coffee/libyuv
1
0
Fork 0
mirror of https://chromium.googlesource.com/libyuv/libyuv synced 2025-12-07 01:06:46 +08:00
Code Issues Packages Projects Releases Wiki Activity
1,226 Commits 3 Branches 0 Tags BSD-3-Clause 60 MiB
Commit Graph

265 Commits

Author SHA1 Message Date
fbarchard@google.com
975dd5a699 macros for storing RGB on windows. 
BUG=403
TESTED=local windows build
R=harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/38119004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1283 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2015-02-14 00:50:48 +00:00
fbarchard@google.com
2f56d2859f Macro to store ARGB value 
BUG=396
TESTED=local windows build
R=harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/38109004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1279 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2015-02-11 18:53:54 +00:00
fbarchard@google.com
d1ac8b17e6 use matrix for win64 version of I420ToARGB 
BUG=396
TESTED=local unittests build/pass
R=harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/41899004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1276 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2015-02-11 00:57:46 +00:00
fbarchard@google.com
3bb829a44f Add a macro for YUV to RGB on Windows. Allows multiple color matrix structures in the future. 
BUG=393
TESTED=local build
R=brucedawson@google.com, harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/38079004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1275 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2015-02-10 23:03:37 +00:00
fbarchard@google.com
0887315390 Remove bayer format support from libyuv. This format is very rare and used on legacy hardware. Its not well optimized and has bugs related to odd widths. Removing the format will allow tests to pass under more circumstances, run faster and allow focus on higher priority quality and performance issues. 
BUG=301
TESTED=local unittests build/pass on windows gyp build.
R=harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/38059004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1270 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2015-02-09 19:58:19 +00:00
fbarchard@google.com
baafc97d6b port YToARGB AVX2 to GCC 
BUG=393
TESTED=untested
R=harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/39819004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1262 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2015-02-05 20:17:27 +00:00
fbarchard@google.com
c4e032c543 change Y multiplier and bias to compensate for 257/256 which makes YToARGB exactly match float math. 
Histogram Before
hist            -3      -2      -1      0       1       2       3
red             0       0       1809408 13140736        1827072 0       0
green           0       0       1679912 13471329        1625975 0       0
blue            168448  994816  1876480 10655488        1893376 1006336 182272
Histogram After
hist            -3      -2      -1      0       1       2       3
red             0       0       558848  15632128        586240  0       0
green           0       0       209907  16350588        216721  0       0
blue            14848   642816  1989376 11363328        2053120 695040  18688
BUG=394
TESTED=more stringent luma tests
R=brucedawson@google.com

Review URL: https://webrtc-codereview.appspot.com/38859004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1259 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2015-02-04 19:45:26 +00:00
fbarchard@google.com
3982998c7c YToARGB AVX2 port from SSE2 
BUG=393
TESTED=YToARGB unittest
R=brucedawson@google.com, harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/41679004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1258 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2015-02-03 01:35:11 +00:00
fbarchard@google.com
29db9b0b89 C version of YToARGB with ubias removed to produce consistent luma ramp. 
BUG=392
TESTED=TestGreyYUV
R=harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/35869004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1251 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2015-01-26 23:07:46 +00:00
fbarchard@google.com
080a316492 port yuv chroma improvements to gcc. YUV to RGB is more accurate using a negative matrix. 2% slower but half as much error. 
BUG=324
TESTED=try bots
R=harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/41629004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1249 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2015-01-26 04:35:51 +00:00
fbarchard@google.com
d12a08712b adjust ubias to minimize error histogram centering error. 
BUG=324
TESTED=TestFullYUV
R=harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/37739004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1248 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2015-01-23 22:16:33 +00:00
fbarchard@google.com
eb8dda3ac7 fix for ybias on YToARGB function. 
BUG=324
TESTED=libyuvTest.YToARGB_Any
R=tpsiaki@google.com

Review URL: https://webrtc-codereview.appspot.com/36939004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1247 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2015-01-23 18:31:29 +00:00
fbarchard@google.com
b114986477 Change YUV to RGB to subtract the chroma contributions from the bias. 
BUG=324
TESTED=win64 build and TestFullYUV
R=harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/33999004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1246 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2015-01-23 04:22:35 +00:00
fbarchard@google.com
c62d30111f adjust bias on Y channel so error histogram is better centered on green channel 
BUG=324
TESTED=FullYUVTest
R=tpsiaki@google.com

Review URL: https://webrtc-codereview.appspot.com/38689004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1245 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2015-01-22 19:43:34 +00:00
fbarchard@google.com
319f047710 Compute chroma using negative coefficients to extend range of U contribution on B to 2 
BUG=324
TESTED=TestI420
R=tpsiaki@google.com

Review URL: https://webrtc-codereview.appspot.com/41569004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1238 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2015-01-21 18:45:13 +00:00
fbarchard@google.com
e7873910df port YUV luma accuracy to posix 
BUG=324
TESTED=try bots
R=harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/33049004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1236 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2015-01-21 00:36:30 +00:00
fbarchard@google.com
c3d09f6021 Improve accuracy of luma channel in YUV to RGB conversion 
BUG=324
TESTED=TestFullYUV
R=harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/36859004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1233 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2015-01-20 23:42:15 +00:00
fbarchard@google.com
b2a6af1be6 Change rectangle low level functions to use more conventional row functions including 'any' variations. Previously the yuv function SetPlane stored 32 bit values. Now a more conventional memset() style function is used for YUV that stores bytes. On Haswell a rep stosb is used for YUV. Overall benefit of this CL is improved performance for 'any' width, and simpler row assembly instead of full image assembly. Previously ARGBRect used a low level function that supported a rectangle in assembly. Now it uses a row function, and relies on row coalesce to combine into a single low level call. 
BUG=371
TESTED=untested
R=brucedawson@google.com, harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/35689004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1222 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2015-01-12 03:58:24 +00:00
fbarchard@google.com
992c3b089a Use HAS_ARGBSETROWS_X86 to detect presence of function. 
BUG=none
TESTED=rectangle unittests
R=harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/35639004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1218 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2015-01-07 00:11:51 +00:00
fbarchard@google.com
966233e5eb Remove sub 16 from yuv conversions and change bias to include it. 
BUG=388
TESTED=out\release\libyuv_unittest --gtest_catch_exceptions=0 --gtest_filter=*420ToARGB_Opt  | sortms
R=harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/34609004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1216 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2014-12-31 01:07:02 +00:00
fbarchard@google.com
7892ea1fe1 Fix for ARGBToUV on AVX2 
BUG=269
TESTED=local testing
R=harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/33669004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1202 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2014-12-15 18:59:23 +00:00
fbarchard@google.com
ddee77cdbd Fix for I422ToRGBA when I422ToARGB is not enabled for AVX2 
BUG=269
TESTED=local windows build
R=harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/32339004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1201 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2014-12-15 18:28:59 +00:00
fbarchard@google.com
f5f5d15dcd Fix register order for ARGBToUV_AVX2 
BUG=269
TESTED=try bots
R=harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/29249004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1200 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2014-12-15 18:07:09 +00:00
fbarchard@google.com
540e8af80c remove add 16 from ARGBToYJ and add rounding, for consistency with Windows version. row.h header macros sorted alphabetically. 
BUG=269
TESTED=untested
R=tpsiaki@google.com

Review URL: https://webrtc-codereview.appspot.com/32579005

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1185 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2014-12-02 22:37:47 +00:00
fbarchard@google.com
c5aac16af9 Remove loop alignment for benefit of modern cpus that dont require alignment. 
BUG=none
TESTED=local libyuv unittest passes
R=brucedawson@google.com, tpsiaki@google.com

Review URL: https://webrtc-codereview.appspot.com/32159004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1180 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2014-11-24 21:26:22 +00:00
fbarchard@google.com
ef14972df0 MergeUV AVX2 use vextractf128 to store results to avoid shuffling. 
BUG=none
TESTED=intel sde on unittests
R=brucedawson@google.com

Review URL: https://webrtc-codereview.appspot.com/33369004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1178 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2014-11-22 03:33:33 +00:00
fbarchard@google.com
ef67597b48 ARGBMirror use SSE2 pshufd instruction instead of SSSE3 pshufb. 
BUG=269
TESTED=local benchmark for ARGBMirror
R=tpsiaki@google.com

Review URL: https://webrtc-codereview.appspot.com/32509004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1176 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2014-11-21 19:25:14 +00:00
fbarchard@google.com
91f240c5db Move sub before branch for loops. 
Remove CopyRow_x86
Add CopyRow_Any versions for AVX, SSE2 and Neon.
BUG=269
TESTED=local build
R=harryjin@google.com, tpsiaki@google.com

Review URL: https://webrtc-codereview.appspot.com/26209004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1175 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2014-11-20 21:14:27 +00:00
fbarchard@google.com
b9d17e1d79 Fix offset in addresses for windows. Wants it within [] now. 
BUG=none
TESTED=local windows build.
R=harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/32479004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1168 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2014-11-17 19:50:42 +00:00
fbarchard@google.com
5822505e0a Remove extra unaligned loop from alphablender. Both aligned and unaligned loops were the same, so remove the extra. 
BUG=none
TESTED=try bots.
R=brucedawson@google.com, harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/29059004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1166 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2014-11-17 18:33:07 +00:00
fbarchard@google.com
1eb636d249 remove initial lea in mirror functions and add the offset in the address mode. 
BUG=none
TESTED=local libyuv unittests on windows
R=harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/26169004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1165 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2014-11-17 18:16:23 +00:00
fbarchard@google.com
35508d0979 Mirror_AVX2 ported to GCC. 
BUG=269
TESTED=try bots
R=harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/32079004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1164 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2014-11-13 23:11:10 +00:00
fbarchard@google.com
91000425a3 ARGBUnattenuate_AVX2 ported to GCC. Minor cleanup of constants to use broadcast to make 16 byte constant instead of 32 byte. 
BUG=269
TESTED=try bots
R=harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/30999004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1163 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2014-11-13 17:57:33 +00:00
fbarchard@google.com
ec1f854f86 Use broadcast to duplicate constants from 16 bytes to 32 bytes to save data space. 
BUG=none
TESTED=intelsde
R=brucedawson@google.com

Review URL: https://webrtc-codereview.appspot.com/32029004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1161 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2014-11-12 01:45:27 +00:00
fbarchard@google.com
ee4bc0d834 vzeroupper moved to just before ret. in one case it was done after ret, which is a bug that would cause a performance stall. 
BUG=none
TESTED=try bots
R=harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/24159004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1149 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2014-10-30 19:27:21 +00:00
fbarchard@google.com
2edea9454d Fix lint extraneous warning on row_win assembly by disabling the warning for those affected lines. 
BUG=none
TESTED=line row_win.cc
R=harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/29969004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1144 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2014-10-27 16:27:48 +00:00
fbarchard@google.com
f2fa453b94 Port I422ToABGR to AVX2. 
BUG=269
TESTED=intelsde on I422ToABGR
R=tpsiaki@google.com

Review URL: https://webrtc-codereview.appspot.com/23149004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1138 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2014-10-23 17:20:22 +00:00
fbarchard@google.com
22eb5965fc Optimize I422ToRGBA for AVX2 by hoisting ymm5 initialization and using different register for output of unpack. 
BUG=269
TESTED=intelsde on I422ToABGR
R=harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/29889004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1137 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2014-10-22 23:39:16 +00:00
fbarchard@google.com
c000955bc0 Port I422ToRGBA to AVX. 
BUG=269
TESTED=intelsde on I422ToRGBA
R=brucedawson@google.com

Review URL: https://webrtc-codereview.appspot.com/28769004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1136 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2014-10-22 22:41:39 +00:00
fbarchard@google.com
af6f25245e Reenable AVX2 scaling with bug fix for any width 
BUG=376
TESTED=unittest on scale functions
R=brucedawson@google.com, harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/30759004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1135 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2014-10-22 01:15:20 +00:00
fbarchard@google.com
4ec55a21cf Use macros to simplify I422ToARGB for AVX code. 
BUG=269
TESTED=local build with Visual C
R=harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/24079004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1133 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2014-10-21 22:48:32 +00:00
fbarchard@google.com
a063a66de4 Change I422ToARGB_AVX2 register usage to match SSSE3. ymm0 = B, ymm1 = G, ymm2 = R. 
BUG=269
TESTED=intelsde passes on unittests.
R=harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/28759004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1132 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2014-10-21 19:02:06 +00:00
fbarchard@google.com
d81dddd3d0 port I420ToBGRA to AVX2. 
BUG=269
TESTED=c:\intelsde\sde -ast -hsw -- out\release\libyuv_unittest.exe --gtest_filter=*I420ToBGRA*
R=brucedawson@google.com, harryjin@google.com, magjed@chromium.org

Review URL: https://webrtc-codereview.appspot.com/26869004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1127 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2014-10-20 19:35:55 +00:00
fbarchard@google.com
3dbaaf0032 switch win64 intrinsics to loadu / storeu for unaligned memory. 
BUG=372
TESTED=untested
R=brucedawson@google.com, harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/30729004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1124 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2014-10-16 23:46:48 +00:00
fbarchard@google.com
205c1440cf Use movdqu then pavgb to allow unaligned memory for rgb subsampling code. Allows this assembly to be used for unaligned pointers as well as aligned ones with no performance hit when memory is aligned on a modern cpu. 
BUG=365
TESTED=libyuvTest.ARGBToI420_Unaligned (453 ms)
R=harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/30679004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1116 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2014-10-07 19:47:06 +00:00
fbarchard@google.com
ca308327d2 Remove unaligned functions, since most function support unaligned memory now. This reduces complexity and improves performance for unaligned cases because C code can be avoided, and overhead is less. Downside is old cpus (core2 and earlier) will be slower for aligned memory case. Except mips, which has alignment requirement, but remove unaligned variant. 
BUG=365
TESTED=unittest builds and passes locally
R=harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/24839004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1113 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2014-10-07 00:59:31 +00:00
fbarchard@google.com
b720049a54 Make row functions used for planarfunctions and convert use movdqu to relax alignment constraint. Step 1 - make functions unaligned. 
BUG=365
TESTED=libyuv_unittest passes
R=harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/26709004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1111 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2014-10-03 21:11:37 +00:00
fbarchard@google.com
d83f63a3b4 InterpolateRow used for scale handle unaligned memory. Remove HalfRow which is not used. 
BUG=367
TESTED=unittest on I422ToI420
R=harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/28639004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1107 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2014-10-03 17:37:11 +00:00
fbarchard@google.com
455ae94c60 Make rotate SIMD allow unaligned pointers. 
BUG=365
TESTED=libyuv_unittest
R=tpsiaki@google.com

Review URL: https://webrtc-codereview.appspot.com/22899004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1102 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2014-10-02 17:56:48 +00:00
fbarchard@google.com
044f914c29 Change scale to unaligned movdqu. 
BUG=365
TESTED=scale unittests
R=tpsiaki@google.com

Review URL: https://webrtc-codereview.appspot.com/22879004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1101 16f28f9a-4ce2-e073-06de-1de4eb20be90
 2014-10-01 01:16:04 +00:00