virtualenv/FFmpeg
1
0
Fork 0
mirror of https://github.com/FFmpeg/FFmpeg.git synced 2025-08-10 06:10:52 +02:00
Code Issues Releases Activity
104,381 Commits 38 Branches 406 Tags
fd3bd5c492834bd100bb2462d1e1dc25a66f28f8
Commit Graph

104381 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Martin Storsjö
fd3bd5c492 aarch64: h264qpel: Do vertical filtering without transposing 
This gives rather big speedups on these functions:

Before:
put_h264_qpel_8_mc01_8_neon:     241.0   131.5   138.7
put_h264_qpel_8_mc02_8_neon:     214.7   121.2   127.5
put_h264_qpel_8_mc03_8_neon:     242.5   131.2   135.7
put_h264_qpel_8_mc11_8_neon:     421.2   218.7   251.0
put_h264_qpel_8_mc12_8_neon:     878.0   509.5   537.5
put_h264_qpel_8_mc13_8_neon:     423.7   217.0   252.0
put_h264_qpel_8_mc21_8_neon:     858.2   479.5   514.0
put_h264_qpel_8_mc22_8_neon:     649.7   385.2   403.0
put_h264_qpel_8_mc23_8_neon:     860.2   476.5   517.7
put_h264_qpel_8_mc31_8_neon:     437.2   219.5   252.5
put_h264_qpel_8_mc32_8_neon:     892.5   510.5   546.0
put_h264_qpel_8_mc33_8_neon:     438.2   218.5   257.0
put_h264_qpel_16_mc01_8_neon:    944.2   509.7   546.7
put_h264_qpel_16_mc02_8_neon:    878.7   469.5   509.7
put_h264_qpel_16_mc03_8_neon:    945.7   510.7   557.0
put_h264_qpel_16_mc11_8_neon:   1663.2   858.5   979.5
put_h264_qpel_16_mc12_8_neon:   3510.2  2027.7  2112.7
put_h264_qpel_16_mc13_8_neon:   1664.7   857.5   980.5
put_h264_qpel_16_mc21_8_neon:   3366.2  1928.5  2030.5
put_h264_qpel_16_mc22_8_neon:   2584.7  1514.7  1590.2
put_h264_qpel_16_mc23_8_neon:   3367.7  1927.7  2035.0
put_h264_qpel_16_mc31_8_neon:   1716.7   849.7   997.0
put_h264_qpel_16_mc32_8_neon:   3564.0  2044.2  3835.2
put_h264_qpel_16_mc33_8_neon:   1717.7   863.0   989.5

After:
put_h264_qpel_8_mc01_8_neon:     136.0    73.7    76.0
put_h264_qpel_8_mc02_8_neon:     108.7    65.0    64.0
put_h264_qpel_8_mc03_8_neon:     137.5    72.7    73.0
put_h264_qpel_8_mc11_8_neon:     316.2   159.0   188.5
put_h264_qpel_8_mc12_8_neon:     653.0   375.5   384.7
put_h264_qpel_8_mc13_8_neon:     318.7   165.5   189.5
put_h264_qpel_8_mc21_8_neon:     739.2   385.7   432.5
put_h264_qpel_8_mc22_8_neon:     530.7   295.5   309.5
put_h264_qpel_8_mc23_8_neon:     741.2   393.7   421.0
put_h264_qpel_8_mc31_8_neon:     332.2   162.5   190.0
put_h264_qpel_8_mc32_8_neon:     667.5   378.2   390.5
put_h264_qpel_8_mc33_8_neon:     332.7   166.5   195.5
put_h264_qpel_16_mc01_8_neon:    524.2   285.2   294.0
put_h264_qpel_16_mc02_8_neon:    454.7   252.2   250.2
put_h264_qpel_16_mc03_8_neon:    525.7   286.0   283.0
put_h264_qpel_16_mc11_8_neon:   1243.2   630.7   726.7
put_h264_qpel_16_mc12_8_neon:   2610.2  1479.7  1481.2
put_h264_qpel_16_mc13_8_neon:   1250.5   631.7   727.7
put_h264_qpel_16_mc21_8_neon:   2890.2  1571.2  1679.7
put_h264_qpel_16_mc22_8_neon:   2108.7  1177.5  1223.5
put_h264_qpel_16_mc23_8_neon:   2891.7  1578.7  1667.7
put_h264_qpel_16_mc31_8_neon:   1296.7   630.5   752.5
put_h264_qpel_16_mc32_8_neon:   2664.0  1483.2  1503.5
put_h264_qpel_16_mc33_8_neon:   1297.7   632.5   747.2

I.e. overall a 20%-60% reduction in runtime of these
functions.

Signed-off-by: Martin Storsjö <martin@martin.st>
 2021-10-18 14:27:58 +03:00
Martin Storsjö
2d5a7f6d00 arm/aarch64: Improve scheduling in the avg form of h264_qpel 
Don't use the loaded registers directly, avoiding stalls on in
order cores. Use vrhadd.u8 with q registers where easily possible.

Signed-off-by: Martin Storsjö <martin@martin.st>
 2021-10-18 14:27:36 +03:00
Gyan Doshi
d04c005021 doc/filters: correct description of select filter variables 2021-10-18 14:28:04 +05:30
Paul B Mahol
bbbf95848b avfilter/vf_w3fdif: do not output extra frame at start with deint=interlaced 2021-10-18 09:29:41 +02:00
Michael Niedermayer
85c169f6a6 avcodec/speexdec: Seed should be unsigned otherwise the operations done on it are undefined 
Fixes: signed integer overflow: 1664525000 + 1013904223 cannot be represented in type 'int'
Fixes: 39865/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_SPEEX_fuzzer-4979694508834816

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
 2021-10-17 22:20:07 +02:00
Limin Wang
77970abb71 avcodec/hevc_filter: Correct indention 
Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
 2021-10-17 16:57:53 +08:00
Limin Wang
5a91850b61 avcodec/hevc_filter: remove unneeded headers 
Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
 2021-10-17 16:57:47 +08:00
Limin Wang
06548e6045 avcodec/hevcdec: remove unused code 
Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
 2021-10-17 16:57:43 +08:00
Limin Wang
fb4f9a2043 avformat/rtpdec_rfc4175: add support for RANGE 
Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
 2021-10-17 16:54:03 +08:00
Limin Wang
bad48dfe9a avformat/rtpdec_rfc4175: add support for colorimetry 
Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
 2021-10-17 16:54:03 +08:00
Limin Wang
ca56fedab5 avformat/rtpdec_rfc4175: add support for TCS 
Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
 2021-10-17 16:54:03 +08:00
Limin Wang
b07437f956 avformat/rtpdec_rfc4175: add support for exactframerate 
Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
 2021-10-17 16:54:03 +08:00
Fei Wang
84c73102d9 avcodec/av1_vaapi: improve decode quality 
- quantizer delta and matrix level specific.
- support loop filter delta.
- support use superres.

Signed-off-by: Fei Wang <fei.w.wang@intel.com>
 2021-10-16 19:00:44 -03:00
Fei Wang
dc94f2eaaf avcodec/av1_vaapi: enable segmentation features 
Signed-off-by: Fei Wang <fei.w.wang@intel.com>
 2021-10-16 19:00:44 -03:00
Fei Wang
7871144cf8 avcodec/av1_vaapi: setting 2 output surface for film grain 
VAAPI needs 2 output surface for film grain frame. One used for
reference and the other used for applying film grain and pushing
to downstream.

Signed-off-by: Fei Wang <fei.w.wang@intel.com>
 2021-10-16 19:00:44 -03:00
Fei Wang
53403158cc avcodec/vaapi: increase av1 decode pool size 
For film grain clip, vaapi_av1 decoder will cache additional 8
surfaces that will be used to store frames which apply film grain.
So increase the pool size by plus 8 to avoid leak of surface.

Signed-off-by: Fei Wang <fei.w.wang@intel.com>
 2021-10-16 19:00:44 -03:00
Tong Wu
4e7a7d75e3 avcodec/dxva2_av1: fix global motion params 
Defined in spec 5.9.24/5.9.25. Since function void
global_motion_params(AV1DecContext *s) already updates
gm type/params, the wminvalid parameter only need to get
the value from cur_frame.gm_invalid.

Signed-off-by: Tong Wu <tong1.wu@intel.com>
 2021-10-16 19:00:44 -03:00
Fei Wang
0d0ea70e7b avcodec/av1_vaapi: add gm params valid check 
Signed-off-by: Fei Wang <fei.w.wang@intel.com>
 2021-10-16 19:00:44 -03:00
Fei Wang
de7475b111 avcodec/av1dec: support setup shear process 
Defined in spec 7.11.3.6/7.11.3.7.

Signed-off-by: Fei Wang <fei.w.wang@intel.com>
 2021-10-16 19:00:44 -03:00
Fei Wang
75de7fe262 avcodec/av1: extend some definitions in spec section 3 
Signed-off-by: Fei Wang <fei.w.wang@intel.com>
 2021-10-16 19:00:44 -03:00
Fei Wang
e7ff5722b1 cbs_av1: fix incorrect data type 
Since order_hint_bits_minus_1 range is 0~7, cur_frame_hint can be
most 128. And similar return value for cbs_av1_get_relative_dist.
So if plus them and use int8_t for the result may lose its precision.

Signed-off-by: Fei Wang <fei.w.wang@intel.com>
 2021-10-16 19:00:43 -03:00
Tsutomu Seki
9b445663a5 avfilter/opencl: Fix program_opencl for source code larger than 64kB 
The condition (pos < len) is always true and the
rest of the OpenCL program code would not be read, while
the maximum number of "rb" is "len - pos - 1", and then, the
maximum number of the "pos" is "len - 1".

Fixes: trac.ffmpeg.org/ticket/9217
 2021-10-16 12:17:23 +02:00
Paul B Mahol
5bcc61ce87 avfilter/vf_v360: add reset_rot option 2021-10-16 11:39:15 +02:00
Niklas Haas
3cc3f5de2a avcodec/hevcdec: apply H.274 film grain 
Similar in spirit and design to 66845cffc3, but slightly simpler due
to the lack of interlaced frames in HEVC. See that commit for more
details.

For the seed value, since no specification for this appears to exist, I
semi-arbitrarily decided to base it off the POC id alone, since there's
no analog of the idr_pic_id in HEVC's I-frames. This design is stable
across remuxes and seeks, but changes for adjacent frames with a period
that's typically long enough not to be noticeable, which makes it
satisfy all of the requirements that a film grain seed should have.

Tested with and without threading, using a patch to insert film grain
metadata artificially (for lack of real files containing film grain).
 2021-10-15 11:55:45 -03:00
Zane van Iperen
5d16660598 avformat/argo_asf: use title metadata when muxing 
Signed-off-by: Zane van Iperen <zane@zanevaniperen.com>
 2021-10-15 23:40:15 +10:00
Zane van Iperen
9a2b9aafba avformat/argo_asf: pass name through as metadata 
Signed-off-by: Zane van Iperen <zane@zanevaniperen.com>
 2021-10-15 23:40:15 +10:00
Zane van Iperen
20fa838da5 avformat/argo_asf: cleanup and NULL-terminate name field in header 
Preparation for metadata changes in the following patches. Saves
having to create an extra buffer.

Signed-off-by: Zane van Iperen <zane@zanevaniperen.com>
 2021-10-15 23:39:47 +10:00
Wu Jianhua
2c734a8496 libswscale/x86/rgb2rgb: add shuffle_bytes avx2 
Performance data(Less is better):
    shuffle_bytes_ssse3   3.64654
    shuffle_bytes_avx2    0.94288

Signed-off-by: Wu Jianhua <jianhua.wu@intel.com>
 2021-10-15 10:59:20 +02:00
Paul B Mahol
767f162432 avfilter/window_func: unify all filters win_func option that use this header 2021-10-15 10:45:50 +02:00
James Almer
39f3c98bb1 x86/vf_lut3d: use three operand form for some instructions 
Fixes compilation with old yasm.

Signed-off-by: James Almer <jamrial@gmail.com>
 2021-10-14 18:09:38 -03:00
Paul B Mahol
890cef1ff6 avfilter/vf_fftfilt: export FFT arrays size 2021-10-14 20:26:23 +02:00
Paul B Mahol
e1b820fa33 avfilter/vf_overlay: unbreak alpha composition with negative y and threads > 1 2021-10-14 20:05:39 +02:00
Martin Storsjö
bb10f8d802 avfilter/vf_fftfilt: Use av_clip_uint8 
The refactoring in 844890b1bc caused
fate-source to point out that this could be av_clip_uintp2 (or
rather av_clip_uint8).

Signed-off-by: Martin Storsjö <martin@martin.st>
 2021-10-14 14:05:39 +03:00
Paul B Mahol
c336c7a9d7 fate: update histogram test results 2021-10-14 12:22:38 +02:00
Paul B Mahol
df05603291 avfilter/vf_histogram: add colors_mode option 2021-10-14 12:16:30 +02:00
Paul B Mahol
7d3a9bb54b avfilter/vf_fftfilt: add gray formats >8 depth support 2021-10-14 10:08:59 +02:00
Pekka Väänänen
4d52e36bd0 avformat/westwood_vqa: Store VQFL codebook chunks 
High color 15-bit VQA3 video streams contain high level chunks with
only codebook updates that shouldn't be considered new frames. Now
the demuxer stores a reference to such VQFL chunks and returns them
later along with a VQFR chunk with full frame data.
 2021-10-14 09:59:52 +02:00
Paul B Mahol
844890b1bc avfilter/vf_fftfilt: add slice threading support 2021-10-14 01:27:16 +02:00
Paul B Mahol
8add1b39e2 avfilter/vf_fftfilt: simplify bits/len calculation 2021-10-14 01:27:16 +02:00
Paul B Mahol
933765aa0e avfilter: add xcorrelate video filter 2021-10-13 19:09:21 +02:00
Paul B Mahol
32eaf4069e avfilter: add limitdiff video filter 2021-10-13 19:02:34 +02:00
Soft Works
73fe19f09c avfilter/vf_palettegen: cosmetic changes 
Signed-off-by: softworkz <softworkz@hotmail.com>
 2021-10-13 18:52:14 +02:00
Soft Works
dea673d0d5 avfilter/vf_palette(gen|use): support palettes with alpha 2021-10-13 18:52:14 +02:00
Mark Reid
3ee7250116 avfilter/vf_lut3d: fix building with --disable-optimizations 2021-10-13 18:01:21 +02:00
Limin Wang
871fee82e1 avcodec/videotoolboxenc: use goto end for memory cleanup 
Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
 2021-10-13 20:12:30 +08:00
Limin Wang
f25871d790 avcodec/avs3_parser: Fix usage of init_get_bits() and use init_get_bits8() 
Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
 2021-10-13 20:12:30 +08:00
Limin Wang
ba03e4ed33 avcodec/audiotoolboxdec: Fix usage of init_get_bits() and use init_get_bits8() 
Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
 2021-10-13 20:12:30 +08:00
Paul B Mahol
13141339c1 avformat/dhav: make duration extraction more robust 2021-10-13 12:14:39 +02:00
Paul B Mahol
6384175d8c avformat/dhav: check if timestamp matches when seeking 2021-10-13 12:14:39 +02:00
Nachiket Tarate
f14adb0516 libavformat/hls: correct indentation 
Signed-off-by: Nachiket Tarate <nachiket.programmer@gmail.com>
Signed-off-by: Steven Liu <lq@chinaffmpeg.org>
 2021-10-13 11:24:02 +08:00