- 97c9626 Test yuv scaler with scale factor of 1 by fbarchard@google.com · 11 years ago
- b911428 Adapt row interpolator to do YUV as well as ARGB without extrude so it can be used in I420Scale. by fbarchard@google.com · 11 years ago
- e24ba5c Fix for llvm GPR register usage by fbarchard@google.com · 11 years ago
- bb92aca Unaligned math effects. by fbarchard@google.com · 11 years ago
- 75a5403 llvm -O0 -fpic 32 bit runs out of registers. Extend the apple fix to llvm. by fbarchard@google.com · 11 years ago
- cd6056c InterpolateAny for unaligned and odd width interpolate. To be used in ARGBScaler in future. by fbarchard@google.com · 11 years ago
- aa7988f Enhanced Rep Mov String version of CopyRow for posix and use cpu detect for ERMS by fbarchard@google.com · 11 years ago
- 55c20a8 Fix movdqu in yuy2 by fbarchard@google.com · 11 years ago
- f8e9017 switch toyuy2 from aligned to unaligned by fbarchard@google.com · 11 years ago
- 050b39a Recomputed JPeg coefficients normalized to 128. Apply to ARGBGray function reusing YJ function/coefficients and rounding. by fbarchard@google.com · 11 years ago
- 4e0d7cc Y coefficients for J420 need to be scaled by 255/219 to full range. by fbarchard@google.com · 11 years ago
- cfaa66c ARGBToJ420 and ARGBToJ400 - Full range YUV Jpeg style. by fbarchard@google.com · 11 years ago
- 9d48df9 Sobel port to posix. Improved unittest for C version. by fbarchard@google.com · 11 years ago
- a3be470 quick fix for sobel compile error by fbarchard@google.com · 11 years ago
- e8df16b Sobel use G channel for consistency on all CPUs, better performance and full range of 0 to 255. by fbarchard@google.com · 11 years ago
- 805fefb Unattenutate with work around for vpgatherdd on avx2 by fbarchard@google.com · 11 years ago
- 1096543 ARGBShuffle AVX2 by fbarchard@google.com · 11 years ago
- 83a63e6 Change YUV_DISABLE_ASM to LIBYUV_DISABLE_NEON, LIBYUV_DISABLE_MIPS, LIBYUV_DISABLE_X86 by fbarchard@google.com · 11 years ago
- 787f827 Unattenuate multiply alpha by 1. by fbarchard@google.com · 11 years ago
- d8b73ca fix ARGBToI444 UV code to use signed math by fbarchard@google.com · 11 years ago
- af137b6 Interpolate for SSE2 with specialization for 1/4, 1/2, 3/4 and 1 by fbarchard@google.com · 11 years ago
- 762c050 ARGBToI444_SSSE3 port to gcc by fbarchard@google.com · 11 years ago
- cde5870 Replace two spaces with one after . by fbarchard@google.com · 11 years ago
- 573a883 Subtract 2 images. by fbarchard@google.com · 11 years ago
- 83e1b17 Add 2 ARGB Images together and store to destination by fbarchard@google.com · 11 years ago
- 8fa7634 Blend style multiple by fbarchard@google.com · 11 years ago
- 8ec6033 ARGBMultiply 2 images together by fbarchard@google.com · 11 years ago
- b94b139 Shade ported to Neon by fbarchard@google.com · 12 years ago
- 1d160cb Attenuate AGRB pixels NEON optimized by fbarchard@google.com · 12 years ago
- 2580ccf Scale benchmark just the OPT and run C version once. by fbarchard@google.com · 12 years ago
- b549175 Neon optimized argb filter row for bilinear scale and Effects Interpolate. by fbarchard@google.com · 12 years ago
- f08ac6b Rename row functions so they are all SomethingRow_CPU by fbarchard@google.com · 12 years ago
- c1f17f1 quick rename bug fix by fbarchard@google.com · 12 years ago
- bdf7cb5 RGB formats converted to YUV with Neon by fbarchard@google.com · 12 years ago
- 30859f7 Neon YToARGB and fix SSE2 to match C version by fbarchard@google.com · 12 years ago
- 00b69a2 I400ToARGB_Neon optimized by fbarchard@google.com · 12 years ago
- dbcabea move rgb565 wrapper to common so it wont inline by fbarchard@google.com · 12 years ago
- 34d2344 noinline attribute for apple by fbarchard@google.com · 12 years ago
- 1544926 NV12ToRGB565 use NV12ToARGB any row function by fbarchard@google.com · 12 years ago
- e0d8648 MergeUV aligned and SplitUV cpu detect combined with width check. by fbarchard@google.com · 12 years ago
- 1dafd44 MergeUV backport to SSE2 by fbarchard@google.com · 12 years ago
- db694ed SplitUV_Any variations for all CPUS. by fbarchard@google.com · 12 years ago
- 9de8867 I422ToYUY2_NEON by fbarchard@google.com · 12 years ago
- 4de0c43 Enable SSE version of I420ToRAW/RGB24 for Linux/Mac/ChromeOS x86 by fbarchard@google.com · 12 years ago
- 8d37dd5 Any conversions fix for RGB 3 bytes. Fix for overread valgrind. Avoid memcpy. Add _Any unittests for all conversions. by fbarchard@google.com · 12 years ago
- e91bdac Move HalfRow to row_win and port to row_neon by fbarchard@google.com · 12 years ago
- 64ce0ab Move Neon source to its own files. by fbarchard@google.com · 12 years ago
- fc7314e Add exports to allow libyuv to be built as a shared lib. by fbarchard@google.com · 12 years ago
- 6fd84a8 NV12 optimization turned back on now that unittest is fixed by fbarchard@google.com · 12 years ago
- 142f6c4 Move row.h to include and remove rotate_priv.h by fbarchard@google.com · 12 years ago
- 9bcc9a2 ARGBInterpolateRow_SSSE3 for motion blur. Used to use bilinear row filter, which extrudes edges. This branches off the code so the extrude can be removed for Interpolate. by fbarchard@google.com · 12 years ago
- a2cc341 ARGBToRGBA implemented as alias by fbarchard@google.com · 12 years ago
- b8eabfe RGBAToARGB conversion. SSSE3 optimized. by fbarchard@google.com · 12 years ago
- d4164fb Clean up posix build of YUY2 by fbarchard@google.com · 12 years ago
- c704f78 YUY2ToARGB and UYVYToARGB use specialized row functions that do not subsample. by fbarchard@google.com · 12 years ago
- 41f24bf Affine function in 64 bit Mac use movd instead of movq for xmm to gpr. by fbarchard@google.com · 12 years ago
- f3181b3 Fix aligned version of SSSE3 alpha blend and re-enable SSE2 alpha blend for performance evaluation on low end CPUs where pshufb is slow. by fbarchard@google.com · 12 years ago
- f877e71 Alpha blend test alignment of source pointer and use movdqa aligned fetches. by fbarchard@google.com · 12 years ago
- e3cc769 4 pixel version of affine for gcc and aligned version of win. by fbarchard@google.com · 12 years ago
- 7344440 AffineRow for GCC. by fbarchard@google.com · 12 years ago
- b0c9797 Update Copyright notice to follow new chromium conventions. by fbarchard@google.com · 12 years ago
- eeac290 ARGBGrayTo takes source and destination pointers so caller does not need to make a copy of the image. by fbarchard@google.com · 12 years ago
- 6398e1d Port ARGBShade to Mac by fbarchard@google.com · 12 years ago
- 4d2cfd3 Change name of variable for mac because static cant be used so variables need to be unique. by fbarchard@google.com · 12 years ago
- 27d42c7 Expose ARGB rotation using scale low levels by fbarchard@google.com · 12 years ago
- 8f439ea ARGBColorMatrixRow_SSSE3 use signed psraw instead of psrlw before doing pack by fbarchard@google.com · 12 years ago
- 81b804e ARGBQuantize to do a posterizing effect. Added random resolution unittest. by fbarchard@google.com · 12 years ago
- e442dc4 ARGBcolorMatrix for applying transforms such as grey and sepia in a more general form. Unittest does sepia for comparison. by fbarchard@google.com · 12 years ago
- 794fe12 alpha blend 4 pixel loop bug fix and blender C code match SSE for better testability and reference code for future optimized code. by fbarchard@google.com · 12 years ago
- ee22088 Fix constant in rowposix, remove addrow and minor cleanups by fbarchard@google.com · 12 years ago
- bac5f2c ARGBBlend functions that handle any alignment and number of pixels by fbarchard@google.com · 12 years ago
- 133adc4 followup cleanup for blur and make compilable with /clr by fbarchard@google.com · 12 years ago
- f38aefe Blur use circular buffer of Cumulative Sums instead of full frame for better cache coherency. by fbarchard@google.com · 12 years ago
- f51e879 Blur functions by fbarchard@google.com · 12 years ago
- 2d9fe08 direct conversion from NV12 to ARGB by fbarchard@google.com · 12 years ago
- 4c416e8 Fix for I444ToARGBRow_Unaligned_SSSE3, I422ToARGBRow_Unaligned_SSSE3, I411ToARGBRow_Unaligned_SSSE3 on Windows using movdqu instead of movdqa. break YUVTORGB into 2 macros - one to fetch pixels, another to do YUV conversion. Less duplicated source and lends itself to future YUV formats. by fbarchard@google.com · 12 years ago
- e214fe3 I411ToARGB doing 2 UV values with 8 Y values by fbarchard@google.com · 12 years ago
- 6d6b770 Fix RGB565ToARGB_SSE2. Expose rgb conversions. Low levels were already there for i420 to rgb, but rgb to rgb functions were not exposed. Added unittests. by fbarchard@google.com · 12 years ago
- 221e602 ARGBSepia planarfunction for converting a region of ARGB image to Sepia tone. ARGBGray optimized weaving of alpha value. 551 ms from 568 ms. by fbarchard@google.com · 12 years ago
- ffaea7e ARGBGray function for convertion a rectangle of ARGB to gray scale by fbarchard@google.com · 12 years ago
- 9c02ba5 ARGBUnattenuateRow_C use same table as ARGBUnattenuateRow_SSE2. omit-frame-pointer for ScaleRowDown38_3_Int_NEON by fbarchard@google.com · 12 years ago
- f2d84dd scale ported to OSX for 32/64 bit. Required breaking up some functions into 2 or 3 asm inlines to set constants. Reworked storing of 6 bytes on all platforms for 3/4 scale by storing first 4 bytes and then storing last 4 bytes, overlapping 2, in order to stay within SSE2 registers. by fbarchard@google.com · 12 years ago
- 5bf29b5 p2align all loops, copy stride to local for scale, and copy last byte in bilinear more efficiently by fbarchard@google.com · 12 years ago
- 5566302 AddRow function to help implement box filter by fbarchard@google.com · 12 years ago
- 5ff3a8f ARGBBlendRow1_SSSE3 added to allow SSSE3 only alpha blending. Saves on SSE2 cpu dispatching by fbarchard@google.com · 12 years ago
- da5cc42 Prune SSE2 code out when building specifically for SSSE3 by fbarchard@google.com · 12 years ago
- 810cd91 ARGBUnattenuateRow_SSE2 use reciprocal table and pmul by fbarchard@google.com · 12 years ago
- 8ed5422 Attenuation ported to SSE2 by fbarchard@google.com · 12 years ago
- 96af870 Port Blend SSSE3 to gcc by fbarchard@google.com · 12 years ago
- 1702ec7 use movdqu on 2nd source for blend by fbarchard@google.com · 12 years ago
- d2f4413 Remove old alpha blend, expose GetARGB2Blend, fix ComputeSumSquareErrorPlane on SSE2 by fbarchard@google.com · 12 years ago
- c757f30 Alpha blend 2 sources and store to a destination. Useful for A under B blending. by fbarchard@google.com · 12 years ago
- 952a507 I420ToARGB without row buffer faster and unlimited in size. Uses SSSE3 unaligned for multiples of 8 and C for remainder by fbarchard@google.com · 12 years ago
- f3fb7b6 Lint cleanup for Neon and Visual C inline. by fbarchard@google.com · 12 years ago
- e14b2ab small count Review URL: https://webrtc-codereview.appspot.com/460007 by fbarchard@google.com · 12 years ago
- 2b9c210 lint fix test by fbarchard@google.com · 12 years ago
- 91ab139 inline version of alpha blend by fbarchard@google.com · 12 years ago
- 18184fd switch looping to jg from ja to allow non-multiple of 16 to underflow to a negative by fbarchard@google.com · 12 years ago
- 2217ced gcc port fixed for alpha by fbarchard@google.com · 12 years ago
- 976423f alpha blend last pixel fix by fbarchard@google.com · 12 years ago