Commit Graph

13193 Commits

Author SHA1 Message Date
James Zern
a98f6c0254 vp9/neon: skip some files in high-bitdepth build
exclude files that only contain functions for non-high-bitdepth builds.
this removes some warnings related to missing prototypes

Change-Id: Ic6642998c46a7b808c6c53b2f9c34bcd4d037abe
2015-03-31 18:06:21 -07:00
James Zern
8845334097 vp9: fix high-bitdepth NEON build
remove incorrect specializations in rtcd and update a configuration
check in partial_idct_test.cc

Change-Id: I20f551f38ce502092b476fb16d3ca0969dba56f0
2015-03-31 17:45:25 -07:00
Yunqing Wang
fc98114761 Merge "Rename vbp thresholds" 2015-03-31 16:33:30 -07:00
Vignesh Venkatasubramanian
639955f66e Merge "webmdec: Fix read_frame return value for calls after EOS" 2015-03-31 16:11:56 -07:00
Marco
c2b8218eba Merge "Set postproc flags in decoder_get_frame." 2015-03-31 15:22:14 -07:00
Yunqing Wang
c28ff1a9de Rename vbp thresholds
Code refactoring

Change-Id: I410fcce1bc6d95c62c474445f4c97ea8469f1e79
2015-03-31 15:14:44 -07:00
Jingning Han
502ac72233 Merge "Tuning SATD rate calculation for speed" 2015-03-31 14:24:26 -07:00
Jingning Han
1c39c5b96f Merge "Use aligned copy in 8x8 Hadamard transform SSE2" 2015-03-31 12:16:47 -07:00
Jingning Han
fa4289522e Merge "Allow block skip coding option in RTC mode" 2015-03-31 12:16:36 -07:00
Jingning Han
1638d7dc96 Merge "Fix 8x8 Hadamard SSE2 implementation" 2015-03-31 12:16:27 -07:00
Alex Converse
9670d766ab Merge "VP9E_GET_ACTIVE_MAP API function." 2015-03-31 11:52:56 -07:00
Jingning Han
531468a07a Tuning SATD rate calculation for speed
This commit allows the encoder to check the eob per transform
block to decide how to compute the SATD rate cost. If the entire
block is quantized to zero, there is no need to add anything; if
only the DC coefficient is non-zero, add its absolute value;
otherwise, sum over the block. This reduces the CPU cycles spent
on vp9_satd_sse2 to one third.

Change-Id: I0d56044b793b286efc0875fafc0b8bf2d2047e32
2015-03-31 11:02:20 -07:00
hui su
d4f2f1dd5b Merge "Move vp9_coef_con_tree to common/" 2015-03-31 10:51:10 -07:00
Jingning Han
014fa45298 Use aligned copy in 8x8 Hadamard transform SSE2
This reduces the 8x8 Hadamard transform cycles by 20%.

Change-Id: If34c5e02f3afa42244c6efabe121f7cf5d2df41b
2015-03-31 10:21:52 -07:00
Jingning Han
db5ec37edc Merge "Enable 16x16 Hadamard transform in SATD based mode decision" 2015-03-31 09:55:41 -07:00
Jingning Han
8c5670bb6f Merge "Use SATD based mode decision for block sizes below 16x16" 2015-03-31 09:47:47 -07:00
Jingning Han
ebe1be9186 Allow block skip coding option in RTC mode
When the estimated rate-distortion cost of skip coding mode is
lower than that of sending quantized coefficients, allow the
encoder to drop these coefficients. This improves the compression
performance of speed -6 by 0.268% and makes the encoding speed
slightly faster.

Change-Id: Idff2d7ba59f27ead33dd5a0e9f68746ed3c2ab68
2015-03-31 09:32:53 -07:00
hui su
302e24cb3e Move vp9_coef_con_tree to common/
This tree should be defined in common/, as it is needed for
both encoder and decoder.

Change-Id: I4f5cbc80025cf2ced14182c98f7c82dc7d0f87db
2015-03-31 09:20:46 -07:00
Marco
385ca8f741 Set postproc flags in decoder_get_frame.
The postproc settings were not set in decoder_get_frame().

Change-Id: I20d23de3ea18f6df061a53d691d4095d5c62532a
2015-03-30 16:15:57 -07:00
Jingning Han
9b99eb2e12 Merge "Reuse inter prediction pixel block for Hadamard transform" 2015-03-30 16:09:38 -07:00
Jingning Han
34a996ac1e Fix 8x8 Hadamard SSE2 implementation
This commit fixes the SSE2 version 8x8 Hadamard transform
alignment and makes it consistent with the C version.

Change-Id: I1304e5f97e0e5ef2d798fe38081609c39f5bfe74
2015-03-30 15:54:08 -07:00
Jingning Han
26d3d3af6a Enable 16x16 Hadamard transform in SATD based mode decision
This commit replaces the 16x16 2D-DCT transform with Hadamard
transform for RTC coding mode. It reduces the CPU cycles cost
on 16x16 transform by 5X. Overall it makes the speed -6 encoding
speed 1.5% faster without compromise on compression performance.

Change-Id: If6c993831dc4c678d841edc804ff395ed37f2a1b
2015-03-30 15:43:31 -07:00
Jingning Han
f0ac5aaa08 Merge "Hadamard transform based coding mode decision process" 2015-03-30 15:43:15 -07:00
Jingning Han
b4b5af6acd Use SATD based mode decision for block sizes below 16x16
This commit makes the encoder to select between SATD/variance as
metric for mode decision. It also allows to account chroma
component costs for mode decision as well. The overall encoding
time increase as compared to variance based mode selection is about
15% for speed -6. The compression performance is on average 2.2%
better than variance based approach, with about 5% compression
performance gains for hard clips (e.g., jimredvga, nikas720p, and
mmmoving) at lower bit-rate range.

Change-Id: I4d04a31d36f4fcb3f5f491dacd6e7fe44cb9d815
2015-03-30 15:20:07 -07:00
Jingning Han
8a927a1b7a Reuse inter prediction pixel block for Hadamard transform
It saves one unnecessary motion compensated prediction constructed
by using 8-tap filter.

Change-Id: I101215131e6f38621d5935885f94cc74de6a5377
2015-03-30 15:04:33 -07:00
Jingning Han
8c411f74e0 Hadamard transform based coding mode decision process
This commit uses Hadamard transform based rate-distortion cost
estimate for rtc coding mode decision. It improves the compression
performance of speed -6 for many hard clips at lower bit-rates.
For example, 5.5% for jimredvga, 6.7% for mmmoving, 6.1% for
niklas720p. This will introduce extra encoding cycle costs at
this point.

Change-Id: Iaf70634fa2417a705ee29f2456175b981db3d375
2015-03-30 14:46:05 -07:00
Vignesh Venkatasubramanian
1f05b19e69 webmdec: Fix read_frame return value for calls after EOS
webm_read_frame assumes that it won't be called once end of file
is reached. But for frame parallel mode that turns out to be not
true. this patch fixes that behavior by checking for EOS and
returning the appropriate value for subsequent calls.

Change-Id: Ie2fddbe00493a0f96c4172c67be1eb719f0fe8ed
2015-03-30 12:58:26 -07:00
Alex Converse
bf7def9a43 Merge "Simplify skip check." 2015-03-30 11:31:45 -07:00
jackychen
b38b32a794 Merge "vp9_postproc.c: eliminate -Wshadow build warnings." 2015-03-30 10:29:39 -07:00
jackychen
68610ae568 vp9_postproc.c: eliminate -Wshadow build warnings.
Change-Id: I6df525a9ad1ae3cfbba8710d21db8fee76e64dbb
2015-03-27 20:27:30 -07:00
Marco
fa20a60f0d Speed 5: use non-rd mode for key frame coding.
Metrics on RTC set go down by ~1.5% on average.
Key frame encoding time goes down by factor of ~5.

Change-Id: Ia83acc55848613870e5ac6efe7f3d904d877febb
2015-03-27 16:19:26 -07:00
hkuang
0c85718954 Merge "Fix the issue that --limit is not working in --frame-parallel mode." 2015-03-27 10:12:45 -07:00
Adrian Grange
553792cee2 Merge "Remove 8-bit array in HBD" 2015-03-26 16:31:27 -07:00
Adrian Grange
300d428ecd Merge "Replace heap with stack memory allocation" 2015-03-26 16:31:06 -07:00
Adrian Grange
9931110971 Merge "Fix use of scaling in joint motion search" 2015-03-26 16:30:35 -07:00
hkuang
ffafcd6281 Fix the issue that --limit is not working in --frame-parallel mode.
The reason is due to early break out before outputting all the frames inside
decoder.

Change-Id: I4a138fba08d12935c39bd7602c95f8c18b474e29
2015-03-26 15:36:22 -07:00
Johann
46ce6954cc Remove duplicate code from merge
Change-Id: I5e2a1270001b7e29f3f198d57ea40e1efccef367
2015-03-26 14:56:24 -07:00
Adrian Grange
ad18b2b641 Remove 8-bit array in HBD
Creating both 8- and 16-bit arrays and then only using one
of them is wasteful.

Change-Id: Ic5b397c283efaff7bcfff2d2413838ba3e065561
2015-03-25 15:37:03 -07:00
Adrian Grange
65df3d138a Replace heap with stack memory allocation
Replaced the dynamic memory allocation of the
second_pred buffer with an allocation on the stack.

Change-Id: I2716c46b71e8587714ca5733a99eca2c68419b23
2015-03-25 15:36:43 -07:00
Adrian Grange
8d8d7bfde5 Fix use of scaling in joint motion search
To enable us to the scale-invariant motion estimation
code during mode selection, each of the reference
buffers is scaled to match the size of the frame
being encoded.

This fix ensures that a unit scaling factor is used in
this case rather than the one calculated assuming that
the reference frame is not scaled.

Change-Id: Id9a5c85dad402f3a7cc7ea9f30f204edad080ebf
2015-03-25 15:35:29 -07:00
Johann
ba13ff8501 Parall -> Parallel
Change-Id: I565fef382fa17a00d5ae54e980ef14d9f0ad4f55
2015-03-25 12:45:36 -07:00
James Zern
e865be95bf Merge "fix static analysis warnings related to CHECK_MEM_ERROR" 2015-03-24 23:56:04 -07:00
Parag Salasakar
84ec68d21a mips msa configuration patch for MIPS SIMD Arch (MSA) P5600 and I6400
For P5600:
CROSS=$MTI/bin/mips-mti-linux-gnu- CFLAGS='-EL' CXXFLAGS='-EL' LDFLAGS='-EL'\
 ../configure --target=mips32-linux-gcc --cpu=p5600 --enable-msa

For I6400:
CROSS=$IMG/bin/mips-img-linux-gnu- CFLAGS='-EL' CXXFLAGS='-EL' LDFLAGS='-EL'\
 ../configure --target=mips64-linux-gcc --cpu=i6400 --enable-msa

Change-Id: Id25f721ea1f1991d5116e04dba713aebd7378f05
2015-03-24 15:18:38 -07:00
paulwilkins
ab788c5380 Merge "Enable group adaptive max q by default." 2015-03-24 15:00:12 -07:00
Alex Converse
4dcb839607 VP9E_GET_ACTIVE_MAP API function.
This is useful when aq mode 3 (cyclic refresh) reactivates segments for refresh.

Change-Id: I3ad1d9410b899ede393d82bb8db14e2da4d84eca
2015-03-24 11:19:47 -07:00
Alex Converse
a1e20ec58f Refactor fast loop filter code to handle 444.
Change-Id: I921b1ebabdf617049f8fa26fbe462c3ff115c1ce
2015-03-24 11:17:50 -07:00
Yaowu Xu
c77d4dcb35 Merge "vp9_pred_mv(): misc fixes and optimizations" 2015-03-24 10:36:51 -07:00
Alex Converse
02697e35dc Merge "A tiny cyclic refresh / active map fix." 2015-03-24 09:43:24 -07:00
paulwilkins
8ea7bafdaa Merge "Revised rd adjustment for variance." 2015-03-24 03:12:56 -07:00
paulwilkins
c0b71cf82f Merge "Experimental rd bias based on source vs recon variance." 2015-03-24 03:12:41 -07:00