Commit Graph

2176 Commits

Author SHA1 Message Date
7f1b897bee Faster stochastic histogram merging.
Constants are such that brute force is sometimes faster for some
data (mostly big images it seems).

Change-Id: I90aef536408683535e3b09ddfa2e77a9834038f6
2016-08-19 14:52:57 +02:00
3884972e3f remove WEBP_FORCE_ALIGNED and use memcpy() instead.
BUG=webp:297

Change-Id: I89a08debec7bb1b3f411c897260ab1bb63f77df2
2016-08-17 20:16:03 -07:00
005e15b10a Merge "{extras,mux}/Makefile.am: add missing -lm" 2016-08-17 18:18:39 +00:00
6ab496ed22 fix some 'unsigned integer overflow' warnings in ubsan
I couldn't find a safe way of fixing VP8GetSigned() so i just
used the big-hammer.

Change-Id: I1039bc00307d1c90c85909a458a4bc70670e48b7
2016-08-16 23:18:27 -07:00
8a4ebc6ab0 Revert "fix 'unsigned integer overflow' warnings in ubsan"
This reverts commit e44f5248ff.

contains unintentional changes in quant.c

Change-Id: I1928f072566788b0c9ea80f6fbc9e571061f9b3e
2016-08-16 16:55:56 -07:00
9d4f209f80 Merge changes I25711dd5,I43188fab
* changes:
  Fix assertions in WebPRescalerExportRow()
  Add descriptions of default configuration in help info.
2016-08-16 22:13:23 +00:00
e44f5248ff fix 'unsigned integer overflow' warnings in ubsan
I couldn't find a safe way of fixing VP8GetSigned() so i just
used the big-hammer.

Change-Id: I1039bc00307d1c90c85909a458a4bc70670e48b7
2016-08-16 15:04:41 -07:00
27b5d991e2 Fix assertions in WebPRescalerExportRow()
Change-Id: I25711dd54e71c90a25f7b18e0ef9155e8151a15e
2016-08-16 14:32:48 -07:00
aaf2530cc3 {extras,mux}/Makefile.am: add missing -lm
+ libwebpmux.pc

quality_estimate.c, anim_encode.c rely on functions from math.h

BUG=webp:306

Change-Id: I3a8eb48febfd52bfbeb04f4dc615ccbed72926f7
2016-08-15 22:32:46 -07:00
1269dc7cfb Refactor VP8LColorCacheContains()
Return key/index if the query is found, and -1 otherwise.
The benefit of this is to save a hashing computation.

Change-Id: Iff056be330f5fb8204011259ac814f7677dd40fe
2016-08-12 15:16:06 -07:00
40872fb2e6 dec_neon,NeedsHev: micro optimization
trade 2 compares + 1 logical or for max + compare

Change-Id: I785ad8efdc64db2d0609456d6e7af795ab2117d8
2016-08-08 20:12:30 -07:00
b551e587b3 cosmetics: add {}s on continued control statements
for consistency within the codebase. in some cases simply join the
lines.

Change-Id: I071f061052e274c8a69f651ed4305befb4414a40
2016-08-03 19:08:59 -07:00
d2e4484ef3 dsp/Makefile.am: put msa source in correct lib
upsampling_msa.c was incorrectly included in the neon convenience lib
+ sort msa sources

Change-Id: I7c4883f16a5c2fed12bfa0e8d8d6a7acd5d4fb84
2016-08-03 17:50:45 -07:00
c7f66c82b3 Merge "utils/thread.c,cosmetics: join a few lines" 2016-07-28 04:08:48 +00:00
39f4ffbcdc utils/thread.c,cosmetics: join a few lines
Change-Id: I94c142fc6f9e6823ce16ca723145354eae4db9af
2016-07-27 18:10:18 -07:00
0b2c58a91c Fix an unsigned integer overflow error in enc/cost.h
Change-Id: I9774b59c417c185f09a61a115364b9642976a100
2016-07-26 13:55:09 -07:00
386e4ba2f0 Reset segment id if we decide not to update segment map
This avoids potential encoder and decoder mismatch.

Change-Id: I5282d3e168afc6193033ad3fce8fbc35618ab2f5
2016-07-25 17:08:10 -07:00
d3ddacb625 Add MSA optimized YUV to RGB upsampling functions
We add the following MSA optimized YUV to RGB upsampling functions:
- UpsampleRgbLinePair
- UpsampleBgrLinePair
- UpsampleRgbaLinePair
- UpsampleBgraLinePair
- UpsampleArgbLinePair
- UpsampleRgba4444LinePair
- UpsampleRgb565LinePair

Change-Id: I7264a615edc7eb376e443e9d38bd8e3c9a2cab1f
2016-07-22 14:28:30 +00:00
45fbeba5cd Merge "Do token recording and counting in a single loop" 2016-07-20 13:07:41 +00:00
ad497fbce3 move src/extras to the top-level
reserve src/ for the code of the main libraries: libwebp, libwebpdemux
and libwebpmux

+ demote this library to internal only; i.e., don't install the header +
lib with make install

Change-Id: I8c9844db8f494be0fa0a2549a5b75b5cebcf666d
2016-07-19 17:23:17 -07:00
0c0fb83211 Do token recording and counting in a single loop
Change-Id: I8afd3c486b210bd67888de03e91dde7f78276f89
2016-07-19 16:28:26 -07:00
9ac74f922e Add MSA optimized rescaling functions
We add the following MSA optimized rescaling functions:
- RescalerExportRowExpand
- RescalerExportRowShrink

Change-Id: Ic1c76065423b02617db94cf0c22bb564219b36e6
2016-07-19 15:52:42 +00:00
cb19dbc1a4 Add MSA optimized color transform functions
We add the following MSA optimized color transform functions:
- TransformColor
- SubtractGreenFromBlueAndRed

Change-Id: Ib182d2b5faa7191f503ce70f0dfde0ac89402fd3
2016-07-18 13:49:24 +00:00
3f4042b52a WebPAnimEncoder: If 'minimize_size' and 'allow_mixed' on, try lossy + lossless.
This improves compression by ~5% at default quality.

If only 'allow_mixed' is on (but 'minimize_size' isn't), we continue to
use a heuristic to try one of the two or both.

Change-Id: Ia573a73ea26ad25f9debff759eed69d2b0449e82
2016-07-15 16:56:20 -07:00
5e2eb89e1f cosmetics,dsp/*msa.c: associate '*' with the type
not the variable

Change-Id: If5823e9731c406655eaf1dc1aaa2e6554ca7daad
2016-07-15 15:40:41 -07:00
5b60db5c9d FastMBAnalyze() for quick i16/i4 decision
The decision is based on the variance between DC values of each
sub-4x4 block. This heuristic is rather ok for predicting whether
the 2nd transform (intra-16) is going to help or not.
The decision threshold varies with quality (=quantization).

It's only used for -m 0 and -m 1, where no full RD-opt is performed.
It actually makes these modes quite faster, with RD curve much
closer to the -m 2 mode.

Change-Id: I15f972db97ba4082cbd1dfd16bee3eb2eca701a8
2016-07-15 11:21:08 -07:00
567e697776 Add MSA optimized CollectHistogram function
We add the following MSA optimized encoder Histogram function:
- CollectHistogram

Change-Id: I28415704ec62c3ad375de06eeef468d9f514bb2d
2016-07-15 22:51:33 +05:30
c54ab8dd1a Add MSA optimized quantization functions
We add the following MSA optimized encoder quantization functions:
- QuantizeBlock
- Quantize2Blocks

Change-Id: Ie32b442afa99eee62d2ef48942b41116a4e157d3
2016-07-15 15:33:47 +00:00
ec6f68c51e Merge "Remove QuantizeBlockWHT() in enc.c" 2016-07-14 18:59:08 +00:00
2a5c417c68 Apply the RLE heuristic to LZ77.
Change-Id: I7317eed7e017ee8981f40fcf1737f97e0e3a238c
2016-07-14 20:12:48 +02:00
91b59e886b Remove QuantizeBlockWHT() in enc.c
QuantizeBlockWHT() is basically identical to QuantizeBlock(),
no need to keep two copies.

Change-Id: I970cb6948da1c750c1339971a55e3b40765cdd01
2016-07-14 10:44:18 -07:00
fe57273736 Add MSA optimized SSE functions
We add the following MSA optimized encoder SSE functions:
- SSE16x16
- SSE16x8
- SSE8x8
- SSE4x4

Change-Id: I9ef9e903019337d9975c83264a652a7282bf5d5b
2016-07-14 15:43:23 +05:30
6b53ca876e cosmetics,(dec|enc)_sse2.c: fix indent
Change-Id: Ic3326136ddd325e911e96c2e5a7f06b3e1d60f66
2016-07-13 16:11:29 -07:00
b15d00d996 Merge "Add MSA optimized encoder IntraChromaPreds function" 2016-07-13 14:54:28 +00:00
afe3cec813 Add MSA optimized encoder IntraChromaPreds function
We add the following MSA optimized intrapred chroma function:
- IntraChromaPreds

Change-Id: I051cd174f5ce675aeb94e648d52c5a340a133ed4
2016-07-13 18:13:51 +05:30
fc8cad9f29 reduce the number of malloc/free cycles in huffman.c
pre-allocating a sorted[] array for most common cases of small
alphabet size cuts a lot of traffic.

Change-Id: I73ff2f6e507f81b0b0bb7d9801a344aa4bcb038a
2016-07-12 12:06:31 -07:00
7b4b05e0dc Add MSA optimized encoder Intra16Preds function
We add the following MSA optimized intrapred 16x16 function:
- Intra16Preds

Change-Id: I89a249e041fbed377cb6a328c0b973add335b980
2016-07-12 14:30:03 +05:30
c18787a0e9 Add MSA optimized encoder Intra4Preds function
We add the following MSA optimized intrapred 4x4 function:
- Intra4Preds

Change-Id: Icf325f3dcbf98bb6210811b666ce632cae575b22
2016-07-12 04:36:56 +00:00
bbb6ecd9b0 Merge "Add MSA optimized distortion functions" 2016-07-07 08:59:40 +00:00
7915396f40 Add MSA optimized distortion functions
We add the following MSA optimized distortion functions:
- Disto4x4
- Disto16x16

Change-Id: I0a545ed0182ea56a0d5f358639f6671c2c21b95c
2016-07-07 07:30:22 +00:00
c0991a141f io,EmitRescaledAlphaYUV: factor out a common expr
+ s/src_a/dst_a/
+ remove unnecessary (void) as expected_num_lines_out is used within the
  function

Change-Id: Ic45f798ef22bd19eaabf1a0512d1cf8a201bb4b5
2016-07-06 20:38:55 -07:00
bfef6c9f82 Merge tag 'v0.5.1'
libwebp-0.5.1
- 6/14/2016: version 0.5.1
  This is a binary compatible release.
  * miscellaneous bug fixes (issues #280, #289)
  * reverted alpha plane encoding with color cache for compatibility with
    libwebp 0.4.0->0.4.3 (issues #291, #298)
  * lossless encoding performance improvements
  * memory reduction in both lossless encoding and decoding
  * force mux output to be in the extended format (VP8X) when undefined chunks
    are present (issue #294)
  * gradle, cmake build support
  * workaround for compiler bug causing 64-bit decode failures on android
    devices using clang-3.8 in the r11c NDK
  * various WebPAnimEncoder improvements

* tag 'v0.5.1': (30 commits)
  update ChangeLog
  Clarify the expected 'config' lifespan in WebPIDecode()
  update ChangeLog
  Fix corner case in CostManagerInit.
  gif2webp: normalize the number of .'s in the help message
  vwebp: normalize the number of .'s in the help message
  cwebp: normalize the number of .'s in the help message
  fix rescaling bug: alpha plane wasn't filled with 0xff
  Improve lossless compression.
  'our bug tracker' -> 'the bug tracker'
  normalize the number of .'s in the help message
  pngdec,ReadFunc: throw an error on invalid read
  decode.h,WebPGetInfo: normalize function comment
  Inline GetResidual for speed.
  Speed-up uniform-region processing.
  free -> WebPSafeFree()
  DecodeImageData(): change the incorrect assert
  Fix a boundary case in BackwardReferencesHashChainDistanceOnly.
  Make sure to consider small distances in LZ77.
  add some asserts to delimit the perimeter of CostManager's operation
  ...

Change-Id: I44cee79fddd43527062ea9d83be67da42484ebfc
2016-07-06 19:31:27 -07:00
deb54d915a Clarify the expected 'config' lifespan in WebPIDecode()
Change-Id: I08c880adc34d651a63fa2d07cfb07b1468a6df1b
2016-07-06 17:17:15 -07:00
435308e029 Add MSA optimized encoder transform functions
We add the following MSA optimized encoder transform functions:
- ITransform
- FTransform
- FTransformWHT

Change-Id: Ia6b17556aba5aff2d7a88208905fb45293d080a8
2016-07-05 14:35:47 +00:00
dce64bfa1b Add MSA optimized alpha filter functions
We add the following MSA optimized alpha filter functions:
- HorizontalFilter
- VerticalFilter
- GradientFilter

Change-Id: I71e2e04050e569b8c0bf086fadf210ee16d50924
2016-07-01 19:58:25 +00:00
429120d0af Add MSA optimized color transform functions
We add the following MSA optimized color transform functions:
- AddGreenToBlueAndRed
- TransformColorInverse

Change-Id: Iceab3813905955aa8b811253df9188512fc7de3f
2016-06-28 20:36:21 +05:30
55b2fede7f normalize the macros' "do {...} while (0)" constructs
(so we're no longer bitten by the extra ';' problem!)

Change-Id: Icf849c97df9a7af135ba15a7906fc28590d7ce77
2016-06-27 15:30:05 -07:00
701c772eed Add MSA optimized colorspace conversion functions
We add the following MSA optimized colorspace conversion functions:
- ConvertBGRAToRGBA
- ConvertBGRAToBGR
- ConvertBGRAToRGB

Change-Id: I76db1c829d593a06d4975d54dbafa385c82b84fb
2016-06-27 21:19:06 +00:00
c7eb06f737 Fix corner case in CostManagerInit.
Change-Id: I91795d05eb78816d6d9a8cadc64d3814650d2aee
2016-06-27 20:01:44 +02:00
f918cb1013 fix rescaling bug: alpha plane wasn't filled with 0xff
(in case no alpha was present in the source .webp,
but user requested some)

Change-Id: I9011d38237907c60d6796a86bd2c72166aa80f27
(cherry picked from commit 06a38c7b1c)
2016-06-27 10:48:31 -07:00