32e34692fb
avcodec/hevc/ps: print the correct unsupported scalability value found
...
With multilayer001.heic, which signals Spatial scalability:
Before:
[hevc @ ...] Scalability type 1 not supported
[hevc @ ...] Ignoring unsupported VPS extension
After:
[hevc @ ...] Scalability type 2 not supported
[hevc @ ...] Ignoring unsupported VPS extension
Signed-off-by: James Almer <jamrial@gmail.com>
2024-10-05 14:36:04 -03:00
099f88b864
avcodec/libx265: unbreak build for X265_BUILD >= 213
...
Earlier, x265 made an API change to support alpha and
other multiple layer pictures. We added guards to accommodate
that in 1f801dfdb5https://bitbucket.org/multicoreware/x265_git/commits/78e5b703b1
Updated our wrapper guards to unbreak build again.
2024-10-05 20:18:47 +05:30
2705c0bd81
configure: Enable -Wno-implicit-const-int-float-conversion if available
...
This silences a lot of compile warnings (around 160 instances at least), when
compiling with Clang.
These warnings look like this:
libavformat/http.c:176:133: warning: implicit conversion from 'long long' to 'double' changes value from 9223372036854775807 to 9223372036854775808 [-Wimplicit-const-int-float-conversion]
176 | { "end_offset", "try to limit the request to bytes preceding this offset", OFFSET(end_off), AV_OPT_TYPE_INT64, { .i64 = 0 }, 0, INT64_MAX, D },
| ~ ^~~~~~~~~
Signed-off-by: Martin Storsjö <martin@martin.st>
2024-10-04 22:57:51 +03:00
5c4ede6b4f
libavcodec: x86: Remove an explicit include of config.asm
...
This file is never included explicitly anywhere else, it's only
included implicitly by passing -Pconfig.asm on the command line.
Signed-off-by: Martin Storsjö <martin@martin.st>
2024-10-04 22:57:38 +03:00
a27f3c590f
av1dec: Don't crash if decoding of some frames have failed
...
If decoding with hwaccel, but decoding fails, these pointers
are null at this point.
Signed-off-by: Martin Storsjö <martin@martin.st>
2024-10-04 22:57:03 +03:00
846fbc395b
avcodec/vvc: simplify priority logical to improve performance for 4K/8K
...
For 4K/8K video processing, it's possible to have over 1,000 tasks pending on the executor.
In such cases, O(n) and O(log(n)) insertion times are too costly.
Reducing this to O(1) will significantly decrease the time spent in critical sections
clip | before | after | delta
------------------------------------------------------------|--------|--------|-------
VVC_HDR_UHDTV2_OpenGOP_7680x4320_50fps_HLG10.bit | 24 | 27 | 12.5%
VVC_HDR_UHDTV2_OpenGOP_7680x4320_50fps_HLG10_HighBitrate.bit| 12 | 17 | 41.7%
tears_of_steel_4k_8M_8bit_2000.vvc | 34 | 102 | 200.0%
VVC_UHDTV1_OpenGOP_3840x2160_60fps_HLG10.bit | 126 | 128 | 1.6%
RitualDance_1920x1080_60_10_420_37_RA.266 | 350 | 378 | 8.0%
NovosobornayaSquare_1920x1080.bin | 341 | 369 | 8.2%
Tango2_3840x2160_60_10_420_27_LD.266 | 69 | 70 | 1.4%
RitualDance_1920x1080_60_10_420_32_LD.266 | 243 | 259 | 6.6%
Chimera_8bit_1080P_1000_frames.vvc | 420 | 392 | -6.7%
BQTerrace_1920x1080_60_10_420_22_RA.vvc | 148 | 144 | -2.7%
2024-10-04 21:58:42 +08:00
40a14ef970
avcodec/executor: remove unused ready callback
...
Due to the nature of multithreading, using a "ready check" mechanism may introduce a deadlock. For example:
Suppose all tasks have been submitted to the executor, and the last thread checks the entire list and finds
no ready tasks. It then goes to sleep, waiting for a new task. However, for some multithreading-related reason,
a task becomes ready after the check. Since no other thread is aware of this and no new tasks are being added to
the executor, a deadlock occurs.
In VVC, this function is unnecessary because we use a scoreboard. All tasks submitted to the executor are ready tasks.
2024-10-04 21:58:42 +08:00
8446e27bf3
avcodec: make a local copy of executor
...
We still need several refactors to improve the current VVC decoder's performance,
which will frequently break the API/ABI. To mitigate this, we've copied the executor from
avutil to avcodec. Once the API/ABI is stable, we will move this class back to avutil
2024-10-04 21:58:42 +08:00
f25c9cc213
fftools: log unconnected filter output label
2024-10-04 14:27:20 +02:00
5beeb3a1f9
fftools: do not access out of bounds filtergraph
...
The log message was logged for `filtergraphs[j]` which would cause a
heap buffer overflow in certain circumstances.
Correctly it should be logged for the current filtergraph, so just
use `fg` here.
2024-10-04 14:27:20 +02:00
ac092c6707
hwcontext_vulkan: guard all uses of new spec defines and fix stray bracket
...
This fixes compilation with less recent Vulkan headers.
2024-10-04 10:41:03 +02:00
a304cbeb8d
vulkan: add profiling debug setting
...
This simply keeps all shader optimizations, but allows debug
data to be generated.
2024-10-04 10:10:46 +02:00
832947be02
vulkan: don't enable GL_EXT_buffer_reference by default
...
Only nlmeans_vulkan uses it.
2024-10-04 10:10:45 +02:00
2bb0b6dd2c
vulkan_shaderc: add more stages, target Vulkan 1.3 and only output debug if needed
2024-10-04 10:10:45 +02:00
a2a983e1d0
vulkan_glslang: add more stages, target Vulkan 1.3 and output debug info
2024-10-04 10:10:45 +02:00
0bc4e03322
lavfi/*vulkan: forward FFVulkanContext to SPIR-V compilers
...
We need to know what extensions are supported.
2024-10-04 10:10:44 +02:00
356d1cc8ff
vulkan: parse instance list and add the DEBUG_UTILS extension
...
Required to let users know whether debugging is active.
2024-10-04 10:10:44 +02:00
e3676d96cb
hwcontext_vulkan: move device feature struct setup to a new function
2024-10-04 10:10:43 +02:00
535e5eb7f3
hwcontext_vulkan: enable VK_KHR_shader_relaxed_extended_instruction
2024-10-04 10:10:43 +02:00
0d5bfd0b21
vulkan_functions: change extension type to a typedef uint64_t
...
We were getting a bit too close for comfort to the 32-bit limit
on enums.
2024-10-04 10:10:42 +02:00
d80f9f55c8
vulkan: always enable GL_EXT_scalar_block_layout
...
This makes std430 (which we use everywhere already) fully match C
layout.
Extension was made mandatory in 1.2.
2024-10-04 10:10:42 +02:00
37d5cb84e8
vulkan: check if current buffer has finished execution before picking another
...
This saves resources, as dependencies are freed/reclaimed with a lower latency,
and provies a speedup.
2024-10-04 10:10:42 +02:00
877c5a9692
vulkan: use shader objects if supported
...
Shader objects finally allow completely independent shaders.
2024-10-04 10:10:41 +02:00
d233d9902f
vulkan: move shader data execution state to execution pools
...
This finally permits using fully compiled shaders across
multiple execution contexts.
2024-10-04 10:10:41 +02:00
0a37d5a3b1
vulkan: merge FFVkSPIRVShader and FFVkPipeline into FFVkShader
...
Pipelines are just shaders. There's no reason to treat them
differently.
This also lets us implement shader objects and is an overall
cleanup.
2024-10-04 10:10:36 +02:00
aad4d5745d
lavfi/vsrc_testsrc: switch to query_func2()
2024-10-04 10:03:34 +02:00
28bde4a141
lavfi/vaf_spectrumsynth: switch to query_func2()
2024-10-04 10:03:34 +02:00
3824ee2faf
lavfi/vf_stack_{vaapi,qsv}: replace query_formats with a pixfmt list
2024-10-04 10:03:34 +02:00
014a4214a9
lavfi/src_movie: switch to query_func2()
2024-10-04 10:03:34 +02:00
55c99fb9c4
lavfi/src_avsynctest: switch to query_func2()
2024-10-04 10:03:34 +02:00
534eef2ace
lavfi/qrencode: switch to query_func2()
2024-10-04 10:03:34 +02:00
f10986a5ec
lavfi/f_streamselect: drop useless query_formats() callback
...
It achieves the same effect as ff_default_query_formats(), which gets
called implicitly.
2024-10-04 10:03:34 +02:00
a6f579544a
lavfi/f_select: switch to query_func2()
...
Also, drop redundant calls that also happen implicitly in generic code.
2024-10-04 10:03:34 +02:00
9e1184b122
lavfi/f_graphmonitor: switch to query_func2()
2024-10-04 10:03:34 +02:00
60192367a8
lavfi/f_ebur128: switch to query_func2()
...
Also, drop redundant calls that also happen implicitly in generic code.
2024-10-04 10:03:34 +02:00
4e66d0c5d0
lavfi/f_drawgraph: switch to query_func2()
2024-10-04 10:03:34 +02:00
670530a927
lavfi/buffersrc: switch to query_func2()
2024-10-04 10:03:34 +02:00
a7fe27f964
lavfi/buffersrc: validate hw context presence in video_init()
...
That is the more appropriate place for it than query_formats().
2024-10-04 10:03:34 +02:00
2fa142f7c0
lavfi/buffersink: switch to query_func2()
2024-10-04 10:03:34 +02:00
b8bf2f4e17
lavfi/buffersink: move channel layout parsing to init
...
That is a more appropriate place for this, and will also be useful in
future commits.
2024-10-04 10:03:34 +02:00
2aad37ffb5
lavfi/buffersink: move the option sanity check to init
...
Options are set before init, so that is the appropriate place to
validate them.
2024-10-04 10:03:34 +02:00
eff9ed7bff
avcodec/mediacodecenc: Fix access of uninitialized value
...
When crop is skipped, av_strlcatf will access `str` which isn't
initialized properly.
Signed-off-by: Zhao Zhili <zhilizhao@tencent.com>
2024-10-04 15:43:57 +08:00
556c767786
avcodec/ffv1enc: Correct error message about unsupported version
...
Sponsored-by: Sovereign Tech Fund
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2024-10-03 22:05:46 +02:00
358fdf3083
avfilter: add missing build deps to msad filter
...
Signed-off-by: James Almer <jamrial@gmail.com>
2024-10-02 12:44:49 -03:00
b094c47ab2
fate/mov: add a test for heic images with cropping and rotation metadata
...
Signed-off-by: James Almer <jamrial@gmail.com>
2024-10-02 12:38:24 -03:00
650ce61745
avformat/mov: don't abort on invalid clap box data
...
Unless explode is requested.
Signed-off-by: James Almer <jamrial@gmail.com>
2024-10-02 12:38:24 -03:00
7573f02840
avformat/mov: get heif image mirroring from imir box
...
Complements 76eb3e5ff3
2024-10-02 12:38:24 -03:00
f225e6f3c6
avformat/mov: parse clap boxes that reference heif items
...
Signed-off-by: James Almer <jamrial@gmail.com>
2024-10-02 12:38:24 -03:00
ba6eeb2c65
avformat/mov: parse colr boxes that reference tile grids
...
Signed-off-by: James Almer <jamrial@gmail.com>
2024-10-02 12:38:24 -03:00
58c265d956
avformat/mov: factorize getting the current item
...
Signed-off-by: James Almer <jamrial@gmail.com>
2024-10-02 12:38:24 -03:00
James Almer
Gyan Doshi
Martin Storsjö
Nuo Mi
Marvin Scholz
Lynne
Anton Khirnov
Zhao Zhili
Michael Niedermayer