Jingning Han [Mon, 23 Nov 2015 20:05:48 +0000 (12:05 -0800)]
Analyze motion field to produce reference motion vectors
This commit allows the codec to analyze the motion field in the
avaiable above and left neighboring area to produce a set of
reference motion vectors for each reference frame. These reference
motion vectors are ranked according to the likelihood that it will
be picked.
Reduces the transform optons for INTRA as well as INTER when
transform size is 16x16 to not use any of the DSTs.
Thus, a total of 10 options are used for 16x16, while 4x4
and 8x8 still uses 17 options.
derflr/hevchd actually improves a little, while hevcmr drops
a little.
Jingning Han [Mon, 23 Nov 2015 17:38:45 +0000 (09:38 -0800)]
Move n8_w and n8_h out of experiment flag
These primitive variables are commonly required by many other
experiments as well. The use of n4_w and n4_h was originally
introduced in the vp9 decoder implementation.
Marco [Fri, 20 Nov 2015 16:44:19 +0000 (08:44 -0800)]
vp9-svc: Fix the setting of is_key_frame.
Change on affects 1 pass CBR.
On key frame, temporal layer_id is reset to 0 for 1 pass CBR,
but since "layer" is reset, the svc.layer_context[layer].is_key_frame
was not correspondingly set properly.
paulwilkins [Tue, 17 Nov 2015 16:20:20 +0000 (16:20 +0000)]
Changes to best quality settings.
Small changes to the best quality default speed trade off.
Some speedup settings are worth while even for best quality as they
have only a very small impact on quality but a significant impact on
encode time.
These changes give as much as a further 50-60% increase in encode
speed for my test animations clip with minimal impact on quality.
For this sequence these changes improve the best quality encode speed
to about the same level as good quality speed 0 in Q3 2015 whilst
retaining the large quality gain of over 1 db
For many natural videos though the quality difference from good 0
to best is much smaller.
Ralph Giles [Fri, 13 Nov 2015 20:56:34 +0000 (12:56 -0800)]
Use Interlocked calls in win32 once() implementation.
This is simpler than the previous scheme, which tried to allocate
the CRITICAL_SECTION struct in a thread-safe manner before it
could use it to run the wrapped function in a thread-safe manner.
paulwilkins [Wed, 14 Oct 2015 09:38:49 +0000 (10:38 +0100)]
Changes to exhaustive motion search.
This change alters the nature and use of exhaustive motion search.
Firstly any exhaustive search is preceded by a normal step search.
The exhaustive search is only carried out if the distortion resulting
from the step search is above a threshold value.
Secondly the simple +/- 64 exhaustive search is replaced by a
multi stage mesh based search where each stage has a range
and step/interval size. Subsequent stages use the best position from
the previous stage as the center of the search but use a reduced range
and interval size.
For example:
stage 1: Range +/- 64 interval 4
stage 2: Range +/- 32 interval 2
stage 3: Range +/- 15 interval 1
This process, especially when it follows on from a normal step
search, has shown itself to be almost as effective as a full range
exhaustive search with step 1 but greatly lowers the computational
complexity such that it can be used in some cases for speeds 0-2.
This patch also removes a double exhaustive search for sub 8x8 blocks
which also contained a bug (the two searches used different distortion
metrics).
For best quality in my test animation sequence this patch has almost
no impact on quality but improves encode speed by more than 5X.
Restricted use in good quality speeds 0-2 yields significant quality gains
on the animation test of 0.2 - 0.5 db with only a small impact on encode
speed. On most clips though the quality gain and speed impact are small.
Jingning Han [Fri, 30 Oct 2015 04:33:03 +0000 (21:33 -0700)]
Alternate reference frame
This commit re-designs the alternate reference frame generation
process. It employs non-local mean approach to produce more stable
pixel estimation for alternate reference frame. It improves the
compression performance gains:
derf 0.5%
hevcmr 0.8%
stdhd 1.3%
hevchr 1.0%
Geza Lore [Wed, 28 Oct 2015 14:35:04 +0000 (14:35 +0000)]
Add AVX vectorized vp9_diamond_search_sad
This function now has an AVX intrinsics version which is about 80%
faster compared to the C implementation. This provides a 2-4% total
speed-up for encode, depending on encoding parameters. The function
utilizes 3 properties of the cost function lookup table, constructed
in 'cal_nmvjointsadcost' and 'cal_nmvsadcosts'.
For the joint cost:
- mvjointsadcost[1] == mvjointsadcost[2] == mvjointsadcost[3]
For the component costs:
- For all i: mvsadcost[0][i] == mvsadcost[1][i]
(equal per component cost)
- For all i: mvsadcost[0][i] == mvsadcost[0][-i]
(Cost function is even)
These must hold, otherwise the AVX version of the function cannot be used.
James Zern [Wed, 11 Nov 2015 01:54:51 +0000 (17:54 -0800)]
libs.mk, testdata: rm redundant test of LIBVPX_TEST_DATA
the return value of enabled, which may be empty, is handled by the for
loop. this avoids making an unnecessarily long command line which may
fail in certain cases.
Jingning Han [Wed, 11 Nov 2015 00:02:33 +0000 (16:02 -0800)]
Fix an encoding failure case when speed features are on
This commit fixes an encoding failure case triggered when early
termination feature is turned on for transform block size search.
It resolves the corresponding enc/dec mismatch issue.
projects / libvpx / log