Jerome Jiang [Tue, 22 Jun 2021 22:13:38 +0000 (22:13 +0000)]
Merge "rc: turn off gf constrain for external RC" into main
Jerome Jiang [Tue, 22 Jun 2021 00:22:51 +0000 (17:22 -0700)]
rc: turn off gf constrain for external RC
Added a new flag in rate control which turns off gf interval constrain
on key frame frequency for external RC.
It remains on for libvpx.
Change-Id: I18bb0d8247a421193f023619f906d0362b873b31
James Zern [Tue, 22 Jun 2021 03:02:58 +0000 (03:02 +0000)]
Merge "test-data.sha1: add missing sha sums" into main
Angie Chiang [Tue, 22 Jun 2021 01:44:02 +0000 (01:44 +0000)]
Merge changes I9f0852a0,Ieecb98a7 into main
* changes:
Add use_simple_encode_api to oxcf
Fix flaky assertions in SimpleEncode
Angie Chiang [Fri, 18 Jun 2021 03:23:30 +0000 (20:23 -0700)]
Add use_simple_encode_api to oxcf
Use this flag to change the encoder behavior when
SimpleEncode APIs are used
BUG=webm:1733
Change-Id: I9f0852a03ff99faa01cdd8eee8ab71718cc58632
Angie Chiang [Fri, 18 Jun 2021 23:09:41 +0000 (16:09 -0700)]
Fix flaky assertions in SimpleEncode
Bug: webm:1731
Change-Id: Ieecb98a7ac19e6291acd5d51432dc6a3789e9552
James Zern [Mon, 21 Jun 2021 20:33:44 +0000 (13:33 -0700)]
test-data.sha1: add missing sha sums
for rc_interface_test_one_layer_vbr and
rc_interface_test_one_layer_vbr_periodic_key added in:
1f45e7b07 vp9 rc: add vbr to rtc rate control library
Change-Id: I8bfa3698284c8ff289e830f7b8fa1ca42b752563
Jerome Jiang [Fri, 18 Jun 2021 23:25:53 +0000 (23:25 +0000)]
Merge "vp9 rc: add vbr to rtc rate control library" into main
Jerome Jiang [Tue, 15 Jun 2021 19:54:13 +0000 (12:54 -0700)]
vp9 rc: add vbr to rtc rate control library
Change-Id: I3d2565572c2b905966d60bcaa6e5e6f057b1bd51
James Zern [Fri, 18 Jun 2021 18:56:27 +0000 (11:56 -0700)]
normalize vp9_calc_[ip]frame declarations and definitions
fixes warnings under visual studio:
vp9\encoder\vp9_ratectrl.c(2012): warning C4028: formal parameter 1
different from declaration
vp9\encoder\vp9_ratectrl.c(2027): warning C4028: formal parameter 1
different from declaration
Change-Id: Ia0740db597fb7a259f90d362b483f58662f9f584
Marco Paniconi [Thu, 17 Jun 2021 19:00:33 +0000 (12:00 -0700)]
vp9: Adjust logic for gf update in 1 pass vbr
This reduces some regression when external RC
is used, for which avg_frame_low_motion is not
set/updated (=0).
Change-Id: I2408e62bd97592e892cefa0f183357c641aa5eea
Chunbo Hua [Wed, 16 Jun 2021 08:51:44 +0000 (01:51 -0700)]
Initialize VP9EncoderConfig profile and bit depth
Change-Id: I5c42013a08677cdef8d47f348458118338ff0138
Jerome Jiang [Tue, 15 Jun 2021 21:55:29 +0000 (14:55 -0700)]
Change the data path in svc rate control test
Change-Id: Iba58e2aa2578964b5c8b48ab0acbee9b44bcdada
Marco Paniconi [Mon, 14 Jun 2021 22:02:52 +0000 (15:02 -0700)]
vp9-rtc: Refactor 1 pass vbr rate control
This refactoring is needed to allow the
RC_rtc library to support VBR.
Change-Id: I863a4a65096fed06b02307098febf7976360e0f3
James Zern [Fri, 11 Jun 2021 23:34:41 +0000 (16:34 -0700)]
Update some comments for rc_target_bitrate
this mirrors the change from libaom:
5b150b150 Update some comments for rc_target_bitrate
Change-Id: Iaabee5924e0320609a29dc8ab71327923fb4c5d2
James Zern [Wed, 9 Jun 2021 22:07:15 +0000 (15:07 -0700)]
simple_encode: fix some -Wsign-compare warnings
Bug: webm:1731
Change-Id: I1db777c0c3a8784fb3dcf7cd39f78ebf833ab915
James Zern [Sun, 6 Jun 2021 02:30:04 +0000 (19:30 -0700)]
simple_encode_test: fix input file path
this allows the file to be located in LIBVPX_TEST_DATA_PATH similar to
other test sources.
Bug: webm:1731
Change-Id: I51606635d91871e7c179aa8d20d4841b0d60b6ad
Cheng Chen [Thu, 27 May 2021 22:38:28 +0000 (15:38 -0700)]
L2E: properly init two pass rc parameters
Two pass rc parameters are only initialized in the second pass
in vp9 normal two pass encoding.
However, the simple_encode API queries the keyframe group, arf group,
and number of coding frames without going throught the two pass
route.
Since recent libvpx rc changes, parameters in the TWO_PASS
struct have a great influence on the determination of the above
information.
We therefore need to properly init two pass rc parameters in
the simple_encode related environment.
Change-Id: Ie14b86d6e7ebf171b638d2da24a7fdcf5a15c3d9
Cheng Chen [Mon, 24 May 2021 22:53:06 +0000 (15:53 -0700)]
Fix simple encode
Properly init and delete cpi struct in simple encode functions.
Change-Id: I6e66bcac852cbb3dec9b754ba3fb01a348ac98b8
Chunbo Hua [Wed, 26 May 2021 09:02:07 +0000 (02:02 -0700)]
Fixed redundant wording for decoder algorithm interface
Change-Id: Id56e03dc9cf6d4e70c4681896f29893a9b4c76f2
James Zern [Tue, 25 May 2021 03:00:47 +0000 (03:00 +0000)]
Merge changes I2e86b005,I971c6261,I87fe4dad
* changes:
Use 'ptrdiff_t' instead of 'int' for pointer offset parameters
Implement vpx_convolve8_avg_vert_neon using SDOT instruction
Merge transpose and permute in Neon SDOT vertical convolution
James Zern [Tue, 25 May 2021 02:37:05 +0000 (02:37 +0000)]
Merge "img_alloc_helper: make align var unsigned"
Jonathan Wright [Mon, 24 May 2021 10:42:09 +0000 (11:42 +0100)]
Use 'ptrdiff_t' instead of 'int' for pointer offset parameters
A number of the load/store functions in mem_neon.h use type 'int' for
the 'stride' pointer offset parameter. This causes Clang to generate
the following warning every time these functions are called with a
wider type passed in for 'stride':
warning: implicit conversion loses integer precision: 'ptrdiff_t'
(aka 'long') to 'int' [-Wshorten-64-to-32]
This patch changes all such instances of 'int' to 'ptrdiff_t'.
Bug: b/
181236880
Change-Id: I2e86b005219e1fbb54f7cf2465e918b7c077f7ee
Jonathan Wright [Sun, 23 May 2021 12:35:15 +0000 (13:35 +0100)]
Implement vpx_convolve8_avg_vert_neon using SDOT instruction
Add an alternative AArch64 implementation of
vpx_convolve8_avg_vert_neon for targets that implement the Armv8.4-A
SDOT (signed dot product) instruction.
The existing MLA-based implementation of vpx_convolve8_avg_vert_neon
is retained and used on target CPUs that do not implement the SDOT
instruction (or CPUs executing in AArch32 mode). The availability of
the SDOT instruction is indicated by the feature macro
__ARM_FEATURE_DOTPROD.
Bug: b/
181236880
Change-Id: I971c626116155e1384bff4c76fd3420312c7a15b
Jonathan Wright [Sat, 22 May 2021 21:07:25 +0000 (22:07 +0100)]
Merge transpose and permute in Neon SDOT vertical convolution
The original dot-product implementation of vpx_convolve8_vert_neon
used a separate transpose before and after the convolution operation.
This patch merges the first transpose with the TBL permute (necessary
before using SDOT to compute the convolution) to significantly reduce
the amount of data re-arrangement. This new approach also allows for
more effective data re-use between loop iterations.
Co-authored by: James Greenhalgh <james.greenhalgh@arm.com>
Bug: b/
181236880
Change-Id: I87fe4dadd312c3ad6216943b71a5410ddf4a1b5b
Jonathan Wright [Mon, 17 May 2021 09:53:07 +0000 (10:53 +0100)]
Implement vpx_convolve8_avg_horiz_neon using SDOT instruction
Add an alternative AArch64 implementation of
vpx_convolve8_avg_horiz_neon for targets that implement the Armv8.4-A
SDOT (signed dot product) instruction.
The existing MLA-based implementation of vpx_convolve8_avg_horiz_neon
is retained and used on target CPUs that do not implement the SDOT
instruction (or CPUs executing in AArch32 mode). The availability of
the SDOT instruction is indicated by the feature macro
__ARM_FEATURE_DOTPROD.
Bug: b/
181236880
Change-Id: Ib435107c47c485f325248da87ba5618d68b0c8ed
Jonathan Wright [Wed, 12 May 2021 15:05:56 +0000 (16:05 +0100)]
Optimize remaining mse and sse functions in variance_neon.c
Implement sum of squared difference calculations in vpx_mse16x16_neon
and vpx_get4x4sse_cs_neon using the ABD and UDOT instructions -
instead of widening subtracts followed by a sequence of MLAs.
The existing implementation is retained for use on CPUs that do not
implement the Armv8.4-A UDOT instruction. This commit also updates
the variable names used in the existing implementations to be more
descriptive.
Bug: b/
181236880
Change-Id: Id4ad8ea7c808af1ac9bb5f1b63327ab487e4b1c7
Jonathan Wright [Tue, 20 Apr 2021 11:03:56 +0000 (12:03 +0100)]
Implement vertical convolution using Neon SDOT instruction
Add an alternative AArch64 implementation of vpx_convolve8_vert_neon
for targets that implement the Armv8.4-A SDOT (signed dot product)
instruction.
The existing MLA-based implementation of vpx_convolve8_vert_neon is
retained and used on target CPUs that do not implement the SDOT
instruction (or CPUs executing in AArch32 mode). The availability of
the SDOT instruction is indicated by the feature macro
__ARM_FEATURE_DOTPROD.
Bug: b/
181236880
Change-Id: Iebb8c77aba1d45b553b5112f3d87071fef3076f0
Jonathan Wright [Tue, 11 May 2021 12:17:44 +0000 (13:17 +0100)]
Implement Neon variance functions using UDOT instruction
Accelerate Neon variance functions by implementing the sum of squares
calculation using the Armv8.4-A UDOT instruction instead of 4 MLAs.
The previous implementation is retained for use on CPUs that do not
implement the Armv8.4-A dot product instructions.
Bug: b/
181236880
Change-Id: I9ab3d52634278b9b6f0011f39390a1195210bc75
Jonathan Wright [Mon, 10 May 2021 11:22:03 +0000 (12:22 +0100)]
Use ABD and UDOT to implement Neon sad_4d functions
Implementing sad16_neon using ABD, UDOT instead of ABAL, ABAL2 saves
a cycle and removes resource contention for a single SIMD pipe on
modern out-of-order Arm CPUs. The UDOT accumulation into 32-bit
elements also allows for a faster reduction at the end of each SAD
function.
The existing implementation is retained for CPUs that do not
implement the Armv8.4-A UDOT instruction, and CPUs executing in
AArch32 mode.
Bug: b/
181236880
Change-Id: Ibd0da46e86751d2f808c7b1e424f82b046a1aa6f
Jonathan Wright [Fri, 7 May 2021 12:25:51 +0000 (13:25 +0100)]
Optimize Neon reductions in sum_neon.h using ADDV instruction
Use the AArch64-only ADDV and ADDLV instructions to accelerate
reductions that add across a Neon vector in sum_neon.h. This commit
also refactors the inline functions to return a scalar instead of a
vector - allowing for optimization of the surrounding code at each
call site.
Bug: b/
181236880
Change-Id: Ieed2a2dd3c74f8a52957bf404141ffc044bd5d79
James Zern [Sat, 8 May 2021 02:35:25 +0000 (19:35 -0700)]
img_alloc_helper: make align var unsigned
quiets an integer sanitizer warning:
vpx/src/vpx_image.c:101:25: runtime error: implicit conversion from
type 'int' of value -2 (32-bit, signed) to type 'unsigned int' changed
the value to
4294967294 (32-bit, unsigned)
Change-Id: Ifeac31cc80811081c1ba10aadaa94dc36cd46efa
Jonathan Wright [Thu, 6 May 2021 14:11:52 +0000 (15:11 +0100)]
Manually unroll the inner loop of Neon sad16x_4d()
Manually unrolling the inner loop is sufficient to stop the compiler
getting confused and emitting inefficient code.
Co-authored by: James Greenhalgh <james.greenhalgh@arm.com>
Bug: b/
181236880
Change-Id: I860768ce0e6c0e0b6286d3fc1b94f0eae95d0a1a
Jonathan Wright [Thu, 6 May 2021 13:51:05 +0000 (14:51 +0100)]
Optimize Neon SAD reductions using wider ADDP instruction
Implement AArch64-only paths for each of the Neon SAD reduction
functions, making use of a wider pairwise addition instruction only
available on AArch64.
This change removes the need for shuffling between high and low
halves of Neon vectors - resulting in a faster reduction that requires
fewer instructions.
Bug: b/
181236880
Change-Id: I1c48580b4aec27222538eeab44e38ecc1f2009dc
James Zern [Wed, 5 May 2021 19:57:10 +0000 (19:57 +0000)]
Merge "Implement horizontal convolution using Neon SDOT instruction"
Jonathan Wright [Sun, 11 Apr 2021 14:20:36 +0000 (15:20 +0100)]
Implement horizontal convolution using Neon SDOT instruction
Add an alternative AArch64 implementation of vpx_convolve8_horiz_neon
for targets that implement the Armv8.4-A SDOT (signed dot product)
instruction.
The existing MLA-based implementation of vpx_convolve8_horiz_neon is
retained and used on target CPUs that do not implement the SDOT
instruction (or CPUs executing in AArch32 mode). The availability of
the SDOT instruction is indicated by the feature macro
__ARM_FEATURE_DOTPROD.
Co-authored by: James Greenhalgh <james.greenhalgh@arm.com>
Change-Id: I5337286b0f5f2775ad7cdbc0174785ae694363cc
James Zern [Tue, 4 May 2021 19:13:17 +0000 (12:13 -0700)]
vp9_denoiser_neon,horizontal_add_s8x16: use vaddlv w/aarch64
this reduces the number of instructions to compute the sum
Change-Id: Icae4d4fb3e343d5b6e5a095c60ac6d171b3e7d54
James Zern [Tue, 4 May 2021 19:10:21 +0000 (12:10 -0700)]
test.mk: enable vp9_denoiser_test w/NEON
this file uses GTEST_ALLOW_UNINSTANTIATED_PARAMETERIZED_TEST so it's
safe to enable unconditionally. the filter check fell out of sync with
the code, there's a sse2 and neon implementation for the filter.
Change-Id: I2a3336ccef3fb524ca5d9b8f88279240c9a276aa
Paul Wilkins [Thu, 29 Apr 2021 16:31:52 +0000 (16:31 +0000)]
Merge "Add assert for zero_motion_factor range"
Paul Wilkins [Thu, 29 Apr 2021 10:06:16 +0000 (11:06 +0100)]
Add assert for zero_motion_factor range
Change clamp to an assert so we are warned if changes to input
ranges or defaults in the future lead to an invalid value.
Change-Id: Idb4e0729f477a519bfff3083cdce3891e2fc6faa
Cheng Chen [Wed, 28 Apr 2021 20:54:07 +0000 (13:54 -0700)]
Bump ABI version
Due to recent changes to command line options for rate control
parameters.
Change-Id: I1de7cb4ff2850a3ed19ec216dd9d07f64a118e92
James Zern [Wed, 28 Apr 2021 17:30:48 +0000 (17:30 +0000)]
Merge changes Iebe9842f,I174b67a5,I80ed1a16
* changes:
vpx_convolve_neon: prefer != 0 to > 0 in tests
vpx_convolve_avg_neon: prefer != 0 to > 0 in tests
vpx_convolve_copy_neon: prefer != 0 to > 0 in tests
James Zern [Wed, 28 Apr 2021 02:27:54 +0000 (02:27 +0000)]
Merge "vp8: enc: Fix valid range for under/over_shoot pct"
James Zern [Wed, 28 Apr 2021 01:02:35 +0000 (18:02 -0700)]
vpx_convolve_neon: prefer != 0 to > 0 in tests
this produces better assembly code; the horizontal convolve is called
with an adjusted intermediate_height where it may over process some rows
so the checks in those functions remain.
Change-Id: Iebe9842f2a13a4960d9a5addde9489452f5ce33a
James Zern [Wed, 28 Apr 2021 01:02:35 +0000 (18:02 -0700)]
vpx_convolve_avg_neon: prefer != 0 to > 0 in tests
this produces better assembly code
Change-Id: I174b67a595d7efeb60c921f066302043b1c7d84e
James Zern [Wed, 28 Apr 2021 01:02:35 +0000 (18:02 -0700)]
vpx_convolve_copy_neon: prefer != 0 to > 0 in tests
this produces better assembly code
Change-Id: I80ed1a165512e941b35a4965faa0c44403357e91
Paul Wilkins [Mon, 26 Apr 2021 14:06:54 +0000 (15:06 +0100)]
Add limits to Vizier input parameters.
Imposed provisional upper and lower limits to each parameter
that can be adjusted in the Vizier ML experiment.
Also in some cases applied secondary limits on on the
range of the final "used" values.
Defaults and limits may well require further tuning after
subsequent rounds of experimentation.
Re-factor get_sr_decay_rate().
Change-Id: I28e804ce3d3710f30cd51a203348e4ab23ef06c0
James Zern [Fri, 23 Apr 2021 23:50:48 +0000 (16:50 -0700)]
sync CONTRIBUTING.md w/libwebm
Change-Id: I63ffea52d079b0d50002526e209ae3fb64811bac
Sreerenj Balachandran [Wed, 21 Apr 2021 18:34:03 +0000 (11:34 -0700)]
vp8: enc: Fix valid range for under/over_shoot pct
The overshoot_pct & undershoot_pct attributes for rate control
are expressed as a percentage of the target bitrate, so the range
should be 0-100.
Change-Id: I67af3c8be7ab814c711c2eaf30786f1e2fa4f5a3
Paul Wilkins [Tue, 20 Apr 2021 16:26:22 +0000 (17:26 +0100)]
Further normalization of Vizier parameters.
Further changes to normalize the Vizier command line parameters.
The intent is that the default behavior for any given parameter
is signaled by the value 1.0 (expressed on the command line as a
rational).
The final values used in the two pass code are obtained by multiplying
the passed in factors by a default values if use_vizier_rc_params is 1.
Where use_vizier_rc_params is 0 the values are explicitly set to
the defaults.
This patch also changes the default value of each parameter to 1.0
even if not set explicitly. This should ensure safe /default behavior
if the user sets use_vizier_rc_params to 1 but does not set all the
the individual parameters.
Change-Id: Ied08b3c22df18f42f446a4cc9363473cad097f69
Cheng Chen [Thu, 15 Apr 2021 05:15:54 +0000 (22:15 -0700)]
Pass vizier rd parameter values
Add command line options for three rd parameters.
They are controlled by --use_vizier_rc_params, together with
other rc parameters.
If not set from command line, current default values will be used.
Change-Id: Ie1b9a98a50326551cc1d5940c4b637cb01a61aa0
Cheng Chen [Wed, 14 Apr 2021 16:31:40 +0000 (16:31 +0000)]
Merge "Set vizier rc parameters"
Cheng Chen [Tue, 13 Apr 2021 18:59:40 +0000 (11:59 -0700)]
Set vizier rc parameters
If pass --use-vizier-rc-params=1, the rc parameters are overwittern
by pass in values. It --use-vizier-rc-params=0, the rc parameters
remain the default values.
Change-Id: I7a3e806e0918f49e8970997379a6e99af6bb7cac
Paul Wilkins [Tue, 13 Apr 2021 19:04:49 +0000 (19:04 +0000)]
Merge "Removed unused constant"
Paul Wilkins [Mon, 12 Apr 2021 12:51:44 +0000 (13:51 +0100)]
Removed unused constant
Deleted #define that is no longer referenced.
Change-Id: If0b132c5a40dd8910f535fffdee7d2d1c7df4748
James Zern [Fri, 9 Apr 2021 00:34:16 +0000 (17:34 -0700)]
vpx_image: clear user provided vpx_image_t early
this avoids uninitialized values and potential misuse of them which
could lead to a crash should the function fail
this is the same fix that was applied in libaom:
d0cac70b5 Fix a free on invalid ptr when img allocation fails
Bug: webm:1722
Change-Id: If7a8d08c4b010f12e2e1d848613c0fa7328f1f9c
Cheng Chen [Wed, 7 Apr 2021 19:03:53 +0000 (19:03 +0000)]
Merge "Fix compilation for CONFIG_RATE_CTRL"
Paul Wilkins [Wed, 7 Apr 2021 12:53:07 +0000 (12:53 +0000)]
Merge "Delete unused constants."
Paul Wilkins [Wed, 7 Apr 2021 12:52:47 +0000 (12:52 +0000)]
Merge "Change zm_factor for Vizier."
Cheng Chen [Fri, 2 Apr 2021 21:27:16 +0000 (14:27 -0700)]
Fix compilation for CONFIG_RATE_CTRL
Recently, some function signatures have been changed.
This change fixes compilation error if --enable-rate-ctrl is used.
Change-Id: Ib8e9cb5e181ba1d4a6969883e377f3dd93e9289a
Cheng Chen [Wed, 7 Apr 2021 00:20:39 +0000 (17:20 -0700)]
Adjust end to end psnr value
A recent change leads to slight difference of encoding results:
d3aaac367 Change calculation of rd multiplier,
which is caught by Jenkins nightly test.
Adjust the threshold to silence the test failure.
BUG=webm:1725
Change-Id: I7e8b3a26b72c831ae4d88d0fca681b354314739d
Paul Wilkins [Tue, 6 Apr 2021 19:05:48 +0000 (20:05 +0100)]
Change zm_factor for Vizier.
Changes the exposed zm_factor parameter.
This patch alters the meaning of the zm_factor
parameter that will be exposed for the Vizier project.
The previous power factor was hard to interpret in terms
of its meaning and effect and has been replaced by a linear factor.
Given that the initial Vizier results suggested a lower zero motion
effect for all formats, the default impact has been reduced.
The patch as it stands gives a modest improvement for PSNR
but is slightly down on some sets for SSIM
(overall psnr, ssim % bdrate change: -ve is better)
lowres -0.111, 0.001
ugc360p -0.282, -0.068
midres2 -0.183, 0.059
hdres2 -0.042, 0.172
Change-Id: Id6566433ceed8470d5fad1f30282daed56de385d
Paul Wilkins [Tue, 6 Apr 2021 19:08:01 +0000 (20:08 +0100)]
Delete unused constants.
Delete some #defines that are no longer needed.
Change-Id: I9e4e4df10716598b0d62b0c70f538d4b78a32296
Cheng Chen [Fri, 2 Apr 2021 17:56:29 +0000 (10:56 -0700)]
Pass vizier rc parameter values with range check
This is similar to the change:
https://chromium-review.googlesource.com/c/webm/libvpx/+/2771081
Which fails libvpx nightly test.
Here we add range check to get rid of the warning of
"divided by zero".
BUG=webm:1723
Change-Id: I7712efe7abd4b11cdb725643d51fd1c0a300d924
Tom Finegan [Fri, 2 Apr 2021 16:40:09 +0000 (09:40 -0700)]
Revert "Pass vizier rc parameter values from command line to twopass"
This reverts commit
f32829a2e5db3cd1624e8a7a530af84c382762ef.
BUG=webm:1723
Change-Id: I866cdf288f9873c350b32091515a6d5f4df362a3
Paul Wilkins [Thu, 1 Apr 2021 13:55:21 +0000 (13:55 +0000)]
Merge "Change calculation of rd multiplier."
James Zern [Wed, 31 Mar 2021 22:49:26 +0000 (22:49 +0000)]
Merge "vp9_ext_ratectrl_test: use uintptr_t for void* value"
Paul Wilkins [Wed, 31 Mar 2021 16:28:30 +0000 (16:28 +0000)]
Merge "Convert Vizier RD parameters to normalized factors"
Paul Wilkins [Wed, 31 Mar 2021 15:58:51 +0000 (16:58 +0100)]
Change calculation of rd multiplier.
Change the way the rd multiplier is adjusted for Q and frame type.
Previously in VP9 the rd multiplier was adjusted based on crude Q bins
and whether the frame was a key frame or inter frame.
The Q bins create some problems as they potentially introduce
discontinuities in the RD curve. For example, rate rising with a
stepwise increase in Q instead of falling. As such, in AV1 they
have been removed.
A further issue was identified when examining the first round of
results from from the Vizier project. Here the multiplier for each Q bin
and each frame type was optimized for a training set, for various video
formats, using single point encodes at the appropriate YT rates.
These initial results appeared to show a trend for increased rd
multiplier at higher Q for key frames. This fits with intuition as in
this encoding context a higher Q indicates that a clip is harder to
encode and frames less well predicted. However, the situation
appeared to reverse for inter frames with higher rd multipliers
chosen at low Q.
My initial suspicion was that this was a result of over fitting, but on
closer analysis I realized that this may be more related to frame type
within the broader inter frame classification. Specifically frames coded
at low Q are predominantly ARF frames, for the mid Q bin there will
likely be a mix of ARF and normal inter frames, and for the high Q bin
the frames will almost exclusively be normal inter frames from difficult
content.
ARF frames are inherently less well predicted than other inter frames
being further apart and not having access to as many prediction modes.
We also know from previous work that ARF frames have a higher
incidence of INTRA coding and may well behave more like key frames
in this context.
This patch replaces the bin based approach with a linear function
that applies a small but smooth Q based adjustment. It also splits
ARF frames and normal inter frames into separate categories.
With this done number of parameters that will be exposed for the
next round of Vizier training is reduced from 7 to 3 (one adjustment
factor each for inter, ARF and key frames)
This patch gives net BDATE gains for our test sets even with the
baseline / default factors as follows: (% BDRATE change in overall
PSNR and SSIM, -ve is better)
LowRes -0.231, -0.050
ugc360p 0.160, -0.315
midres2 -0.348, -1.170
hdres2 -0.407, -0.691
Change-Id: I46dd2fea77b1c2849c122f10fd0df74bbd3fcc7f
Cheng Chen [Wed, 17 Mar 2021 05:42:40 +0000 (22:42 -0700)]
Pass vizier rc parameter values from command line to twopass
Change-Id: I02eabeccf2fe4604875820d38e23c2586a63e290
Cheng Chen [Wed, 17 Mar 2021 04:27:18 +0000 (21:27 -0700)]
Add command line options for a few rc parameters
These rate control parameters are for the Vizier experiment.
They are defined as rational numbers.
Change-Id: I23f382dd49158db463b75b5ad8a82d8e0d536308
James Zern [Fri, 26 Mar 2021 17:55:29 +0000 (10:55 -0700)]
vp9_ext_ratectrl_test: use uintptr_t for void* value
this avoids a warning about differences in size between void* and
unsigned int under msvc:
vp9_ext_ratectrl_test.cc(40,3): warning C4312: 'reinterpret_cast':
conversion from 'const unsigned int' to 'void *' of greater size
Change-Id: I5a412ec785ddcaeff2ec71bb83a6048505400293
Jerome Jiang [Wed, 24 Mar 2021 23:39:19 +0000 (16:39 -0700)]
Merge tag 'v1.10.0'
Release v1.10.0 Ruddy Duck
2021-03-09 v1.10.0 "Ruddy Duck"
This maintenance release adds support for darwin20 and new codec controls, as
well as numerous bug fixes.
- Upgrading:
New codec control is added to disable loopfilter for VP9.
New encoder control is added to disable feature to increase Q on overshoot
detection for CBR.
Configure support for darwin20 is added.
New codec control is added for VP9 rate control. The control ID of this
interface is VP9E_SET_EXTERNAL_RATE_CONTROL. To make VP9 use a customized
external rate control model, users will have to implement each callback
function in vpx_rc_funcs_t and register them using libvpx API
vpx_codec_control_() with the control ID.
- Enhancement:
Use -std=gnu++11 instead of -std=c++11 for c++ files.
- Bug fixes:
Override assembler with --as option of configure for MSVS.
Fix several compilation issues with gcc 4.8.5.
Fix to resetting rate control for temporal layers.
Fix to the rate control stats of SVC example encoder when number of spatial
layers is 1.
Fix to reusing motion vectors from the base spatial layer in SVC.
2 pass related flags removed from SVC example encoder.
Bug: webm:1712
Change-Id: I4d807da7aee5a4d9d7a7af66b927983622e9cefa
Paul Wilkins [Mon, 22 Mar 2021 19:45:10 +0000 (19:45 +0000)]
Convert Vizier RD parameters to normalized factors
This patch converts the Vizier custom RD multipliers, to factors
that adjust each RD multiplier either side of its default value, where
a factor of 1.0 will give the previous default behavior.
Ultimately I would like to replace the multiple RD multipliers
triggered at different Q thresholds (eg, low, medium, high q)
with a function that adjusts the rd behavior smoothly as Q
changes.
Vizier could then be presented with a single adjustment control
for each of key frame and inter frame rd.
The current behavior is problematic.
Firstly having hard threshold Q values at which rd behavior changes
may cause anomalies in the rate distortion curve, where in some
situations, raising Q, for example, may not cause the expected drop
in rate and rise in distortion, because we have crossed a threshold
where the rate distortion multiplier changes sharply and this alters
the balance of bits spent in the prediction and residual parts of the
signal.
Having a single value that is used for a range of Q index values
(eg 0-64), (65-128) may also cause problems and over-fitting in
the context of the Vizier ML project. This project tries to optimize
the values for each Q range, for various YT formats, but does so
by analyzing the results of single point encodes on a set of clips.
For a given format all the clips are encoded with the same parameters
(target rate etc) so there is likely to be clustering in regards to the
Q values used. For example the training set may give a new value
for the Q range 0-64 but most of the data points used may have Q
close 64.
It will likely require several iterations working with the Vizier team
to get this right. This patch just gives an initial framework for
testing.
Change-Id: Iaa4cd5561b95a202bcae7a1d876c4f40ef444fa2
Paul Wilkins [Fri, 19 Mar 2021 19:44:38 +0000 (19:44 +0000)]
Merge "Change SR_diff calculation and representation"
Adam B. Goode [Wed, 17 Mar 2021 19:11:57 +0000 (14:11 -0500)]
Msvc builds convert to windows path w/msys env
Bug: webm:1720
Change-Id: I56689ad408f8086c511e1711dfa9c8d404727b2e
(cherry picked from commit
04086a30664d2a3e89d6a6e4e1c18f1a82c8f958)
Paul Wilkins [Wed, 10 Mar 2021 14:39:43 +0000 (14:39 +0000)]
Change SR_diff calculation and representation
This patch changes the way prediction decay is calculated.
We expect that frames that are further from an ALT-REF frame (or Golden
Frame) will be less well predicted by that ALT-REF frame. As such it is
desirable that they should contribute less to the boost calculation used
to assign bits to the ALT_REF.
This code looks at the reduction in prediction quality between the last
frame and the second reference frame (usually two frames old). We make
the assumption that we can accumulate this to get a proxy for the likely
loss of prediction quality over multiple frames.
Previously the calculation looked at the absolute difference in the
coded errors. The issue here is that the meaning of a unit difference
is not the same for very complex frames as it is for easy frames.
In this patch we scale the decay value based on how the error difference
compares to the overall frame complexity as represented by the intra
coding error.
This was tuned experimentally to give test results that
were approximately neutral for our various test sets. There was
a slight drop in Overall PSNR but a consistent improvement in
SSIM. This balance may be improved with tuning further as it is
noteworthy that it was much better on the hd_res set.
Results (Overall PSNR, SSIM -ve better) for low_res, ugc360, midres2,
ugc480P and hd_res are as follows:
0.173 -0.688
0.118 -0.153
0.132 -0.239
0.261 -0.405
-0.305 -1.109
As part of this adjustment the contribution of motion amplitude was
removed.
This patch also changes the control mechanism that will be exposed
on the command line for use by the Vizier project. The control is now
a linear factor which defaults to 1.0, where values < 1.0 mean a lower
decay rate and values > 1.0 mean an increased decay rate.
This presents a more easily understandable interface for use in
optimizing the decay behavior for various formats, where it is clear
what a passed in value means relative to the default.
With the new decay mechanism the current values for various formats
are almost certainly wrong and we still need to define sensible upper
and lower bounds for use during future training.
Change-Id: Ib1074bbea97c725cdbf25772ee8ed66831461ce3
Adam B. Goode [Wed, 17 Mar 2021 19:11:57 +0000 (14:11 -0500)]
Msvc builds convert to windows path w/msys env
Bug: webm:1720
Change-Id: I56689ad408f8086c511e1711dfa9c8d404727b2e
James Zern [Fri, 12 Mar 2021 19:44:42 +0000 (19:44 +0000)]
Merge "vp8: restrict 1st pass cpu_used range"
Marco Paniconi [Fri, 12 Mar 2021 17:54:03 +0000 (09:54 -0800)]
vp9-rtc: Add postencode_drop control to sample encoder
Change-Id: I1c989f26b0a7b9239adf37df8d96776f33b89a8b
Jerome Jiang [Tue, 9 Mar 2021 00:07:02 +0000 (16:07 -0800)]
Prepare for v1.10.0 release.
Update CHANGELOG, AUTHORS, README, libs.mk
Bug: webm:1712
Change-Id: Ic99de12b91a92c32f8a9485dcb759c48bc3eccd6
Paul Wilkins [Tue, 9 Mar 2021 15:11:41 +0000 (15:11 +0000)]
Vizer: Added in experimental max KF boost values.
Added the experimental max per frame KF boost values derived from
the Vizier experiments.
These are still all off by default.
When enabled I expect these to cause significant regression as they
fluctuate wildly and in a way that makes no sense from format to format.
I suspect these values reflect over fitting perhaps from a subset of
training clips with more frequent mid chunk key frames and or short key
frame groups.
Also fixed incorrect value for gf boost for one format.
Experiment to moderate these values and use different values for first
and subsequent KF groups to follow.
Change-Id: Ibeb4268957f2edacdb4549d74930255a22a2fcc5
Paul Wilkins [Tue, 9 Mar 2021 14:47:25 +0000 (14:47 +0000)]
Vizier: Add in field for min kf frame boost.
Added kf_frame_min_boost field to hold the minimum per frame
boost in key frame boost calculations. Replaces hard wired value.
To be used in conjunction with and tied to the maximum value.
Change-Id: I67a39ecb3f21b5918512a5ccd9a1b214d7971e45
Paul Wilkins [Wed, 10 Mar 2021 14:35:53 +0000 (14:35 +0000)]
Merge "Vizier: Add defaults for > 1080P"
Cheng Chen [Wed, 10 Mar 2021 05:50:50 +0000 (05:50 +0000)]
Merge "L2E: let vp9 encoder respect external max frame size constraint"
Paul Wilkins [Tue, 9 Mar 2021 17:30:04 +0000 (17:30 +0000)]
Merge "Further integration for Vizier."
Paul Wilkins [Tue, 9 Mar 2021 14:07:48 +0000 (14:07 +0000)]
Vizier: Add defaults for > 1080P
Previous code did not have sensible defaults for larger image formats.
Added defaults for Vizier RD parameters for sizes > 1080P and changed
the first pass parameters for large formats to use the 1080P values.
No supplied value for rd_mult_q_sq_key_high_qp case yet so set to
old hard wired default value.
If the Vizier parameters were enabled the lack of sensible defaults
caused a large regression for 2K clips in one of our test sets.
Change-Id: I306c0cd76eab00d50880c91fadb5842faf6661ff
Cheng Chen [Mon, 8 Mar 2021 20:25:31 +0000 (12:25 -0800)]
L2E: let vp9 encoder respect external max frame size constraint
Change-Id: Ib926e694d4bc4675af1435a32f6316a587756380
Paul Wilkins [Thu, 4 Mar 2021 17:10:09 +0000 (17:10 +0000)]
Further integration for Vizier.
Further integration of Vizier adjustable parameters,
This patch connects up additional configurable two pass rate control
parameters for the Vizier project. This still needs to be connected up
to a command line interface and at the moment should still be using
default values that match previous behavior.
Do not submit until verified that defaults are all working correctly.
Change-Id: If1241c2dba6759395e6efa349c4659a0c345361d
James Touton [Fri, 5 Mar 2021 02:43:29 +0000 (18:43 -0800)]
Check for _WIN32 instead of WIN32.
_WIN32 is predefined for the Windows platform in MSVC, whereas WIN32 is not, and WIN32 is also not defined in the makefiles.
Change-Id: I8b58e42d891608dbe1e1313dc9629c2be588d9ec
Paul Wilkins [Thu, 4 Mar 2021 21:17:28 +0000 (21:17 +0000)]
Merge "Add fields into RC for Vizier ML experiments."
Jerome Jiang [Thu, 4 Mar 2021 20:30:02 +0000 (20:30 +0000)]
Merge "override assembler with --as option on msvs"
Jerome Jiang [Thu, 4 Mar 2021 01:33:30 +0000 (17:33 -0800)]
Use -std=gnu++11 instead of -std=c++11
Cygwin and msys2 have stricter compliance requirement over standard c
headers.
Bug: webm:1708
Change-Id: I676b1227b9dd304149e50016468df0f057c6a78f
Jerome Jiang [Wed, 3 Mar 2021 22:46:48 +0000 (14:46 -0800)]
override assembler with --as option on msvs
Bug: webm:1709
Change-Id: I962a64c00042fe95cc1cd845b187f71ad6cfd1b7
Paul Wilkins [Wed, 3 Mar 2021 16:45:42 +0000 (16:45 +0000)]
Add fields into RC for Vizier ML experiments.
This patch adds fields into the RC data structure for the Vizier.
The added fields allow control of some extra rate control parameters
and rate distortion.
This patch also adds functions to initialize the various parameters
though many are not yet used / wired in and for now all are set to
default values. Ultimately many will be set through new command
line options.
Change-Id: I41591bb627d3837d2104fb363845adedbddf2e02
Wan-Teh Chang [Sat, 27 Feb 2021 02:02:24 +0000 (18:02 -0800)]
Remove comments for removed 'active_map' parameter
Change-Id: I8635f6121e13089c25e201df033d5bc68e2862b4
Jerome Jiang [Thu, 18 Feb 2021 01:34:27 +0000 (17:34 -0800)]
Remove two pass related code from svc sample encoder.
SVC sample encoder is only supposed to be used for realtime SVC.
Bug: webm:1705
Change-Id: I5c0c3491732db3e148073aaf7f90ee8d662b57b5
James Zern [Fri, 5 Feb 2021 18:21:39 +0000 (10:21 -0800)]
vp8_denoiser_sse2_test: disable BitexactCheck w/gcc-8+
this test fails under gcc 8-10, but not with other compilers
Bug: webm:1718
Change-Id: I8c6c7a25c4aaf019a7f91f835a1a2c9a731cfadc
Marco Paniconi [Thu, 4 Feb 2021 06:09:24 +0000 (22:09 -0800)]
svc: Fix an existing unittest for flexible mode
The flag update_pattern_ was being set to 0
(because it was set before reset) instead of 1.
And the example flexible mode pattern was not setting
non-reference frame on top temporal top spatial.
Change-Id: I8aee56ce13cc4e0d614126592f9d0f691fe527b0
Cheng Chen [Wed, 3 Feb 2021 23:57:00 +0000 (23:57 +0000)]
Merge "L2E: let external rate control pass in a max frame size"