History | View | Annotate | Download (169 KB)
support doing motion estimation and compensation without any residual transform or codingthis is usefull for filters which need ME/MC
Originally committed as revision 5456 to svn://svn.ffmpeg.org/ffmpeg/trunk
Snow: don't try to encode lossless with 9/7 wavelet, because it isn't lossless.
Originally committed as revision 5433 to svn://svn.ffmpeg.org/ffmpeg/trunk
Snow: cosmetics
Originally committed as revision 5432 to svn://svn.ffmpeg.org/ffmpeg/trunk
Snow: scale predicted mv based on which reference frame the neighbors used.
Originally committed as revision 5431 to svn://svn.ffmpeg.org/ffmpeg/trunk
Originally committed as revision 5428 to svn://svn.ffmpeg.org/ffmpeg/trunk
Snow: fix subband weighting in wavelet cmp functions. use 32x32 cmp in iterative motion estimation.
Originally committed as revision 5427 to svn://svn.ffmpeg.org/ffmpeg/trunk
Snow: 10l. 1pass ratecontrol failed to set snow's internal quant.
Originally committed as revision 5426 to svn://svn.ffmpeg.org/ffmpeg/trunk
Snow multiple reference frames(bitstream is not backwards compatible, even if refs aren't used)
Originally committed as revision 5417 to svn://svn.ffmpeg.org/ffmpeg/trunk
Snow 1pass ratecontrol
Originally committed as revision 5416 to svn://svn.ffmpeg.org/ffmpeg/trunk
sanity checks, some might have been exploitable ...
Originally committed as revision 5369 to svn://svn.ffmpeg.org/ffmpeg/trunk
Snow mmx + sse2 part 2
Patch by Robert Edele, yartrebo <<at>> earthlink <<dot>> net
Originally committed as revision 5184 to svn://svn.ffmpeg.org/ffmpeg/trunk
First part of a series of speed-enchancing patches.This one sets up a snow.h and makes snow use the dsputil function pointerframework to access the three functions that will be implemented in asmin the other parts of the patchset.Patch by Robert Edele < yartrebo AH earthlink POIS net>...
dont pass intra flag if its always 0
Originally committed as revision 5164 to svn://svn.ffmpeg.org/ffmpeg/trunk
seed iterative_me with mvs from conventional search.+15% speed, +.1 dB
Originally committed as revision 5158 to svn://svn.ffmpeg.org/ffmpeg/trunk
obey vqscale in 2nd pass
Originally committed as revision 5157 to svn://svn.ffmpeg.org/ffmpeg/trunk
av_free -> av_freep
Originally committed as revision 5145 to svn://svn.ffmpeg.org/ffmpeg/trunk
fix %8 != 0 w/h
Originally committed as revision 5144 to svn://svn.ffmpeg.org/ffmpeg/trunk
add static keyword to some functionspatch by Dieter <freebsd at sopwith solgatos com>
Originally committed as revision 4913 to svn://svn.ffmpeg.org/ffmpeg/trunk
Remove all stray tabs and trailing whitespace, this time for good.
Originally committed as revision 4891 to svn://svn.ffmpeg.org/ffmpeg/trunk
fix convergence of snow iterative_me
Originally committed as revision 4876 to svn://svn.ffmpeg.org/ffmpeg/trunk
revert 1.83 because it's not needed
Originally committed as revision 4875 to svn://svn.ffmpeg.org/ffmpeg/trunk
detect when iterative_me fails to converge, and stop searching
Originally committed as revision 4868 to svn://svn.ffmpeg.org/ffmpeg/trunk
obmc-aware 4mv
Originally committed as revision 4864 to svn://svn.ffmpeg.org/ffmpeg/trunk
Update licensing information: The FSF changed postal address.
Originally committed as revision 4842 to svn://svn.ffmpeg.org/ffmpeg/trunk
bugs found by valgrind
Originally committed as revision 4841 to svn://svn.ffmpeg.org/ffmpeg/trunk
faster iterative_me: avoid duplicate mc of neighboring blocks.
Originally committed as revision 4839 to svn://svn.ffmpeg.org/ffmpeg/trunk
use h264 MC functions for 2xX Xx2 blocks in snow too
Originally committed as revision 4824 to svn://svn.ffmpeg.org/ffmpeg/trunk
use h264 MC functions for rectangular blocks too
Originally committed as revision 4821 to svn://svn.ffmpeg.org/ffmpeg/trunk
10l (assertion failure)
Originally committed as revision 4819 to svn://svn.ffmpeg.org/ffmpeg/trunk
cleanupfix 4mv playback distortions
Originally committed as revision 4813 to svn://svn.ffmpeg.org/ffmpeg/trunk
:%s/check_block_2p/check_block_inter/g
Originally committed as revision 4794 to svn://svn.ffmpeg.org/ffmpeg/trunk
this fixes compilation with gcc 2.95 which seems to not like array casting... special case inline of check_block which makes the array and passes it
Originally committed as revision 4793 to svn://svn.ffmpeg.org/ffmpeg/trunk
mem corruption fix
Originally committed as revision 4789 to svn://svn.ffmpeg.org/ffmpeg/trunk
4mv + iter ME 10l fix (still not bugfree but better)
Originally committed as revision 4783 to svn://svn.ffmpeg.org/ffmpeg/trunk
10l
Originally committed as revision 4782 to svn://svn.ffmpeg.org/ffmpeg/trunk
iterative overlapped block based motion estimation for snow
Originally committed as revision 4781 to svn://svn.ffmpeg.org/ffmpeg/trunk
COSMETICS: tabs --> spaces, some prettyprinting
Originally committed as revision 4764 to svn://svn.ffmpeg.org/ffmpeg/trunk
Originally committed as revision 4758 to svn://svn.ffmpeg.org/ffmpeg/trunk
COSMETICS: Remove all trailing whitespace.
Originally committed as revision 4749 to svn://svn.ffmpeg.org/ffmpeg/trunk
check block_max_depth (fixes crash)
Originally committed as revision 4718 to svn://svn.ffmpeg.org/ffmpeg/trunk
make selftest code compile again
Originally committed as revision 4717 to svn://svn.ffmpeg.org/ffmpeg/trunk
fix unused variable warnings
Originally committed as revision 4609 to svn://svn.ffmpeg.org/ffmpeg/trunk
reducing minimum supported picture size
Originally committed as revision 4535 to svn://svn.ffmpeg.org/ffmpeg/trunk
#defines for strict_std_compliance and split between inofficial extensions and non standarized things
Originally committed as revision 4205 to svn://svn.ffmpeg.org/ffmpeg/trunk
PIX_FMT_NONE and related fixes
Originally committed as revision 4161 to svn://svn.ffmpeg.org/ffmpeg/trunk
tighter snow slicing patch by (Yartrebo || yartrebo earthlink net)
Originally committed as revision 4159 to svn://svn.ffmpeg.org/ffmpeg/trunk
Ministry of English Composition, reporting for duty (and the word is "skipped", not "skiped"; "skiped" would rhyme with "hyped")
Originally committed as revision 4153 to svn://svn.ffmpeg.org/ffmpeg/trunk
fix overflow in mc_block() patch by (Yartrebo )yartrebo earthlink net
Originally committed as revision 4132 to svn://svn.ffmpeg.org/ffmpeg/trunk
store the number of runs to avoid storing the last run valueabout 10% lower bitrate for -qscale 32 (forman & some music video)worst case bitrate increase <0.1% (lossless or low qscale)and now the bad news, even though this just adds a single subtraction and an if() into the medium sized unpack_coeffs() loop and the if() will only be false once per unpac_coeff() call, gcc produces 50% slower code, i didnt look at the generated asm yet, not sure if i want to ......
allow 2pass ratecontrol. also fixes psnr displayed by mencoder.
Originally committed as revision 4130 to svn://svn.ffmpeg.org/ffmpeg/trunk
replace complicated pointer dereference + index stuff by pointers in unpack_coeffs()
Originally committed as revision 4129 to svn://svn.ffmpeg.org/ffmpeg/trunk
minor optimization
Originally committed as revision 4128 to svn://svn.ffmpeg.org/ffmpeg/trunk
optimizing unpack_coeffs()
Originally committed as revision 4127 to svn://svn.ffmpeg.org/ffmpeg/trunk
Snow segfault bug in revision 1.43 patch by (Yartrebo /yartrebo earthlink net/)
Originally committed as revision 4126 to svn://svn.ffmpeg.org/ffmpeg/trunk
rescale coefficients during IDWT, that way the lifting steps are much simpler and faster
Originally committed as revision 4119 to svn://svn.ffmpeg.org/ffmpeg/trunk
increasing precission of the quantization parameterthis is needed as the quantization stepsize for each subband is also in this precission and insignificant changes to the wavelet like scaling its coefficients slightly differently would lead to wildly variing PSNR and bitrate...
fix QROOT != 8
Originally committed as revision 4114 to svn://svn.ffmpeg.org/ffmpeg/trunk
10% faster unpack_coeffs
Originally committed as revision 4108 to svn://svn.ffmpeg.org/ffmpeg/trunk
fixing lossless snow
Originally committed as revision 4107 to svn://svn.ffmpeg.org/ffmpeg/trunk
merge vertical lifting steps and a little cleanup
Originally committed as revision 4104 to svn://svn.ffmpeg.org/ffmpeg/trunk
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
Originally committed as revision 4101 to svn://svn.ffmpeg.org/ffmpeg/trunk
slice-based idwt (10% faster decoding)
Originally committed as revision 4040 to svn://svn.ffmpeg.org/ffmpeg/trunk
-strict -1 comment fix
Originally committed as revision 3976 to svn://svn.ffmpeg.org/ffmpeg/trunk
avoid buf_size == 0 checks in every decoder
Originally committed as revision 3872 to svn://svn.ffmpeg.org/ffmpeg/trunk
10l patch by (matthieu castet <castet.matthieu free fr>)
Originally committed as revision 3842 to svn://svn.ffmpeg.org/ffmpeg/trunk
various security fixes and precautionary checks
Originally committed as revision 3822 to svn://svn.ffmpeg.org/ffmpeg/trunk
Disable encoders patch by (Gianluigi Tiesi <mplayer netfarm it>)
Originally committed as revision 3797 to svn://svn.ffmpeg.org/ffmpeg/trunk
fix for build on IRIX by (Michel Bardiaux {mbardiaux peaktime be})
Originally committed as revision 3713 to svn://svn.ffmpeg.org/ffmpeg/trunk
CABAC -> range coder
Originally committed as revision 3655 to svn://svn.ffmpeg.org/ffmpeg/trunk
unused
Originally committed as revision 3502 to svn://svn.ffmpeg.org/ffmpeg/trunk
merge predict_plane() with DWTELEM->8bit conversation (21% faster)
Originally committed as revision 3499 to svn://svn.ffmpeg.org/ffmpeg/trunk
8 -> FRAC_BITS
Originally committed as revision 3489 to svn://svn.ffmpeg.org/ffmpeg/trunk
consistent use of types patch by (D Richard Felker III <dalias at aerifal dot cx>)
Originally committed as revision 3480 to svn://svn.ffmpeg.org/ffmpeg/trunk
simplify getsymbol patch by (Loren Merritt <lorenm at u dot washington dot edu>)
Originally committed as revision 3476 to svn://svn.ffmpeg.org/ffmpeg/trunk
per picture psnr
Originally committed as revision 3465 to svn://svn.ffmpeg.org/ffmpeg/trunk
100l (forgot to set the 8x8 mc functions, so ME did use the mpeg ones ...)
Originally committed as revision 3456 to svn://svn.ffmpeg.org/ffmpeg/trunk
use h264 MC code if possible50% faster predict_plane() if mmx2/3dnow is available0.1% bitrate increase
Originally committed as revision 3455 to svn://svn.ffmpeg.org/ffmpeg/trunk
use pointer arithmetic in mc_block(), 25% faster
Originally committed as revision 3452 to svn://svn.ffmpeg.org/ffmpeg/trunk
20% faster predict_plane()moion vector max/min fix
Originally committed as revision 3434 to svn://svn.ffmpeg.org/ffmpeg/trunk
10l (keyframes and context resets)
Originally committed as revision 3426 to svn://svn.ffmpeg.org/ffmpeg/trunk
cleanup
Originally committed as revision 3425 to svn://svn.ffmpeg.org/ffmpeg/trunk
10lmore benchmarking codeinline
Originally committed as revision 3424 to svn://svn.ffmpeg.org/ffmpeg/trunk
merge decode_subband() and dequantize()
Originally committed as revision 3422 to svn://svn.ffmpeg.org/ffmpeg/trunk
11% faster decode_subband()
Originally committed as revision 3421 to svn://svn.ffmpeg.org/ffmpeg/trunk
5% faster decode_subband()a little cleanup
Originally committed as revision 3420 to svn://svn.ffmpeg.org/ffmpeg/trunk
46% faster decode_subband()
Originally committed as revision 3419 to svn://svn.ffmpeg.org/ffmpeg/trunk
50% faster decode_subband()
Originally committed as revision 3418 to svn://svn.ffmpeg.org/ffmpeg/trunk
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)different intra block coding (previous was just an ugly hack)1.8% bitrate reduction -0.01PSNR (foreman@352x288 qscale=8)1.5% bitrate reduction +0.05PSNR (foreman@352x288 qscale=1)...
edge bugfix1-1.3% lower bitrate and +0.01 PSNR (foremen@352x288 qscale=1,8)
Originally committed as revision 3414 to svn://svn.ffmpeg.org/ffmpeg/trunk
Originally committed as revision 3377 to svn://svn.ffmpeg.org/ffmpeg/trunk
lossless support
Originally committed as revision 3374 to svn://svn.ffmpeg.org/ffmpeg/trunk
Originally committed as revision 3373 to svn://svn.ffmpeg.org/ffmpeg/trunk
improved magnitude coding, 0.2% lower bitrate (foreman@352x288 qscale 1 and 8)
Originally committed as revision 3372 to svn://svn.ffmpeg.org/ffmpeg/trunk
Originally committed as revision 3367 to svn://svn.ffmpeg.org/ffmpeg/trunk
various subband encoders (all either worse or complicated so they are commented out)different 0-context encoding, 1.2% lower bitrate (foreman@352x288 qscale=8) no significant change at qscale=1
Originally committed as revision 3366 to svn://svn.ffmpeg.org/ffmpeg/trunk
1/0 fix
Originally committed as revision 3357 to svn://svn.ffmpeg.org/ffmpeg/trunk
dont segfault without -strict -1
Originally committed as revision 3356 to svn://svn.ffmpeg.org/ffmpeg/trunk
optimize quantizaton (about 3x faster)further opt is easily possible but could lead to overflows depening upon coefficient range, so this wont be done yet as it would make the code somewhat less flexible
Originally committed as revision 3354 to svn://svn.ffmpeg.org/ffmpeg/trunk
cleanupremove ugly and slow vertical scaning code0.1% bitrate increase (foreman@352x288 qp=1)0.3% bitrate increase (foreman@352x288 qp=8)
Originally committed as revision 3353 to svn://svn.ffmpeg.org/ffmpeg/trunk
use parent sample to predict significance & magnitude1% bitrate reduction (foreman@352x288 qscale 1 & 8)
Originally committed as revision 3352 to svn://svn.ffmpeg.org/ffmpeg/trunk