> -Original Message-
> From: ffmpeg-devel On Behalf Of Zhao
> Zhili
> Sent: Sunday, April 28, 2024 2:47 PM
> To: ffmpeg-devel@ffmpeg.org
> Cc: Zhao Zhili
> Subject: [FFmpeg-devel] [PATCH WIP v2 1/9] avfilter/dnn: Refactor DNN
> parameter configuration system
>
> From: Zhao Zhili
> +
>
> > On Apr 29, 2024, at 18:29, Guo, Yejun
> wrote:
> >
> >
> >
> >> -Original Message-
> >> From: ffmpeg-devel On Behalf Of
> Zhao
> >> Zhili
> >> Sent: Sunday, April 28, 2024 6:55 PM
> >> To: FFmpeg development discussions and patches >> de...@ffmpeg.org>
> >> Subject: Re:
> > On Apr 28, 2024, at 18:58, Paul B Mahol wrote:
> >
> > Extremely low quality filters, both in source code quality and
> > performance/security and output quality should be queued for removal.
These filters cannot be removed because there are users using them.
What are your suggestions to
> >> -Original Message-
> >> From: ffmpeg-devel On Behalf Of
> >> wenbin.chen-at-intel@ffmpeg.org
> >> Sent: Monday, March 25, 2024 10:15 AM
> >> To: ffmpeg-devel@ffmpeg.org
> >> Subject: [FFmpeg-devel] [PATCH v2] doc: Add libtoch backend option to
> >> dnn_processing
>
> Typo in
> On date Wednesday 2024-03-20 16:01:36 +0800, wenbin.chen-at-
> intel@ffmpeg.org wrote:
> > From: Wenbin Chen
> >
> > Signed-off-by: Wenbin Chen
> > ---
> > Changelog | 1 +
> > 1 file changed, 1 insertion(+)
> >
> > diff --git a/Changelog b/Changelog
> > index e3ca52430c..d0c41887f3
> > -Original Message-
> > From: ffmpeg-devel On Behalf Of
> > wenbin.chen-at-intel@ffmpeg.org
> > Sent: Monday, March 11, 2024 1:02 PM
> > To: ffmpeg-devel@ffmpeg.org
> > Subject: [FFmpeg-devel] [PATCH v5] libavfi/dnn: add LibTorch as one of DNN
> > backend
> >
> > From: Wenbin Chen
>
> > On Feb 20, 2024, at 7:07 PM, wenbin.chen-at-intel@ffmpeg.org wrote:
> >
> > From: Wenbin Chen
> >
> > PyTorch is an open source machine learning framework that accelerates
> > the path from research prototyping to production deployment. Official
> > website: https://pytorch.org/. We
> Hello,
>
> On Tue, 20 Feb 2024, at 05:48, wenbin.chen-at-intel@ffmpeg.org wrote:
> > From: Wenbin Chen
> >
> > PyTorch is an open source machine learning framework that accelerates
>
> OK for me
>
> > the path from research prototyping to production deployment. Official
> > websit:
> > Hello,
> >
> > On Fri, 2 Feb 2024, at 08:26, wenbin.chen-at-intel@ffmpeg.org wrote:
> > > +static void infer_completion_callback(void *args) {
> > > +THRequestItem *request = (THRequestItem*)args;
> > > +LastLevelTaskItem *lltask = request->lltask;
> > > +TaskItem *task =
> Hello,
>
> On Fri, 2 Feb 2024, at 08:26, wenbin.chen-at-intel@ffmpeg.org wrote:
> > +static void infer_completion_callback(void *args) {
> > +THRequestItem *request = (THRequestItem*)args;
> > +LastLevelTaskItem *lltask = request->lltask;
> > +TaskItem *task = lltask->task;
> >
> On Mon, 27 Nov 2023 02:10:11 +0000 "Chen, Wenbin" intel@ffmpeg.org> wrote:
> > > > From: Niklas Haas
> > > >
> > > > This logic was incongruent with logic used elsewhere, where floating
> > > > point formats are explici
> > > From: Niklas Haas
> > >
> > > This logic was incongruent with logic used elsewhere, where floating
> > > point formats are explicitly exempted from range conversion. Fixes an
> > > issue where floating point formats were not going through special
> > > unscaled converters even when it was
> > From: Niklas Haas
> >
> > This logic was incongruent with logic used elsewhere, where floating
> > point formats are explicitly exempted from range conversion. Fixes an
> > issue where floating point formats were not going through special
> > unscaled converters even when it was otherwise
> From: Niklas Haas
>
> This logic was incongruent with logic used elsewhere, where floating
> point formats are explicitly exempted from range conversion. Fixes an
> issue where floating point formats were not going through special
> unscaled converters even when it was otherwise possible.
>
>
> Will apply soon.
>
Hi Niklas:
This patchset causes a regression.
The command: "ffmpeg -i input.png -vf format=grayf32,format=gray8 output.png"
reports error.
If I configure with "--disable-sse2", the error is unseen.
Thanks
Wenbin
> ___
>
> > On Sep 20, 2023, at 10:26, wenbin.chen-at-intel@ffmpeg.org wrote:
> >
> > From: Wenbin Chen
> >
> > Signed-off-by: Wenbin Chen
> > ---
> > libavfilter/dnn/dnn_backend_tf.c | 3 ++-
> > 1 file changed, 2 insertions(+), 1 deletion(-)
> >
> > diff --git a/libavfilter/dnn/dnn_backend_tf.c
>
> wenbin.chen-at-intel@ffmpeg.org:
> > From: Wenbin Chen
> >
> > According to https://github.com/Intel-Media-
> SDK/MediaSDK/blob/master/doc/mediasdk-man.md#configuration-change.
> > Before calling MFXVideoENCODE_Reset, The application needs to retrieve
> > any cached frames in the SDK
> From: Wenbin Chen
>
> When encode RGB frame, Intel driver convert RGB to YUV, so we cannot
> set RGB colorspace to VPL/MSDK.
>
> Signed-off-by: Wenbin Chen
> ---
> libavcodec/qsvenc.c | 7 ++-
> 1 file changed, 6 insertions(+), 1 deletion(-)
>
> diff --git a/libavcodec/qsvenc.c
> On Do, 2022-12-01 at 10:48 +0800, wenbin.chen-at-intel@ffmpeg.org
> wrote:
> > From: Wenbin Chen
> >
> > When process yuv420 frames, FFmpeg uses same alignment on Y/U/V
> > planes. VPL and MSDK use Y plane's pitch / 2 as U/V planes's
> > pitch, which makes U/V planes 16-bytes aligned. We
> On 11/29/2022 5:07 AM, wenbin.chen-at-intel@ffmpeg.org wrote:
> > From: Wenbin Chen
> >
> > When process yuv420 frames, FFmpeg use same alignment on Y/U/V
> > planes. VPL and MSDK use Y plane's pitch / 2 as U/V planes's
> > pitch, which make U/V planes 16-bytes aligned. We need to set
> > a
> From: Haihao Xiang
>
> The current pbc might be small for an obu frame, so a new pbc is
> required then parse this obu frame again. Because
> CodedBitstreamAV1Context has already been updated for this obu frame,
> we
> need to restore CodedBitstreamAV1Context, otherwise
>
> Wenbin Chen:
> > Unset qsv_h264 and qsv_hevc's default settings. Let runtime to decide
> > these parameters, so that it can choose the best parameter and ffmpeg-qsv
> > can keep up with runtime's update.
> >
> > Signed-off-by: Wenbin Chen
> > ---
> > libavcodec/qsvenc_h264.c | 4 ++--
> >
> On Wed, 2022-09-21 at 17:41 +0800, Wenbin Chen wrote:
> > cbs_av1_write_obu() check pbc size after parsing obu frame, and return
> > AVERROR(ENOSPC) if pbc is small. pbc will be reallocated and this obu
> > frame will be parsed again, but this may cause error because
> > CodedBitstreamAV1Context
> Fixes: 005c7a4 ("libavcodec/qsvenc: Add max/min qp reset support in
> qsvenc")
> CC: Wenbin Chen
> Signed-off-by: Dmitry Rogozhkin
> ---
> libavcodec/qsvenc.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/libavcodec/qsvenc.c b/libavcodec/qsvenc.c
> index
> > cbs_av1_write_unit() check pbc size after parsing obu frame, and return
>
> It is cbs_av1_write_obu(), not cbs_av1_write_unit(), right ?
>
> Thanks
> Haihao
>
Sorry, it is typo. cbs_av1_write_obu is assigned to write_unit function pointer.
I will fix it and resubmit patch.
Thanks
Wenbin
> On Thu, 2022-08-18 at 14:59 +0800, Wenbin Chen wrote:
> > Signed-off-by: Wenbin Chen
> > ---
> > doc/encoders.texi | 3 +++
> > libavcodec/qsvenc.c | 18 +-
> > libavcodec/qsvenc.h | 2 ++
> > 3 files changed, 22 insertions(+), 1 deletion(-)
> >
> > diff --git
> > cbs_av1_write_unit() check pbc size after parsing obu frame, and return
> > AVERROR(ENOSPC) if pbc is small. pbc will be reallocated and this obu
> > frame will be parsed again, but this may cause error because
> > CodedBitstreamAV1Context has already been updated, for example
> >
> cbs_av1_write_unit() check pbc size after parsing obu frame, and return
> AVERROR(ENOSPC) if pbc is small. pbc will be reallocated and this obu
> frame will be parsed again, but this may cause error because
> CodedBitstreamAV1Context has already been updated, for example
> ref_order_hint is
> Using parameter from AVCodecContext to reset qsv codec is more suitable
> for MFXVideoENCODE_Reset()'s usage. Per-frame metadata is more suitable
> for the usage of mfxEncodeCtrl being passed to
> MFXVideoENCODE_EncodeFrameAsync(). Now change it to use the value
> from AVCodecContext.
>
>
> > On Jul 7, 2022, at 1:57 PM, Wenbin Chen intel@ffmpeg.org> wrote:
> >
> > Add a line feed to fix a log format issue.
> >
> > Signed-off-by: Wenbin Chen
> > ---
> > libavcodec/qsvenc.c | 1 +
> > 1 file changed, 1 insertion(+)
> >
> > diff --git a/libavcodec/qsvenc.c b/libavcodec/qsvenc.c
>
> Quoting Chen, Wenbin (2022-07-04 08:33:49)
> > > Why is this using frame metadata rather than the AVVideoEncParams
> side
> > > data?
> >
> > The usage of AVVideoEncParams relates to the "qp" variable in
> mfxEncodeCtrl which is passed
> &g
> Quoting Wenbin Chen (2022-06-23 07:32:42)
> > From: Yue Heng
> >
> > Enable dynamic QP configuration in runtime on qsv encoder. Through
> > AVFrame->metadata, we can set key "qsv_config_qp" to change QP
> > configuration when we encode video in CQP mode.
> >
> > Signed-off-by: Yue Heng
> >
> On Fri, Jul 01, 2022 at 01:34:34PM +0800, Wenbin Chen wrote:
> > For 422 frames we should not use hard coded 8 to calculate mb size for
> > uv plane. Chroma shift should be taken into consideration to be
> > compatiple with different sampling format.
> >
> > The error is reported by fate test
> On Fri, Jul 1, 2022 at 10:00 AM Xiang, Haihao <
> haihao.xiang-at-intel@ffmpeg.org> wrote:
>
> > On Fri, 2022-07-01 at 13:34 +0800, Wenbin Chen wrote:
> > > For 422 frames we should not use hard coded 8 to calculate mb size for
> > > uv plane. Chroma shift should be taken into consideration
> On Wed, 2022-06-15 at 15:25 +0800, Wenbin Chen wrote:
> > From: Yue Heng
> >
> > Enable dynamic QP configuration in runtime on qsv encoder. Through
> > AVFrame->metadata, we can set key "qsv_config_qp" to change QP
> > configuration when we encode video in CQP mode.
> >
> > Signed-off-by: Yue
> Use The mfxEncoderCtrl parameter to enable ROI. Get side data
> "AVRegionOfInterest" and use it to configure "mfxExtEncoderROI" which is
> the MediaSDK's ROI configuration.
>
> Signed-off-by: Wenbin Chen
> ---
> libavcodec/qsv_internal.h | 4 ++
> libavcodec/qsvenc.c | 86
>
> On Fri, 2022-05-06 at 13:49 +0800, Wenbin Chen wrote:
> > Use the mfxEncoderCtrl parameter to enable ROI. Get side data
> > "AVRegionOfInterest"
> > from filter "addroi" and use it to configure "mfxExtEncoderROI" which is
> > the MediaSDK's ROI configuration.
> >
> > Signed-off-by: Wenbin Chen
> On Fri, 2022-05-06 at 10:19 +0800, Wenbin Chen wrote:
> > From: Yue Heng
> >
> > To do more accurate QP control, add min/max QP control on I/P/B frame
> > separately to qsv encoder. qmax and qmin still work but newly-added
> > options have higher priority.
> >
> > Signed-off-by: Yue Heng
> >
> On 29/03/2022 09:29, Wenbin Chen wrote:
> > cbs_av1_write_unit() check pbc size after parsing obu frame, and return
> > AVERROR(ENOSPC) if pbc is small. pbc will be reallocated and this obu
> > frame will be parsed again, but this may cause error because
> > CodedBitstreamAV1Context has already
> On Tue, 2022-04-26 at 18:00 +0800, Wenbin Chen wrote:
> > Use h264_sei to parse SEI data got from MediaSDK. Extract frame
> > packing arrangement information from SEI data and config AVStereo3D
> > side data for decoded frame.
> >
> > Signed-off-by: Wenbin Chen
> > Signed-off-by: Tong Wu
> >
> On Sat, 2022-04-02 at 17:35 +0800, Wenbin Chen wrote:
> > Qsv decoder only supports directly output nv12 and p010 to system
> > memory. For other format, we need to download frame from qsv format
> > to system memory. Now add other supported format to qsvdec.
> >
> > Signed-off-by: Wenbin Chen
> On Wed, 2022-03-30 at 14:43 +0800, Wenbin Chen wrote:
> > Qsv decoder only supports directly output nv12 and p010 to system
> > memory. For other format, we need to download frame from qsv format
> > to system memory. Now add other supported format to qsvdec.
> >
> > Signed-off-by: Wenbin Chen
> On Tue, 2022-02-15 at 15:00 +0800, Wenbin Chen wrote:
> > Add support for hevc_qsv to input RGB format frame. It will
> > transform frame to yuv inside MediaSDK instead of using auto
> > scale. Now hevc_qsv supports directly encoding BGRA and X2RGB10
> > format. X2RGB10 is only supported in
> Wenbin Chen:
> > Add support for hevc_qsv to input RGB format frame. It will
> > transform frame to yuv inside MediaSDK instead of using auto
> > scale. Now hevc_qsv supports directly encoding BGRA and X2RGB10
> > format. X2RGB10 is only supported in VDENC (-low_power 1).
> > The X2RGB10
> On Tue, 2022-02-08 at 11:05 +0800, Wenbin Chen wrote:
> > Add vaSyncBuffer to VAAPI encoder. Old version API vaSyncSurface wait
> > surface to complete. When surface is used for multiple operation, it
> > waits all operations to finish. vaSyncBuffer only wait one channel to
> > finish.
> >
> >
> Add async_depth to increase encoder's performance. Reuse encode_fifo as
> async buffer. Encoder puts all reordered frame to HW and then check
> fifo size. If fifo < async_depth and the top frame is not ready, it will
> return AVERROR(EAGAIN) to require more frames.
>
> 1080p transcoding (no B
> On Wed, 2022-01-05 at 10:48 +0800, Wenbin Chen wrote:
> > Fix: #7706. After commit 5fdcf85bbffe7451c2, vaapi encoder's performance
> > decrease. The reason is that vaRenderPicture() and vaSyncBuffer() are
> > called at the same time (vaRenderPicture() always followed by a
> > vaSyncBuffer()).
> On Mon, 2022-01-24 at 10:59 +0800, Wenbin Chen wrote:
> > Add intra refresh support to hevc_qsv as well.
> > Add an new intra refresh type: "horizontal", and an new param
> > ref_cycle_dist. This param specify the distance between the
> > beginnings of the intra-refresh cycles in frames.
> >
> >
> For vaapi if the init_pool_size is not zero, the pool size is fixed.
> This means max surfaces is init_pool_size, but when mapping vaapi
> frame to qsv frame, the init_pool_size < nb_surface. The cause is that
> vaapi_decode_make_config() config the init_pool_size and it is called
> twice. The
> From: nyanmisaka
>
> mfxHDLPair was added to qsv, so modify qsv->opencl map function as well.
> Now the following commandline works:
>
> ffmpeg -v verbose -init_hw_device vaapi=va:/dev/dri/renderD128 \
> -init_hw_device qsv=qs@va -init_hw_device opencl=ocl@va -
> filter_hw_device ocl \
>
> On Thu, 2022-01-13 at 13:12 +0800, Wenbin Chen wrote:
> > Add transform_skip option to hevc_qsv. By enabling this option,
> > the transform_skip_enabled_flag in PPS will be set to 1.
> > This option is supported on the platform equal or newer than CNL.
> >
> > Signed-off-by: Wenbin Chen
> > ---
> On Thu, 2022-01-13 at 13:12 +0800, Wenbin Chen wrote:
> > Add dblk_idc option to 264_qsv and hevc_qsv. Turining on this opion can
> > disable deblocking.
> >
> > Signed-off-by: Wenbin Chen
> > ---
> > doc/encoders.texi | 6 ++
> > libavcodec/qsvenc.c | 8
> > libavcodec/qsvenc.h
> On Wed, 2022-01-12 at 13:28 +0800, Wenbin Chen wrote:
> > Add dblk_idc option to 264_qsv and hevc_qsv. Turining on this opion can
> > disable deblocking.
> >
> > Signed-off-by: Wenbin Chen
> > ---
> > doc/encoders.texi | 6 ++
> > libavcodec/qsvenc.c | 8
> > libavcodec/qsvenc.h
> On Fri, 2022-01-07 at 15:36 +0800, Wenbin Chen wrote:
> > From: nyanmisaka
> >
> > mfxHDLPair was added to qsv, so modify qsv->opencl map function as well.
> > Now the following commandline works:
> >
> > ffmpeg -v verbose -init_hw_device vaapi=va:/dev/dri/renderD128 \
> > -init_hw_device
> Wenbin Chen:
> > Fix: #7706. After commit 5fdcf85bbffe7451c2, vaapi encoder's performance
> > decrease. The reason is that vaRenderPicture() and vaSyncBuffer() are
> > called at the same time (vaRenderPicture() always followed by a
> > vaSyncBuffer()). When we encode stream with B frames, we
> On 10/12/2021 16:05, Wenbin Chen wrote:
> > ffmpeg | branch: master | Wenbin Chen | Tue
> Dec 7 17:05:50 2021 +0800| [f3c9847c2754b7a43eb721c95e356a53085c2491]
> | committer: Lynne
> >
> > libavutil/hwcontext_vaapi: Add a new nv12 format map to support vulkan
> frame
> >
> > Vulkan will map
> On 27/10/2021 09:57, Wenbin Chen wrote:
> > Fix: #7706. After commit 5fdcf85bbffe7451c2, vaapi encoder's performance
> > decrease. The reason is that vaRenderPicture() and vaSyncSurface() are
> > called at the same time (vaRenderPicture() always followed by a
> > vaSyncSurface()). When we encode
> On Wed, 2021-10-27 at 16:57 +0800, Wenbin Chen wrote:
> > Fix: #7706. After commit 5fdcf85bbffe7451c2, vaapi encoder's performance
> > decrease. The reason is that vaRenderPicture() and vaSyncSurface() are
> > called at the same time (vaRenderPicture() always followed by a
> > vaSyncSurface()).
> On Fri, 2021-12-10 at 13:38 +0800, Wenbin Chen wrote:
> > From: nyanmisaka
> >
> > mfxHDLPair was added to qsv, so modify qsv->opencl map function as well.
> > Now the following commandline works:
> >
> > ffmpeg -v verbose -init_hw_device vaapi=va:/dev/dri/renderD128 \
> > -init_hw_device
> On Thu, 2021-12-02 at 15:40 +0800, Wenbin Chen wrote:
> > When we upload a frame that is not padded as MSDK requires, we create a
> > new AVFrame to copy data. The frame's padding data is uninitialized so
> > it brings run to run problem. For example, If we run the following
> > command serveral
> > -Original Message-
> > From: ffmpeg-devel On Behalf Of
> Wenbin Chen
> > Sent: Friday, December 10, 2021 3:22 AM
> > To: ffmpeg-devel@ffmpeg.org
> > Cc: Wenbin Chen
> > Subject: [FFmpeg-devel] [PATCH] libavfilter/vf_overlay_qsv: Use format of
> > first input to set output format for
> > -Original Message-
> > From: ffmpeg-devel On Behalf Of
> Wenbin Chen
> > Sent: Friday, December 10, 2021 3:22 AM
> > To: ffmpeg-devel@ffmpeg.org
> > Cc: Wenbin Chen
> > Subject: [FFmpeg-devel] [PATCH] libavfilter/vf_overlay_qsv: Use format of
> > first input to set output format for
> 7 Dec 2021, 10:05 by wenbin.c...@intel.com:
>
> > When vulkan image exports to drm, the tilling need to be
> > VK_IMAGE_TILING_DRM_FORMAT_MODIFIER_EXT. Now add code to
> create vulkan
> > image using this format.
> >
> > Now the following command line works:
> >
> > ffmpeg -hwaccel vaapi
> Quoting Wenbin Chen (2021-11-30 09:44:04)
> > For vaapi if the init_pool_size is not zero, the pool size is fixed.
> > This means max surfaces is init_pool_size, but when mapping vaapi
> > frame to qsv frame, the init_pool_size < nb_surface. The cause is that
> > vaapi_decode_make_config()
> 2 Dec 2021, 02:43 by wenbin.c...@intel.com:
>
> > When vulkan image exports to drm, the tilling need to be
> > VK_IMAGE_TILING_DRM_FORMAT_MODIFIER_EXT. Now add code to
> create vulkan
> > image using this format.
> >
> > Now the following command line works:
> >
> > ffmpeg -hwaccel vaapi
> 2 Dec 2021, 02:53 by wenbin.c...@intel.com:
>
> >> The vaapi can import external frame, but the planes of the external
> >> frames should be in the same drm object. A new option
> "contiguous_planes"
> >> is added to device. This flag tells device to allocate places in one
> >> memory. When
> The vaapi can import external frame, but the planes of the external
> frames should be in the same drm object. A new option "contiguous_planes"
> is added to device. This flag tells device to allocate places in one
> memory. When device is derived from vaapi this flag will be enabled.
> A new
> Quoting Wenbin Chen (2021-11-30 07:28:13)
> > diff --git a/libavutil/hwcontext_vulkan.h b/libavutil/hwcontext_vulkan.h
> > index fdf2a60156..c485ee7437 100644
> > --- a/libavutil/hwcontext_vulkan.h
> > +++ b/libavutil/hwcontext_vulkan.h
> > @@ -35,6 +35,17 @@
> > * with the data pointer set to
> The vaapi can import external frame, but the planes of the external
> frames should be in the same drm object. A new option "contiguous_planes"
> is added to device. This flag tells device to allocate places in one
> memory. When device is derived from vaapi this flag will be enabled.
> A new
> Quoting Wenbin Chen (2021-11-16 09:16:23)
> > From: nyanmisaka
> >
> > mfxHDLPair was added to qsv, so modify qsv->opencl map function as well.
> > Now the following commandline works:
> >
> > ffmpeg -v verbose -init_hw_device vaapi=va:/dev/dri/renderD128 \
> > -init_hw_device qsv=qs@va
> 26 Nov 2021, 03:54 by wenbin.c...@intel.com:
>
> > The vaapi can import external frame, but the planes of the external
> > frames should be in the same drm object. A new option
> "contiguous_planes"
> > is added to device. This flag tells device to allocate places in one
> > memory. When device
> 24 Nov 2021, 06:28 by wenbin.c...@intel.com:
>
> > The vaapi can import external frame, but the planes of the external
> > frames should be in the same drm object. A new option
> "contiguous_planes"
> > is added to device. This flag tells device to allocate places in one
> > memory. When device
> 24 Nov 2021, 06:28 by wenbin.c...@intel.com:
>
> > From: Bas Nieuwenhuizen
> >
> > This way we can pass explicit modifiers in. Sometimes the
> > modifier matters for the number of memory planes that
> > libva accepts, in particular when dealing with
> > driver-compressed textures. Furthermore
> From: nyanmisaka
>
> mfxHDLPair was added to qsv, so modify qsv->opencl map function as well.
> Now the following commandline works:
>
> ffmpeg -v verbose -init_hw_device vaapi=va:/dev/dri/renderD128 \
> -init_hw_device qsv=qs@va -init_hw_device opencl=ocl@va -
> filter_hw_device ocl \
>
> From: nyanmisaka
>
> The data stored in data[3] in VAAPI AVFrame is VASurfaceID while
> the data stored in pair->first is the pointer of VASurfaceID, so
> we need to do cast to make following commandline works:
>
> ffmpeg -hwaccel vaapi -hwaccel_device /dev/dri/renderD128 \
>
> For vaapi if the init_pool_size is not zero, the pool size is fixed.
> This means max surfaces is init_pool_size, but when mapping vaapi
> frame to qsv frame, the init_pool_size < nb_surface. The cause is that
> vaapi_decode_make_config() config the init_pool_size and it is called
> twice. The
> > -Original Message-
> > From: ffmpeg-devel On Behalf Of
> > Xiang, Haihao
> > Sent: Monday, October 18, 2021 6:48 AM
> > To: ffmpeg-devel@ffmpeg.org
> > Subject: Re: [FFmpeg-devel] [PATCH v3 1/1] avutils/hwcontext: When
> > deriving a hwdevice, search for existing device in both
> 19 Nov 2021, 19:13 by d...@lynne.ee:
>
> > 19 Nov 2021, 18:59 by d...@lynne.ee:
> >
> >> 15 Nov 2021, 08:25 by wenbin.c...@intel.com:
> >>
> 9 Nov 2021, 10:18 by wenbin.c...@intel.com:
>
> > The vaapi can import external frame, but the planes of the external
> > frames
> 9 Nov 2021, 10:18 by wenbin.c...@intel.com:
>
> > When vulkan image exports to drm, the tilling need to be
> > VK_IMAGE_TILING_DRM_FORMAT_MODIFIER_EXT. Now add code to
> create vulkan
> > image using this format.
> >
> > Now the following command line works:
> >
> > ffmpeg -hwaccel vaapi
> 9 Nov 2021, 10:18 by wenbin.c...@intel.com:
>
> > The vaapi can import external frame, but the planes of the external
> > frames should be in the same drm object. I add a new function to
> > allocate vkFrame in one memory and vulkan device will choose a way
> > to allocate memory according to
> > -Original Message-
> > From: ffmpeg-devel On Behalf Of
> > Chen, Wenbin
> > Sent: Wednesday, November 10, 2021 4:03 AM
> > To: ffmpeg-devel@ffmpeg.org
> > Subject: Re: [FFmpeg-devel] [PATCH 1/4] libavutil/hwcontext_d3d11va:
> >
> > -Original Message-
> > From: ffmpeg-devel On Behalf Of
> > Chen, Wenbin
> > Sent: Wednesday, November 10, 2021 4:03 AM
> > To: ffmpeg-devel@ffmpeg.org
> > Subject: Re: [FFmpeg-devel] [PATCH 2/4] libavutil/hwcontext_qsv: fix
> &g
> 9 Nov 2021, 10:18 by wenbin.c...@intel.com:
>
> > sem_sig_val is wrongly assigned to pWaitSemaphoreValues when export
> drm. Now fix
> > it.
> >
> > Signed-off-by: Wenbin Chen <> wenbin.c...@intel.com> >
> >
>
> Thanks for spotting this, I fixed that in my patchset and updated branch.
>
> From: nyanmisaka
>
> mfxHDLPair was added to qsv, so modify qsv->opencl map function as well.
> Now the following commandline works:
>
> ffmpeg -v verbose -init_hw_device vaapi=va:/dev/dri/renderD128 \
> -init_hw_device qsv=qs@va -init_hw_device opencl=ocl@va -
> filter_hw_device ocl \
>
> From: nyanmisaka
>
> The data stored in data[3] in VAAPI AVFrame is VASurfaceID while
> the data stored in pair->first is the pointer of VASurfaceID, so
> we need to do cast to make following commandline works:
>
> ffmpeg -hwaccel vaapi -hwaccel_device /dev/dri/renderD128 \
>
> This commandline cause core dumped:
> ffmpeg -hwaccel vaapi -hwaccel_device /dev/dri/renderD128 \
> -hwaccel_output_format vaapi -i input.264 \
> -vf "hwmap=derive_device=qsv,format=qsv" \
> -c:v h264_qsv output.264
>
> reason: We use nb_surfaces to assign surface to handle_pairs_internal
> but
> Adding nb_surfaces in AVD3D11VAFramesContext in the end of the structure
> to support flexible size of this arrays and align to
> AVDXVA2FramesContext and AVVAAPIFramesContext.
>
> Signed-off-by Wenbin Chen
> ---
> libavutil/hwcontext_d3d11va.c | 3 +--
> libavutil/hwcontext_d3d11va.h | 2 ++
> 9 Sept 2021, 07:44 by jianhua...@intel.com:
>
> > This commit adds a powerful and customizable gblur Vulkan filter,
> > which provides a maximum 127x127 kernel size of Gaussian Filter.
> > The size could be adjusted by requirements on quality or performance.
> >
> > The following command is on
> Add async_depth to increase encoder's performance. Reuse encode_fifo as
> async buffer. Encoder puts all reordered frame to HW and then check
> fifo size. If fifo < async_depth and the top frame is not ready, it will
> return AVERROR(EAGAIN) to require more frames.
>
> 1080p transcoding (no B
> Add vaSyncBuffer to VAAPI encoder. Old version API vaSyncSurface wait
> surface to complete. When surface is used for multiple operation, it
> wait all operation to finish. vaSyncBuffer only wait one channel to
> finish.
>
> Add wait param to vaapi_encode_wait() to prepare for the async_depth
>
> Fix: #7706. After commit 5fdcf85bbffe7451c2, vaapi encoder's performance
> decrease. The reason is that vaRenderPicture() and vaSyncSurface() are
> called at the same time (vaRenderPicture() always followed by a
> vaSyncSurface()). When we encode stream with B frames, we need buffer to
> reorder
> Command below failed.
> ffmpeg -v verbose -init_hw_device vaapi=va:/dev/dri/renderD128
> -init_hw_device qsv=qs@va -hwaccel qsv -hwaccel_device qs
> -filter_hw_device va -c:v h264_qsv
> -i 1080P.264 -vf "hwmap,format=vaapi" -c:v h264_vaapi output.264
>
> Cause: Assign pair->first directly to
> Command below failed.
> ffmpeg -v verbose -init_hw_device vaapi=va:/dev/dri/renderD128
> -init_hw_device qsv=qs@va -hwaccel qsv -hwaccel_device qs
> -filter_hw_device va -c:v h264_qsv
> -i 1080P.264 -vf "hwmap,format=vaapi" -c:v h264_vaapi output.264
>
> Cause: Assign pair->first directly to
> On Fri, 2021-09-10 at 02:19 +0000, Chen, Wenbin wrote:
> > > -Original Message-
> > > From: ffmpeg-devel On Behalf Of
> > > James Almer
> > > Sent: Thursday, September 9, 2021 8:48 PM
> > > To: ffmpeg-devel@ffmpeg.org
> > > Su
> -Original Message-
> From: ffmpeg-devel On Behalf Of
> James Almer
> Sent: Thursday, September 9, 2021 8:48 PM
> To: ffmpeg-devel@ffmpeg.org
> Subject: Re: [FFmpeg-devel] [PATCH] libavutil/hwcontext_qsv: fix a bug for
> mapping qsv frame to vaapi
>
> On 9/9/2021 6:13 AM, Wenbin Chen
a
> hwdevice, search for existing device in both directions
>
>
>
> > -Original Message-
> > From: ffmpeg-devel On Behalf Of
> > Chen, Wenbin
> > Sent: Wednesday, 8 September 2021 07:58
> > To: FFmpeg development discussions and patches > de...@ffmpeg
Hi softworkz:
Any updates on this patch?
I want to submit pathes to qsv, but they depends on this patch
Thanks
Wenbin
> -Original Message-
> From: ffmpeg-devel On Behalf Of Soft
> Works
> Sent: Thursday, August 19, 2021 3:52 PM
> To: FFmpeg development discussions and patches
> -Original Message-
> From: Lynne
> Sent: Wednesday, September 1, 2021 7:33 PM
> To: Chen, Wenbin
> Subject: RE: [FFmpeg-devel] [V2 08/10] libavutil/hwcontext_vulkan: fix wrong
> offset of plane
>
> 1 Sept 2021, 03:40 by wenbin.c...@intel.com:
>
> &g
> -Original Message-
> From: ffmpeg-devel On Behalf Of
> Lynne
> Sent: Tuesday, August 31, 2021 7:26 PM
> To: FFmpeg development discussions and patches de...@ffmpeg.org>
> Subject: Re: [FFmpeg-devel] [V2 03/10] libavfilter/vulkan: Fix the way to use
> sem
>
> 31 Aug 2021, 09:24 by
> -Original Message-
> From: ffmpeg-devel On Behalf Of
> Lynne
> Sent: Tuesday, August 31, 2021 7:42 PM
> To: FFmpeg development discussions and patches de...@ffmpeg.org>
> Subject: Re: [FFmpeg-devel] [V2 08/10] libavutil/hwcontext_vulkan: fix wrong
> offset of plane
>
> 31 Aug 2021,
1 - 100 of 115 matches
Mail list logo
