view release on metacpan or  search on metacpan

src/Source/LibWebP/ChangeLog  view on Meta::CPAN

57897ba Merge "lossless_neon: use vcreate_*() where appropriate"
6aa4777 Merge "(enc|dec)_neon: use vcreate_*() where appropriate"
0d346e4 Always reinit VP8TransformWHT instead of hard-coding
7d039fc cwebp: fallback to native webp decode in WIC builds
d471f42 cwebp: add some missing newlines in longhelp output
bf0e003 lossless_neon: use vcreate_*() where appropriate
9251c2f (enc|dec)_neon: use vcreate_*() where appropriate
399b916 lossy decoding: correct alpha-rescaling for YUVA format
78c12ed Merge "Makefile.vc: add rudimentary avx2 support"
dc5b122 try to remove the spurious warning for static analysis
ddfefd6 Makefile.vc: add rudimentary avx2 support
a891164 Merge "simplify VP8LInitBitReader()"
fdbcd44 simplify VP8LInitBitReader()
7c00428 makefile.unix: add rudimentary avx2 support
515e35c Merge "add stub dsp/enc_avx2.c"
a05dc14 SSE2: yuv->rgb speed-up for point-sampling
178e9a6 add stub dsp/enc_avx2.c
1b99c09 Merge "configure: add a test for -mavx2"
fe72807 configure: add a test for -mavx2
e46a247 cpu: fix check for __cpuidex availability
176fda2 fix the bit-writer for lossless in 32bit mode
541784c dsp.h: add a check for AVX2 / define WEBP_USE_AVX2
bdb151e dsp/cpu: add AVX2 detection
ab9f2f8 Merge "revamp the point-sampling functions by processing a full plane"
a2f8b28 revamp the point-sampling functions by processing a full plane
ef07602 use decoder's DSP functions for autofilter
2b5cb32 Merge "dsp/cpu: add AVX detection"
df08e67 dsp/cpu: add AVX detection
e2f405c Merge "clean-up and slight speed-up in-loop filtering SSE2"
f60957b clean-up and slight speed-up in-loop filtering SSE2
9fc3ae4 .gitattributes: treat .ppm as binary
3da924b Merge "dsp/WEBP_USE_NEON: test for __aarch64__"
c716449 Android.mk: always include *_neon.c in the build
a577b23 dsp/WEBP_USE_NEON: test for __aarch64__
54bfffc move RemapBitReader() from idec.c to bit_reader code
34168ec Merge "remove all unused layer code"
f1e7717 remove all unused layer code
b0757db Code cleanup for VP8LGetHistoImageSymbols.
5fe628d make the token page size be variable instead of fixed 8192
f948d08 memory debug: allow setting pre-defined malloc failure points
ca3d746 use block-based allocation for backward refs storage, and free-lists
1ba61b0 enable NEON intrinsics in aarch64 builds
b9d2bb6 dsp/neon.h: coalesce intrinsics-related defines
b5c7525 iosbuild: add support for iOSv7/aarch64
9383afd Reduce number of memory allocations while decoding lossless.
888e63e Merge "dsp/lossless: prevent signed int overflow in left shift ops"
8137f3e Merge "instrument memory allocation routines for debugging"
2aa1873 instrument memory allocation routines for debugging
d3bcf72 Don't allocate VP8LHashChain, but treat like automatic object
bd6b861 dsp/lossless: prevent signed int overflow in left shift ops
b7f19b8 Merge "dec/vp8l: prevent signed int overflow in left shift ops"
29059d5 Merge "remove some uint64_t casts and use."
e69a1df dec/vp8l: prevent signed int overflow in left shift ops
cf5eb8a remove some uint64_t casts and use.
38e2db3 MIPS: MIPS32r1: Added optimization for HistogramAdd.
e0609ad dwebp: fix exit code on webp load failure
bbd358a Merge "example_util.h: avoid forward declaring enums"
8955da2 example_util.h: avoid forward declaring enums
6d6865f Added SSE2 variants for Average2/3/4
b3a616b make HistogramAdd() a pointer in dsp
c8bbb63 dec_neon: relocate some inline-asm defines
4e393bb dec_neon: enable intrinsics-only functions
ba99a92 dec_neon: use positive tests for USE_INTRINSICS
69058ff Merge "example_util: add ExUtilDecodeWebPIncremental"
a7828e8 dec_neon: make WORK_AROUND_GCC conditional on version
3f3d717 Merge "enc_neon: enable intrinsics-only functions"
de3cb6c Merge "move LOCAL_GCC_VERSION def to dsp.h"
1b2fe14 example_util: add ExUtilDecodeWebPIncremental
ca49e7a Merge "enc_neon: move Transpose4x4 to dsp/neon.h"
ad900ab Merge "fix warning about size_t -> int conversion"
4825b43 fix warning about size_t -> int conversion
42b35e0 enc_neon: enable intrinsics-only functions
f937e01 move LOCAL_GCC_VERSION def to dsp.h
5e1a17e enc_neon: move Transpose4x4 to dsp/neon.h
c7b92a5 dec_neon: (WORK_AROUND_GCC) delete unused Load4x8
8e5f90b Merge "make ExUtilLoadWebP() accept NULL bitstream param."
05d4c1b Merge "cwebp: add webpdec"
ddeb6ac cwebp: add webpdec
35d7d09 Merge "Reduce memory footprint for encoding WebP lossless."
0b89610 Reduce memory footprint for encoding WebP lossless.
f0b65c9 make ExUtilLoadWebP() accept NULL bitstream param.
9c0a60c Merge "dwebp: move webp decoding to example_util"
1d62acf MIPS: MIPS32r1: Added optimization for HuffmanCost functions.
4a0e739 dwebp: move webp decoding to example_util
c022046 Merge "Bugfix: Incremental decode of lossy-alpha"
8c7cd72 Bugfix: Incremental decode of lossy-alpha
7955152 MIPS: fix error with number of registers.
b1dabe3 Merge "Move the HuffmanCost() function to dsp lib"
75b1200 Move the HuffmanCost() function to dsp lib
2772b8b MIPS: fix assembler error revealed by clang's debug build
6653b60 enc_mips32: fix unused symbol warning in debug
8dec120 enc_mips32: disable ITransform(One) in debug builds
98519dd enc_neon: convert Disto4x4 to intrinsics
fe9317c cosmetics:
953b074 enc_neon: cosmetics
a9fc697 Merge "WIP: extract the float-calculation of HuffmanCost from loop"
3f84b52 Merge "replace some mult-long (vmull_u8) with mult-long-accumulate (vmlal_u8)"
4ae0533 MIPS: MIPS32r1: Added optimizations for ExtraCost functions.
b30a04c WIP: extract the float-calculation of HuffmanCost from loop
a8fe8ce Merge "NEON intrinsics version of CollectHistogram"
95203d2 NEON intrinsics version of CollectHistogram
7ca2e74 replace some mult-long (vmull_u8) with mult-long-accumulate (vmlal_u8)
41c6efb fix lossless_neon.c
8ff96a0 NEON intrinsics version of FTransform
0214f4a Merge "MIPS: MIPS32r1: Added optimizations for FastLog2"
baabf1e MIPS: MIPS32r1: Added optimizations for FastLog2
3d49871 NEON functions for lossless coding
3fe0291 MIPS: MIPS32r1: Added optimizations for SSE functions.
c503b48 Merge "fix the gcc-4.6.0 bug by implementing alternative method"
abe6f48 fix the gcc-4.6.0 bug by implementing alternative method
5598bde enc_mips32.c: fix file mode
2b1b4d5 MIPS: MIPS32r1: Add optimization for GetResidualCost
f0a1f3c Merge "MIPS: MIPS32r1: Added optimization for FTransform"
7231f61 MIPS: MIPS32r1: Added optimization for FTransform
869eaf6  ~30% encoding speedup: use NEON for QuantizeBlock()
f758af6 enc_neon: convert FTransformWHT to intrinsics
7dad095 MIPS: MIPS32r1: Added optimization for Disto4x4 (TTransform)
2298d5f MIPS: MIPS32r1: Added optimization for QuantizeBlock
e88150c Merge "MIPS: MIPS32r1: Add optimization for ITransform"
de693f2 lossless_neon: disable VP8LConvert* functions
4143332 NEON intrinsics for encoding
0ca2914 MIPS: MIPS32r1: Add optimization for ITransform
71bca5e dec_neon: use vst_lane instead of vget_lane
bf06105 Intrinsics NEON version of TransformOne
19c6f1b Merge "dec_neon: use vld?_lane instead of vset?_lane"
7a94c0c upsampling_neon: drop NEON suffix from local functions
d14669c upsampling_sse2: drop SSE2 suffix from local functions
2ca42a4 enc_sse2: drop SSE2 suffix from local functions
d038e61 dec_sse2: drop SSE2 suffix from local functions
fa52d75 dec_neon: use vld?_lane instead of vset?_lane
c520e77 cosmetic: fix long line
4b0f2da Merge "add intrinsics NEON code for chroma strong-filtering"
e351ec0 add intrinsics NEON code for chroma strong-filtering
aaf734b Merge "Add SSE2 version of forward cross-color transform"
c90a902 Add SSE2 version of forward cross-color transform
bc374ff Use histogram_bits to initalize transform_bits.
2132992 Merge "Add strong filtering intrinsics (inner and outer edges)"
5fbff3a Add strong filtering intrinsics (inner and outer edges)
d4813f0 Add SSE2 function for Inverse Cross-color Transform
2602956 dec_neon: add strong loopfilter intrinsics
cca7d7e Merge "add intrinsics version of SimpleHFilter16NEON()"
1a05dfa windows: fix dll builds
d6c50d8 Merge "add some colorspace conversion functions in NEON"
4fd7c82 SSE2 variants of Subtract-Green: Rectify loop condition
97e5fac add some colorspace conversion functions in NEON
b9a7a45 add intrinsics version of SimpleHFilter16NEON()
daccbf4 add light filtering NEON intrinsics
af44460 fix typo in STORE_WHT
6af6b8e Tune HistogramCombineBin for large images.
af93bdd use WebPSafe[CM]alloc/WebPSafeFree instead of [cm]alloc/free
51f406a lossless_sse2: relocate VP8LDspInitSSE2 proto
0f4f721 separate SSE2 lossless functions into its own file
514fc25 VP8LConvertFromBGRA: use conversion function pointers
6d2f352 dsp/dec: TransformDCUV: use VP8TransformDC
defc8e1 Merge "fix out-of-bound read during alpha-plane decoding"
fbed364 Merge "dsp: reuse wht transform from dec in encoder"
d846708 Merge "Add SSE2 version of ARGB -> BGR/RGB/... conversion functions"
207d03b fix out-of-bound read during alpha-plane decoding
d1b33ad 2-5% faster trellis with clang/MacOS (and ~2-3% on ARM)
369c26d Add SSE2 version of ARGB -> BGR/RGB/... conversion functions
df230f2 dsp: reuse wht transform from dec in encoder
80e218d Android.mk: fix build with APP_ABI=armeabi-v7a-hard
59daf08 Merge "cosmetics:"
5362200 cosmetics:
3e7f34a AssignSegments: quiet array-bounds warning
3c2ebf5 Merge "UpdateHistogramCost: avoid implicit double->float"
cf821c8 UpdateHistogramCost: avoid implicit double->float
312e638 Extend the search space for GetBestGreenRedToBlue
1c58526 Fix few nits
fef2270 Optimize and re-structure VP8LGetHistoImageSymbols
068b14a Optimize lossless decoding.
5f0cfa8 Do a binary search to get the optimum cache bits.
24ca367 Merge "allow 'cwebp -o -' to emit output to stdout"
e12f874 allow 'cwebp -o -' to emit output to stdout
2bcad89 allow some more stdin/stout I/O
84ed4b3 fix cwebp.1 typos after patch #69199
65b99f1 add a -z option to cwebp, and WebPConfigLosslessPreset() function
3017661 4-5% faster trellis by removing some unneeded calculations.
687a58e histogram.c: reindent after b33e8a0
06d456f Merge "~3-4% faster lossless encoding"
c60de26 ~3-4% faster lossless encoding
42eb06f Merge "few cosmetics after patch #69079"
82af826 few cosmetics after patch #69079
b33e8a0 Refactor code for HistogramCombine.
ca1bfff Merge "5-10% encoding speedup with faster trellis (-m 6)"
5aeeb08 5-10% encoding speedup with faster trellis (-m 6)
82ae1bf cosmetics: normalize VP8GetCPUInfo checks
e3dd924 Merge "Refactor GetBestPredictorForTile for future tuning."
206cc1b Refactor GetBestPredictorForTile for future tuning.
3cb8406 Merge "speed-up trellis quant (~5-10% overall speed-up)"
b66f222 Merge "lossy encoding: ~3% speed-up"
4287d0d speed-up trellis quant (~5-10% overall speed-up)
390c8b3 lossy encoding: ~3% speed-up
9a463c4 Merge "dec_neon: convert TransformWHT to intrinsics"
e8605e9 Merge "dec_neon: add ConvertU8ToS16"
4aa3e41 MIPS: MIPS32r1: rescaler bugfix
c16cd99 Speed up lossless encoder.
9d6b5ff dec_neon: convert TransformWHT to intrinsics
2ff0aae dec_neon: add ConvertU8ToS16
77a8f91 fix compilation with USE_YUVj flag
4acbec1 Merge changes I3b240ffb,Ia9370283,Ia2d28728
2719bb7 dec_neon: TransformAC3: work on packed vectors
b7b60ca dec_neon: add SaturateAndStore4x4
b7685d7 Rescale: let ImportRow / ExportRow be pointer-to-function
e02f16e dec_neon.c: convert TransformDC to intrinsics
9cba963 add missing file
8992ddb use static clipping tables
0235d5e 1-2% faster quantization in SSE2
b2fbc36 fix VC12-x64 warning
6e37cb9 Merge "cosmetics: backward_references.c: reindent after a7d2ee3"
a42ea97 cosmetics: backward_references.c: reindent after a7d2ee3