virtualenv/FFmpeg
1
0
Fork 0
mirror of https://github.com/FFmpeg/FFmpeg.git synced 2024-12-28 20:53:54 +02:00
Code Issues Releases Activity
74,204 Commits 37 Branches 400 Tags 467 MiB
f080a01fd3
Commit Graph

155 Commits

Author SHA1 Message Date
James Almer
d9e10af547 x86/vf_interlace: add missing colon to labels 
Silences warnings with Nasm

Signed-off-by: James Almer <jamrial@gmail.com>
 2015-07-26 02:50:50 -03:00
James Almer
e3851169ee x86/vf_ssim: add ff_ssim_4x4_line_xop 
~20% faster than ssse3. Also enabled for x86_32

Reviewed-by: Ronald S. Bultje <rsbultje@gmail.com>
Signed-off-by: James Almer <jamrial@gmail.com>
 2015-07-20 13:18:05 -03:00
James Almer
e1778fb657 x86/vf_ssim: fix some instruction comments 
Reviewed-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: James Almer <jamrial@gmail.com>
 2015-07-20 13:17:58 -03:00
Paul B Mahol
eea08efc0d avfilter/x86/vf_psnr.asm: split one line of license text into two 
Signed-off-by: Paul B Mahol <onemda@gmail.com>
 2015-07-14 23:54:26 +00:00
James Darnley
bff7242608 avfilter/vf_removegrain: add x86 and x86_64 SSE2 functions 
Speed of all modes increased by a factor between 7.4 and 19.8 largely depending
on whether bytes are unpacked into words.  Modes 2, 3, and 4 have been sped-up
by a factor of 43 (thanks quick sort!)

All modes are available on x86_64 but only modes 1, 10, 11, 12, 13, 14, 19, 20,
21, and 22 are available on x86 due to the number of SIMD registers used.

With a contribution from James Almer <jamrial@gmail.com>
 2015-07-14 23:50:50 +00:00
Ronald S. Bultje
ae4c9ddebc vf_psnr: sse2 optimizations for sum-squared-error. 
The internal line accumulator for 16bit can overflow, so I changed that
from int to uint64_t in the C code. The matching assembly looks a little
weird but output looks correct.

(avx2 should be trivial to add later.)

Reviewed-by: Paul B Mahol <onemda@gmail.com>
Reviewed-by: James Almer <jamrial@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
 2015-07-14 17:57:14 +02:00
Ronald S. Bultje
dfc58584b4 vf_ssim: x86 simd for ssim_4x4xN and ssim_endN. 
Both are 2-2.5x faster than their C counterpart.

Reviewed-by: Paul B Mahol <onemda@gmail.com>
Reviewed-by: James Almer <jamrial@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
 2015-07-14 05:07:07 +02:00
James Almer
c16e99e3b3 x86: check for AV_CPU_FLAG_AVXSLOW where useful 
Signed-off-by: James Almer <jamrial@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2015-06-01 00:15:35 +02:00
Michael Niedermayer
52fc3e372f avfilter/x86/vf_hqdn3d: Fix register types 
Fixes Ticket4301

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2015-05-27 05:18:55 +02:00
Michael Niedermayer
5bc2c39527 avfilter/x86/vf_fspp: Fix invalid combination of opcode and operands 
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2015-05-26 01:43:47 +02:00
Michael Niedermayer
a6f9a5d0f6 avfilter/x86/vf_fspp: Fix loop condition for column_fidct() 
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2015-01-28 17:23:27 +01:00
Michael Niedermayer
f5b3257c50 avfilter/vf_eq: mark src as const 
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2015-01-27 01:46:08 +01:00
Michael Niedermayer
530bf8ece6 avfilter/vf_eq: Fix clipping code 
Found-by: Christophe Gisquet <christophe.gisquet@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2015-01-26 23:46:44 +01:00
Arwa Arif
4c38e960d0 avfilter: Port mp=eq/eq2 to lavfi 
Code adapted from James Darnley's port
Some fixes from Paul B Mahol <onemda@gmail.com>

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2015-01-26 00:14:04 +01:00
James Almer
da02ee127a x86/vf_pp7: port dctB_mmx to yasm 
Reviewed-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: James Almer <jamrial@gmail.com>
 2015-01-09 20:02:27 -03:00
Arwa Arif
a299cd5ab3 lavfi: port mp=pp7 to libavfilter 
The only difference with mp=pp7 is that default mode is "medium", as stated
in the MPlayer docs, rather than "hard".

Signed-off-by: Stefano Sabatini <stefasab@gmail.com>
 2015-01-09 17:26:31 +01:00
James Almer
a4f876a1a2 x86/vf_fspp: move pxor in store slice functions out of the loop 
m7 is not overwritten, so we only need to clear it once.
Found by Christophe Gisquet.

Signed-off-by: James Almer <jamrial@gmail.com>
 2014-12-26 17:15:34 -03:00
James Almer
466e32bf25 x86/vf_fspp: port inline asm to yasm 
Reviewed-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: James Almer <jamrial@gmail.com>
 2014-12-26 15:39:51 -03:00
James Almer
b94e85453e avfilter/vf_fspp: add missing inline asm guards 2014-12-24 15:44:06 -03:00
Arwa Arif
bdc4db0ee3 lavfi: port mp=fspp to a native libavfilter filter 
Signed-off-by: Stefano Sabatini <stefasab@gmail.com>
 2014-12-24 16:29:18 +01:00
Michael Niedermayer
6706a2986c avfilter/vf_spp: Fix overflow in 8bit store slice 
Fixes regression with
ffplay -f lavfi -i testsrc=640x480  -vf format=gray,boxblur=20:10,geq="'mod(lum(X,Y),16)*15'",boxblur=10,geq="'abs(mod(lum(X,Y),15)-7)*32'",spp=4:40

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2014-12-21 01:48:19 +01:00
Michael Niedermayer
838aa08d75 avfilter/vf_spp: support 10bit per sample 
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2014-12-15 18:49:35 +01:00
Michael Niedermayer
30d2ac4bf9 avfilter/vf_spp: change temporary to unsigned 
More consistent with uspp and allows for future 10bit support

Found-by: ubitux
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2014-12-12 13:34:18 +01:00
Michael Niedermayer
ca59b5b6ec avfilter/x86/vf_interlace: remove redundant instructions 
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2014-11-25 12:37:19 +01:00
Michael Niedermayer
3fe3c8abb1 Merge commit 'ca5c3ff90972a5c97aabda2ace57ba72dcd7d83b' 
* commit 'ca5c3ff90972a5c97aabda2ace57ba72dcd7d83b':
  vf_interlace: x86: improve asm performance

Conflicts:
	libavfilter/x86/vf_interlace.asm

See: 05e4b25e9b
Merged-by: Michael Niedermayer <michaelni@gmx.at>
 2014-11-25 12:31:45 +01:00
Michael Niedermayer
ca5c3ff909 vf_interlace: x86: improve asm performance 
4775 decicycles -> 3688 decicycles
 2014-11-25 02:00:06 +00:00
Michael Niedermayer
05e4b25e9b avfilter/x86/vf_interlace: rewrite asm 
4775 decicycles -> 3688 decicycles

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2014-11-15 04:09:03 +01:00
Michael Niedermayer
fb3eb57369 avfilter/tinterlace: add Support for ff_lowpass_line_avx() & ff_lowpass_line_sse2() 
Based-on: 2e1704059a by Kieran Kunhya

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2014-11-15 04:02:33 +01:00
Michael Niedermayer
6f373d75e8 Merge commit '2e1704059ae8625beda2ffde847ad22c5ba416dc' 
* commit '2e1704059ae8625beda2ffde847ad22c5ba416dc':
  vf_interlace: Add SIMD for lowpass filter

Conflicts:
	libavfilter/vf_interlace.c
	libavfilter/x86/Makefile

Merged-by: Michael Niedermayer <michaelni@gmx.at>
 2014-11-15 02:39:49 +01:00
Kieran Kunhya
2e1704059a vf_interlace: Add SIMD for lowpass filter 
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
 2014-11-15 00:35:31 +01:00
James Almer
864f9326fb x86/vf_noise: move asm code to a separate file 
Reviewed-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: James Almer <jamrial@gmail.com>
 2014-10-17 00:44:35 -03:00
Pascal Massimino
649b7a9946 av_filter/x86/idet: use HADDD where appropriate 
Signed-off-by: James Almer <jamrial@gmail.com>
 2014-09-09 19:02:49 -03:00
Pascal Massimino
e3fd6a3a4e av_filter/x86/idet: MMX/SSE2 implementation of 16bits filter_line() 
tested on http://ps-auxw.de/10bit-h264-sample/10bit-eldorado.mkv
MMX: ~30% faster decoding overall
SSE2:~40% faster

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2014-09-09 16:47:22 +02:00
James Darnley
db8970d7b6 vfi/x86/vf_idet: fix incorrect use of paddq 
paddq is an SSE2 instruction so it cannot be used for MMX.

This was probably just a typo because the sums are dwords anyway.

Reviewed-by: Pascal Massimino <pascal.massimino@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2014-09-05 12:49:34 +02:00
Pascal Massimino
161fc0f463 avfilter/x86/idet: fix license header (GPL -> LGPL) 
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2014-09-05 12:22:36 +02:00
skal
406a9ccffe avfilter/vf_idet: MMX/MMXEXT/SSE2 implementation of idet's filter_line() 
integration by Neil Birkbeck, with help from Vitor Sessak.
core SSE2 loop by Skal (pascal.massimino@gmail.com)

Reviewed-by: Clément Bœsch <u@pkh.me>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2014-09-04 22:19:00 +02:00
Andreas Cadhalpun
39a6e02fd4 fix spelling errors 
Reviewed-by: Timothy Gu <timothygu99@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2014-07-12 22:33:27 +02:00
James Almer
ddea3b7106 x86/yadif-10: remove duplicate ABS macro 
And use the x86util ones instead, which are optimized for mmxext/sse2.
About ~1% increase in performance on pre SSSE3 processors.

Signed-off-by: James Almer <jamrial@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2014-07-10 21:06:51 +02:00
Michael Niedermayer
a348f4befe avfilter/x86/vf_pullup: fix "invalid combination of opcode and operands" with nasm 
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2014-06-28 16:06:00 +02:00
Michael Niedermayer
b8255a4c70 avfilter/x86/vf_pullup: fix old typo 
This makes C and MMX match, no change to fate as the differences where
apparently not sufficient to show up in fate

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2014-06-25 18:22:48 +02:00
Michael Niedermayer
6dffc8f5aa avfilter/vf_pullup: use ptrdiff_t as stride argument for dsp functions 
This should avoid issues on x86_64

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2014-06-25 18:22:31 +02:00
Christophe Gisquet
9107612818 x86util: add and use RSHIFT/LSHIFT macros 
Those macros take a byte number as shift argument, as this argument
differs between MMX and SSE2 instructions.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2014-06-15 13:19:27 +02:00
Michael Niedermayer
ebb21887b8 Merge commit '01c5779f56cf708e6cb88b11cfdc248cae7e2ee8' 
* commit '01c5779f56cf708e6cb88b11cfdc248cae7e2ee8':
  x86: Drop some unnecessary YASM ifdefs

Conflicts:
	libavfilter/x86/vf_yadif_init.c

Merged-by: Michael Niedermayer <michaelni@gmx.at>
 2014-04-05 19:16:39 +02:00
Diego Biurrun
01c5779f56 x86: Drop some unnecessary YASM ifdefs 
Dead code elimination is enough to avoid undefined references in these cases.
 2014-04-04 19:08:05 +02:00
Robert Krüger
194ef56ba7 Change license of yadif from GPL to LGPL 
Signed-off-by: Robert Krüger <krueger@lesspain.de>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2014-01-14 14:19:15 +01:00
Robert Krüger
4a38eeec38 Revert "Revert "vf_yadif: move x86 init code to x86/yadif.c"" 
This reverts commit 975110a85e.

Signed-off-by: Robert Krüger <krueger@lesspain.de>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2014-01-14 14:19:14 +01:00
Robert Krüger
d8e763fda7 vf_yadif: Relicense from GPL to LGPL 
All copyright holders have agreed to the relicensing.
 2014-01-14 00:04:59 +01:00
Michael Niedermayer
975110a85e Revert "vf_yadif: move x86 init code to x86/yadif.c" 
This reverts commit a87b17f328.
This reduces the amount of non LGPL code, making a relicensing to LGPL
easier

Conflicts:

	libavfilter/vf_yadif.c
	libavfilter/x86/yadif.c
	libavfilter/x86/yadif_template.c
	libavfilter/yadif.h

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2013-12-01 20:26:26 +01:00
Clément Bœsch
969329fe11 Revert "Merge commit 'ed1a11ed52bbd1f15bb9b0416d69b7924bee3191'" 
This reverts commit fc5fe4804f, reversing
changes made to ffe3350098.

The factoring is broken; it's not calling the ssse3 code anymore, and
calling the mmx2 code with bad alignment. It also broke some FATE
instances.

Conflicts:
	libavfilter/x86/vf_gradfun_init.c
 2013-11-01 14:28:08 +01:00
Michael Niedermayer
c6125f5e1c avfilter/x86/vf_gradfun_init: fix some consts & related warnings 
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2013-11-01 14:20:10 +01:00