]> granicus.if.org Git - libvpx/log
libvpx
5 years agoMerge "Modified key frame detection."
Paul Wilkins [Wed, 7 Nov 2018 17:03:44 +0000 (17:03 +0000)]
Merge "Modified key frame detection."

5 years agoMerge "vp8: remove VP8_ENTROPY_STATS code"
Johann Koenig [Wed, 7 Nov 2018 15:41:22 +0000 (15:41 +0000)]
Merge "vp8: remove VP8_ENTROPY_STATS code"

5 years agoMerge "vp9: postencode drop frame for screen content in CBR."
Jerome Jiang [Wed, 7 Nov 2018 00:33:58 +0000 (00:33 +0000)]
Merge "vp9: postencode drop frame for screen content in CBR."

5 years agoMerge "vp9 screen-content: Adjustments for screen content."
Marco Paniconi [Tue, 6 Nov 2018 23:45:30 +0000 (23:45 +0000)]
Merge "vp9 screen-content: Adjustments for screen content."

5 years agovp9: postencode drop frame for screen content in CBR.
Jerome Jiang [Tue, 28 Aug 2018 21:24:53 +0000 (14:24 -0700)]
vp9: postencode drop frame for screen content in CBR.

Encode the next frame at max q.

For layers: post_encode_drop is only check on base
spatial layer, and if base is post-encoded-dropped,
then whole superframe is dropped.

Added API to guard postencode dropping. Turned off by default.

Added unittest.

BUG=b/112990050
Change-Id: I42fee279014aca616f7a4d9b582cb2bf5da2f2e7

5 years agoMerge "vp8: Increase rate correction threshold for drop-overshoot"
Marco Paniconi [Tue, 6 Nov 2018 22:43:16 +0000 (22:43 +0000)]
Merge "vp8: Increase rate correction threshold for drop-overshoot"

5 years agoMerge "Refactor Hadamard tests and add highbd tests"
Sai Deng [Tue, 6 Nov 2018 19:03:17 +0000 (19:03 +0000)]
Merge "Refactor Hadamard tests and add highbd tests"

5 years agovp9 screen-content: Adjustments for screen content.
Marco Paniconi [Tue, 6 Nov 2018 01:49:39 +0000 (17:49 -0800)]
vp9 screen-content: Adjustments for screen content.

Increase search area, use NSTEP, and in some cases avoid
bsize below 16x16. This for base spatial layer when many blocks
in the frame have motion (from scene detection analysis).

Improves quality for scrolling motion.

Change-Id: If77b43e738a6c43610d4727a95712667088db564

5 years agoMerge "vp8 dec: only compute ref frame buffer pointer for non intra"
Jerome Jiang [Tue, 6 Nov 2018 18:47:16 +0000 (18:47 +0000)]
Merge "vp8 dec: only compute ref frame buffer pointer for non intra"

5 years agoRefactor Hadamard tests and add highbd tests
sdeng [Thu, 1 Nov 2018 17:21:13 +0000 (10:21 -0700)]
Refactor Hadamard tests and add highbd tests

Change-Id: I306083f233e53884ac21fb4621066713edddc8f7

5 years agoMerge "Track maximum layer depth in a GOP"
Jingning Han [Tue, 6 Nov 2018 17:10:24 +0000 (17:10 +0000)]
Merge "Track maximum layer depth in a GOP"

5 years agoMerge "Fix gf_group->frame_end assignment"
Jingning Han [Tue, 6 Nov 2018 17:10:18 +0000 (17:10 +0000)]
Merge "Fix gf_group->frame_end assignment"

5 years agoMerge "Refactor define_gf_group_structure()"
Jingning Han [Tue, 6 Nov 2018 17:10:10 +0000 (17:10 +0000)]
Merge "Refactor define_gf_group_structure()"

5 years agoMerge "Remove redundant assignments in define_gf_group_structure()"
Jingning Han [Tue, 6 Nov 2018 17:10:03 +0000 (17:10 +0000)]
Merge "Remove redundant assignments in define_gf_group_structure()"

5 years agoMerge "Refactor find_arf_order()"
Jingning Han [Tue, 6 Nov 2018 17:09:54 +0000 (17:09 +0000)]
Merge "Refactor find_arf_order()"

5 years agoModified key frame detection.
Paul Wilkins [Fri, 26 Oct 2018 11:12:57 +0000 (12:12 +0100)]
Modified key frame detection.

Address poor key frame detection in some content.

This patch improves on poor key frame / scene cut detection observed
with some test content. The content in question was letter boxed film
style material and also had quite low contrast. For both 1080P and 4K
multiple genuine scene cuts were being missed.

The changes alter the conditions for marking a transition as a "flash" rather
than a scene change. The new code still deals well with genuine flashes as
observed in the "crew" test clip, without falsely flagging some of the
the scene cuts in the "film" test clip.

The new film test clip also had some "flash" frames caused by a lightning
effect and in one case a flash occurred right before a scene change. This
caused a misplacement of the key frame but has been addressed by a new
clause that requires the coded error for the next frame after a candidate
key frame to be lower than the current frame.

The patch also changes the way in which neutral blocks (similar inter and
inter error) are handled in the candidate key frame decision in a way which
hopefully handles the letter boxed format better.

During wider testing some film clips still had missed key frames but this
patch does improve things. In the case of the initial test clip the encoder
correctly marks all 3 scene cuts vs 0 before the patch.

Testing with our standard (mainly short single kf) derf and NF test clips
is neutral.

Change-Id: I3b7dcfe7b2fb13fd0816ea46acc3e69c8bc581b3

5 years agoMerge "vpx_codec_enc_config_default: disable 'usage'"
Johann Koenig [Tue, 6 Nov 2018 02:34:04 +0000 (02:34 +0000)]
Merge "vpx_codec_enc_config_default: disable 'usage'"

6 years agovp8 dec: only compute ref frame buffer pointer for non intra
Jerome Jiang [Tue, 6 Nov 2018 01:46:03 +0000 (17:46 -0800)]
vp8 dec: only compute ref frame buffer pointer for non intra

When ref frame is INTRA_FRAME, pre buffer shouldn't be used.

This CL copies behavior in single thread. That should apply to
multithreading case too.

BUG=webm:1496

Change-Id: Ibe9ab8ea9dc664151fa7ebac529d5fd1a481b4a3

6 years agoTrack maximum layer depth in a GOP
Jingning Han [Tue, 6 Nov 2018 00:21:49 +0000 (16:21 -0800)]
Track maximum layer depth in a GOP

Track the effective maximum layer depth in a given group of
pictures. Keep it in the GF_GROUP data structure.

Change-Id: If777c4e0f4a871c7226a91e3871f445e92f18b24

6 years agoFix gf_group->frame_end assignment
Jingning Han [Mon, 5 Nov 2018 23:42:20 +0000 (15:42 -0800)]
Fix gf_group->frame_end assignment

The previous value was set off by 1. Use the correct value.

Change-Id: I1ce53cc99063ce31e7ab1c43c6e444cb9a1972db

6 years agoRefactor define_gf_group_structure()
Jingning Han [Mon, 5 Nov 2018 23:38:10 +0000 (15:38 -0800)]
Refactor define_gf_group_structure()

Make it a standalone operation unit. Refactor to cut off unnecessary
dependency between define_gf_group_structure() and
allocate_gf_group_bits().

Change-Id: I954fd4e96152471a994f2ffd38a72061ab517ddd

6 years agoMerge changes I774a0711,I0b4fd670,Ia09935e5
Johann Koenig [Mon, 5 Nov 2018 23:32:11 +0000 (23:32 +0000)]
Merge changes I774a0711,I0b4fd670,Ia09935e5

* changes:
  Fix compilation on OS/2
  Use wcslen() instead of std::wcslen()
  Fix compilation on OS/2

6 years agoMerge "clang-tidy: fix vpx_dsp parameters"
Johann Koenig [Mon, 5 Nov 2018 23:29:06 +0000 (23:29 +0000)]
Merge "clang-tidy: fix vpx_dsp parameters"

6 years agovpx_codec_enc_config_default: disable 'usage'
Johann [Thu, 1 Nov 2018 20:02:45 +0000 (13:02 -0700)]
vpx_codec_enc_config_default: disable 'usage'

Found with clang-tidy. This value is unused in libvpx.

There is an existing test which ensures this is not used:
test/encode_api_test.cc:
    EXPECT_EQ(VPX_CODEC_INVALID_PARAM,
              vpx_codec_enc_config_default(kCodecs[i], &cfg, 1));

Change-Id: I94bd0663c6652b4267204c02c3921972c854d0b0

6 years agoRemove redundant assignments in define_gf_group_structure()
Jingning Han [Mon, 5 Nov 2018 19:15:24 +0000 (11:15 -0800)]
Remove redundant assignments in define_gf_group_structure()

The functionality has been covered in the above
set_gf_overlay_frame_type() call.

Change-Id: Id4049cd9a1a5a9bad7ea62c412fcb557afa9a572

6 years agoclang-tidy: fix vpx_img_wrap align
Johann [Thu, 1 Nov 2018 21:45:38 +0000 (14:45 -0700)]
clang-tidy: fix vpx_img_wrap align

This function specifically only aligns the stride and not the base buffer
like vpx_img_alloc does.

BUG=webm:1444

Change-Id: I3092827eeec3c9e16306a3973534d3a362a337e8

6 years agoRefactor find_arf_order()
Jingning Han [Mon, 5 Nov 2018 19:12:24 +0000 (11:12 -0800)]
Refactor find_arf_order()

Make the maximum layer depth allowed a control parameter in
GF_GROUP. No coding stats would change.

Change-Id: I9d17167da322831e7013d761980e1c16375a161b

6 years agovp8: Increase rate correction threshold for drop-overshoot
Marco Paniconi [Mon, 5 Nov 2018 17:35:47 +0000 (09:35 -0800)]
vp8: Increase rate correction threshold for drop-overshoot

For 1 pass cbr encoding mode, with frame-dropping on:
increase the rate correction threshold for drop-overshoot detection,
to better capture cases of large overshoot.

Change-Id: I1153b1b71cf106749dd985074d6bc8f37d163c7e

6 years agoMerge "vpx postproc: rewrite in intrinsics"
Johann Koenig [Fri, 2 Nov 2018 18:13:14 +0000 (18:13 +0000)]
Merge "vpx postproc: rewrite in intrinsics"

6 years agoMerge "Add highbd Hadamard transform C implementations"
Sai Deng [Fri, 2 Nov 2018 16:46:46 +0000 (16:46 +0000)]
Merge "Add highbd Hadamard transform C implementations"

6 years agoMerge "fix snprintf error on windows"
Johann Koenig [Fri, 2 Nov 2018 16:23:23 +0000 (16:23 +0000)]
Merge "fix snprintf error on windows"

6 years agoMerge "clang-tidy: normalize variance functions"
Johann Koenig [Fri, 2 Nov 2018 14:41:24 +0000 (14:41 +0000)]
Merge "clang-tidy: normalize variance functions"

6 years agofix snprintf error on windows
Johann [Fri, 2 Nov 2018 14:34:12 +0000 (07:34 -0700)]
fix snprintf error on windows

Include vpx_ports/msvc.h to handle snprintf on older
versions of Visual Studio

Change-Id: I06cd99b32bbae82b3df079d41ff20a9a07f6fe1c

6 years agovp8: remove VP8_ENTROPY_STATS code
Johann [Wed, 31 Oct 2018 21:40:44 +0000 (14:40 -0700)]
vp8: remove VP8_ENTROPY_STATS code

Does not compile. Noticed while cleaning up un-namespaced functions

Change-Id: I4a9048e66d051397f652e7b5412606a5e234f61f

6 years agoAdd highbd Hadamard transform C implementations
sdeng [Tue, 30 Oct 2018 22:35:44 +0000 (15:35 -0700)]
Add highbd Hadamard transform C implementations

Change-Id: Ibec078c80ca1dfe6fbbc4288db89d719dac453a7

6 years agoclang-tidy: fix vpx_dsp parameters
Johann [Tue, 30 Oct 2018 19:59:46 +0000 (12:59 -0700)]
clang-tidy: fix vpx_dsp parameters

BUG=webm:1444

Change-Id: Iee19be068afc6c81396c79218a89c469d2e66207

6 years agoMerge "vp8 boolcoder: normalize to "bc""
Johann Koenig [Wed, 31 Oct 2018 22:39:08 +0000 (22:39 +0000)]
Merge "vp8 boolcoder: normalize to "bc""

6 years agoMerge "vp8: fix to address overflow in decoder."
Jerome Jiang [Wed, 31 Oct 2018 22:27:44 +0000 (22:27 +0000)]
Merge "vp8: fix to address overflow in decoder."

6 years agoMerge "vp8dx_get_quantizer: normalize VP8D_COMP"
Johann Koenig [Wed, 31 Oct 2018 22:19:43 +0000 (22:19 +0000)]
Merge "vp8dx_get_quantizer: normalize VP8D_COMP"

6 years agoclang-tidy: normalize variance functions
Johann [Tue, 30 Oct 2018 21:43:36 +0000 (14:43 -0700)]
clang-tidy: normalize variance functions

Always use src/ref and _ptr/_stride suffixes.

Normalize to [xy]_offset and second_pred.

Drop some stray source/recon_strides.

BUG=webm:1444

Change-Id: I32362a50988eb84464ab78686348610ea40e5c80

6 years agoMerge "clang-tidy: fix vp9/encoder parameters"
Johann Koenig [Wed, 31 Oct 2018 21:43:05 +0000 (21:43 +0000)]
Merge "clang-tidy: fix vp9/encoder parameters"

6 years agoMerge "clang-tidy: fix vp9/decoder parameters"
Johann Koenig [Wed, 31 Oct 2018 21:42:16 +0000 (21:42 +0000)]
Merge "clang-tidy: fix vp9/decoder parameters"

6 years agoMerge "clang-tidy: fix vp9/common parameters"
Johann Koenig [Wed, 31 Oct 2018 21:42:07 +0000 (21:42 +0000)]
Merge "clang-tidy: fix vp9/common parameters"

6 years agoMerge "clang-tidy: fix vp8/encoder parameters"
Johann Koenig [Wed, 31 Oct 2018 21:41:53 +0000 (21:41 +0000)]
Merge "clang-tidy: fix vp8/encoder parameters"

6 years agovp8 boolcoder: normalize to "bc"
Johann [Wed, 31 Oct 2018 21:24:31 +0000 (14:24 -0700)]
vp8 boolcoder: normalize to "bc"

"bc" maps to BOOL_CODER better than "br"

Change-Id: Idefd03e79ccc1851a1b26f8206a159b0e5c5fb2d

6 years agoMerge "clang-tidy: fix vp8/decoder parameters"
Johann Koenig [Wed, 31 Oct 2018 21:14:46 +0000 (21:14 +0000)]
Merge "clang-tidy: fix vp8/decoder parameters"

6 years agovp8dx_get_quantizer: normalize VP8D_COMP
Johann [Wed, 31 Oct 2018 21:13:45 +0000 (14:13 -0700)]
vp8dx_get_quantizer: normalize VP8D_COMP

Use "pbi" like the rest of the functions

Change-Id: I5f3036b8f8361c30353be378d83455b83b82ac9f

6 years agoMerge "Add SSE2 support for hbd 4-tap interpolation filter."
Chi Yo Tsai [Wed, 31 Oct 2018 20:38:52 +0000 (20:38 +0000)]
Merge "Add SSE2 support for hbd 4-tap interpolation filter."

6 years agovp8: fix to address overflow in decoder.
Jerome Jiang [Tue, 7 Aug 2018 18:10:26 +0000 (11:10 -0700)]
vp8: fix to address overflow in decoder.

Can't call internal error from the decoder thread.

Add vpx_internal_error_info to MACROBLOCKD. When corrupted frame
detected, the decoder thread returns to its own context and signal
completion of decoding for current frame.

The main decoding thread will detect error too and return error code to
decoding API call.

Each thread will signal end of decoding of the frame. Main thread waits
for the signal of all other threads to start decoding next frame.

BUG=875626,webm:1496
Change-Id: Icd05fbc558893a4e7d8532c1e7177e7550283a64

6 years agoMerge "clang-tidy: fix vp8/common parameters"
Johann Koenig [Tue, 30 Oct 2018 22:11:59 +0000 (22:11 +0000)]
Merge "clang-tidy: fix vp8/common parameters"

6 years agoclang-tidy: fix vp9/encoder parameters
Johann [Tue, 30 Oct 2018 19:46:39 +0000 (12:46 -0700)]
clang-tidy: fix vp9/encoder parameters

BUG=webm:1444

Change-Id: I6823635eb1a99c3fcca0a8f091878e3ab2fdd2ac

6 years agoclang-tidy: fix vp9/decoder parameters
Johann [Tue, 30 Oct 2018 19:17:22 +0000 (12:17 -0700)]
clang-tidy: fix vp9/decoder parameters

BUG=webm:1444

Change-Id: I9c7c0a4161aaf52436bd5c01d30b035b2ff5508c

6 years agoAdd SSE2 support for hbd 4-tap interpolation filter.
chiyotsai [Mon, 29 Oct 2018 23:12:05 +0000 (16:12 -0700)]
Add SSE2 support for hbd 4-tap interpolation filter.

Unit test performance on bitdepth 10:
    | 4X4 | 8X8 |16X16|64X64|
 2D |1.582|1.461|1.425|1.572|
HORZ|1.643|1.247|1.346|1.345|
VERT|1.378|1.695|2.020|1.763|

Unit test performance on bitdepth 12:

    | 4X4 | 8X8 |16X16|64X64|
 2D |1.578|1.409|1.426|1.497|
HORZ|1.625|1.153|1.323|1.259|
VERT|1.392|1.707|2.030|1.787|

Change-Id: I6df85330ac33fcb17d46e4302b41415dda1219f5

6 years agoclang-tidy: fix vp9/common parameters
Johann [Tue, 30 Oct 2018 18:56:17 +0000 (11:56 -0700)]
clang-tidy: fix vp9/common parameters

BUG=webm:1444

Change-Id: I1a14ad119b3bcbaddcf2291a7521513cf6425635

6 years agoclang-tidy: fix vp8/encoder parameters
Johann [Tue, 30 Oct 2018 18:24:40 +0000 (11:24 -0700)]
clang-tidy: fix vp8/encoder parameters

BUG=webm:1444

Change-Id: I57a305cdab0d62b0745116272fbd5d9257c6e679

6 years agoclang-tidy: fix vp8/decoder parameters
Johann [Tue, 30 Oct 2018 17:55:21 +0000 (10:55 -0700)]
clang-tidy: fix vp8/decoder parameters

BUG=webm:1444

Change-Id: I3dfc56f7f6430d36a1c447d8999733015a001101

6 years agoclang-tidy: fix vp8/common parameters
Johann [Tue, 30 Oct 2018 17:21:58 +0000 (10:21 -0700)]
clang-tidy: fix vp8/common parameters

Match function definitions to declarations

BUG=webm:1444

Change-Id: Ib96d3b735eaf81cece5406c89cc5156bc2cde462

6 years agoMerge "Add AVX2 support for hbd 4-tap interpolation filter."
Chi Yo Tsai [Tue, 30 Oct 2018 16:50:00 +0000 (16:50 +0000)]
Merge "Add AVX2 support for hbd 4-tap interpolation filter."

6 years agoMerge "Properly space qp in q mode for multi-layer ARF"
Jingning Han [Tue, 30 Oct 2018 04:44:09 +0000 (04:44 +0000)]
Merge "Properly space qp in q mode for multi-layer ARF"

6 years agovpx postproc: rewrite in intrinsics
Johann [Thu, 25 Oct 2018 20:37:50 +0000 (13:37 -0700)]
vpx postproc: rewrite in intrinsics

About ~10% faster on 64bit but ~10% slower on 32

Removes the assembly usage of vpx_rv.

Change-Id: I214698fb5677f615dee0a8f5f5bb8f64daf2565e

6 years agoProperly space qp in q mode for multi-layer ARF
Jingning Han [Mon, 29 Oct 2018 21:08:20 +0000 (14:08 -0700)]
Properly space qp in q mode for multi-layer ARF

Space the quantization parameter distribution according to the
layer depth for multi-layer ARF coding structure. This allows
lower layers to have relatively smaller quantization parameters
than higher layers. It improves the compression performance
in constant q mode for multi-layer ARF system:

        avg PSNR      overall PSNR      SSIM
lowres  -0.33%         -0.31%          -1.44%
midres  -0.29%         -0.38%          -1.14%
hdres   -0.27%         -0.49%          -1.02%

Change-Id: I9cfe2f27e6c0029c30614970a46de3045840264e

6 years agoMerge "vp8 bilinear: ensure non-16x16 arrays are aligned"
Johann Koenig [Mon, 29 Oct 2018 22:26:42 +0000 (22:26 +0000)]
Merge "vp8 bilinear: ensure non-16x16 arrays are aligned"

6 years agoAdd AVX2 support for hbd 4-tap interpolation filter.
chiyotsai [Fri, 26 Oct 2018 21:14:28 +0000 (14:14 -0700)]
Add AVX2 support for hbd 4-tap interpolation filter.

Speed gain:

BIT DEPTH | 8TAP FPS | 4TAP FPS | PCT INC |
    10    |   1.69   |   1.85   |  9.46%  |
    12    |   1.64   |   1.78   |  8.54%  |

Speed test is done on jet.y4m on speed 1 profile 2 over 100 frame with
br=500.

Change-Id: I411e122553e2c466be7a26e64b4dd144efb884a9

6 years agovp8 bilinear: ensure non-16x16 arrays are aligned
Johann Koenig [Mon, 29 Oct 2018 18:59:56 +0000 (18:59 +0000)]
vp8 bilinear: ensure non-16x16 arrays are aligned

The 16x16 array was changed to aligned. The 8xN and 4x4 functions
use aligned loads/stores on their internal arrays as well.

BUG=webm:1570

Change-Id: I9cfe53d7c8ed76e8854c2688eb9a509b876471d8

6 years agoMerge "vp8 bilinear: ensure temp array is aligned"
Johann Koenig [Mon, 29 Oct 2018 18:55:52 +0000 (18:55 +0000)]
Merge "vp8 bilinear: ensure temp array is aligned"

6 years agoMerge "Enable 10 bit tpl support"
Sai Deng [Mon, 29 Oct 2018 17:14:09 +0000 (17:14 +0000)]
Merge "Enable 10 bit tpl support"

6 years agovp8 bilinear: ensure temp array is aligned
Johann [Mon, 29 Oct 2018 16:21:15 +0000 (09:21 -0700)]
vp8 bilinear: ensure temp array is aligned

Loads and stores to this array require 16 byte alignment.

BUG=webm:1570

Change-Id: I82c7d21c9539a108930fd030d79caaa0bcd1eeb3

6 years agoMerge "remove "register" keyword"
Johann Koenig [Mon, 29 Oct 2018 02:07:20 +0000 (02:07 +0000)]
Merge "remove "register" keyword"

6 years agoMerge "Remove unused macros from vp9_firstpass.c"
Jingning Han [Sat, 27 Oct 2018 03:56:18 +0000 (03:56 +0000)]
Merge "Remove unused macros from vp9_firstpass.c"

6 years agoEnable 10 bit tpl support
sdeng [Wed, 24 Oct 2018 23:23:24 +0000 (16:23 -0700)]
Enable 10 bit tpl support

         lowres_bd10   midres_bd10
avg_psnr      -0.897        -1.261
ovr_psnr      -0.975        -1.349

Change-Id: Id54f2c419f4edaa91e89ffea52b4038b1d94e563

6 years agoremove "register" keyword
Johann [Fri, 26 Oct 2018 21:55:26 +0000 (14:55 -0700)]
remove "register" keyword

This has been deprecated for a long time. c++17 is trying to recover the name.

Change-Id: Iade6bebce03a50b76061695f9e634a107cd989cd

6 years agoMerge "Add Memory to Enable Row Decode"
Harish Mahendrakar [Fri, 26 Oct 2018 18:31:41 +0000 (18:31 +0000)]
Merge "Add Memory to Enable Row Decode"

6 years agoRemove unused macros from vp9_firstpass.c
Jingning Han [Fri, 26 Oct 2018 18:03:31 +0000 (11:03 -0700)]
Remove unused macros from vp9_firstpass.c

Change-Id: If5267a8c71113b171b7bddda5b49f0326c4266b8

6 years agovp8 bilinear: rewrite 4x4
Johann [Thu, 25 Oct 2018 19:23:03 +0000 (12:23 -0700)]
vp8 bilinear: rewrite 4x4

~20% faster than the MMX. Removes the last usage of
vp8_bilinear_filters_x86_[48].

Change-Id: Iee976fab9655d0020440f26c4403ce50103af913

6 years agoMerge "vp8 bilinear: rewrite 16x16"
Johann Koenig [Thu, 25 Oct 2018 19:59:06 +0000 (19:59 +0000)]
Merge "vp8 bilinear: rewrite 16x16"

6 years agoMerge "Add AVX2 support for 4-tap interpolation filter."
Chi Yo Tsai [Thu, 25 Oct 2018 18:25:09 +0000 (18:25 +0000)]
Merge "Add AVX2 support for 4-tap interpolation filter."

6 years agovp8 bilinear: rewrite 16x16
Johann [Wed, 24 Oct 2018 22:48:32 +0000 (15:48 -0700)]
vp8 bilinear: rewrite 16x16

Marginally faster. Most importantly it drops a dependency on an
external symbol (vp8_bilinear_filters_x86_8).

Change-Id: Iff022e718720f1f0eeced6201a1ad69a9c9c4f45

6 years agoMerge "vp8 bilinear: rewrite in intrinsics"
Johann Koenig [Thu, 25 Oct 2018 17:13:50 +0000 (17:13 +0000)]
Merge "vp8 bilinear: rewrite in intrinsics"

6 years agoAdd Memory to Enable Row Decode
Ritu Baldwa [Wed, 10 Oct 2018 10:55:51 +0000 (16:25 +0530)]
Add Memory to Enable Row Decode

Row based multi-thread needs extra memory to store the parsed
co-efficients, partitions and eob. This commit adds memory for the same.

Change-Id: I13fa4a6ada2ec3048bc973e465055b832429388f

6 years agoMerge "Enable tpl model to support multi-layer ARF"
Jingning Han [Thu, 25 Oct 2018 00:01:52 +0000 (00:01 +0000)]
Merge "Enable tpl model to support multi-layer ARF"

6 years agoMerge "Reset frame udpate flags after qp estimate in tpl"
Jingning Han [Thu, 25 Oct 2018 00:01:46 +0000 (00:01 +0000)]
Merge "Reset frame udpate flags after qp estimate in tpl"

6 years agoMerge "Bypass processing on use existing frame"
Jingning Han [Thu, 25 Oct 2018 00:01:41 +0000 (00:01 +0000)]
Merge "Bypass processing on use existing frame"

6 years agoMerge "Fix frame offset computation for GOP extension"
Jingning Han [Thu, 25 Oct 2018 00:01:35 +0000 (00:01 +0000)]
Merge "Fix frame offset computation for GOP extension"

6 years agoMerge "Refactor gop_length use case in tpl model"
Jingning Han [Thu, 25 Oct 2018 00:01:29 +0000 (00:01 +0000)]
Merge "Refactor gop_length use case in tpl model"

6 years agovp8 bilinear: rewrite in intrinsics
Johann [Wed, 24 Oct 2018 19:22:35 +0000 (12:22 -0700)]
vp8 bilinear: rewrite in intrinsics

8x8 is 15% faster than the assembly. 8x4 is 200% faster than MMX.

Remove MMX version.

Change-Id: I55642ebd276db265911f2c79616177a3a9a7e04f

6 years agoMerge "Clean up vpx_dsp/x86/convolve_sse2.h"
Chi Yo Tsai [Wed, 24 Oct 2018 16:36:20 +0000 (16:36 +0000)]
Merge "Clean up vpx_dsp/x86/convolve_sse2.h"

6 years agoEnable tpl model to support multi-layer ARF
Jingning Han [Wed, 24 Oct 2018 03:30:35 +0000 (20:30 -0700)]
Enable tpl model to support multi-layer ARF

Enable temporal dependency model for the base layer ARF. It
improves the multi-layer ARF compression performance (results
are tested in speed 0 vbr mode):

         avg PSNR    overall PSNR     SSIM
lowres   -0.40%       -0.46%         -0.32%
midres   -0.59%       -0.68%         -0.45%
720p     -0.55%       -0.59%         -1.07%

Change-Id: I7790b89ccfb6e61f9b7965f34d348c7440220dd0

6 years agoAdd AVX2 support for 4-tap interpolation filter.
chiyotsai [Tue, 23 Oct 2018 19:42:21 +0000 (12:42 -0700)]
Add AVX2 support for 4-tap interpolation filter.

Performance:
     | 4X4 | 8X8 |16X16|64X64|
2 DIM|1.491|1.902|1.772|1.479|
 HORZ|1.145|1.521|1.757|1.497|
 VERT|1.176|1.614|1.707|1.467|

Each number in the chart above is 8-tap function time / 4-tap function time.

The framerate tested on jets.y4m for 100 frames on speed 1 increased from 3.72
fps to 3.91 fps (about 5% increase).

Change-Id: Ic0ad275cf32fafeefd0a89811badd8adff2134a0

6 years agoClean up vpx_dsp/x86/convolve_sse2.h
chiyotsai [Thu, 18 Oct 2018 16:51:56 +0000 (09:51 -0700)]
Clean up vpx_dsp/x86/convolve_sse2.h

Removes unnecesssary includes and reword some functions/comments.

Change-Id: Ied557d7faa9d845d38255e6e3e0e3fe1395276e1

6 years agoReset frame udpate flags after qp estimate in tpl
Jingning Han [Tue, 23 Oct 2018 23:24:28 +0000 (16:24 -0700)]
Reset frame udpate flags after qp estimate in tpl

After the frame quantizer estimate run in tpl model, reset the
actual value assigned to the current coding frame. This would
avoid certain frame update flags being overwritten by different
frame types' update.

Change-Id: Idde2ba1108f1f68747b14149b211f882965c99f0

6 years agoMerge "Use 8-tap interp filter in temporal filtering"
Yunqing Wang [Tue, 23 Oct 2018 22:29:46 +0000 (22:29 +0000)]
Merge "Use 8-tap interp filter in temporal filtering"

6 years agoUse 8-tap interp filter in temporal filtering
Yunqing Wang [Tue, 23 Oct 2018 19:30:13 +0000 (12:30 -0700)]
Use 8-tap interp filter in temporal filtering

Used 8-tap interp filter in temporal filtering to achieve more accurate
motion search result. Using 8-tap sharp gave slight better result than
using 8-tap regular.

Speed 0 borg test showed that
        avg_psnr:  ovr_psnr:    ssim:
hdres:  -0.160      -0.157     -0.173
midres: -0.083      -0.061     -0.183
lowres: -0.077      -0.099     -0.204

Speed test didn't see noticeable encoder time changes.

Change-Id: I97dc3c4864b5a5675a6c1e3952799b81eedd7d93

6 years agoMerge "Remove empty else branch in mode_estimation"
Jingning Han [Tue, 23 Oct 2018 19:16:41 +0000 (19:16 +0000)]
Merge "Remove empty else branch in mode_estimation"

6 years agoBypass processing on use existing frame
Jingning Han [Mon, 22 Oct 2018 23:25:43 +0000 (16:25 -0700)]
Bypass processing on use existing frame

The use of show existing frame requries no further operation on
that coding frame. Bypass the corresponding process.

Change-Id: Ia092027a8a543be0ca54c00b4d51e453039712b8

6 years agoFix frame offset computation for GOP extension
Jingning Han [Mon, 22 Oct 2018 21:21:48 +0000 (14:21 -0700)]
Fix frame offset computation for GOP extension

Properly compute the extended GOP frames' buffer offsets.

Change-Id: I9aed14f4b8d623f1832e782828dce07aa546507d

6 years agoRefactor gop_length use case in tpl model
Jingning Han [Mon, 22 Oct 2018 17:37:09 +0000 (10:37 -0700)]
Refactor gop_length use case in tpl model

Make it support both single- and multi-layer ARF GOP structure.

Change-Id: I760a95804d1b583b057120f6d6be65195a0e6c19

6 years agoRemove empty else branch in mode_estimation
Jingning Han [Tue, 23 Oct 2018 05:51:48 +0000 (22:51 -0700)]
Remove empty else branch in mode_estimation

Change-Id: Iefa184aae80b920b054e3e922a77244c2b0d4b61

6 years agoMerge "Use the proper gfu_boost factor to compute rd_mult"
Jingning Han [Tue, 23 Oct 2018 02:28:22 +0000 (02:28 +0000)]
Merge "Use the proper gfu_boost factor to compute rd_mult"

6 years agoUse the proper gfu_boost factor to compute rd_mult
Jingning Han [Mon, 22 Oct 2018 16:28:04 +0000 (09:28 -0700)]
Use the proper gfu_boost factor to compute rd_mult

Update the Lagrangian multiplier according to the gfu_boost factor
assigned per frame. It improves the multi-layer ARF compression
performance (results below shown for speed 0):

         avg PSNR      overall PSNR      SSIM
lowres    -0.08%          0.02%         -0.28%
midres    -0.08%          0.03%         -0.22%
hdres     -0.19%         -0.10%         -0.39%
nflx2k    -0.29%         -0.18%         -0.85%

Change-Id: Ifeb4b14918f880ba011ea41c1454ab00504f8855

6 years agoMerge "ML_VAR_PARTITION: enable at speed 5"
Hui Su [Fri, 19 Oct 2018 16:48:40 +0000 (16:48 +0000)]
Merge "ML_VAR_PARTITION: enable at speed 5"