The implementation is based on:
https://learn.microsoft.com/en-us/windows/win32/medfound/direct3d-12-video-overview
With the Direct3D 12 video decoding support, we can render or process
the decoded images by the pixel shaders or compute shaders directly
without the extra copy overhead, which is beneficial especially if you
are trying to render or post-process a 4K or 8K video.
The command below is how to enable d3d12va:
ffmpeg -hwaccel d3d12va -i input.mp4 output.mp4
Signed-off-by: Wu Jianhua <toqsxw@outlook.com>
Signed-off-by: Tong Wu <tong1.wu@intel.com>
In close_output(), a dummy frame is created with format NONE passed
to enc_open(), which isn't prepared for it. The NULL pointer
dereference happened at
av_pix_fmt_desc_get(enc_ctx->pix_fmt)->comp[0].depth.
When fgt.graph is NULL, skip fg_output_frame() since there is
nothing to output.
frame #0: 0x0000005555bc34a4 ffmpeg_g`enc_open(opaque=0xb400007efe2db690, frame=0xb400007efe2d9f70) at ffmpeg_enc.c:235:44
frame #1: 0x0000005555bef250 ffmpeg_g`enc_open(sch=0xb400007dde2d4090, enc=0xb400007e4e2daad0, frame=0xb400007efe2d9f70) at ffmpeg_sched.c:1462:11
frame #2: 0x0000005555bee094 ffmpeg_g`send_to_enc(sch=0xb400007dde2d4090, enc=0xb400007e4e2daad0, frame=0xb400007efe2d9f70) at ffmpeg_sched.c:1571:19
frame #3: 0x0000005555bee01c ffmpeg_g`sch_filter_send(sch=0xb400007dde2d4090, fg_idx=0, out_idx=0, frame=0xb400007efe2d9f70) at ffmpeg_sched.c:2154:12
frame #4: 0x0000005555bcf124 ffmpeg_g`close_output(ofp=0xb400007e4e2d85b0, fgt=0x0000007d1790eb08) at ffmpeg_filter.c:2225:15
frame #5: 0x0000005555bcb000 ffmpeg_g`fg_output_frame(ofp=0xb400007e4e2d85b0, fgt=0x0000007d1790eb08, frame=0x0000000000000000) at ffmpeg_filter.c:2317:16
frame #6: 0x0000005555bc7e48 ffmpeg_g`filter_thread(arg=0xb400007eae2ce7a0) at ffmpeg_filter.c:2836:15
frame #7: 0x0000005555bee568 ffmpeg_g`task_wrapper(arg=0xb400007d8e2db478) at ffmpeg_sched.c:2200:21
Signed-off-by: Zhao Zhili <zhilizhao@tencent.com>
From AOSP doc, these values are device and codec specific, but lower
values generally result in more efficient (smaller-sized) encoding.
For example, global_quality 50 on Pixel 6 results a 1080P 30 FPS
HEVC with 3744 kb/s, while global_quality 80 results 28178 kb/s.
Fix#10689
Signed-off-by: Zhao Zhili <zhilizhao@tencent.com>
The ffmpeg coding style doesn't usually use const on scalar
parameters (or on the pointer values - as opposed to the type
that is pointed to, where it has a semantic meaning), contrary
to the dav1d coding style (where this was imported from).
This avoids warnings about differences in the type signatures
between declaration and definition of this function, with older
versions of MSVC.
The issue was observed with one version of MSVC 2017,
19.16.27024.1, with warnings like these:
src/tests/checkasm/checkasm.c(969): warning C4028: formal parameter 3 different from declaration
The warning itself is bogus as the const here is harmless, and
newer versions of MSVC no longer warn about this.
Signed-off-by: Martin Storsjö <martin@martin.st>
This can be used to run tests multple times, with e.g. differing
QEMU settings, by adding something like this to the FATE configuration
file:
target_exec="qemu-aarch64-static"
fate_targets="fate-checkasm fate-cpu"
fate_environments="sve128 sve256 sve512"
sve128_env="QEMU_CPU=max,sve128=on"
sve256_env="QEMU_CPU=max,sve256=on"
sve512_env="QEMU_CPU=max,sve512=on"
It's also possible to customize the target_exec command further
by injecting a sufficiently quoted variable into it, which then can
be updated for each run, e.g. target_exec="\$(CUR_EXEC_CMD)".
For each of the environment names in fate_environments, the tests
that are run get the name suffixed on the fate tests in the
test log and fate report, e.g. "fate-checkasm-h264dsp_sve128".
Signed-off-by: Martin Storsjö <martin@martin.st>
This can be useful if doing testing of uncommon CPU extensions by
running tests with QEMU (by configuring with e.g.
"target_exec=qemu-aarch64"), by only running the checkasm tests,
to get a reasonable test coverage without excessive test runtime.
For such a config, setting fate_targets="fate-checkasm fate-cpu"
can be a good tradeoff.
Signed-off-by: Martin Storsjö <martin@martin.st>
Reduce false positives for VVC files by adding additional checks in
`vvc_probe`. Specifically, `nuh_temporal_id_plus1` is tested for valid
values in extra cases depending on the NAL unit type, as per ITU-T H.266
section 7.4.2.2.
Resolves trac #10703.
Signed-off-by: Frank Plowman <post@frankplowman.com>
Signed-off-by: Zhao Zhili <zhilizhao@tencent.com>
Converting from an integer to HWND (which is a pointer) requires
an explicit cast, otherwise Clang errors out like this:
src/libavdevice/gdigrab.c:280:14: error: incompatible integer to pointer conversion assigning to 'HWND' (aka 'struct HWND__ *') from 'long' [-Wint-conversion]
280 | hwnd = strtol(name, &p, 0);
| ^ ~~~~~~~~~~~~~~~~~~~
(With GCC and MSVC, this was a mere warning, but with recent Clang,
this is an error.)
After adding a cast, all compilers also warn something like this:
src/libavdevice/gdigrab.c:280:16: warning: cast to 'HWND' (aka 'struct HWND__ *') from smaller integer type 'long' [-Wint-to-pointer-cast]
280 | hwnd = (HWND) strtol(name, &p, 0);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~
On Windows, long types are 32 bit, so to get a usable pointer, we
need to use long long. And interpret it as unsigned long long
while at it - i.e. using strtoull.
Finally, right above it, the code triggered the following warning:
src/libavdevice/gdigrab.c:278:15: warning: mixing declarations and code is incompatible with standards before C99 [-Wdeclaration-after-statement]
278 | char *p;
| ^
Signed-off-by: Martin Storsjö <martin@martin.st>
With the way the runtime checks are currently set up, every single
openh264 release, no matter how minor, is considered an ABI break and
requires ffmpeg recompilation. This is unnecessarily strict because it
doesn't allow downstream distributions to ship any openh264 bug fix
version updates without breaking ffmpeg's openh264 support.
Years ago, at the time when ffmpeg's openh264 support was merged,
openh264 releases were done without a versioned soname (the library was
just libopenh264.so, unversioned). Since then, starting with version
1.3.0, openh264 has started using versioned sonames and the intent has
been to bump the soname every time there's a new release with an ABI
change.
This patch drops the exact version check and instead adds a minimum
requirement on 1.3.0 to the configure script.
Signed-off-by: Kalev Lember <klember@redhat.com>
Signed-off-by: Martin Storsjö <martin@martin.st>
During a resampling operation where
1) user has specified first_pts
2) SWR_FLAG_RESAMPLE is not set initially (directly or otherwise)
3) first_pts has been fulfilled (always using hard compensation)
then upon first encountering a delay where a soft compensation is
required, swr_set_compensation will lead to another init of swr which
will reset outpts to the specified firstpts thus leading to an output
frame having its pts = firstpts. When the next input frame is received,
swr will see a large delay and inject silence from firstpts to the
current frame's pts. This can lead to severe desync and in worst case,
loss of audio playback.
Parameter firstpts initialized to AV_NOPTS_VALUE in swr_alloc and then
checked in swr_init to avoid resetting outpts, thus avoiding reapplication
of firstpts.
Fixes#4131.
The {br}/{abr} directives are not limited to post-encoding, they can
also be used pre-muxing. The already-present {packet} tag describes this
more accurately, so just drop the assertions.
They cannot store 1 as signed, only 0 and -1.
Avoids warnings such as:
implicit truncation from 'int' to a one-bit wide bit-field changes value from 1 to -1 [-Wsingle-bit-bitfield-constant-conversion]
The names of the cpu flags, when parsed from a string with
av_parse_cpu_caps, are parsed by the libavutil eval functions. These
interpret dashes as subtractions. Therefore, these previous cpu flag
names haven't been possible to set.
Use the official names for these extensions, as the previous ad-hoc
names wasn't parseable.
libavutil/tests/cpu tests that the cpu flags can be set, and prints
the detected flags.
Acked-by: Rémi Denis-Courmont <remi@remlab.net>
Signed-off-by: Martin Storsjö <martin@martin.st>
x11grab can capture windows by their ID, but gdigrab can only capture
windows by their names, internally calling FindWindowW to lookup its
handle.
This patch simply allows the user to specify a window handle directly.
Signed-off-by: Lena <lena@nihil.gay>
The 8x4 and 4x4 use a needlessly large multiplier (unless/until we care
about embedded 64-bit-vector hardware). This is merely suboptimal.
The 8x4 case also uses an incorrect vector length, which leads to incorrect
behaviour on future/hypothetical hardware with 256-bit or larger vectors.
Pointed-out-by: Martin Storsjö <martin@martin.st>
It's available on macOS since 10.8, but not available on iOS until
16.0.
Reported and tested by xufuji456.
Signed-off-by: Zhao Zhili <zhilizhao@tencent.com>
We can't call ff_get_rv_vlenb() if we don't have RVV available
at all.
Acked-by: Rémi Denis-Courmont <remi@remlab.net>
Signed-off-by: Martin Storsjö <martin@martin.st>
The difference of yolov4 is that sigmoid function needed to be applied
on x, y coordinates. Also make it compatiple with NHWC output as the
yolov4 model from openvino model zoo has NHWC output layout.
Model refer to: https://github.com/openvinotoolkit/open_model_zoo/tree/master/models/public/yolo-v4-tf
Signed-off-by: Wenbin Chen <wenbin.chen@intel.com>
Reviewed-by: Guo Yejun <yejun.guo@intel.com>
Add yolov3 support. The difference of yolov3 is that it has multiple
outputs in different scale to perform better on both large and small
object.
The model detail refer to: https://github.com/openvinotoolkit/open_model_zoo/tree/master/models/public/yolo-v3-tf
Signed-off-by: Wenbin Chen <wenbin.chen@intel.com>
Reviewed-by: Guo Yejun <yejun.guo@intel.com>
Add input pad to get model input resolution. Detection models always
have fixed input size. And the output coordinators are based on the
input resolution, so we need to get input size to map coordinators to
our real output frames.
Signed-off-by: Wenbin Chen <wenbin.chen@intel.com>
Reviewed-by: Guo Yejun <yejun.guo@intel.com>
Add multiple output support to openvino backend. You can use '&' to
split different output when you set output name using command line.
Signed-off-by: Wenbin Chen <wenbin.chen@intel.com>
Reviewed-by: Guo Yejun <yejun.guo@intel.com>
Avoids a -Wstringop-truncation warning by using av_strlcopy instead of
strncpy. Additionally, prints a warning to the log context if this
truncation occurred.
Signed-off-by: Leo Izen <leo.izen@gmail.com>
The loop iterates over the length of the vector, not the order. This is
to avoid reloading the same data for each lag value. However this means
the loop only works if the maximum order is no larger than VLENB.
The loop is roughly equivalent to:
for (size_t j = 0; j < lag; j++)
autoc[j] = 1.;
while (len > lag) {
for (ptrdiff_t j = 0; j < lag; j++)
autoc[j] += data[j] * *data;
data++;
len--;
}
while (len > 0) {
for (ptrdiff_t j = 0; j < len; j++)
autoc[j] += data[j] * *data;
data++;
len--;
}
Since register pressure is only at 50%, it should be possible to implement
the same loop for order up to 2xVLENB. But this is left for future work.
Performance numbers are all over the place from ~1.25x to ~4x speedups,
but at least they are always noticeably better than nothing.
This test verifies the parser's handling of multiframe JXL files that
have an entropy-encoded permuted table of contents for each frame. The
testcase is actually six JXL codestreams concatenated together, and the
parser needs to be able to find the boundaries.
Signed-off-by: Leo Izen <leo.izen@gmail.com>
Wu Jianhua
Zhao Zhili
Martin Storsjö
James Almer
Frank Plowman
Michael Niedermayer
Kalev Lember
Gyan Doshi
Anton Khirnov
hung kuishing
Lena
Rémi Denis-Courmont
Wenbin Chen
Leo Izen