2016-05-28 16:23:17 +00:00
|
|
|
/*
|
|
|
|
* This file is part of mpv.
|
|
|
|
*
|
|
|
|
* mpv is free software; you can redistribute it and/or
|
|
|
|
* modify it under the terms of the GNU Lesser General Public
|
|
|
|
* License as published by the Free Software Foundation; either
|
|
|
|
* version 2.1 of the License, or (at your option) any later version.
|
|
|
|
*
|
|
|
|
* mpv is distributed in the hope that it will be useful,
|
|
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
|
|
* GNU Lesser General Public License for more details.
|
|
|
|
*
|
|
|
|
* You should have received a copy of the GNU Lesser General Public
|
|
|
|
* License along with mpv. If not, see <http://www.gnu.org/licenses/>.
|
|
|
|
*/
|
|
|
|
|
|
|
|
#include <assert.h>
|
|
|
|
#include <windows.h>
|
|
|
|
#include <d3d11.h>
|
|
|
|
|
2017-12-02 03:33:43 +00:00
|
|
|
#include <libavutil/hwcontext.h>
|
|
|
|
#include <libavutil/hwcontext_d3d11va.h>
|
|
|
|
|
2016-05-28 16:23:17 +00:00
|
|
|
#include "common/common.h"
|
|
|
|
#include "osdep/timer.h"
|
|
|
|
#include "osdep/windows_utils.h"
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
#include "filters/f_autoconvert.h"
|
|
|
|
#include "filters/filter.h"
|
|
|
|
#include "filters/filter_internal.h"
|
|
|
|
#include "filters/user_filters.h"
|
2016-05-28 16:23:17 +00:00
|
|
|
#include "refqueue.h"
|
|
|
|
#include "video/hwdec.h"
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
#include "video/mp_image.h"
|
2016-05-28 16:23:17 +00:00
|
|
|
#include "video/mp_image_pool.h"
|
|
|
|
|
|
|
|
// missing in MinGW
|
2016-07-15 17:48:58 +00:00
|
|
|
#define D3D11_VIDEO_PROCESSOR_PROCESSOR_CAPS_DEINTERLACE_BLEND 0x1
|
2016-05-28 16:23:17 +00:00
|
|
|
#define D3D11_VIDEO_PROCESSOR_PROCESSOR_CAPS_DEINTERLACE_BOB 0x2
|
2016-07-15 17:48:58 +00:00
|
|
|
#define D3D11_VIDEO_PROCESSOR_PROCESSOR_CAPS_DEINTERLACE_ADAPTIVE 0x4
|
|
|
|
#define D3D11_VIDEO_PROCESSOR_PROCESSOR_CAPS_DEINTERLACE_MOTION_COMPENSATION 0x8
|
|
|
|
#define D3D11_VIDEO_PROCESSOR_PROCESSOR_CAPS_INVERSE_TELECINE 0x10
|
|
|
|
#define D3D11_VIDEO_PROCESSOR_PROCESSOR_CAPS_FRAME_RATE_CONVERSION 0x20
|
2016-05-28 16:23:17 +00:00
|
|
|
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
struct opts {
|
|
|
|
int deint_enabled;
|
|
|
|
int interlaced_only;
|
|
|
|
int mode;
|
|
|
|
};
|
|
|
|
|
|
|
|
struct priv {
|
|
|
|
struct opts *opts;
|
|
|
|
|
2016-05-28 16:23:17 +00:00
|
|
|
ID3D11Device *vo_dev;
|
|
|
|
|
|
|
|
ID3D11DeviceContext *device_ctx;
|
|
|
|
ID3D11VideoDevice *video_dev;
|
|
|
|
ID3D11VideoContext *video_ctx;
|
|
|
|
|
|
|
|
ID3D11VideoProcessor *video_proc;
|
|
|
|
ID3D11VideoProcessorEnumerator *vp_enum;
|
|
|
|
D3D11_VIDEO_FRAME_FORMAT d3d_frame_format;
|
|
|
|
|
video: remove d3d11 video processor use from OpenGL interop
We now have a video filter that uses the d3d11 video processor, so it
makes no sense to have one in the VO interop code. The VO uses it for
formats not directly supported by ANGLE (so the video data is converted
to a RGB texture, which ANGLE can take in).
Change this so that the video filter is automatically inserted if
needed. Move the code that maps RGB surfaces to its own inteorp backend.
Add a bunch of new image formats, which are used to enforce the new
constraints, and to automatically insert the filter only when needed.
The added vf mechanism to auto-insert the d3d11vpp filter is very dumb
and primitive, and will work only for this specific purpose. The format
negotiation mechanism in the filter chain is generally not very pretty,
and mostly broken as well. (libavfilter has a different mechanism, and
these mechanisms don't match well, so vf_lavfi uses some sort of hack.
It only works because hwaccel and non-hwaccel formats are strictly
separated.)
The RGB interop is now only used with older ANGLE versions. The only
reason I'm keeping it is because it's relatively isolated (uses only
existing mechanisms and adds no new concepts), and because I want to be
able to compare the behavior of the old code with the new one for
testing. It will be removed eventually.
If ANGLE has NV12 interop, P010 is now handled by converting to NV12
with the video processor, instead of converting it to RGB and using the
old mechanism to import that as a texture.
2016-05-29 15:13:22 +00:00
|
|
|
DXGI_FORMAT out_format;
|
|
|
|
|
|
|
|
bool require_filtering;
|
|
|
|
|
|
|
|
struct mp_image_params params, out_params;
|
2016-05-28 16:23:17 +00:00
|
|
|
int c_w, c_h;
|
|
|
|
|
|
|
|
struct mp_image_pool *pool;
|
|
|
|
|
|
|
|
struct mp_refqueue *queue;
|
|
|
|
};
|
|
|
|
|
|
|
|
static void release_tex(void *arg)
|
|
|
|
{
|
|
|
|
ID3D11Texture2D *texture = arg;
|
|
|
|
|
|
|
|
ID3D11Texture2D_Release(texture);
|
|
|
|
}
|
|
|
|
|
video: remove d3d11 video processor use from OpenGL interop
We now have a video filter that uses the d3d11 video processor, so it
makes no sense to have one in the VO interop code. The VO uses it for
formats not directly supported by ANGLE (so the video data is converted
to a RGB texture, which ANGLE can take in).
Change this so that the video filter is automatically inserted if
needed. Move the code that maps RGB surfaces to its own inteorp backend.
Add a bunch of new image formats, which are used to enforce the new
constraints, and to automatically insert the filter only when needed.
The added vf mechanism to auto-insert the d3d11vpp filter is very dumb
and primitive, and will work only for this specific purpose. The format
negotiation mechanism in the filter chain is generally not very pretty,
and mostly broken as well. (libavfilter has a different mechanism, and
these mechanisms don't match well, so vf_lavfi uses some sort of hack.
It only works because hwaccel and non-hwaccel formats are strictly
separated.)
The RGB interop is now only used with older ANGLE versions. The only
reason I'm keeping it is because it's relatively isolated (uses only
existing mechanisms and adds no new concepts), and because I want to be
able to compare the behavior of the old code with the new one for
testing. It will be removed eventually.
If ANGLE has NV12 interop, P010 is now handled by converting to NV12
with the video processor, instead of converting it to RGB and using the
old mechanism to import that as a texture.
2016-05-29 15:13:22 +00:00
|
|
|
static struct mp_image *alloc_pool(void *pctx, int fmt, int w, int h)
|
2016-05-28 16:23:17 +00:00
|
|
|
{
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
struct mp_filter *vf = pctx;
|
|
|
|
struct priv *p = vf->priv;
|
2016-05-28 16:23:17 +00:00
|
|
|
HRESULT hr;
|
|
|
|
|
|
|
|
ID3D11Texture2D *texture = NULL;
|
|
|
|
D3D11_TEXTURE2D_DESC texdesc = {
|
|
|
|
.Width = w,
|
|
|
|
.Height = h,
|
video: remove d3d11 video processor use from OpenGL interop
We now have a video filter that uses the d3d11 video processor, so it
makes no sense to have one in the VO interop code. The VO uses it for
formats not directly supported by ANGLE (so the video data is converted
to a RGB texture, which ANGLE can take in).
Change this so that the video filter is automatically inserted if
needed. Move the code that maps RGB surfaces to its own inteorp backend.
Add a bunch of new image formats, which are used to enforce the new
constraints, and to automatically insert the filter only when needed.
The added vf mechanism to auto-insert the d3d11vpp filter is very dumb
and primitive, and will work only for this specific purpose. The format
negotiation mechanism in the filter chain is generally not very pretty,
and mostly broken as well. (libavfilter has a different mechanism, and
these mechanisms don't match well, so vf_lavfi uses some sort of hack.
It only works because hwaccel and non-hwaccel formats are strictly
separated.)
The RGB interop is now only used with older ANGLE versions. The only
reason I'm keeping it is because it's relatively isolated (uses only
existing mechanisms and adds no new concepts), and because I want to be
able to compare the behavior of the old code with the new one for
testing. It will be removed eventually.
If ANGLE has NV12 interop, P010 is now handled by converting to NV12
with the video processor, instead of converting it to RGB and using the
old mechanism to import that as a texture.
2016-05-29 15:13:22 +00:00
|
|
|
.Format = p->out_format,
|
2016-05-28 16:23:17 +00:00
|
|
|
.MipLevels = 1,
|
|
|
|
.ArraySize = 1,
|
|
|
|
.SampleDesc = { .Count = 1 },
|
|
|
|
.Usage = D3D11_USAGE_DEFAULT,
|
|
|
|
.BindFlags = D3D11_BIND_RENDER_TARGET | D3D11_BIND_SHADER_RESOURCE,
|
|
|
|
};
|
video: remove d3d11 video processor use from OpenGL interop
We now have a video filter that uses the d3d11 video processor, so it
makes no sense to have one in the VO interop code. The VO uses it for
formats not directly supported by ANGLE (so the video data is converted
to a RGB texture, which ANGLE can take in).
Change this so that the video filter is automatically inserted if
needed. Move the code that maps RGB surfaces to its own inteorp backend.
Add a bunch of new image formats, which are used to enforce the new
constraints, and to automatically insert the filter only when needed.
The added vf mechanism to auto-insert the d3d11vpp filter is very dumb
and primitive, and will work only for this specific purpose. The format
negotiation mechanism in the filter chain is generally not very pretty,
and mostly broken as well. (libavfilter has a different mechanism, and
these mechanisms don't match well, so vf_lavfi uses some sort of hack.
It only works because hwaccel and non-hwaccel formats are strictly
separated.)
The RGB interop is now only used with older ANGLE versions. The only
reason I'm keeping it is because it's relatively isolated (uses only
existing mechanisms and adds no new concepts), and because I want to be
able to compare the behavior of the old code with the new one for
testing. It will be removed eventually.
If ANGLE has NV12 interop, P010 is now handled by converting to NV12
with the video processor, instead of converting it to RGB and using the
old mechanism to import that as a texture.
2016-05-29 15:13:22 +00:00
|
|
|
hr = ID3D11Device_CreateTexture2D(p->vo_dev, &texdesc, NULL, &texture);
|
2016-05-28 16:23:17 +00:00
|
|
|
if (FAILED(hr))
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
struct mp_image *mpi = mp_image_new_custom_ref(NULL, texture, release_tex);
|
2023-01-10 17:59:21 +00:00
|
|
|
MP_HANDLE_OOM(mpi);
|
2016-05-28 16:23:17 +00:00
|
|
|
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
mp_image_setfmt(mpi, IMGFMT_D3D11);
|
2016-05-28 16:23:17 +00:00
|
|
|
mp_image_set_size(mpi, w, h);
|
video: remove d3d11 video processor use from OpenGL interop
We now have a video filter that uses the d3d11 video processor, so it
makes no sense to have one in the VO interop code. The VO uses it for
formats not directly supported by ANGLE (so the video data is converted
to a RGB texture, which ANGLE can take in).
Change this so that the video filter is automatically inserted if
needed. Move the code that maps RGB surfaces to its own inteorp backend.
Add a bunch of new image formats, which are used to enforce the new
constraints, and to automatically insert the filter only when needed.
The added vf mechanism to auto-insert the d3d11vpp filter is very dumb
and primitive, and will work only for this specific purpose. The format
negotiation mechanism in the filter chain is generally not very pretty,
and mostly broken as well. (libavfilter has a different mechanism, and
these mechanisms don't match well, so vf_lavfi uses some sort of hack.
It only works because hwaccel and non-hwaccel formats are strictly
separated.)
The RGB interop is now only used with older ANGLE versions. The only
reason I'm keeping it is because it's relatively isolated (uses only
existing mechanisms and adds no new concepts), and because I want to be
able to compare the behavior of the old code with the new one for
testing. It will be removed eventually.
If ANGLE has NV12 interop, P010 is now handled by converting to NV12
with the video processor, instead of converting it to RGB and using the
old mechanism to import that as a texture.
2016-05-29 15:13:22 +00:00
|
|
|
mpi->params.hw_subfmt = p->out_params.hw_subfmt;
|
2016-05-28 16:23:17 +00:00
|
|
|
|
2017-05-03 23:13:03 +00:00
|
|
|
mpi->planes[0] = (void *)texture;
|
|
|
|
mpi->planes[1] = (void *)(intptr_t)0;
|
2016-05-28 16:23:17 +00:00
|
|
|
|
|
|
|
return mpi;
|
|
|
|
}
|
|
|
|
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
static void flush_frames(struct mp_filter *vf)
|
2016-05-28 16:23:17 +00:00
|
|
|
{
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
struct priv *p = vf->priv;
|
2016-05-28 16:23:17 +00:00
|
|
|
mp_refqueue_flush(p->queue);
|
|
|
|
}
|
|
|
|
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
static void destroy_video_proc(struct mp_filter *vf)
|
2016-05-28 16:23:17 +00:00
|
|
|
{
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
struct priv *p = vf->priv;
|
2016-05-28 16:23:17 +00:00
|
|
|
|
|
|
|
if (p->video_proc)
|
|
|
|
ID3D11VideoProcessor_Release(p->video_proc);
|
|
|
|
p->video_proc = NULL;
|
|
|
|
|
|
|
|
if (p->vp_enum)
|
|
|
|
ID3D11VideoProcessorEnumerator_Release(p->vp_enum);
|
|
|
|
p->vp_enum = NULL;
|
|
|
|
}
|
|
|
|
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
static int recreate_video_proc(struct mp_filter *vf)
|
2016-05-28 16:23:17 +00:00
|
|
|
{
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
struct priv *p = vf->priv;
|
2016-05-28 16:23:17 +00:00
|
|
|
HRESULT hr;
|
|
|
|
|
|
|
|
destroy_video_proc(vf);
|
|
|
|
|
|
|
|
D3D11_VIDEO_PROCESSOR_CONTENT_DESC vpdesc = {
|
|
|
|
.InputFrameFormat = p->d3d_frame_format,
|
|
|
|
.InputWidth = p->c_w,
|
|
|
|
.InputHeight = p->c_h,
|
|
|
|
.OutputWidth = p->params.w,
|
|
|
|
.OutputHeight = p->params.h,
|
|
|
|
};
|
|
|
|
hr = ID3D11VideoDevice_CreateVideoProcessorEnumerator(p->video_dev, &vpdesc,
|
|
|
|
&p->vp_enum);
|
|
|
|
if (FAILED(hr))
|
|
|
|
goto fail;
|
|
|
|
|
|
|
|
D3D11_VIDEO_PROCESSOR_CAPS caps;
|
|
|
|
hr = ID3D11VideoProcessorEnumerator_GetVideoProcessorCaps(p->vp_enum, &caps);
|
|
|
|
if (FAILED(hr))
|
|
|
|
goto fail;
|
|
|
|
|
2016-07-15 17:48:58 +00:00
|
|
|
MP_VERBOSE(vf, "Found %d rate conversion caps. Looking for caps=0x%x.\n",
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
(int)caps.RateConversionCapsCount, p->opts->mode);
|
2016-06-16 14:32:38 +00:00
|
|
|
|
2016-05-28 16:23:17 +00:00
|
|
|
int rindex = -1;
|
|
|
|
for (int n = 0; n < caps.RateConversionCapsCount; n++) {
|
|
|
|
D3D11_VIDEO_PROCESSOR_RATE_CONVERSION_CAPS rcaps;
|
|
|
|
hr = ID3D11VideoProcessorEnumerator_GetVideoProcessorRateConversionCaps
|
|
|
|
(p->vp_enum, n, &rcaps);
|
|
|
|
if (FAILED(hr))
|
|
|
|
goto fail;
|
2016-06-16 14:32:38 +00:00
|
|
|
MP_VERBOSE(vf, " - %d: 0x%08x\n", n, (unsigned)rcaps.ProcessorCaps);
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
if (rcaps.ProcessorCaps & p->opts->mode) {
|
2016-06-16 14:32:38 +00:00
|
|
|
MP_VERBOSE(vf, " (matching)\n");
|
|
|
|
if (rindex < 0)
|
|
|
|
rindex = n;
|
2016-05-28 16:23:17 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (rindex < 0) {
|
2016-07-15 17:48:58 +00:00
|
|
|
MP_WARN(vf, "No fitting video processor found, picking #0.\n");
|
2016-06-16 15:17:54 +00:00
|
|
|
rindex = 0;
|
2016-05-28 16:23:17 +00:00
|
|
|
}
|
|
|
|
|
2016-07-15 17:48:58 +00:00
|
|
|
// TOOD: so, how do we select which rate conversion mode the processor uses?
|
|
|
|
|
2016-05-28 16:23:17 +00:00
|
|
|
hr = ID3D11VideoDevice_CreateVideoProcessor(p->video_dev, p->vp_enum, rindex,
|
|
|
|
&p->video_proc);
|
|
|
|
if (FAILED(hr)) {
|
|
|
|
MP_ERR(vf, "Failed to create D3D11 video processor.\n");
|
|
|
|
goto fail;
|
|
|
|
}
|
|
|
|
|
|
|
|
// Note: libavcodec does not support cropping left/top with hwaccel.
|
|
|
|
RECT src_rc = {
|
|
|
|
.right = p->params.w,
|
|
|
|
.bottom = p->params.h,
|
|
|
|
};
|
|
|
|
ID3D11VideoContext_VideoProcessorSetStreamSourceRect(p->video_ctx,
|
|
|
|
p->video_proc,
|
|
|
|
0, TRUE, &src_rc);
|
|
|
|
|
|
|
|
// This is supposed to stop drivers from fucking up the video quality.
|
|
|
|
ID3D11VideoContext_VideoProcessorSetStreamAutoProcessingMode(p->video_ctx,
|
|
|
|
p->video_proc,
|
|
|
|
0, FALSE);
|
|
|
|
|
|
|
|
ID3D11VideoContext_VideoProcessorSetStreamOutputRate(p->video_ctx,
|
|
|
|
p->video_proc,
|
|
|
|
0,
|
|
|
|
D3D11_VIDEO_PROCESSOR_OUTPUT_RATE_NORMAL,
|
|
|
|
FALSE, 0);
|
|
|
|
|
|
|
|
D3D11_VIDEO_PROCESSOR_COLOR_SPACE csp = {
|
2016-06-29 07:16:13 +00:00
|
|
|
.YCbCr_Matrix = p->params.color.space != MP_CSP_BT_601,
|
|
|
|
.Nominal_Range = p->params.color.levels == MP_CSP_LEVELS_TV ? 1 : 2,
|
2016-05-28 16:23:17 +00:00
|
|
|
};
|
|
|
|
ID3D11VideoContext_VideoProcessorSetStreamColorSpace(p->video_ctx,
|
|
|
|
p->video_proc,
|
|
|
|
0, &csp);
|
2019-10-16 21:37:06 +00:00
|
|
|
ID3D11VideoContext_VideoProcessorSetOutputColorSpace(p->video_ctx,
|
|
|
|
p->video_proc,
|
|
|
|
&csp);
|
2016-05-28 16:23:17 +00:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
fail:
|
|
|
|
destroy_video_proc(vf);
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
static struct mp_image *render(struct mp_filter *vf)
|
2016-05-28 16:23:17 +00:00
|
|
|
{
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
struct priv *p = vf->priv;
|
2016-05-28 16:23:17 +00:00
|
|
|
int res = -1;
|
|
|
|
HRESULT hr;
|
|
|
|
ID3D11VideoProcessorInputView *in_view = NULL;
|
|
|
|
ID3D11VideoProcessorOutputView *out_view = NULL;
|
|
|
|
struct mp_image *in = NULL, *out = NULL;
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
out = mp_image_pool_get(p->pool, IMGFMT_D3D11, p->params.w, p->params.h);
|
|
|
|
if (!out) {
|
|
|
|
MP_WARN(vf, "failed to allocate frame\n");
|
2016-05-28 16:23:17 +00:00
|
|
|
goto cleanup;
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
}
|
2016-05-28 16:23:17 +00:00
|
|
|
|
2017-05-03 23:13:03 +00:00
|
|
|
ID3D11Texture2D *d3d_out_tex = (void *)out->planes[0];
|
2016-05-28 16:23:17 +00:00
|
|
|
|
|
|
|
in = mp_refqueue_get(p->queue, 0);
|
|
|
|
if (!in)
|
|
|
|
goto cleanup;
|
2017-05-03 23:13:03 +00:00
|
|
|
ID3D11Texture2D *d3d_tex = (void *)in->planes[0];
|
|
|
|
int d3d_subindex = (intptr_t)in->planes[1];
|
2016-05-28 16:23:17 +00:00
|
|
|
|
|
|
|
mp_image_copy_attributes(out, in);
|
|
|
|
|
|
|
|
D3D11_VIDEO_FRAME_FORMAT d3d_frame_format;
|
2016-07-15 17:52:32 +00:00
|
|
|
if (!mp_refqueue_should_deint(p->queue)) {
|
2016-05-28 16:23:17 +00:00
|
|
|
d3d_frame_format = D3D11_VIDEO_FRAME_FORMAT_PROGRESSIVE;
|
|
|
|
} else if (mp_refqueue_top_field_first(p->queue)) {
|
|
|
|
d3d_frame_format = D3D11_VIDEO_FRAME_FORMAT_INTERLACED_TOP_FIELD_FIRST;
|
|
|
|
} else {
|
|
|
|
d3d_frame_format = D3D11_VIDEO_FRAME_FORMAT_INTERLACED_BOTTOM_FIELD_FIRST;
|
|
|
|
}
|
|
|
|
|
|
|
|
D3D11_TEXTURE2D_DESC texdesc;
|
|
|
|
ID3D11Texture2D_GetDesc(d3d_tex, &texdesc);
|
|
|
|
if (!p->video_proc || p->c_w != texdesc.Width || p->c_h != texdesc.Height ||
|
|
|
|
p->d3d_frame_format != d3d_frame_format)
|
|
|
|
{
|
|
|
|
p->c_w = texdesc.Width;
|
|
|
|
p->c_h = texdesc.Height;
|
|
|
|
p->d3d_frame_format = d3d_frame_format;
|
|
|
|
if (recreate_video_proc(vf) < 0)
|
|
|
|
goto cleanup;
|
|
|
|
}
|
|
|
|
|
2016-07-15 17:52:32 +00:00
|
|
|
if (!mp_refqueue_should_deint(p->queue)) {
|
2016-05-28 16:23:17 +00:00
|
|
|
d3d_frame_format = D3D11_VIDEO_FRAME_FORMAT_PROGRESSIVE;
|
|
|
|
} else if (mp_refqueue_is_top_field(p->queue)) {
|
|
|
|
d3d_frame_format = D3D11_VIDEO_FRAME_FORMAT_INTERLACED_TOP_FIELD_FIRST;
|
|
|
|
} else {
|
|
|
|
d3d_frame_format = D3D11_VIDEO_FRAME_FORMAT_INTERLACED_BOTTOM_FIELD_FIRST;
|
|
|
|
}
|
|
|
|
|
|
|
|
ID3D11VideoContext_VideoProcessorSetStreamFrameFormat(p->video_ctx,
|
|
|
|
p->video_proc,
|
|
|
|
0, d3d_frame_format);
|
|
|
|
|
|
|
|
D3D11_VIDEO_PROCESSOR_INPUT_VIEW_DESC indesc = {
|
|
|
|
.ViewDimension = D3D11_VPIV_DIMENSION_TEXTURE2D,
|
|
|
|
.Texture2D = {
|
|
|
|
.ArraySlice = d3d_subindex,
|
|
|
|
},
|
|
|
|
};
|
|
|
|
hr = ID3D11VideoDevice_CreateVideoProcessorInputView(p->video_dev,
|
|
|
|
(ID3D11Resource *)d3d_tex,
|
|
|
|
p->vp_enum, &indesc,
|
|
|
|
&in_view);
|
|
|
|
if (FAILED(hr)) {
|
|
|
|
MP_ERR(vf, "Could not create ID3D11VideoProcessorInputView\n");
|
|
|
|
goto cleanup;
|
|
|
|
}
|
|
|
|
|
|
|
|
D3D11_VIDEO_PROCESSOR_OUTPUT_VIEW_DESC outdesc = {
|
|
|
|
.ViewDimension = D3D11_VPOV_DIMENSION_TEXTURE2D,
|
|
|
|
};
|
|
|
|
hr = ID3D11VideoDevice_CreateVideoProcessorOutputView(p->video_dev,
|
|
|
|
(ID3D11Resource *)d3d_out_tex,
|
|
|
|
p->vp_enum, &outdesc,
|
|
|
|
&out_view);
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
if (FAILED(hr)) {
|
|
|
|
MP_ERR(vf, "Could not create ID3D11VideoProcessorOutputView\n");
|
2016-05-28 16:23:17 +00:00
|
|
|
goto cleanup;
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
}
|
2016-05-28 16:23:17 +00:00
|
|
|
|
|
|
|
D3D11_VIDEO_PROCESSOR_STREAM stream = {
|
|
|
|
.Enable = TRUE,
|
|
|
|
.pInputSurface = in_view,
|
|
|
|
};
|
|
|
|
int frame = mp_refqueue_is_second_field(p->queue);
|
|
|
|
hr = ID3D11VideoContext_VideoProcessorBlt(p->video_ctx, p->video_proc,
|
|
|
|
out_view, frame, 1, &stream);
|
|
|
|
if (FAILED(hr)) {
|
|
|
|
MP_ERR(vf, "VideoProcessorBlt failed.\n");
|
|
|
|
goto cleanup;
|
|
|
|
}
|
|
|
|
|
|
|
|
res = 0;
|
|
|
|
cleanup:
|
|
|
|
if (in_view)
|
|
|
|
ID3D11VideoProcessorInputView_Release(in_view);
|
|
|
|
if (out_view)
|
|
|
|
ID3D11VideoProcessorOutputView_Release(out_view);
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
if (res < 0)
|
|
|
|
TA_FREEP(&out);
|
|
|
|
return out;
|
2016-05-28 16:23:17 +00:00
|
|
|
}
|
|
|
|
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
static void vf_d3d11vpp_process(struct mp_filter *vf)
|
2016-05-28 16:23:17 +00:00
|
|
|
{
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
struct priv *p = vf->priv;
|
|
|
|
|
|
|
|
struct mp_image *in_fmt = mp_refqueue_execute_reinit(p->queue);
|
|
|
|
if (in_fmt) {
|
|
|
|
mp_image_pool_clear(p->pool);
|
|
|
|
|
|
|
|
destroy_video_proc(vf);
|
|
|
|
|
|
|
|
p->params = in_fmt->params;
|
|
|
|
p->out_params = p->params;
|
|
|
|
|
2019-10-16 21:37:06 +00:00
|
|
|
p->out_params.hw_subfmt = IMGFMT_NV12;
|
|
|
|
p->out_format = DXGI_FORMAT_NV12;
|
2016-05-28 16:23:17 +00:00
|
|
|
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
p->require_filtering = p->params.hw_subfmt != p->out_params.hw_subfmt;
|
|
|
|
}
|
2016-05-28 16:23:17 +00:00
|
|
|
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
if (!mp_refqueue_can_output(p->queue))
|
|
|
|
return;
|
video: remove d3d11 video processor use from OpenGL interop
We now have a video filter that uses the d3d11 video processor, so it
makes no sense to have one in the VO interop code. The VO uses it for
formats not directly supported by ANGLE (so the video data is converted
to a RGB texture, which ANGLE can take in).
Change this so that the video filter is automatically inserted if
needed. Move the code that maps RGB surfaces to its own inteorp backend.
Add a bunch of new image formats, which are used to enforce the new
constraints, and to automatically insert the filter only when needed.
The added vf mechanism to auto-insert the d3d11vpp filter is very dumb
and primitive, and will work only for this specific purpose. The format
negotiation mechanism in the filter chain is generally not very pretty,
and mostly broken as well. (libavfilter has a different mechanism, and
these mechanisms don't match well, so vf_lavfi uses some sort of hack.
It only works because hwaccel and non-hwaccel formats are strictly
separated.)
The RGB interop is now only used with older ANGLE versions. The only
reason I'm keeping it is because it's relatively isolated (uses only
existing mechanisms and adds no new concepts), and because I want to be
able to compare the behavior of the old code with the new one for
testing. It will be removed eventually.
If ANGLE has NV12 interop, P010 is now handled by converting to NV12
with the video processor, instead of converting it to RGB and using the
old mechanism to import that as a texture.
2016-05-29 15:13:22 +00:00
|
|
|
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
if (!mp_refqueue_should_deint(p->queue) && !p->require_filtering) {
|
|
|
|
// no filtering
|
|
|
|
struct mp_image *in = mp_image_new_ref(mp_refqueue_get(p->queue, 0));
|
|
|
|
if (!in) {
|
|
|
|
mp_filter_internal_mark_failed(vf);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
mp_refqueue_write_out_pin(p->queue, in);
|
video: remove d3d11 video processor use from OpenGL interop
We now have a video filter that uses the d3d11 video processor, so it
makes no sense to have one in the VO interop code. The VO uses it for
formats not directly supported by ANGLE (so the video data is converted
to a RGB texture, which ANGLE can take in).
Change this so that the video filter is automatically inserted if
needed. Move the code that maps RGB surfaces to its own inteorp backend.
Add a bunch of new image formats, which are used to enforce the new
constraints, and to automatically insert the filter only when needed.
The added vf mechanism to auto-insert the d3d11vpp filter is very dumb
and primitive, and will work only for this specific purpose. The format
negotiation mechanism in the filter chain is generally not very pretty,
and mostly broken as well. (libavfilter has a different mechanism, and
these mechanisms don't match well, so vf_lavfi uses some sort of hack.
It only works because hwaccel and non-hwaccel formats are strictly
separated.)
The RGB interop is now only used with older ANGLE versions. The only
reason I'm keeping it is because it's relatively isolated (uses only
existing mechanisms and adds no new concepts), and because I want to be
able to compare the behavior of the old code with the new one for
testing. It will be removed eventually.
If ANGLE has NV12 interop, P010 is now handled by converting to NV12
with the video processor, instead of converting it to RGB and using the
old mechanism to import that as a texture.
2016-05-29 15:13:22 +00:00
|
|
|
} else {
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
mp_refqueue_write_out_pin(p->queue, render(vf));
|
video: remove d3d11 video processor use from OpenGL interop
We now have a video filter that uses the d3d11 video processor, so it
makes no sense to have one in the VO interop code. The VO uses it for
formats not directly supported by ANGLE (so the video data is converted
to a RGB texture, which ANGLE can take in).
Change this so that the video filter is automatically inserted if
needed. Move the code that maps RGB surfaces to its own inteorp backend.
Add a bunch of new image formats, which are used to enforce the new
constraints, and to automatically insert the filter only when needed.
The added vf mechanism to auto-insert the d3d11vpp filter is very dumb
and primitive, and will work only for this specific purpose. The format
negotiation mechanism in the filter chain is generally not very pretty,
and mostly broken as well. (libavfilter has a different mechanism, and
these mechanisms don't match well, so vf_lavfi uses some sort of hack.
It only works because hwaccel and non-hwaccel formats are strictly
separated.)
The RGB interop is now only used with older ANGLE versions. The only
reason I'm keeping it is because it's relatively isolated (uses only
existing mechanisms and adds no new concepts), and because I want to be
able to compare the behavior of the old code with the new one for
testing. It will be removed eventually.
If ANGLE has NV12 interop, P010 is now handled by converting to NV12
with the video processor, instead of converting it to RGB and using the
old mechanism to import that as a texture.
2016-05-29 15:13:22 +00:00
|
|
|
}
|
2016-05-28 16:23:17 +00:00
|
|
|
}
|
|
|
|
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
static void uninit(struct mp_filter *vf)
|
2016-05-28 16:23:17 +00:00
|
|
|
{
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
struct priv *p = vf->priv;
|
2016-05-28 16:23:17 +00:00
|
|
|
|
|
|
|
destroy_video_proc(vf);
|
|
|
|
|
|
|
|
flush_frames(vf);
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
talloc_free(p->queue);
|
2016-05-28 16:23:17 +00:00
|
|
|
talloc_free(p->pool);
|
|
|
|
|
|
|
|
if (p->video_ctx)
|
|
|
|
ID3D11VideoContext_Release(p->video_ctx);
|
|
|
|
|
|
|
|
if (p->video_dev)
|
|
|
|
ID3D11VideoDevice_Release(p->video_dev);
|
|
|
|
|
|
|
|
if (p->device_ctx)
|
|
|
|
ID3D11DeviceContext_Release(p->device_ctx);
|
|
|
|
|
|
|
|
if (p->vo_dev)
|
|
|
|
ID3D11Device_Release(p->vo_dev);
|
|
|
|
}
|
|
|
|
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
static const struct mp_filter_info vf_d3d11vpp_filter = {
|
|
|
|
.name = "d3d11vpp",
|
|
|
|
.process = vf_d3d11vpp_process,
|
|
|
|
.reset = flush_frames,
|
|
|
|
.destroy = uninit,
|
|
|
|
.priv_size = sizeof(struct priv),
|
|
|
|
};
|
|
|
|
|
|
|
|
static struct mp_filter *vf_d3d11vpp_create(struct mp_filter *parent,
|
|
|
|
void *options)
|
2016-05-28 16:23:17 +00:00
|
|
|
{
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
struct mp_filter *f = mp_filter_create(parent, &vf_d3d11vpp_filter);
|
|
|
|
if (!f) {
|
|
|
|
talloc_free(options);
|
|
|
|
return NULL;
|
video: remove d3d11 video processor use from OpenGL interop
We now have a video filter that uses the d3d11 video processor, so it
makes no sense to have one in the VO interop code. The VO uses it for
formats not directly supported by ANGLE (so the video data is converted
to a RGB texture, which ANGLE can take in).
Change this so that the video filter is automatically inserted if
needed. Move the code that maps RGB surfaces to its own inteorp backend.
Add a bunch of new image formats, which are used to enforce the new
constraints, and to automatically insert the filter only when needed.
The added vf mechanism to auto-insert the d3d11vpp filter is very dumb
and primitive, and will work only for this specific purpose. The format
negotiation mechanism in the filter chain is generally not very pretty,
and mostly broken as well. (libavfilter has a different mechanism, and
these mechanisms don't match well, so vf_lavfi uses some sort of hack.
It only works because hwaccel and non-hwaccel formats are strictly
separated.)
The RGB interop is now only used with older ANGLE versions. The only
reason I'm keeping it is because it's relatively isolated (uses only
existing mechanisms and adds no new concepts), and because I want to be
able to compare the behavior of the old code with the new one for
testing. It will be removed eventually.
If ANGLE has NV12 interop, P010 is now handled by converting to NV12
with the video processor, instead of converting it to RGB and using the
old mechanism to import that as a texture.
2016-05-29 15:13:22 +00:00
|
|
|
}
|
2016-05-28 16:23:17 +00:00
|
|
|
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
mp_filter_add_pin(f, MP_PIN_IN, "in");
|
|
|
|
mp_filter_add_pin(f, MP_PIN_OUT, "out");
|
video: remove d3d11 video processor use from OpenGL interop
We now have a video filter that uses the d3d11 video processor, so it
makes no sense to have one in the VO interop code. The VO uses it for
formats not directly supported by ANGLE (so the video data is converted
to a RGB texture, which ANGLE can take in).
Change this so that the video filter is automatically inserted if
needed. Move the code that maps RGB surfaces to its own inteorp backend.
Add a bunch of new image formats, which are used to enforce the new
constraints, and to automatically insert the filter only when needed.
The added vf mechanism to auto-insert the d3d11vpp filter is very dumb
and primitive, and will work only for this specific purpose. The format
negotiation mechanism in the filter chain is generally not very pretty,
and mostly broken as well. (libavfilter has a different mechanism, and
these mechanisms don't match well, so vf_lavfi uses some sort of hack.
It only works because hwaccel and non-hwaccel formats are strictly
separated.)
The RGB interop is now only used with older ANGLE versions. The only
reason I'm keeping it is because it's relatively isolated (uses only
existing mechanisms and adds no new concepts), and because I want to be
able to compare the behavior of the old code with the new one for
testing. It will be removed eventually.
If ANGLE has NV12 interop, P010 is now handled by converting to NV12
with the video processor, instead of converting it to RGB and using the
old mechanism to import that as a texture.
2016-05-29 15:13:22 +00:00
|
|
|
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
struct priv *p = f->priv;
|
|
|
|
p->opts = talloc_steal(p, options);
|
2016-05-28 16:23:17 +00:00
|
|
|
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
// Special path for vf_d3d11_create_outconv(): disable all processing except
|
|
|
|
// possibly surface format conversions.
|
|
|
|
if (!p->opts) {
|
|
|
|
static const struct opts opts = {0};
|
|
|
|
p->opts = (struct opts *)&opts;
|
|
|
|
}
|
2016-05-28 16:23:17 +00:00
|
|
|
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
p->queue = mp_refqueue_alloc(f);
|
2017-12-02 03:33:43 +00:00
|
|
|
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
struct mp_stream_info *info = mp_filter_find_stream_info(f);
|
|
|
|
if (!info || !info->hwdec_devs)
|
|
|
|
goto fail;
|
2016-05-28 16:23:17 +00:00
|
|
|
|
2022-03-20 19:31:38 +00:00
|
|
|
struct hwdec_imgfmt_request params = {
|
|
|
|
.imgfmt = IMGFMT_D3D11,
|
|
|
|
.probing = false,
|
|
|
|
};
|
|
|
|
hwdec_devices_request_for_img_fmt(info->hwdec_devs, ¶ms);
|
2016-05-28 16:23:17 +00:00
|
|
|
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
struct mp_hwdec_ctx *hwctx =
|
2022-03-24 21:18:59 +00:00
|
|
|
hwdec_devices_get_by_imgfmt(info->hwdec_devs, IMGFMT_D3D11);
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
if (!hwctx || !hwctx->av_device_ref)
|
|
|
|
goto fail;
|
|
|
|
AVHWDeviceContext *avhwctx = (void *)hwctx->av_device_ref->data;
|
|
|
|
AVD3D11VADeviceContext *d3dctx = avhwctx->hwctx;
|
2017-12-02 03:33:43 +00:00
|
|
|
|
|
|
|
p->vo_dev = d3dctx->device;
|
2016-05-28 16:23:17 +00:00
|
|
|
ID3D11Device_AddRef(p->vo_dev);
|
|
|
|
|
|
|
|
HRESULT hr;
|
|
|
|
|
|
|
|
hr = ID3D11Device_QueryInterface(p->vo_dev, &IID_ID3D11VideoDevice,
|
|
|
|
(void **)&p->video_dev);
|
|
|
|
if (FAILED(hr))
|
|
|
|
goto fail;
|
|
|
|
|
|
|
|
ID3D11Device_GetImmediateContext(p->vo_dev, &p->device_ctx);
|
|
|
|
if (!p->device_ctx)
|
|
|
|
goto fail;
|
|
|
|
hr = ID3D11DeviceContext_QueryInterface(p->device_ctx, &IID_ID3D11VideoContext,
|
|
|
|
(void **)&p->video_ctx);
|
|
|
|
if (FAILED(hr))
|
|
|
|
goto fail;
|
|
|
|
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
p->pool = mp_image_pool_new(f);
|
|
|
|
mp_image_pool_set_allocator(p->pool, alloc_pool, f);
|
|
|
|
mp_image_pool_set_lru(p->pool);
|
|
|
|
|
|
|
|
mp_refqueue_add_in_format(p->queue, IMGFMT_D3D11, 0);
|
|
|
|
|
|
|
|
mp_refqueue_set_refs(p->queue, 0, 0);
|
|
|
|
mp_refqueue_set_mode(p->queue,
|
|
|
|
(p->opts->deint_enabled ? MP_MODE_DEINT : 0) |
|
|
|
|
MP_MODE_OUTPUT_FIELDS |
|
|
|
|
(p->opts->interlaced_only ? MP_MODE_INTERLACED_ONLY : 0));
|
|
|
|
|
|
|
|
return f;
|
2016-05-28 16:23:17 +00:00
|
|
|
|
|
|
|
fail:
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
talloc_free(f);
|
|
|
|
return NULL;
|
2016-05-28 16:23:17 +00:00
|
|
|
}
|
|
|
|
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
#define OPT_BASE_STRUCT struct opts
|
2016-05-28 16:23:17 +00:00
|
|
|
static const m_option_t vf_opts_fields[] = {
|
options: change option macros and all option declarations
Change all OPT_* macros such that they don't define the entire m_option
initializer, and instead expand only to a part of it, which sets certain
fields. This requires changing almost every option declaration, because
they all use these macros. A declaration now always starts with
{"name", ...
followed by designated initializers only (possibly wrapped in macros).
The OPT_* macros now initialize the .offset and .type fields only,
sometimes also .priv and others.
I think this change makes the option macros less tricky. The old code
had to stuff everything into macro arguments (and attempted to allow
setting arbitrary fields by letting the user pass designated
initializers in the vararg parts). Some of this was made messy due to
C99 and C11 not allowing 0-sized varargs with ',' removal. It's also
possible that this change is pointless, other than cosmetic preferences.
Not too happy about some things. For example, the OPT_CHOICE()
indentation I applied looks a bit ugly.
Much of this change was done with regex search&replace, but some places
required manual editing. In particular, code in "obscure" areas (which I
didn't include in compilation) might be broken now.
In wayland_common.c the author of some option declarations confused the
flags parameter with the default value (though the default value was
also properly set below). I fixed this with this change.
2020-03-14 20:28:01 +00:00
|
|
|
{"deint", OPT_FLAG(deint_enabled)},
|
|
|
|
{"interlaced-only", OPT_FLAG(interlaced_only)},
|
|
|
|
{"mode", OPT_CHOICE(mode,
|
|
|
|
{"blend", D3D11_VIDEO_PROCESSOR_PROCESSOR_CAPS_DEINTERLACE_BLEND},
|
|
|
|
{"bob", D3D11_VIDEO_PROCESSOR_PROCESSOR_CAPS_DEINTERLACE_BOB},
|
|
|
|
{"adaptive", D3D11_VIDEO_PROCESSOR_PROCESSOR_CAPS_DEINTERLACE_ADAPTIVE},
|
|
|
|
{"mocomp", D3D11_VIDEO_PROCESSOR_PROCESSOR_CAPS_DEINTERLACE_MOTION_COMPENSATION},
|
|
|
|
{"ivctc", D3D11_VIDEO_PROCESSOR_PROCESSOR_CAPS_INVERSE_TELECINE},
|
|
|
|
{"none", 0})},
|
2016-05-28 16:23:17 +00:00
|
|
|
{0}
|
|
|
|
};
|
|
|
|
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
const struct mp_user_filter_entry vf_d3d11vpp = {
|
|
|
|
.desc = {
|
|
|
|
.description = "D3D11 Video Post-Process Filter",
|
|
|
|
.name = "d3d11vpp",
|
|
|
|
.priv_size = sizeof(OPT_BASE_STRUCT),
|
|
|
|
.priv_defaults = &(const OPT_BASE_STRUCT) {
|
|
|
|
.deint_enabled = 1,
|
2018-02-12 17:23:22 +00:00
|
|
|
.interlaced_only = 0,
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
.mode = D3D11_VIDEO_PROCESSOR_PROCESSOR_CAPS_DEINTERLACE_BOB,
|
|
|
|
},
|
|
|
|
.options = vf_opts_fields,
|
2016-05-28 16:23:17 +00:00
|
|
|
},
|
video: rewrite filtering glue code
Get rid of the old vf.c code. Replace it with a generic filtering
framework, which can potentially handle more than just --vf. At least
reimplementing --af with this code is planned.
This changes some --vf semantics (including runtime behavior and the
"vf" command). The most important ones are listed in interface-changes.
vf_convert.c is renamed to f_swscale.c. It is now an internal filter
that can not be inserted by the user manually.
f_lavfi.c is a refactor of player/lavfi.c. The latter will be removed
once --lavfi-complex is reimplemented on top of f_lavfi.c. (which is
conceptually easy, but a big mess due to the data flow changes).
The existing filters are all changed heavily. The data flow of the new
filter framework is different. Especially EOF handling changes - EOF is
now a "frame" rather than a state, and must be passed through exactly
once.
Another major thing is that all filters must support dynamic format
changes. The filter reconfig() function goes away. (This sounds complex,
but since all filters need to handle EOF draining anyway, they can use
the same code, and it removes the mess with reconfig() having to predict
the output format, which completely breaks with libavfilter anyway.)
In addition, there is no automatic format negotiation or conversion.
libavfilter's primitive and insufficient API simply doesn't allow us to
do this in a reasonable way. Instead, filters can use f_autoconvert as
sub-filter, and tell it which formats they support. This filter will in
turn add actual conversion filters, such as f_swscale, to perform
necessary format changes.
vf_vapoursynth.c uses the same basic principle of operation as before,
but with worryingly different details in data flow. Still appears to
work.
The hardware deint filters (vf_vavpp.c, vf_d3d11vpp.c, vf_vdpaupp.c) are
heavily changed. Fortunately, they all used refqueue.c, which is for
sharing the data flow logic (especially for managing future/past
surfaces and such). It turns out it can be used to factor out most of
the data flow. Some of these filters accepted software input. Instead of
having ad-hoc upload code in each filter, surface upload is now
delegated to f_autoconvert, which can use f_hwupload to perform this.
Exporting VO capabilities is still a big mess (mp_stream_info stuff).
The D3D11 code drops the redundant image formats, and all code uses the
hw_subfmt (sw_format in FFmpeg) instead. Although that too seems to be a
big mess for now.
f_async_queue is unused.
2018-01-16 10:53:44 +00:00
|
|
|
.create = vf_d3d11vpp_create,
|
2016-05-28 16:23:17 +00:00
|
|
|
};
|