] vaapi_encode: Initialize the pointer

On Tue, 2018-05-08 at 22:34 +0100, Mark Thompson wrote: > On 08/05/18 03:35, Xiang, Haihao wrote: > > On Mon, 2018-05-07 at 21:46 +0100, Mark Thompson wrote: > > > On 04/05/18 15:41, Haihao Xiang wrote: > > > > Otherwise it might use unitialized last_pic in av_assert0(last_pic) > > > > > > > > Sig

Re: [FFmpeg-devel] [PATCH 2/3] hwcontext_vaapi: Return error if can not find a VA RT format

On Tue, 2018-05-08 at 22:37 +0100, Mark Thompson wrote: > On 08/05/18 03:53, Xiang, Haihao wrote: > > On Mon, 2018-05-07 at 21:48 +0100, Mark Thompson wrote: > > > On 04/05/18 15:41, Haihao Xiang wrote: > > > > Otherwise va_rt_format might be unitialized > > > > > > > > Signed-off-by: Haihao Xiang

Re: [FFmpeg-devel] [PATCH 2/4] vaapi_encode_h265: Insert mastering display colour colume if needed

On Tue, 2018-05-08 at 22:51 +0100, Mark Thompson wrote: > On 08/05/18 04:54, Xiang, Haihao wrote: > > On Mon, 2018-05-07 at 22:03 +0100, Mark Thompson wrote: > > > On 04/05/18 09:54, Xiang, Haihao wrote: > > > > On Thu, 2018-05-03 at 22:43 +0100, Mark Thompson wrote: > > > > > On 03/05/18 04:07, Ha

[FFmpeg-devel] [PATCH 1/2] avcodec/vaapi: slice_vertical_position starts from zero for the second field

Contrary to VC-1 spec, VAAPI expects the row address of the first macroblock row in the first slice to start from zero for the second field in a field interlaced picture. Signed-off-by: Jerome Borsboom --- This patch set adds support for hardware decoding multi-slice field interlaced pictures. Wi

  1   2   >