Luc Trudeau [Thu, 14 Jun 2018 15:21:05 +0000 (11:21 -0400)]
include msvc.h for snprintf support in benchmarks
include vpx_ports/msvc.h to avoid issues with snprintf issues with MSVC.
Change-Id: Ida09cff8ee3b84e09fd61de131f84b32c113fa1a
Zoe Liu [Sun, 17 Jun 2018 12:42:31 +0000 (05:42 -0700)]
Add update types for hierarchical refs
Change-Id: I0cd91187e1efc1441086772e5683fbf72d9371cf
Jerome Jiang [Sat, 16 Jun 2018 00:12:44 +0000 (00:12 +0000)]
Merge "VP9 HBD: Fix integer overflow problem in variance calc."
Jerome Jiang [Fri, 15 Jun 2018 23:33:15 +0000 (23:33 +0000)]
Merge "vp9 svc: add tests for inter layer prediction."
Jerome Jiang [Fri, 15 Jun 2018 23:03:00 +0000 (16:03 -0700)]
VP9 HBD: Fix integer overflow problem in variance calc.
BUG=webm:1534
Change-Id: I535ac48e3dd2454cc7088c4f9a1e08ea74107da6
Tom Finegan [Fri, 15 Jun 2018 22:25:25 +0000 (22:25 +0000)]
Merge "Clean up avx512 compiler support test."
Jingning Han [Fri, 15 Jun 2018 18:05:58 +0000 (18:05 +0000)]
Merge changes I3436302c,I8969f5c3
* changes:
Prepare motion estimation process for temporal dependency model
Construct temporal dependency building system
Tom Finegan [Fri, 15 Jun 2018 17:58:11 +0000 (10:58 -0700)]
Clean up avx512 compiler support test.
Moves the check into a function, check_gcc_avx512_compiles,
that behaves somewhat similarly to check_gcc_machine_options.
Change-Id: I2bef3ddd98e636eef12d9d5e548c43282fac7826
Jingning Han [Wed, 23 May 2018 19:13:50 +0000 (12:13 -0700)]
Prepare motion estimation process for temporal dependency model
Set up needed stack for the motion estimation process to build up
the temporal dependency model.
Change-Id: I3436302c916a686e8c82572ffc106bf8023404b6
Jingning Han [Wed, 23 May 2018 16:07:45 +0000 (09:07 -0700)]
Construct temporal dependency building system
Schedule the frame processing to construct temporal dependency
statistics within a group of pictures. Align the corresponding
reference frames.
Change-Id: I8969f5c335a4a5c2614f4530b636fe13a25a8a98
Jingning Han [Fri, 15 Jun 2018 01:33:41 +0000 (01:33 +0000)]
Merge "Allocate tpl_dep_stats frame buffer"
Zoe Liu [Fri, 15 Jun 2018 00:00:13 +0000 (00:00 +0000)]
Merge "Separate GF structure defining from bit allocation"
Jingning Han [Thu, 14 Jun 2018 23:06:14 +0000 (23:06 +0000)]
Merge "Add temporal model control as a speed feature"
Jerome Jiang [Thu, 14 Jun 2018 19:30:38 +0000 (12:30 -0700)]
vp9 svc: add tests for inter layer prediction.
Change-Id: Ic8e07b07790e067c014677cf33c3b016fcf4cb39
Zoe Liu [Thu, 14 Jun 2018 18:09:41 +0000 (11:09 -0700)]
Separate GF structure defining from bit allocation
This CL separates the defining of the GF group structure from the
handling of its bitrate allocation. The encoder performance should stay
unchanged.
Change-Id: Ib77967757702bb4b284034e429d4c41ae86d0838
Jingning Han [Mon, 21 May 2018 17:13:53 +0000 (10:13 -0700)]
Allocate tpl_dep_stats frame buffer
Allocate buffers to support gather temporal dependency statistics
at the encoder.
Change-Id: I97d4594913a2423e8a916f20caf82ab0f5836961
Jingning Han [Thu, 14 Jun 2018 17:04:03 +0000 (10:04 -0700)]
Add temporal model control as a speed feature
The model construction would incur 15% slowdown for speed 2. The
speed change on speed 0 is unnoticeable.
The current speed features set up would DISABLE temporal dependency
model for all speed settings.
Change-Id: Ic45dd962f3a54a8f5f0452502dc05e352dc09ca1
Jingning Han [Mon, 21 May 2018 16:31:13 +0000 (09:31 -0700)]
Add data structure for frame dependent mode decision
Add block and frame level data structures to support frame
dependent mode decision.
Change-Id: I996fc84155fcba8e2ec2a114bb0799d6aa5539dd
Zoe Liu [Thu, 14 Jun 2018 16:18:51 +0000 (16:18 +0000)]
Merge "Unify frame_index in defining GF group structure"
Luc Trudeau [Thu, 14 Jun 2018 16:13:02 +0000 (16:13 +0000)]
Merge changes I51776f0e,I843f3b34
* changes:
[VSX] Optimize PROCESS16 macro
VSX Version of SAD8xN
Jerome Jiang [Thu, 14 Jun 2018 06:56:40 +0000 (06:56 +0000)]
Merge "vp8: remove assertion in tree coder."
Luc Trudeau [Wed, 13 Jun 2018 19:24:54 +0000 (15:24 -0400)]
[VSX] Optimize PROCESS16 macro
The PROCESS16 macro now uses 8-bit lanes instead of 16-bit lanes.
SADTest Speed Test (POWER8 Model 2.1)
16x8 Old VSX time = 16.7 ms, new VSX time = 9.1 ms [1.8x]
16x16 Old VSX time = 15.7 ms, new VSX time = 7.9 ms [2.0x]
16x32 Old VSX time = 14.4 ms, new VSX time = 7.2 ms [2.0x]
32x16 Old VSX time = 14.0 ms, new VSX time = 7.4 ms [1.9x]
32x32 Old VSX time = 13.4 ms, new VSX time = 6.5 ms [2.0x]
32x64 Old VSX time = 12.7 ms, new VSX time = 6.3 ms [2.0x]
64x32 Old VSX time = 12.6 ms, new VSX time = 6.3 ms [2.0x]
64x64 Old VSX time = 12.7 ms, new VSX time = 6.2 ms [2.0x]
Change-Id: I51776f0e428162e78edde8eac47f30ffd2379873
Zoe Liu [Thu, 14 Jun 2018 00:33:57 +0000 (17:33 -0700)]
Unify frame_index in defining GF group structure
Following are completed in defining GF group structure in firstpass:
1. Remove redundant alt_frame_index;
2. Remove hard coded index value with the variable of frame_index.
Change-Id: I7b56e454559bbf704afc7410ea9832b20ffcd57e
Jerome Jiang [Wed, 13 Jun 2018 19:55:17 +0000 (12:55 -0700)]
vp8: remove assertion in tree coder.
Cast the counter to uint64_t in case it overflows.
The assert was to prevent c[0] * Pfac being overflow beyong unsigned int
since Pfac could be 2^8. Thus c[0] needs to be smaller than 2^24.
In VP9, the assert was removed and c[0] was casted to uint64_t.
Bug: 805277
Change-Id: Ic46a3c5b4af2f267de4e32c1518b64e8d6e9d856
Luc Trudeau [Wed, 13 Jun 2018 17:39:04 +0000 (13:39 -0400)]
VSX Version of SAD8xN
VSX versions of the SAD functions of width 8.
SADTest Speed Test (POWER8 Model 2.1)
8x4 C time = 68.7 ms (±0.3 ms), VSX time = 31.8 ms (±0.1 ms) [2.2x]
8x8 C time = 55.6 ms (±0.3 ms), VSX time = 18.3 ms (±0.1 ms) [3.0x]
8x16 C time = 46.5 ms (±0.1 ms), VSX time = 15.6 ms (±0.1 ms) [3.0x]
Change-Id: I843f3b34e103b72deeade4a939193d8b53cee460
Luc Trudeau [Wed, 13 Jun 2018 17:36:17 +0000 (13:36 -0400)]
Add Speed Tests for the SADTest test suite.
Speed tests are added for the SADTest test suite. These test use the
AbstractBench and print the median run time of SAD operations. Speed
tests are disabled by default.
Change-Id: I5d0957248f9b5b307ae2d757d5f8d4761a1dd712
Tom Finegan [Wed, 13 Jun 2018 17:50:39 +0000 (17:50 +0000)]
Merge "Fix avx512 related MSVC build failure."
Jingning Han [Wed, 13 Jun 2018 17:32:12 +0000 (17:32 +0000)]
Merge "Remove duplicate vp9_twopass_postencode_update def"
Tom Finegan [Wed, 13 Jun 2018 16:39:28 +0000 (09:39 -0700)]
Fix avx512 related MSVC build failure.
Check GCC specific AVX512 flags only when GCC is enabled.
Change-Id: I15dc2a0dbf8bce37f4364fedfd34a0a34882104b
Jingning Han [Tue, 12 Jun 2018 20:46:43 +0000 (13:46 -0700)]
Remove duplicate vp9_twopass_postencode_update def
Change-Id: I370f37c85a02c032a8ba266b9b9445ee38eb0756
Marco Paniconi [Tue, 12 Jun 2018 05:40:32 +0000 (05:40 +0000)]
Merge "vp9 svc: Denoise golden when it's a temporal ref."
Jerome Jiang [Mon, 11 Jun 2018 18:05:36 +0000 (11:05 -0700)]
vp9 svc: Denoise golden when it's a temporal ref.
When golden was the inter-layer reference, a block that selected the golden ref
would not be denoised.
But when golden is used as a second temporal reference then we should denoise
blocks that select the golden reference.
This changes allows for that.
Change-Id: Ifdea2ac88f6a74f73520fedcd7fec2f32c559ec9
Luc Trudeau [Sat, 9 Jun 2018 02:53:30 +0000 (22:53 -0400)]
VSX Version of vp9_quantize_fp_32x32
Low bit depth version only. Passes the VP9QuantizeTest test suite.
VP9QuantizeTest Speed Test (POWER8 Model 2.1)
32x32 C time = 93.1 ms (±0.4 ms), VSX time = 6.5 ms (±0.2 ms) [14.4x]
Change-Id: I7f1fd0fc987af86baf2b74147a25aee811289112
Luc Trudeau [Thu, 7 Jun 2018 19:30:23 +0000 (15:30 -0400)]
VSX Version of vp9_quantize_fp
Low bit depth version only. Passes the VP9QuantizeTest test suite.
VP9QuantizeTest Speed Test (POWER8 Model 2.1)
4x4 C time = 86.3 ms (±0.7 ms), VSX time = 18.2 ms (±0.0 ms) [ 4.7x]
8x8 C time = 57.7 ms (±0.3 ms), VSX time = 7.6 ms (±0.0 ms) [ 7.6x]
16x16 C time = 50.7 ms (±0.1 ms), VSX time = 4.9 ms (±0.0 ms) [10.3x]
Change-Id: Ic09bc786c57cc89bba14624064216b52996075eb
Hui Su [Mon, 11 Jun 2018 18:07:18 +0000 (18:07 +0000)]
Merge "Small speedup of ml_pruning_partition()"
Marco Paniconi [Mon, 11 Jun 2018 16:54:43 +0000 (16:54 +0000)]
Merge "vp9-svc: Fix to frames_since_golden update for SVC."
Jerome Jiang [Mon, 11 Jun 2018 16:43:44 +0000 (16:43 +0000)]
Merge "vp9 svc: clean up first_spatial_layer_to_encode."
Marco Paniconi [Thu, 7 Jun 2018 23:35:44 +0000 (16:35 -0700)]
vp9-svc: Fix to frames_since_golden update for SVC.
When the second (gf) temporal reference is used in SVC:
the reference is refreshed on base TL superframes, and so
the rc->frames_since_golden counter was also only updated on
base TL frames. But this was disabling the golden reference
from being used as a temporal reference for TL > 0 frames
(since frames_since_golden was 0/not updated on TL > 0 frames).
Fix is to copy the update of rc->frames_since_golden to all
upper temporal layers. This allows TL > 0 frames to test the
golden inter mode.
Gain on RTC set: ~2%, ~8% on desktop_vga clip.
Encode time increase ~5-8% on linux, 3SL-3TL run with 1 thread.
For now keep this off for TL > 0 frames in speed features, so
this change does not change current behavior for speed >= 7.
Change-Id: I405708f3f80039ae47bd64ec53e66f92160acd9e
Jerome Jiang [Fri, 8 Jun 2018 17:53:34 +0000 (10:53 -0700)]
vp9 svc: clean up first_spatial_layer_to_encode.
Change-Id: I3c9aefd3ea5028797b9105d7e49b1cb2f762a9fc
Hui Su [Fri, 8 Jun 2018 20:41:05 +0000 (13:41 -0700)]
Small speedup of ml_pruning_partition()
Terminate early and skip neural net model when linear score is already
high enough, which indicates that we should not skip split and
rectangular partitions.
No changes on compression; encoding speed improves slightly.
Change-Id: I4e0995090200eb4889344da905d2f7048673af5f
James Zern [Fri, 8 Jun 2018 17:44:43 +0000 (17:44 +0000)]
Merge "vp9_subtract_test,cosmetics: fix class order, casts"
Jingning Han [Fri, 8 Jun 2018 17:18:33 +0000 (17:18 +0000)]
Merge "Localize variable scope in vp9_rc_get_second_pass_params()"
Tom Finegan [Fri, 8 Jun 2018 15:11:13 +0000 (15:11 +0000)]
Merge "Add avx512 compile test."
James Zern [Fri, 8 Jun 2018 06:19:50 +0000 (23:19 -0700)]
vp9_subtract_test,cosmetics: fix class order, casts
+ remove obsolete FIXME
Change-Id: I97ceb94b0e7860167e9c8cc6900bec8d155f0e8f
James Zern [Fri, 8 Jun 2018 06:15:24 +0000 (06:15 +0000)]
Merge changes I89ce12b6,Id91b52d6,Icd7d4453
* changes:
Implement subtract_block for VSX
Cast bsize as int to print a meaninful debug info
Speed test for subtract_block
Marco Paniconi [Thu, 7 Jun 2018 22:07:57 +0000 (15:07 -0700)]
vp9-svc: Adjust some logic on gf temporal reference.
For the feature of using second temporal reference (when
inter-layer is off): move the buffer_idx assignement and
refresh flag settings further down to vp9_rc_get_svc_params(),
since is_key_frame is set there for every frame/layer.
Otherwise it was using the setting from the previous frame/layer.
This makes the refresh more consistent for both layers for
2 spatial layers case.
Small/negligible change in metrics.
Change-Id: I88279243bc27898448e8891dba38143d936cf6d5
Luca Barbato [Wed, 6 Jun 2018 21:10:18 +0000 (21:10 +0000)]
Implement subtract_block for VSX
~2x speedup or better.
[ RUN ] C/VP9SubtractBlockTest.Speed/0
[ BENCH ] 4x4 365.1 ms ( ±2.2 ms )
[ BENCH ] 8x4 258.5 ms ( ±0.3 ms )
[ BENCH ] 4x8 202.7 ms ( ±0.2 ms )
[ BENCH ] 8x8 162.2 ms ( ±0.5 ms )
[ BENCH ] 16x8 138.8 ms ( ±0.3 ms )
[ BENCH ] 8x16 121.5 ms ( ±0.4 ms )
[ BENCH ] 16x16 110.2 ms ( ±0.5 ms )
[ BENCH ] 32x16 104.8 ms ( ±0.1 ms )
[ BENCH ] 16x32 32.7 ms ( ±0.1 ms )
[ BENCH ] 32x32 30.0 ms ( ±0.0 ms )
[ BENCH ] 64x32 28.7 ms ( ±0.0 ms )
[ BENCH ] 32x64 20.1 ms ( ±0.0 ms )
[ BENCH ] 64x64 19.3 ms ( ±0.0 ms )
[ RUN ] VSX/VP9SubtractBlockTest.Speed/0
[ BENCH ] 4x4 155.3 ms ( ±0.9 ms )
[ BENCH ] 8x4 99.3 ms ( ±0.4 ms )
[ BENCH ] 4x8 77.2 ms ( ±0.1 ms )
[ BENCH ] 8x8 45.7 ms ( ±0.0 ms )
[ BENCH ] 16x8 34.1 ms ( ±0.0 ms )
[ BENCH ] 8x16 29.5 ms ( ±0.0 ms )
[ BENCH ] 16x16 19.9 ms ( ±0.0 ms )
[ BENCH ] 32x16 15.1 ms ( ±0.0 ms )
[ BENCH ] 16x32 16.7 ms ( ±0.0 ms )
[ BENCH ] 32x32 14.1 ms ( ±0.0 ms )
[ BENCH ] 64x32 12.6 ms ( ±0.0 ms )
[ BENCH ] 32x64 12.0 ms ( ±0.0 ms )
[ BENCH ] 64x64 11.2 ms ( ±0.0 ms )
Change-Id: I89ce12b6475871dc9e8fde84d0b6fe5c420c28c7
Luca Barbato [Wed, 6 Jun 2018 22:49:39 +0000 (22:49 +0000)]
Cast bsize as int to print a meaninful debug info
cout helpfully decides to print the bsize value as non-printable char
otherwise.
Change-Id: Id91b52d6475ae9f869365468d1d56d94b2e10ecb
Luca Barbato [Wed, 6 Jun 2018 22:16:28 +0000 (22:16 +0000)]
Speed test for subtract_block
Change-Id: Icd7d4453f0ee699635a2a1d484d24cba71d748de
Tom Finegan [Thu, 7 Jun 2018 19:35:05 +0000 (12:35 -0700)]
Add avx512 compile test.
Some compiler releases allow the -mavx512f arg without actually
implementing support. Test for this situation, and disable avx512
when it is detected by configure.
BUG=webm:1536
Change-Id: I63952153bb4b24aa9f25267ed47a0fe845d61f8b
Jerome Jiang [Thu, 7 Jun 2018 06:02:26 +0000 (23:02 -0700)]
vp9 svc: add control to set using second temporal ref.
Bump up ABI version.
Change-Id: I4498d7ea4ed72994c5f847aa98e75b0150dd7f82
Marco Paniconi [Thu, 7 Jun 2018 17:52:09 +0000 (10:52 -0700)]
vp9-svc: Allow second temporal reference for next highest layer.
When inter-layer prediction is disabled on INTER frames, allow
for next highest resolution to have second temporal reference.
Current code allowed for only top/highest spatial layer.
Change-Id: I102137273e3e4d57512a13d95e8ccb9c5b0a7b4b
Marco Paniconi [Thu, 7 Jun 2018 15:57:42 +0000 (15:57 +0000)]
Merge "VP9 SVC: Write out svc src for all spatial layers."
Marco Paniconi [Thu, 7 Jun 2018 05:42:38 +0000 (22:42 -0700)]
vp9-svc: Modify choose_partitioning for second temporal ref
For mode where second temporal reference is used in SVC: allow
for using/testing this reference (golden ref) in the variance
partition scheme (choose_partitioning).
Small positive gain (~0.25%) on metrics for 3 layer SVC,
negligible change in speed.
Change-Id: I29b8315da530e60db3d6c90faa8fb178d9f2de26
Jerome Jiang [Thu, 7 Jun 2018 00:02:48 +0000 (00:02 +0000)]
Merge "VP9: fix unsigned integer overflow in decoder."
Jerome Jiang [Wed, 6 Jun 2018 22:44:01 +0000 (15:44 -0700)]
VP9 SVC: Write out svc src for all spatial layers.
Change-Id: Ie78676e4df75f3f870ee2de0c87a8167b7ec68e0
Marco Paniconi [Wed, 6 Jun 2018 22:16:14 +0000 (15:16 -0700)]
vp9-svc: Enable use of second temporal reference for SVC.
When inter-layer is disabled on INTER frames, this will allow
use of a second (longer term) temporal reference for SVC.
Only enabled on highest resolution spatial layer.
Average gains of ~4% on RTC set, speed decrease of about ~2%.
Change-Id: I3c2d415653c448eb7269c828e120fe8bb2ef3f97
Marco Paniconi [Wed, 6 Jun 2018 19:14:59 +0000 (12:14 -0700)]
vp9-svc: Add a buffer_idx is_used parameter for SVC.
For the case where a second (long term) temoral reference is
used in the SVC: this additional parameter is to make sure the
buffer slot selected for this reference is available for usage,
i.e., it is never used for any of the 3 references set for the
fixed SVC patterns.
And some code cleanup (replace cpi->svc).
No change in behavior.
Change-Id: Icba46edfbbefb94d5ea8e2d5c24cccd85a406ee6
Jingning Han [Wed, 6 Jun 2018 19:52:45 +0000 (12:52 -0700)]
Localize variable scope in vp9_rc_get_second_pass_params()
Remove unnecessary definitions.
Change-Id: Ie540aaed5f3ed3768eff4e6563455666aef9c9e8
Jerome Jiang [Thu, 24 May 2018 21:07:28 +0000 (14:07 -0700)]
VP9: fix unsigned integer overflow in decoder.
The difference of two size_t variables.
Change-Id: I73f35cdafc2ba64a9ddaf855cc6a410cfb63b8da
Jerome Jiang [Tue, 5 Jun 2018 22:21:29 +0000 (15:21 -0700)]
vp9: Move up reset of cyclic refresh under dynamic resize.
When resize happens and cyclic refresh is not applied on the
current (resized) frame, the sb_index is not reset and then
might be out of boundary on future frames when the
cyclic refresh is applied.
Change-Id: I05282fc4bc2323522d60e019ed0790d69221a2f7
James Zern [Wed, 6 Jun 2018 04:54:41 +0000 (04:54 +0000)]
Merge changes I3ba75c45,I97d26285
* changes:
force-inline the convolve functions
Unbreak the force inline directive for gcc
Luca Barbato [Fri, 25 May 2018 20:22:42 +0000 (22:22 +0200)]
force-inline the convolve functions
Change-Id: I3ba75c459ed7c9591b7892e9f8f108146c04507d
Marco Paniconi [Tue, 5 Jun 2018 03:24:52 +0000 (03:24 +0000)]
Merge "vp9-svc: Allow usage of second (long term) temporal reference."
James Zern [Sat, 2 Jun 2018 06:22:48 +0000 (23:22 -0700)]
test,cosmetics: fix func/member naming, decl order
functions: upper camelcase
members: lowercase with trailing '_'
decl order: functions (overrides marked virtual), members
after:
656e8ac61 VSX version of vpx_post_proc_down_and_across_mb_row
766d875b9 VSX version of vpx_mbpost_proc_ip
35e98a70b VSX version of vpx_mbpost_proc_down
b2898a9ad Bench Class For More Robust Speed Tests
Change-Id: Ib257bd607c5c1248d30e619ec9e8a47cc629825b
Jerome Jiang [Fri, 1 Jun 2018 21:27:34 +0000 (14:27 -0700)]
vp9-svc: Allow usage of second (long term) temporal reference.
Allow for second temporal reference for top spatial layer in SVC,
when inter-layer prediction is disabled on INTER frames.
The second temporal reference is labelled as the golden reference
and the update/refresh of this reference buffer is only on base
temporal layer superframes. For now the period of refresh is
fixed at every 20 TL0 superframes.
Average gain is ~4% on RTC set, several clips up
by ~8-12%. Speed loss is about ~2% on mac.
Feature is disabled as default for now.
Change-Id: I2e5db5052c62dbe958a3b14be97d043823b7a529
Luca Barbato [Mon, 28 May 2018 14:11:17 +0000 (16:11 +0200)]
Unbreak the force inline directive for gcc
Change-Id: I97d26285ec146628cbafd3573ca812c630c6687d
Jerome Jiang [Fri, 1 Jun 2018 23:04:35 +0000 (23:04 +0000)]
Merge "VP9: exclude speed 9 from VBR datarate tests."
Jerome Jiang [Fri, 1 Jun 2018 18:36:39 +0000 (11:36 -0700)]
VP9: exclude speed 9 from VBR datarate tests.
Change-Id: I4c4d31d013cb45e20918f4ef83ce32811d76e02b
Jerome Jiang [Thu, 31 May 2018 23:00:32 +0000 (23:00 +0000)]
Merge "VP9: Allow for bilinear subpel interp at speed 9 for high motion."
Marco Paniconi [Thu, 31 May 2018 22:22:57 +0000 (22:22 +0000)]
Merge "vp9-svc: Fix to some frame metrics for real-time mode."
Jerome Jiang [Wed, 23 May 2018 18:52:37 +0000 (11:52 -0700)]
VP9: Allow for bilinear subpel interp at speed 9 for high motion.
Fixed some settings in nonrd pick mode to allow for frame-level bilinear
to be set.
On Galaxy S8+ it has 4% speed up on high motion clips. Almost the same
for low motion.
0.17% quality loss on RTC.
Change-Id: I044a7de020183754ba08bb6c96c5a78ba5c7fea2
Hui Su [Thu, 31 May 2018 21:04:40 +0000 (21:04 +0000)]
Merge "Improve the ML based partition pruning"
Marco Paniconi [Thu, 31 May 2018 16:59:15 +0000 (09:59 -0700)]
vp9-svc: Fix to some frame metrics for real-time mode.
Add condition of LAST frame to the consec_zeromv and
avg_frame_low_motion metrics. This is needed for SVC as
the golden reference is a spatial reference and should
not be included in the metric computation.
Small/negligible change in metrics on RTC set.
Change-Id: I6ea16298fae566bb288c34cf50d120b509146eee
Luc Trudeau [Wed, 30 May 2018 02:08:18 +0000 (22:08 -0400)]
VSX version of vpx_post_proc_down_and_across_mb_row
Low bit depth version only. Passes the VpxPostProcDownAndAcrossMbRowTest
VpxMbPostProcAcrossIpTest Speed Test (POWER8 Model 2.1)
C time = 121.3 ms (±4.0 ms), VSX time = 9.4 ms (±0.3 ms) [12.9x]
Change-Id: I28300779e197ea3855cf30867d17a2805388b447
Hui Su [Mon, 14 May 2018 21:35:25 +0000 (14:35 -0700)]
Improve the ML based partition pruning
Add a neural net model that uses the same features as the existing
linear model. Make the pruning decision based on both the linear
and the neural net model. It provides more accurate predictions,
and may improve compression and/or encoding speed.
This only affects speed 0.
Coding gain:
0.37% on midres
0.34% on hdres
0.50% on jvet8b720p
Encoding speed impact(average over locally tested 20 clips from midres
and hdres):
QP=20: down by 2.5%.
QP=30: down by 3.9%.
QP=40: donw by 4.5%.
QP=50: up by 5.2%.
Change-Id: I402ec799745ad3b74abf0789fa5e124fe64e704d
Marco Paniconi [Thu, 31 May 2018 05:44:05 +0000 (05:44 +0000)]
Merge "vp9-svc: Fix to compute some metrics on top spatiail layer."
James Zern [Thu, 31 May 2018 04:01:02 +0000 (04:01 +0000)]
Merge "libs.mk: expose libvpx.{ver,syms} in all configs"
Marco Paniconi [Wed, 30 May 2018 23:01:53 +0000 (16:01 -0700)]
vp9-svc: Fix to compute some metrics on top spatiail layer.
The avg_frame_low_motion and consec_zeromv are frame-level
metrics that are updated on every frame. For SVC these should be
updated on top spatial layer (full resolution).
Small/negligible change in metrics.
Change-Id: Ibe14f05be3b82daa9dd60378097ff11a27f1b95e
Marco Paniconi [Wed, 30 May 2018 20:55:28 +0000 (13:55 -0700)]
vp9: Refactor code for q adjustment in CBR mode.
Move the adjustment code to separate function.
Change-Id: I876b246a5c26095f262bb9a19f03d1f17077225d
James Zern [Wed, 30 May 2018 21:25:55 +0000 (21:25 +0000)]
Merge "Revert 3 slide show coding changes"
James Zern [Tue, 29 May 2018 23:38:46 +0000 (16:38 -0700)]
Revert 3 slide show coding changes
This is a combination of the following 3 reverts. The changes cause
issues on certain hardware devices. We'll pull them for now to allow for
further investigation.
Revert "Experiment regarding playback problems on Bravia TVs."
This reverts commit
624f8105f5ae7e0bc82bdc7e10a4253295134a8b.
Revert "Improved slide show coding"
This reverts commit
f4091bc30eef0fcacb5d5bd74ab7cfbc3c8aab5f.
Revert "Improved coding on slide show content."
This reverts commit
2fa333c2ae1b6fcc1864de12a7ad344a16e2ac0a.
BUG=b/
77492144
Change-Id: Ifba937792d644a9286307262f050216408e8ecf4
Jim Bankoski [Tue, 29 May 2018 22:23:48 +0000 (15:23 -0700)]
tiny_ssim: fix for odd image sizes.
Change-Id: I7dd1e37c5de3efccc07fcdc877653d4873a88266
Marco Paniconi [Wed, 30 May 2018 03:24:57 +0000 (03:24 +0000)]
Merge "vp9-svc: Add frame dropper control to sample encoder."
Marco Paniconi [Tue, 29 May 2018 22:50:00 +0000 (15:50 -0700)]
vp9-svc: Add frame dropper control to sample encoder.
Disabled as default as enc_cfg.rc_dropframe_thresh is
set to 0 as default.
Change-Id: Ia888aa16b1a86a716ec33ea041e8b16b19bf93be
Marco Paniconi [Wed, 30 May 2018 02:25:17 +0000 (02:25 +0000)]
Merge "vp9: Adjust cyclic refresh and limit frame-level q."
James Zern [Wed, 30 May 2018 01:21:01 +0000 (18:21 -0700)]
libs.mk: expose libvpx.{ver,syms} in all configs
this allows the targets to be used explicitly in builds configured with
--enable-external-build
Change-Id: Id7db309a39a73cfd8f15f74430b17b317c0a847f
Marco Paniconi [Tue, 29 May 2018 04:08:57 +0000 (21:08 -0700)]
vp9: Adjust cyclic refresh and limit frame-level q.
For CBR mode with aq-mode=3: reduce delta-q for second
segment and limit how much the frame-level q can decreae
from one frame to the next.
Reduces bitrate spikes in slide/sreen content.
Change-Id: Id9ac4b7270f07e09690380755cfbef4aec5c26dc
Luc Trudeau [Mon, 28 May 2018 19:04:57 +0000 (15:04 -0400)]
VSX version of vpx_mbpost_proc_ip
Low bit depth version only. Passes the VpxMbPostProcAcrossIpTest.
VpxMbPostProcAcrossIpTest Speed Test (POWER8 Model 2.1)
C time = 188.5ms (±0.2ms), VSX time = 65.2ms (±0.1ms) [2.9x]
Change-Id: I1cf72365d94a9d7f1e9323925a87a30e3bd5cfe2
Luc Trudeau [Fri, 18 May 2018 17:38:48 +0000 (13:38 -0400)]
VSX version of vpx_mbpost_proc_down
Low bit depth version only. Passes the VpxMbPostProcDownTest.
VpxMbPostProcDownTest Speed Test (POWER8 Model 2.1)
Full calculations:
C time = 195.4 ms, VSX time = 33.7 ms (5.8x)
Change-Id: If1aca7c135de036a1ab7923c0d1e6733bfe27ef7
Luc Trudeau [Tue, 22 May 2018 18:16:15 +0000 (14:16 -0400)]
Bench Class For More Robust Speed Tests
To make speed testing more robust, the AbstractBench runs the
desired code multiple times and report the median run time with
mean absolute deviation around the median.
To use the AbstractBench, simply add it as a parent to your test
class, and implement the run() method (with the code you want to
benchmark).
Sample output for VP9QuantizeTest
[ BENCH ] Bypass calculations 4x4 165.8 ms ( ±1.0 ms )
[ BENCH ] Full calculations 4x4 165.8 ms ( ±0.9 ms )
[ BENCH ] Bypass calculations 8x8 129.7 ms ( ±0.9 ms )
[ BENCH ] Full calculations 8x8 130.3 ms ( ±1.4 ms )
[ BENCH ] Bypass calculations 16x16 110.3 ms ( ±1.4 ms )
[ BENCH ] Full calculations 16x16 110.1 ms ( ±0.9 ms )
Change-Id: I1dd649754cb8c4c621eee2728198ea6a555f38b3
Marco Paniconi [Sat, 26 May 2018 04:51:33 +0000 (21:51 -0700)]
vp9-realtime: Move frame dropper to after scene detection.
Move frame dropper to after scene detection and noise estimation.
Scene detection and noise estimation operate on source data and
update metrics along sequence, so they should be moved before
the frame dropper.
Also we don't want to drop on scene change, as the scene detection
and (possible) re-encode step will be missed.
Change-Id: I3d9e16d785bd5ace6707db2abce77ddc110bfef4
Marco Paniconi [Tue, 29 May 2018 02:32:01 +0000 (19:32 -0700)]
vp9-svc: Fix to allowed value of max_consec_drop.
For the max_consec_drop parameter in svc frame drop:
since passing value 0 in the control would completely
disable the dropper, only allow for values >= 1 to be set.
Change-Id: I6b74ec9cc08a638fa571d6246a021dab9c811d14
Jerome Jiang [Fri, 25 May 2018 17:48:28 +0000 (17:48 +0000)]
Merge "VP8: Fix use-after-free in postproc."
Jerome Jiang [Wed, 23 May 2018 22:43:00 +0000 (15:43 -0700)]
VP8: Fix use-after-free in postproc.
The pointer in vp8 postproc refers to show_frame_mi which is only
updated on show frame. However, when there is a no-show frame which also
changes the size (thus new frame buffers allocated), show_frame_mi is
not updated with new frame buffer memory.
Change the pointer in postproc to mi which is always updated.
Bug: 842265
Change-Id: I33874f2112b39f74562cba528432b5f239e6a7bd
Marco Paniconi [Fri, 25 May 2018 15:50:19 +0000 (15:50 +0000)]
Merge "VP9: Fix issues with high bitdepth in real-time."
Jerome Jiang [Thu, 24 May 2018 21:03:43 +0000 (14:03 -0700)]
VP9: Fix issues with high bitdepth in real-time.
Disable denoiser, skin detection and aq-mode for high bitdepth for now.
BUG=webm:1534
Change-Id: I361a4e20b2319041148af497bf7043bfd5c5f589
Marco Paniconi [Thu, 24 May 2018 18:13:36 +0000 (11:13 -0700)]
vp9-svc: Add max_consec_drop to SVC frame drop.
For any spatial, limits the amount of consecutive frame drop.
Change-Id: I692d90363f329f571f2b59e12cc680ad2e76065d
Marco Paniconi [Wed, 23 May 2018 21:51:00 +0000 (14:51 -0700)]
vp9: Rate control adjustments for screen content.
For screen content mode: changes to reduce occurence of
significant QP decrease (from one frame to next),
which can cause large frames (overshoot/delay).
-cap the buffer increase to optimal level for frame drop
mode where full superframe can drop
-reduce the max_adjustment_down due to buffer overflow
-reduce qp threshold to trigger re-encode on large frame
Change-Id: I3e30e4814192b5f728abff3f7359eb64f561b8f0
Paul Wilkins [Wed, 23 May 2018 11:03:12 +0000 (11:03 +0000)]
Merge "Experiment regarding playback problems on Bravia TVs."