FFmpeg

Commit Graph

Author	SHA1	Message	Date
Ganesh Ajjanagadde	83a04f103d	lavc: move exp2fi to ff_exp2fi in internal.h Reviewed-by: Michael Niedermayer <michael@niedermayer.cc> Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>	9 years ago
Ganesh Ajjanagadde	42868ca569	avcodec/jpeg2000: replace naive pow call with smarter exp2fi pow is a very wasteful function for this purpose. A low hanging fruit would be simply to replace with exp2f, and that does yield some speedup. However, there are 2 drawbacks of this: 1. It does not exploit the integer nature of the argument. 2. (minor) Some platforms lack a proper exp2f routine, making benefits available only to non broken libm. 3. exp2f does not solve the same issue that plagues pow, namely terrible worst case performance. This is a fundamental issue known as the "table-maker's dilemma" recognized by Prof. Kahan himself and subsequently elaborated and researched by many others. All this is clear from benchmarks below. This exploits the IEEE-754 format to get very good performance even in the worst case for integer powers of 2. This solves all the issues noted above. Function tested with clang usan over [-1000, 1000] (beyond range of relevance for this, which is [-255, 255]), patch itself with FATE. Benchmarks obtained on x86-64, Haswell, GNU-Linux via 10^5 iterations of the pow call, START/STOP, and command ffplay ~/samples/jpeg2000/chiens_dcinema2K.mxf. Low number of runs also given to prove the point about worst case: pow: 216270 decicycles in pow, 1 runs, 0 skips 110175 decicycles in pow, 2 runs, 0 skips 56085 decicycles in pow, 4 runs, 0 skips 29013 decicycles in pow, 8 runs, 0 skips 15472 decicycles in pow, 16 runs, 0 skips 8689 decicycles in pow, 32 runs, 0 skips 5295 decicycles in pow, 64 runs, 0 skips 3599 decicycles in pow, 128 runs, 0 skips 2748 decicycles in pow, 256 runs, 0 skips 2304 decicycles in pow, 511 runs, 1 skips 2072 decicycles in pow, 1022 runs, 2 skips 1963 decicycles in pow, 2044 runs, 4 skips 1894 decicycles in pow, 4091 runs, 5 skips 1860 decicycles in pow, 8184 runs, 8 skips exp2f: 134140 decicycles in pow, 1 runs, 0 skips 68110 decicycles in pow, 2 runs, 0 skips 34530 decicycles in pow, 4 runs, 0 skips 17677 decicycles in pow, 8 runs, 0 skips 9175 decicycles in pow, 16 runs, 0 skips 4931 decicycles in pow, 32 runs, 0 skips 2808 decicycles in pow, 64 runs, 0 skips 1747 decicycles in pow, 128 runs, 0 skips 1208 decicycles in pow, 256 runs, 0 skips 952 decicycles in pow, 512 runs, 0 skips 822 decicycles in pow, 1024 runs, 0 skips 765 decicycles in pow, 2047 runs, 1 skips 722 decicycles in pow, 4094 runs, 2 skips 693 decicycles in pow, 8190 runs, 2 skips exp2fi: 2740 decicycles in pow, 1 runs, 0 skips 1530 decicycles in pow, 2 runs, 0 skips 955 decicycles in pow, 4 runs, 0 skips 622 decicycles in pow, 8 runs, 0 skips 477 decicycles in pow, 16 runs, 0 skips 368 decicycles in pow, 32 runs, 0 skips 317 decicycles in pow, 64 runs, 0 skips 291 decicycles in pow, 128 runs, 0 skips 277 decicycles in pow, 256 runs, 0 skips 268 decicycles in pow, 512 runs, 0 skips 265 decicycles in pow, 1024 runs, 0 skips 263 decicycles in pow, 2048 runs, 0 skips 263 decicycles in pow, 4095 runs, 1 skips 260 decicycles in pow, 8191 runs, 1 skips Reviewed-by: Michael Niedermayer <michael@niedermayer.cc> Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>	9 years ago
Michael Niedermayer	0afdfbe116	avcodec/jpeg2000: fix type of arguments of tag_tree_size Fixes: out of array read Fixes: 36b8096fefab16c4c9326a508053e95c/signal_sigsegv_1d9ce18_3233_1a55196b018106dfabeace071a432d9e.r3d Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>	9 years ago
Michael Niedermayer	0eb7de1973	avcodec/jpeg2000: Change coord to 32bit to support larger than 32k width or height Fixes: 03e0abe721b1174856d41a1eb5d6a896/signal_sigabrt_7ffff6ae7cc9_3813_e71bf3541abed3ccba031cd5ba0269a4.avi Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>	9 years ago
Michael Niedermayer	a1a8cbcb35	avcodec/jpeg2000: Check comp coords to be within the supported size Fixes assertion failure Fixes: 03e0abe721b1174856d41a1eb5d6a896/signal_sigabrt_7ffff6ae7cc9_3813_e71bf3541abed3ccba031cd5ba0269a4.avi This fix is choosen to be simple to backport, better solution for master is planed Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>	9 years ago
Michael Niedermayer	016fd413f9	avcodec/jpeg2000: Use av_image_check_size() in ff_jpeg2000_init_component() Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>	9 years ago
Luca Barbato	570fcaf332	jpeg2000: Factor out prec init Makes debugging a little simpler.	9 years ago
Luca Barbato	95a41311ac	jpeg2000: Factor out band stepsize initialization	9 years ago
Luca Barbato	7fb93eae43	jpeg2000: Factor out band initialization	9 years ago
Ganesh Ajjanagadde	36f7a37895	avcodec/jpeg2000: comment out unused variable Fixes -Wunused-const-variable from http://fate.ffmpeg.org/report.cgi?time=20150820031140&slot=arm64-darwin-clang-apple-5.1 Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com> Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>	9 years ago
Michael Niedermayer	96e73fa649	avcodec/jpeg2000dec: Merge i_stepsize/2 decoder special case into dequant Signed-off-by: Michael Niedermayer <michaelni@gmx.at>	10 years ago
Michael Niedermayer	317be53ca6	avcodec/jpeg2000: Support 32 decomposition levels Fixes Ticket4680 Signed-off-by: Michael Niedermayer <michaelni@gmx.at>	10 years ago
Michael Niedermayer	c72a831931	avcodec/jpeg2000dec: Support progression order changes Fixes Ticket4657 Signed-off-by: Michael Niedermayer <michaelni@gmx.at>	10 years ago
Michael Niedermayer	f1e173049e	avcodec/jpeg2000: Remove CBLK limit This also reduces the amount of memory needed Fixes Ticket4672 The new code seems slightly faster as well, probably due to better cache usage Signed-off-by: Michael Niedermayer <michaelni@gmx.at>	10 years ago
Michael Niedermayer	67fc898243	avcodec/jpeg2000: Improve right bottom precinct positions Fixes Ticket4656 Signed-off-by: Michael Niedermayer <michaelni@gmx.at>	10 years ago
Michael Niedermayer	4e926fb969	avcodec/jpeg2000: Move L band scaling from the 9/7f wavelet to quantization stage This reduces the number of operations Its not done for 9/7i as that would overflow thanks to JPEG2000 allowing 32 decomposition levels Signed-off-by: Michael Niedermayer <michaelni@gmx.at>	10 years ago
Michael Niedermayer	6c7b1597c7	avcodec/jpeg2000: Move H band scaling from wavelet into quantization code This reduces the number of operations done and is equivalent except for rounding Signed-off-by: Michael Niedermayer <michaelni@gmx.at>	10 years ago
Michael Niedermayer	1c495b0bf6	avcodec/jpeg2000: Move gainb handling into the quantization code thats how the specification defines it, this also improves numerical accuracy of the integer wavelet implementation. It otherwise should be equivalent, in case of overflows this can be reverted. Signed-off-by: Michael Niedermayer <michaelni@gmx.at>	10 years ago
Michael Niedermayer	ec5164aa31	avcodec/jpeg2000: Fix regression with multiple tiles Fixes Ticket 4624 Signed-off-by: Michael Niedermayer <michaelni@gmx.at>	10 years ago
Michael Niedermayer	a206c13289	avcodec/jpeg2000: Change codeblock coords to same origin as bands Signed-off-by: Michael Niedermayer <michaelni@gmx.at>	10 years ago
Michael Niedermayer	d9841e53e2	avcodec/jpeg2000: Change precinct coords to same origin as bands Signed-off-by: Michael Niedermayer <michaelni@gmx.at>	10 years ago
Michael Niedermayer	59c107292c	avcodec/jpeg2000: Fix band coords Signed-off-by: Michael Niedermayer <michaelni@gmx.at>	10 years ago
James Almer	9f815bc2c2	avcodec/jpeg200dsp: add ff_rct_int_{sse2,avx2} Reviewed-by: Michael Niedermayer <michaelni@gmx.at> Signed-off-by: James Almer <jamrial@gmail.com>	10 years ago
James Almer	7912a6830d	avcodec/jpeg200dsp: add ff_ict_float_{sse,avx} Original intrinsics version by Nicolas Bertrand. Reviewed-by: Michael Niedermayer <michaelni@gmx.at> Signed-off-by: James Almer <jamrial@gmail.com>	10 years ago
Michael Niedermayer	dc73c7adc0	avcodec/jpeg2000dec: Fix Selective arithmetic coding bypass and Multiple codeword segments These 2 are highly related so they are in the same commit Fixes part of Ticket4605 Fixes p0_04.j2k Signed-off-by: Michael Niedermayer <michaelni@gmx.at>	10 years ago
Michael Niedermayer	dbbb31e5bb	avcodec/jpeg2000: remove unused variable Signed-off-by: Michael Niedermayer <michaelni@gmx.at>	10 years ago
Michael Niedermayer	44327cbc9a	avcodec/jpeg2000: Disable special case for JPEG2000_QSTY_SI The code gave apparently completely wrong values Fixes Ticket2872 Signed-off-by: Michael Niedermayer <michaelni@gmx.at>	10 years ago
Michael Niedermayer	074159ed70	avcodec/jpeg2000dec: Fix subsampled decoding Fixes part of Ticket3619 Signed-off-by: Michael Niedermayer <michaelni@gmx.at>	10 years ago
Vittorio Giovara	a2448cfe16	jpeg2000: do not compute the same value twice CC: libav-stable@libav.org Bug-Id: CID 1026771 / CID 1026772 / CID 1026773	10 years ago
Vittorio Giovara	10306e9c5f	jpeg2000: fix dereferencing invalid pointers during cleanup CC: libav-stable@libav.org Found-by: Laurent Butti <laurentb@gmail.com> Signed-off-by: Vittorio Giovara <vittorio.giovara@gmail.com>	11 years ago
Diego Biurrun	fe4d5fe936	jpeg2000: Mark static data init functions as av_cold	11 years ago
Lou Logan	88f2586adb	fix various typos Signed-off-by: Lou Logan <lou@lrcd.com> Signed-off-by: Michael Niedermayer <michaelni@gmx.at>	11 years ago
Michael Niedermayer	e54f4510aa	avcodec/jpeg2000: zero i/f_data Fixes uninitialized use of variables Fixes part of Ticket2921 Signed-off-by: Michael Niedermayer <michaelni@gmx.at>	11 years ago
Michael Niedermayer	9e477a3770	jpeg2000: fix null pointer dereference in case of malloc failure Signed-off-by: Michael Niedermayer <michaelni@gmx.at>	11 years ago
Michael Niedermayer	09927f3eaa	jpeg2000: zero reslevel array on allocation prevent use of uninitialized values Signed-off-by: Michael Niedermayer <michaelni@gmx.at>	11 years ago
Michael Niedermayer	912ce9dd20	jpeg2000: fix dereferencing invalid pointers Found-by: Laurent Butti <laurentb@gmail.com> Signed-off-by: Michael Niedermayer <michaelni@gmx.at>	11 years ago
Michael Niedermayer	f57119b8e5	jpeg2000: fix overflow in dequantization Fixes decoding of file generated with: ffmpeg -f lavfi -i smptehdbars=hd720 -pix_fmt rgb48 /tmp/o.jp2 Reviewed-by: Nicolas BERTRAND <nicoinattendu@gmail.com> Signed-off-by: Michael Niedermayer <michaelni@gmx.at>	12 years ago
Luca Barbato	daeb4e3042	jpeg2000: Proper cleanup on failure in decode_frame() Signed-off-by: Luca Barbato <lu_zero@gentoo.org>	12 years ago
Michael Niedermayer	db13e02c6f	jpeg2000: Compute corrections of sub-band coordinates Signed-off-by: Luca Barbato <lu_zero@gentoo.org>	12 years ago
Michael Niedermayer	5bf208f659	jpeg2000: Use separate fields for int and float codepaths Split stepsize and data into int and float variants. Eliminates a number of casts and simplifies spotting errors. Signed-off-by: Luca Barbato <lu_zero@gentoo.org>	12 years ago
Michael Niedermayer	690c8af97b	jpeg2000: Remove unneeded variable initialization Signed-off-by: Luca Barbato <lu_zero@gentoo.org>	12 years ago
Michael Niedermayer	4cbd5ed11b	jpeg2000: Fix compute precedence error in lut_gain index Signed-off-by: Luca Barbato <lu_zero@gentoo.org>	12 years ago
Michael Niedermayer	952f7ed3c0	jpeg2000: Compute quantization for 'scalar derived' in the correct case. Scalar derived case is represented by the JPEG2K_QSTY_SI define. Signed-off-by: Luca Barbato <lu_zero@gentoo.org>	12 years ago
Michael Niedermayer	b44925ae6b	jpeg2000: Initialize code blocks structures in precincts to 0 Prevent use of uninitialized memory / valgrind failure. Found-by: ubitux Signed-off-by: Michael Niedermayer <michaelni@gmx.at> Signed-off-by: Luca Barbato <lu_zero@gentoo.org>	12 years ago
Michael Niedermayer	2c3901b2c3	jpeg2000: Remove unneeded returns Signed-off-by: Luca Barbato <lu_zero@gentoo.org>	12 years ago
Michael Niedermayer	78962d3df4	jpeg2000: Check that nreslevels2decode has been initialized before use Avoid buffer overread. Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind Signed-off-by: Luca Barbato <lu_zero@gentoo.org>	12 years ago
Michael Niedermayer	4e11b15534	jpeg2000: Calculate code-block coord in ff_jpeg2000_init_component() Signed-off-by: Luca Barbato <lu_zero@gentoo.org> Signed-off-by: Nicolas Bertrand <nicoinattendu@gmail.com> Signed-off-by: Luca Barbato <lu_zero@gentoo.org>	12 years ago
Michael Niedermayer	ffb1051b9b	jpeg2000: Zero prec->cblk Prevent use of uninitialized memory / valgrind failure Found-by: ubitux Signed-off-by: Michael Niedermayer <michaelni@gmx.at>	12 years ago
Michael Niedermayer	6e9bfc19bd	jpeg2000: check that nreslevels2decode has been initialized before use Fixes out of array accesses Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind Signed-off-by: Michael Niedermayer <michaelni@gmx.at>	12 years ago
Michael Niedermayer	234f76ae7d	jpeg2000: Fix unintended sign extension in malloc arguments of prec Fixes: CID1026772/3 Signed-off-by: Michael Niedermayer <michaelni@gmx.at>	12 years ago

1 2

79 Commits (954f865c8ee291d7c4c83a03815711321e931ada)