The rounding and arithmetic is not the same as previously, to prevent overflow cases for large upscale factors.
We still rely on 32b x 32b -> 64b multiplies. Raised the fixed-point precision to 32b
so that we have some nice shifts from epi64 to epi32.
Changed rescaler_t type to 'uint32_t' in order to squeeze in all the precision required.
The MIPS code has been disabled because it's now out-of-sync. Will be fixed in
a subsequent CL when the dust settles.
~30-35% faster
Change-Id: I32e4ddc00933f1b1aa3463403086199fd5dad07b
* vertical expansion now uses bilinear interpolation
* heavily assumes that the alpha plane is decoded in full, not row-by-row
* split the RescalerExportRow and RescalerImportRow methods into Shrink
and Expand variants.
* MIPS implementation of ExportRowExpand is missing.
There's room for extra speed optim and code re-org, but let's keep that for later patches.
addresses https://code.google.com/p/webp/issues/detail?id=254
Change-Id: I8f12b855342bf07dd467fe85e4fde5fd814effdb
allows the values to be used in preproc checks, fixing a
-Wunreachable-code warning in 64-bit builds where VP8L_WRITER_BITS != 16
Change-Id: Ie98dff4e8ef896436557c64d5da2c5d70228a730
the x_add/x_sub increments were wrong for u/v in the upscaling case.
They shouldn't be left to the caller's discretion, but set up by
WebPRescalerInit to their exact necessary values.
-> Cleaned-up WebPRescalerInit() param list.
-> added safety asserts
-> removed the mips32/mips_r2 variant of "ImportRow" which were buggy prior
Change-Id: I347c75804d835811e7025de92a0758d7929dfc09
valgrind --tool=callgrind shows a 9 % speedup: 1021201984 ticks before vs.
927917709 after
-q 0 -m 0 -lossless ~/alpi/1.png
22.040 MP/s before
24.796 MP/s after
Change-Id: Iaab928167b3e20fb0d9401c6f8317a26c5a610b4
_BitScanReverse() takes an unsigned long*
http://msdn.microsoft.com/en-us/library/fbxyd7zd.aspx
fixes:
C4057: 'function': 'unsigned long *' differs in indirection to slightly
different base types from 'uint32_t *'
fixes issue #253
Change-Id: I0101ef7be18c7ed188b35e9b17e7f71290953786
this should be used in preference to free() for releasing memory
returned from WebPDecode*() / WebPEncode*(). this simplifies memory
management when working through language bindings
Change-Id: I15eb538a45390efc552fda8e5c251a3fbdc13c29
* make VP8LPrefetchBits() safe wrt past-EOS reads
* set 'BitReader::bits_" to a safe shifting value upon EOS
no visible performance difference on x86
Change-Id: I0a4177928cfa81d5dfc9054b36a686eaa1bf8c65
quiets an msvs /analyze warning, the cast is due to an earlier msvs
build warning.
C6297: Arithmetic overflow: 32-bit value is shifted, then cast to
64-bit value. Results might not be an expected value.
Change-Id: I0bb6cda57879f2fbd1e3515f6753a11bc08d14ac
histogram.c:
- Verified (earlier) that there's low correlation between Red & Blue colors
(particularly after applying Cross-color transform). The Bin based histogram
merge, bins on three entropies viz literal, red & blue symbols. Removing
either of blue or red increases the compression density. So keeping the bins
for red & blue sybmols.
- Keeping the compact bins method as-is. This way it's simpler to read.
huffman_encode.h: Added field comments for struct HuffmanTree and removed the TODO.
Change-Id: Ia76f7bc730079d1b3b644038c5d9931db3797f0e
* We don't need to change DecodeAlpha, since incremental
decoding is not useful for Alpha (we already decode
progressively along the RGB)
* Similarly, we don't do incremental decoding for level>0 planes:
the metadata don't turn into visible pixel (only the ones in level0), so...
(No visible speed change)
Change-Id: I2fd4b9ba227561a7dbede647686584b752be7baa
Optimize the decoding for region that have trivial literal codes.
The trivial literal is defined as huffman image with Red, Blue and Alpha
huffman trees with only single code values.
This speeds up lossless decoding by 3%
Change-Id: I0204949917836f74c0eb4ba5a7f4052a4797833b
ReadSymbol() finishes with a VP8LSetBitPos() call only and could miss an eos_ during the decode loop.
Things are faster because of inlining too.
Change-Id: I2d2a275f38834ba005bc767d45c5de72d032103e
eos_ needs to be set only when superfluous bits have actually
been requested.
Earlier, we were assuming pre-mature end-of-stream to be an error.
Now, more precisely, we mark error when we have encountered end-of-stream *and*
we attempt to read more bits after that.
This handles cases where image data requires no bits to be read
Change-Id: I628e2c39c64f10c443fb51f86b1f5919cc9fd299
speed-up is ~1.6% for photographic image to 10% for graphical image
(1000 images corpus was sped up by 5.8 %)
Code by akramarz@google.com and jyrki@google.com
Change-Id: Iceb2e50e6cc761b9315a3865d22ec9d19b8011c6
if a thread was still doing work when End() was called there'd be a race
on worker->status_. in these cases, however, the specific value is
meaningless as it would be >= OK and the thread would have been shut
down properly, but we'll check 'impl_' instead to avoid any potential
TSan/DRD reports.
Change-Id: Ib93cbc226a099f07761f7bad765549dffb8054b1
defines HAVE_BUILTIN_BSWAP16/32/64
updated endian_inl.h to have a non-configure fallback for gcc and clang
BSwap16() now uses __builtin_bswap16 if available
Change-Id: Ia04ee07b39303c4b247df96d84f298fb8a81f389
also reduce the load size from 64 to 32 bits as the top 32 bits are
being shifted away in the operation.
the change is neutral speed-wise on x86_64 as is the change in load size
on x86, but it gives a slight improvement on 32-bit arm.
x86 is improved ~13%, 32-bit arm ~3.7%
aarch64 is untested but will likely benefit as well.
Change-Id: Ibcb02a70f46f2651105d7ab571afe352673bef48
forces aligned memory reads (via memcpy) in the VP8 bit reader, useful
for platforms that don't support unaligned loads.
Change-Id: Ifa44a9a1677fbdc6a929520f9340b7e3fcbd6692
this defines WORDS_BIGENDIAN, replacing uses of
__BIG_ENDIAN__/__BYTE_ORDER__ with it
+ fixes lossless BGRA output with big-endian toolchains
that do not define __BIG_ENDIAN__ (codesourcery mips gcc)
Change-Id: Ieaccd623292d235343b5e34b7a720fc251c432d7
moves the following to this header:
- htole*() definitions from bit_writer.c
- __BIG_ENDIAN__ fallback define from bit_reader_inl.h
Change-Id: I7fff59543f08a70bf8f9ddac849b72ed290471b1
mostly to balance the use of bswap64, some gcc platforms are already
interpreting the default case the same
Change-Id: Icf860f55b3f16bea349a7d721e6d6abeeb4e5cf3
new options:
dwebp -alpha_dither
vwebp -noalphadither
When the source was marked as quantized, we use a threshold-averaging
filter to smooth the decoded alpha plane.
Note: this option forces the decoding of alpha data in one pass, and
might slow the decoding a bit.
The new field in WebPDecoderOptions struct is 'alpha_dithering_strength'
(0 by default, means: off). Max strength value is '100'.
Change-Id: I218e21af96360d4781587fede95f8ea4e2b7287a
only 1 of <lib>_CPPFLAGS and AM_CPPFLAGS is used, with the former
getting precedence when it's defined. configure's DEFAULT_INCLUDES is
covering what's necessary given the include paths are all source
relative.
Change-Id: I7d14076acd266b28a88a3d92bcc3d7165284d5f3
this change has the side-effect of using directory names in the
include, silencing a lint warning.
Change-Id: Ib91cf63a90534e32fadfa5c2372bfdb29f854d02
* remove LEFT/RIGHT_JUSTIFY distinction. It's all RIGHT_JUSTIFY now.
* simplify VP8GetSigned(), and add some masking branch-less code. Much
faster on ARM (~13% speed-up). 8% on x86-64, 5% on MacBook.
* split critical implementation into separate bit_reader_inl.h file that
is only included where needed (vp8.c / tree.c / bit_reader.c)
* bumped BITS value from 16 to 24 for x86-32b too, since it's a bit faster.
Change-Id: If41ca1da3e5c3dadacf2379d1ba419b151e7fce8
The luminance needs to be pre- and post- multiplied by
the alpha value in case of rescaling, for proper averaging.
Also:
- removed util/alpha_processing and moved it to dsp/
- removed WebPInitPremultiply() which was mostly useless
and merged it with the new function WebPInitAlphaProcessing()
Change-Id: If089cefd4ec53f6880a791c476fb1c7f7c5a8e60
gcc was generating very complex code, one for each case of br->len_ values!
also, pretty-fy the mask constants
Change-Id: If62b1e8266f3fe5334517305113038d2ea8a6b42
Sometimes, we can write 18bit or more at time, and it would
overflow the 32bit accumulator.
Also clarified the num-bits limitations (and exposed
VP8L_MAX_NUM_BIT_READ in bit_reader.h)
fixes http://code.google.com/p/webp/issues/detail?id=200
Seems a bit faster (use of local fields for bits_ / used_)
also: added the __QNX__ bswap while at it.
Change-Id: I876db93a931db15b083cf1d838c70105effa7167
MALLOC_FAIL_AT flag can be used to set-up a pre-determined failure
point during malloc calls. The counter value is retrieved using
getenv().
Example usage: export MALLOC_FAIL_AT=37 && cwebp input.png
will make 'cwebp' report a memory allocation error the 37th time
malloc() or calloc() is called.
MALLOC_MEM_LIMIT can be used similarly to prevent allocating more
than a given amount of memory. This is usually less convenient to
use than MALLOC_FAIL_AT since one has to know in advance the typical
memory size allocated.
Both these flags are meant to be used for debugging only!
Also: added a 'total_mem_allocated' to record the overall memory allocated
Change-Id: I9d408095ee7d76acba0f3a31b1276fc36478720a
This change reduces the number of calls to WebPSafeMalloc from 200 to
100. The overall memory consumption is down 3% for Lenna image.
Change-Id: I1b351a1f61abf2634c035ef1ccb34050b7876bdd
Some tracing code is activated by PRINT_MEM_INFO flag.
For debugging only! (not thread-safe, and slow).
Change-Id: I282c623c960f97d474a35b600981b761ef89ace9
* merged the two HistogramAdd/AddEval() into a single call
(with detection of special case when b==out)
* added a SSE2 variant
* harmonize the histogram type to 'uint32_t' instead
of just 'int'. This has a lot of ripples on signatures.
* 1-2% faster
Change-Id: I10299ff300f36cdbca5a560df1ae4d4df149d306
Reduce calls to Malloc (WebPSafeMalloc/WebPSafeCalloc) for:
- Building HashChain data-structure used in creating the backward references.
- Creating Backward references for LZ77 or RLE coding.
- Creating Huffman tree for encoding the image.
For the above mentioned code-paths, allocate memory once and re-use it
subsequently.
Reduce the foorprint of VP8LHistogram struct by changing the Struct
field 'literal_' from an array of constant size to dynamically allocated
buffer based on the input parameter cache_bits.
Initialize BitWriter buffer corresponding to 16bpp (2*W*H).
There are some hard-files that are compressed at 12 bpp or more. The
realloc is costly and can be avoided for most of the WebP lossless
images by allocating some extra memory at the encoder initializaiton.
Change-Id: I1ea8cf60df727b8eb41547901f376c9a585e6095
When remapping buffer, br->eos_ was wrongly being set to true for
certain
images.
Also, refactored the end-of-stream detection as a function.
Reported in http://crbug.com/364830
Change-Id: I716ce082ef2b505fe24246b9c14912d8e97b5d84
.set at - Indicates that macro expansions may clobber
the assembler temporary ($at or $28) register.
Some macros may not be expanded without this
and will generate an error message if noat
is in effect.
"at" also added to the clobber list.
Change-Id: I67feebbd9f2944fc7f26c28496e49e1e2348529d
there's still some malloc/free in the external example
This is an encoder API change because of the introduction
of WebPMemoryWriterClear() for symmetry reasons.
The MemoryWriter object should probably go in examples/ instead
of being in the main lib, though.
mux_types.h stil contain some inlined free()/malloc() that are
harder to remove (we need to put them in the libwebputils lib
and make sure link is ok). Left as a TODO for now.
Also: WebPDecodeRGB*() function are still returning a pointer
that needs to be free()'d. We should call WebPSafeFree() on
these, but it means exposing the whole mechanism. TODO(later).
Change-Id: Iad2c9060f7fa6040e3ba489c8b07f4caadfab77b
Separate the C version from the MIPS32 version and have run-time
initialization during RescalerInit()
Change-Id: I93cfa5691c073a099fe62eda1333ad2bb749915b
* simplify the endian logic
* remove the need for memset()
* write 16 or 32 at a time (likely aligned)
Makes the code a bit faster on ARM (~1%)
Change-Id: I650bc5654e8d0b0454318b7a78206b301c5f6c2c
Even at high quality setting, the U/V quantizer step is limited
to 4 which can lead to banding on gradient.
This option allows to selectively apply some randomness to
potentially flattened-out U/V blocks and attenuate the banding.
This option is off by default in 'dwebp', but set to -dither 50
by default in 'vwebp'.
Note: depending on the number of blocks selectively dithered,
we can have up to a 10% slow-down in decoding speed it seems.
Change-Id: Icc2446007f33ddacb60b3a80a9e63f2d5ad162de
Earlier we were only testing for bit_pos == LBITS. But this is not
sufficient,
as bit_pos can jump from < LBITS to > LBITS.
This was resulting in some bit-stream truncation errors not being
caught.
Note: Not a security bug though, as br->pos wasn't incremented in such
cases
and so we weren't reading beyond the buffer.
Change-Id: Idadcdcbc6a5713f8fac3470f907fa37a63074836
in_bits is const. Trying to apply bswap on it, one gets the error message:
error: read-only variable 'in_bits' used as 'asm' output
Change-Id: I0bef494b822c83d8ea87b1938b0e486d94de4742
speeds up those codes that are not part of the main lookup.
This gives a 10 % speedup for a photographic image.
Change-Id: Ief54b0ad77db790a01314402ad351b40ac9a7be4
rather than symlink the webm/vpx terms, use the same header as libvpx to
reference in-tree files
based on the discussion in:
https://codereview.chromium.org/12771026/
Change-Id: Ia3067ecddefaa7ee01550136e00f7b3f086d4af4
subdirectories with more than one target can have the install targets
run in parallel with make -jN. group the shared headers in one place to
produce a common install target.
Change-Id: I1f3aa338a8ee6d681de1e5d0b2c6244d2c3d5451
-> split libraries further into decoder / encoder
-> add libwebpdecoder.a in Makefile.unix
-> make dwebp link against libwebpdecoder.a in Makefile.unix
also: in makefile.unix, pass EXTRA_FLAGS to LDFLAGS too
(otherwise, -m32 wouldn't work, e.g.)
Change-Id: Ief3da02a729dd86bbaf949ed048836716941657f
Simplify and re-organize the VP8L bit-reader functions
(e.g.: the 40-bit look-ahead code was helping much)
Speed-up with LBITS=64, on arm7-a:
=> before:
./dwebp_justify_24_neon -v bryce_ll.webp
Time to decode picture: 11.393s
File bryce_ll.webp can be decoded (dimensions: 11158 x 2156).
...
=> after (LBITS=64): Time to decode picture: 9.953s
making the VP8L bit-reader in 32 bit mode is going to be
harder (because we need to be able to read two symbols
at a time, each with max length 15 bits)
Change-Id: I89746fb103b87b5e2fd40a3208a6fbc584b88297
This flag will make the code use no uint64, no asm, and no fancy
trick, but instead aim at being as simple and straightforward as
possible.
Main use is to help emscripten generate proper JS code.
More code needs to be simplified later.
Also: tune the BITS values to be 24 and make use of WEBP_RIGHT_JUSTIFY
Here are the typical timing for decoding a large image:
ARM7-a:
dwebp_justify_32_neon Time to decode picture: 3.280s
dwebp_justify_24_neon Time to decode picture: 2.640s
dwebp_justify_16_neon Time to decode picture: 2.723s
dwebp_justify_8_neon Time to decode picture: 2.802s
dwebp_justify_32 Time to decode picture: 4.264s
dwebp_justify_24 Time to decode picture: 3.696s
dwebp_justify_16 Time to decode picture: 3.779s
dwebp_justify_8 Time to decode picture: 3.834s
dwebp_32_neon Time to decode picture: 4.010s
dwebp_24_neon Time to decode picture: 2.725s
dwebp_16_neon Time to decode picture: 2.852s
dwebp_8_neon Time to decode picture: 2.778s
dwebp_32 Time to decode picture: 4.587s
dwebp_24 Time to decode picture: 3.800s
dwebp_16 Time to decode picture: 3.902s
dwebp_8 Time to decode picture: 3.815s
REFERENCE (HEAD) Time to decode picture: 3.818s
x86_64:
dwebp_justify_32 Time to decode picture: 0.473s
dwebp_justify_24 Time to decode picture: 0.434s
dwebp_justify_16 Time to decode picture: 0.450s
dwebp_justify_8 Time to decode picture: 0.467s
dwebp_32 Time to decode picture: 0.474s
dwebp_24 Time to decode picture: 0.468s
dwebp_16 Time to decode picture: 0.468s
dwebp_8 Time to decode picture: 0.481s
REFERENCE (HEAD) Time to decode picture: 0.436s
i386:
dwebp_justify_32 Time to decode picture: 0.723s
dwebp_justify_24 Time to decode picture: 0.618s
dwebp_justify_16 Time to decode picture: 0.626s
dwebp_justify_8 Time to decode picture: 0.651s
dwebp_32 Time to decode picture: 0.744s
dwebp_24 Time to decode picture: 0.627s
dwebp_16 Time to decode picture: 0.642s
dwebp_8 Time to decode picture: 0.642s
Change-Id: Ie56c7235733a24f94fbfc2e4351aae36ec39c225
When the config option '--enable-libwebpdecoder' is specified, the
lean decoder library 'libwebpdecoder' will be created in addition to
libwebp. Also dwebp binary will be linked to libwebpdecoder, if this
config option is specified.
Change-Id: I9de3e149b59c9a8390fae2ba660941749640e54a
Actually, it turns out we now should never call these functions
with a zero size, otherwise something is wrong in the logic.
Change-Id: Ie414fcbec95486c169190470a71f2cff0843782a
also change lossless encoder logic, which was relying on explicit
NULL return from WebPSafeMalloc(0)
renamed function to CheckSizeArgumentsOverflow() explicitly
addresses issue #138
Change-Id: Ibbd51cc0281e60e86dfd4c5496274399e4c0f7f3
The main advantage is that you can avoid the use of uint64_t
some times, sticking to 32bit only.
Default still is BITS=32, this is mainly "in case".
Change-Id: Id694028793117ba822c37d46ef6c52fa0afed4ac
- Separate out mux.h and demux.h
- muxtypes.h: new header for data types common to mux/demux
- Move some misc read/write utilities to utils/utils.h
- Remove some duplicate methods.
- Separate out mux/demux libraries
Change-Id: If9b9569b10d55d922ad9317ef51710544315d6de
Returning 0 (equal) can lead to undefined behaviour.
And, in our cases we'll never have equal keys (added asserts for that)
Change-Id: Ifaf202df321d3f877ad2a03de42e0d6cdd1b2388
SBITS=8 is reported 20-30% faster on ARM (where 64bit ops
are expensive).
Also use 32bits for i32.
Change-Id: Id6a7197d805061aeb8832f20432512d0d930ebfa
no speed diff observed by removing the test before calling BitWriterResize().
+ remove some unnecessary memset() in VP8LBitWriter
+ fix mixed code/variable-decl in BIG_ENDIAN mode
Change-Id: I36be61f83d10a43e4682b680c2dae0e494da4218
* ~1-4% faster
* if it's not used, don't use it
* remove the special handling of cache_bits = 0
* remove some tests in the loops
Change-Id: I19d87c3ca731052ff532ea8b2d8e89816507b75f
