1
0
mirror of https://github.com/FFmpeg/FFmpeg.git synced 2024-12-07 11:13:41 +02:00
Commit Graph

75 Commits

Author SHA1 Message Date
sfan5
a428f2fcd9 libavcodec/hevc_filter: move AVDISCARD_NONREF switch-case into function
In preparation for implementation of skip_frame.

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2017-12-08 20:02:16 +01:00
sfan5
942eafcf08 libavcodec/hevc_filter: support for all skip_loop_filter levels.
Continues where commit 52c75d486e left off.

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2017-12-04 19:05:36 +01:00
Michael Niedermayer
d7b3d5c3f2 avcodec/hevc_filter: Fix invalid shift
Fixes: runtime error: left shift of negative value -1

Fixes: 2299/clusterfuzz-testcase-minimized-4843509351710720

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2017-06-21 02:25:23 +02:00
Clément Bœsch
038e6aef7a Merge commit '4abe3b049d987420eb891f74a35af2cebbf52144'
* commit '4abe3b049d987420eb891f74a35af2cebbf52144':
  hevc: rename hevc.[ch] to hevcdec.[ch]

Merged-by: Clément Bœsch <u@pkh.me>
2017-03-23 16:48:41 +01:00
James Almer
1e18548826 Merge commit 'ba479f3daafc7e4359ec1212164569ebe59f0bb7'
* commit 'ba479f3daafc7e4359ec1212164569ebe59f0bb7':
  hevc: Change type of array stride parameters to ptrdiff_t

Merged-by: James Almer <jamrial@gmail.com>
2017-03-21 15:49:09 -03:00
Anton Khirnov
4abe3b049d hevc: rename hevc.[ch] to hevcdec.[ch]
This is more consistent with the rest of libav and frees up the hevc.h
name for decoder-independent shared declarations.
2016-10-16 20:26:17 +02:00
Diego Biurrun
ba479f3daa hevc: Change type of array stride parameters to ptrdiff_t
ptrdiff_t is the correct type for array strides and similar.
2016-09-29 17:54:23 +02:00
Clément Bœsch
bef74ef367 Merge commit '4f81f8dba735c212efae077c4fec8ad4fe53b352'
* commit '4f81f8dba735c212efae077c4fec8ad4fe53b352':
  Drop unnecessary golomb.h #includes

Merged-by: Clément Bœsch <clement@stupeflix.com>
2016-06-29 11:54:54 +02:00
Diego Biurrun
4f81f8dba7 Drop unnecessary golomb.h #includes 2016-06-08 18:51:57 +02:00
Michael Niedermayer
29d147c94d Merge commit '059a934806d61f7af9ab3fd9f74994b838ea5eba'
* commit '059a934806d61f7af9ab3fd9f74994b838ea5eba':
  lavc: Consistently prefix input buffer defines

Conflicts:
	doc/examples/decoding_encoding.c
	libavcodec/4xm.c
	libavcodec/aac_adtstoasc_bsf.c
	libavcodec/aacdec.c
	libavcodec/aacenc.c
	libavcodec/ac3dec.h
	libavcodec/asvenc.c
	libavcodec/avcodec.h
	libavcodec/avpacket.c
	libavcodec/dvdec.c
	libavcodec/ffv1enc.c
	libavcodec/g2meet.c
	libavcodec/gif.c
	libavcodec/h264.c
	libavcodec/h264_mp4toannexb_bsf.c
	libavcodec/huffyuvdec.c
	libavcodec/huffyuvenc.c
	libavcodec/jpeglsenc.c
	libavcodec/libxvid.c
	libavcodec/mdec.c
	libavcodec/motionpixels.c
	libavcodec/mpeg4videodec.c
	libavcodec/mpegvideo.c
	libavcodec/noise_bsf.c
	libavcodec/nuv.c
	libavcodec/nvenc.c
	libavcodec/options.c
	libavcodec/parser.c
	libavcodec/pngenc.c
	libavcodec/proresenc_kostya.c
	libavcodec/qsvdec.c
	libavcodec/svq1enc.c
	libavcodec/tiffenc.c
	libavcodec/truemotion2.c
	libavcodec/utils.c
	libavcodec/utvideoenc.c
	libavcodec/vc1dec.c
	libavcodec/wmalosslessdec.c
	libavformat/adxdec.c
	libavformat/aiffdec.c
	libavformat/apc.c
	libavformat/apetag.c
	libavformat/avidec.c
	libavformat/bink.c
	libavformat/cafdec.c
	libavformat/flvdec.c
	libavformat/id3v2.c
	libavformat/isom.c
	libavformat/matroskadec.c
	libavformat/mov.c
	libavformat/mpc.c
	libavformat/mpc8.c
	libavformat/mpegts.c
	libavformat/mvi.c
	libavformat/mxfdec.c
	libavformat/mxg.c
	libavformat/nutdec.c
	libavformat/oggdec.c
	libavformat/oggparsecelt.c
	libavformat/oggparseflac.c
	libavformat/oggparseopus.c
	libavformat/oggparsespeex.c
	libavformat/omadec.c
	libavformat/rawdec.c
	libavformat/riffdec.c
	libavformat/rl2.c
	libavformat/rmdec.c
	libavformat/rtpdec_latm.c
	libavformat/rtpdec_mpeg4.c
	libavformat/rtpdec_qdm2.c
	libavformat/rtpdec_svq3.c
	libavformat/sierravmd.c
	libavformat/smacker.c
	libavformat/smush.c
	libavformat/spdifenc.c
	libavformat/takdec.c
	libavformat/tta.c
	libavformat/utils.c
	libavformat/vqf.c
	libavformat/westwood_vqa.c
	libavformat/xmv.c
	libavformat/xwma.c
	libavformat/yop.c

Merged-by: Michael Niedermayer <michael@niedermayer.cc>
2015-07-27 23:15:19 +02:00
Nicolas George
52c75d486e lavc/hevc: rudimentary support for skip_loop_filter.
+~9% speed on Core i5 on test sample.

All frames are treated as ref frames, skipping only applies
at level "all". The following mail contains information on
how to improve that:
http://ffmpeg.org/pipermail/ffmpeg-devel/2015-July/176116.html
2015-07-23 16:26:14 +02:00
Michael Niedermayer
d5fcca83b9 Merge commit 'b11acd57326db6c2cc1475dd0bea2a06fbc85aa2'
* commit 'b11acd57326db6c2cc1475dd0bea2a06fbc85aa2':
  hevc: remove HEVCContext usage from hevc_ps

Conflicts:
	libavcodec/hevc.c
	libavcodec/hevc_cabac.c
	libavcodec/hevc_filter.c
	libavcodec/hevc_mvs.c
	libavcodec/hevc_ps.c
	libavcodec/hevc_refs.c
	libavcodec/hevcpred_template.c

Merged-by: Michael Niedermayer <michael@niedermayer.cc>
2015-07-12 22:28:54 +02:00
Anton Khirnov
b11acd5732 hevc: remove HEVCContext usage from hevc_ps
Factor out the parameter sets into a separate struct and use it instead.

This will allow us to reuse this code in the parser.
2015-07-12 18:15:39 +02:00
Christophe Gisquet
efd3f407e5 hevc/sao: use unaligned copy
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-03-02 19:32:23 +01:00
Peter Meerwald
eea769df32 hevc: Use generic av_clip function, not C implementation
hevc seems to be the only place where the C implementation
of the av_clip function is explicitly selected, precluding
platform-specific optimizations

Signed-off-by: Peter Meerwald <pmeerw@pmeerw.net>
Signed-off-by: Anton Khirnov <anton@khirnov.net>
2015-02-13 08:37:13 +01:00
Christophe Gisquet
97996eff4f hevc/sao: do in-place band filtering when possible
The copies are only needed when data must be restored, so skip them
when it must not be.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-02-07 18:04:25 +01:00
James Almer
042c1159fc x86/hevcdsp: add ff_hevc_sao_edge_filter_8_{ssse3,avx2}
Original x86 intrinsics code and initial yasm port by Pierre-Edouard Lepere.
Refactoring and optimizations by James Almer.

Benchmarks of BQTerrace_1920x1080_60_qp22.bin with an Intel Core i5-4200U

Width 32
158583 decicycles in edge, sao_edge_filter_8 runs, 0 skips
5205 decicycles in ff_hevc_sao_edge_filter_32_8_ssse3, 32767 runs, 1 skips
2942 decicycles in ff_hevc_sao_edge_filter_32_8_avx2, 32767 runs, 1 skips

Width 64
705639 decicycles in sao_edge_filter_8, 262144 runs, 0 skips
19224 decicycles in ff_hevc_sao_edge_filter_64_8_ssse3, 262111 runs, 33 skips
10433 decicycles in ff_hevc_sao_edge_filter_64_8_avx2, 262115 runs, 29 skips

Signed-off-by: James Almer <jamrial@gmail.com>
2015-02-05 15:02:27 -03:00
James Almer
1f1c7c8a57 hevcdsp: remove compilation-time-fixed parameter from sao_edge_filter
The stride_src parameter is always 2 * MAX_PB_SIZE + FF_INPUT_BUFFER_PADDING_SIZE.

Signed-off-by: James Almer <jamrial@gmail.com>
2015-02-05 15:02:22 -03:00
James Almer
7457afc64d hevcdsp: replace the SAOParams struct parameter from sao_edge_filter
As with sao_band_filter, pass instead the two variables from the struct needed in the function.
This simplifies writing asm optimized versions.

Reviewed-by: Mickaël Raulet <mraulet@insa-rennes.fr>
Signed-off-by: James Almer <jamrial@gmail.com>
2015-02-04 17:53:04 -03:00
Seppo Tomperi
4386e1fd94 hevcdsp: simplified sao_edge_filter
Reviewed-by: Christophe Gisquet <christophe.gisquet@gmail.com>
Reviewed-by: Mickaël Raulet <mraulet@insa-rennes.fr>
2015-02-04 17:52:54 -03:00
Seppo Tomperi
74d7faf400 hevcdsp: separated sao edge filter and pixel restore funcs
Reviewed-by: Michael Niedermayer <michaelni@gmx.at>
Reviewed-by: Christophe Gisquet <christophe.gisquet@gmail.com>
Reviewed-by: Mickaël Raulet <mraulet@insa-rennes.fr>
2015-02-04 17:52:49 -03:00
Christophe Gisquet
52f2adc015 avcodec/hevc: Update the USE_SAO_SMALL_BUFFER case for the alignment requirements in FFmpeg
Use edge emu buffers
And enable the code unconditionally

Speed difference without USE_SAO_SMALL_BUFFER and with the new code:
Decicycles: 26772->26220 (BO32),  83803->80942 (BO64)

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-02-04 01:54:22 +01:00
Fabrice Bellard
da81cc38e8 avcodec/hevc: reduce memory for SAO
cherry picked from commit 5d9f79edef2c11b915bdac3a025b59a32082f409

SAO edge filter uses pre-SAO pixel data on the left and top of the ctb, so
this data must be kept available. This was done previously by having 2
copies of the frame, one before and one after SAO.

This commit reduces the storage to just that, instead of the previous whole
frame.

Commit message taken from patch by Christophe Gisquet <christophe.gisquet@gmail.com>

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-02-04 01:54:01 +01:00
Michael Niedermayer
d525b45fde avcodec/hevc_filter: Fix indention
Found-by: Timothy Gu <timothygu99@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-02-02 19:56:44 +01:00
Mickaël Raulet
7cf6a67ef9 avcodec/hevc: adding support for monochrome sequences in hevc
cherry picked from commit 8e50557707d2ec11ccad657470b2e140f314348e

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-02-02 18:10:57 +01:00
Christophe Gisquet
6a6aeb538b hevc/sao: use aligned copies
For band filter, source and destination are aligned (except for 16x16 ctbs),
and otherwise, they are most often aligned. Overall, the total width is also
too small for amortizing memcpy.

Timings (using an intrinsic version of edge filters):
          B/32     B/64     E/32     E/64
Before:  32045    93952    38925    126896
After:   26772    83803    33942    117182
2015-02-01 20:23:03 -03:00
James Almer
fa3eccb4f9 x86/hevc: add ff_hevc_sao_band_filter_{8,10,12}_{sse2,avx,avx2}
Original x86 intrinsics code and initial 8bit yasm port by Pierre-Edouard Lepere.
10/12bit yasm ports, refactoring and optimizations by James Almer

Benchmarks of BQTerrace_1920x1080_60_qp22.bin with an Intel Core i5-4200U

width 32
40338 decicycles in sao_band_filter_0_8, 2048 runs, 0 skips
8056 decicycles in ff_hevc_sao_band_filter_8_32_sse2, 2048 runs, 0 skips
7458 decicycles in ff_hevc_sao_band_filter_8_32_avx, 2048 runs, 0 skips
4504 decicycles in ff_hevc_sao_band_filter_8_32_avx2, 2048 runs, 0 skips

width 64
136046 decicycles in sao_band_filter_0_8, 16384 runs, 0 skips
28576 decicycles in ff_hevc_sao_band_filter_8_32_sse2, 16384 runs, 0 skips
26707 decicycles in ff_hevc_sao_band_filter_8_32_avx, 16384 runs, 0 skips
14387 decicycles in ff_hevc_sao_band_filter_8_32_avx2, 16384 runs, 0 skips

Reviewed-by: Christophe Gisquet <christophe.gisquet@gmail.com>
Signed-off-by: James Almer <jamrial@gmail.com>
2015-02-01 20:22:35 -03:00
James Almer
2929e56006 hevcdsp: replace the SAOParams struct parameter from sao_band_filter
Pass instead the two variables from the struct needed in the function.
This simplifies writing asm optimized versions of the function

Signed-off-by: James Almer <jamrial@gmail.com>
2015-02-01 15:45:20 -03:00
James Almer
65e6ab0c5a hevcdsp: remove unused parameter from sao_band_filter
Signed-off-by: James Almer <jamrial@gmail.com>
2015-02-01 15:45:14 -03:00
Michael Niedermayer
499ff6a052 Merge commit '7acdd3a1275bcd9cad48f9632169f6bbaeb39d84'
* commit '7acdd3a1275bcd9cad48f9632169f6bbaeb39d84':
  hevc_filter: avoid excessive calls to ff_hevc_get_ref_list()

Conflicts:
	libavcodec/hevc_filter.c

Merged-by: Michael Niedermayer <michaelni@gmx.at>
2014-08-09 21:49:55 +02:00
Michael Niedermayer
8d7c4cc082 Merge commit 'a7a17e3f1915ce69b787dc58c5d8dba0910fc0a4'
* commit 'a7a17e3f1915ce69b787dc58c5d8dba0910fc0a4':
  hevc_filter: move some conditions out of loops

Conflicts:
	libavcodec/hevc_filter.c

This is possibly less readable than the variant used before.
Thus please take a look and if people agree its worse, dont
hesitate to revert.

See: 83976e40e8
Merged-by: Michael Niedermayer <michaelni@gmx.at>
2014-08-09 21:29:33 +02:00
Anton Khirnov
7acdd3a127 hevc_filter: avoid excessive calls to ff_hevc_get_ref_list()
1) each of the loops run within a single CTB, so the relevant reference
list is constant
2) when that CTB is, or lies on the same slice as, the current one, we
can use a simple access instead of a relatively expensive call to
ff_hevc_get_ref_list()
2014-08-09 16:15:51 +00:00
Anton Khirnov
a7a17e3f19 hevc_filter: move some conditions out of loops 2014-08-09 16:15:11 +00:00
Anton Khirnov
70211539a3 hevc: deobfuscate slice/tile boundary handling for DBF
Use named constants instead of magic numbers, avoid using variables with
inverse meaning from what their name implies.
2014-08-09 16:13:48 +00:00
Anton Khirnov
5501971578 hevc_filter: drop more redundant checks
The if() around those loops ensures this condition is always false.
2014-08-09 16:13:37 +00:00
Anton Khirnov
65b8b6c476 hevc_filter: drop redundant checks
ff_hevc_deblocking_boundary_strengths() is never called if the
deblocking filter is disabled for the slice.
2014-08-09 16:13:24 +00:00
Christophe Gisquet
6786848585 hevc_deblock: change tc type
The x86 asm expects int32_t so use that type.

Reviewed-by: Mickaël Raulet <mraulet@insa-rennes.fr>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2014-08-06 12:38:26 +02:00
Christophe Gisquet
65746bfbae hevc_filter: run vertical and horizontal together
This should help cache locality. On win64:
Before: 1397x cycles, 16216 bytes
After:  1369x cycles, 16040 bytes

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2014-07-29 03:56:43 +02:00
Anton Khirnov
a5c621aa85 hevc: rename variable in boundary strength to b more explicit
Signed-off-by: Mickaël Raulet <mraulet@insa-rennes.fr>

cherry picked from commit 348bebedc0012aae201419669fca1eb61ec93ca6
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2014-07-28 14:20:19 +02:00
Mickaël Raulet
3b777db132 hevc: remove non necessary parameters to ff_hevc_set_qpy
cherry picked from commit 6f58c111ad9920d983bb18eacf901193bac5d937

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2014-07-28 00:39:13 +02:00
Michael Niedermayer
226b290f9f Merge commit '73bb8f61d48dbf7237df2e9cacd037f12b84b00a'
* commit '73bb8f61d48dbf7237df2e9cacd037f12b84b00a':
  hevcdsp: remove an unneeded variable in the loop filter

Conflicts:
	libavcodec/hevc_filter.c

See: d7e162d46b
Merged-by: Michael Niedermayer <michaelni@gmx.at>
2014-07-26 23:53:31 +02:00
Anton Khirnov
73bb8f61d4 hevcdsp: remove an unneeded variable in the loop filter
beta0 and beta1 will always be the same
2014-07-26 15:00:11 +00:00
Anton Khirnov
d7e162d46b hevcdsp: remove an unneeded variable in the loop filter
beta0 and beta1 will always be the same within a CU

Signed-off-by: Mickaël Raulet <mraulet@insa-rennes.fr>

cherry picked from commit 4a23d824741a289c7d2d2f2871d1e2621b63fa1b
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2014-07-22 16:27:26 +02:00
Mickaël Raulet
fa0506e413 hevc/rext: update boundary strength for chroma deblocking filter
cherry picked from commit 7d05c95ac5a63d7675bf645e74b4cf1fffff4796

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2014-07-22 13:00:14 +02:00
Christophe Gisquet
5e55c7e1bc hevc: report more precise progress
There's a lag of one CTB line for SAO behind deblocking filter, except for
last line. However, once SAO has been completed on a line, all its pixels,
i.e. up to y+ctb_size are filtered and ready to be used as reference.

Without SAO, when deblocking filter finishes a CTB line, only the bottom
bottom 4 pixels may be filtered when next CTB is process by the deblocing.

The await_progess for hevc then checks whether the bottom pixels of a PU
requires access beyond that point, so the reporting should effectively
report up to the the above limits.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2014-07-19 14:20:10 +02:00
Mickaël Raulet
e1e3ec9b02 hevc: fix transquant_bypass
cherry picked from commit 4a16cb2c70728a55d2fd723aff01b13ea259c4df
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2014-07-19 00:11:05 +02:00
Mickaël Raulet
d249e6828e hevc/sao: optimze sao implementation
- adding one extra pixel all around the frame
- do not copy when SAO is not applied

5% improvement

cherry picked from commit 10fc29fc19a12c4d8168fbe1a954b76386db12d0
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2014-07-18 22:46:50 +02:00
Mickaël Raulet
5a41999d81 hevc/rext: basic infrastructure for supporting range extension
- support for 4:2:2 and 4:4:4 up to 12 bits
- add a new profile for range extension
(cherry picked from commit d3c067fa65bbc871758d28aa07f54123430ca346)

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2014-07-15 13:47:35 +02:00
Mickaël Raulet
f5beda3bfd hevc: move restore_tqb where it should be.
(cherry picked from commit 8fafc96a9805d11bfe32537c8f78a294a5844065)

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2014-07-15 13:25:55 +02:00
Mickaël Raulet
1241eb8870 hevc: simplify SAO computation, delay from one row its computation
(cherry picked from commit f2c5f647cec786df26f442a85e6d685a131a50c9)

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2014-07-15 13:11:33 +02:00