Commit Graph

5384 Commits

Author SHA1 Message Date
Dmitry Kovalev
b35b426e58 Merge "Fixing constant value used to calculate frame pts and duration." 2014-05-13 15:10:45 -07:00
Dmitry Kovalev
32c51e8222 Merge "Inlining setup_key_frame() and setup_inter_frame()." 2014-05-13 15:00:22 -07:00
Adrian Grange
9e0b5abc30 Merge "Move frame stats output to after update" 2014-05-13 10:07:36 -07:00
Dmitry Kovalev
889887aa68 Inlining setup_key_frame() and setup_inter_frame().
Change-Id: I6c0f0f91c189c041559f724f4faeada144390353
2014-05-12 17:13:07 -07:00
Yaowu Xu
38f3cf125e Merge "vp9_pickmode.c: Removed unused function parameters" 2014-05-12 14:56:54 -07:00
Yaowu Xu
69d2fb856b Merge "Change to use proper type" 2014-05-12 14:05:52 -07:00
Yaowu Xu
c80287f02d Merge "fixed comparison of different int types." 2014-05-12 14:05:47 -07:00
Yaowu Xu
3e81ee26cc vp9_pickmode.c: Removed unused function parameters
Change-Id: I4ec07d3935dc56ca16ea4ba1e5730b09f1bf1f21
2014-05-12 12:10:23 -07:00
Yaowu Xu
26cb7bee7d Change to use proper type
Change-Id: If1afb9f3eaec88079d1d97907870409bce691c2a
2014-05-12 11:56:17 -07:00
Yaowu Xu
77c754df79 Merge "Cleanup unused parameters" 2014-05-12 11:05:54 -07:00
Yaowu Xu
817ac44aa2 Merge "Fixed comparison of different int types" 2014-05-12 11:05:48 -07:00
Yaowu Xu
7145e3f797 fixed comparison of different int types.
Change-Id: Iaa5543eb9e17589d46d9ed2bc363e2646ed4e01e
2014-05-12 10:01:00 -07:00
Johann
dbf61d2f01 Merge "Only build neon assembly for armv7 targets" 2014-05-12 09:57:20 -07:00
Yaowu Xu
d22faee98d Cleanup unused parameters
Remove unused function parameters from:
vp9_encodemb.c
vp9_encodeframe.c

Change-Id: I07725e5099cf98486e70c9c70babb0cd785c58a9
2014-05-12 09:53:11 -07:00
Yaowu Xu
eccad3897e Fixed comparison of different int types
Change-Id: I499635967faaf4ec1c498e88d09b2d3f2a4badba
2014-05-12 09:40:11 -07:00
Johann
ce23931a3f Only build neon assembly for armv7 targets
Allow selectively building just the intrinsics for armv8

Change-Id: I2f29b2e4508b8b8e5649c2906b3159ad1d4ec477
2014-05-12 08:52:02 -07:00
Paul Wilkins
702e25839e Clean up of firstpass.c
Re-factor duplicate code.
Add two pass check for use of section_intra_rating as
it is un-initialised in the 1 pass and rt case.

Change-Id: I93120796f07961b8a21fb26e1a9f0d3d13949994
2014-05-12 12:32:42 +01:00
Paul Wilkins
00d54aa331 First pass clean up.
One of a series of changes to clean up two pass
allocation as precursor to support for multiple arf
or boosted frames per GF/ARF group.

This change pulls out the calculation of the total bits
allocated to a GF/ARF group into a function, to aid
readability and reduce the line count for define_gf_group().

This change should have no material impact on output.

Change-Id: I716fba08e26f9ddde3257e7d9b188453791883a3
2014-05-12 12:31:02 +01:00
Adrian Grange
3305909bc2 Merge "Fix check of debug counts for corrupt frame" 2014-05-11 09:13:55 -07:00
James Zern
e975791564 Merge "vp9_dx_iface: subtract ptrs to validate frame_size" 2014-05-10 11:05:31 -07:00
Jingning Han
6d065ba3cb Merge "Apply constrained partition search range to non-RD mode decision" 2014-05-10 10:39:26 -07:00
Dmitry Kovalev
1b60a14ef0 Merge "Removing VP9DecoderConfig." 2014-05-09 19:13:08 -07:00
Jingning Han
734c5ffa2c Apply constrained partition search range to non-RD mode decision
This commit enables a chessboard pattern for partition search. All
the black blocks run regular partition search ranging from 8x8 to
32x32. The rest white blocks take the nearby blocks' information
to adaptively decide the effective search range.

The compression performance for rtc set at speed -5 is down by 1.5%.
For pedestrian 1080p at speed -5, the runtime goes from 41594 ms to
39697 ms, i.e., about 5% faster.

Change-Id: Ia4b96e237abfaada487c743bca08fe1afd298685
2014-05-09 17:07:19 -07:00
Alex Converse
ec8a3272fa Merge "Add an x86inc MMX fwht4x4." 2014-05-09 13:48:49 -07:00
Alex Converse
eb8a348672 Merge "Allow speed 5 lossless" 2014-05-09 13:48:42 -07:00
Adrian Grange
99892e85d5 Fix check of debug counts for corrupt frame
Fixes the idecoder in the case where:
  cm->error_resilient_mode == 0, and
  cm->frame_parallel_decoding_mode == 0, but
  new_fb->corrupted == 1.

The assert in debug_check_frame_counts fails to
take into account the case of a corrupt frame.

Change-Id: Idf318a68458cc88d65d6f3f408a10d8ffe87e43f
2014-05-09 10:12:23 -07:00
Jingning Han
9412785b02 Merge changes I3edd4b95,I4514f974,Ie7fa4386
* changes:
  Turn on unit tests for SSSE3 8x8 forward and inverse 2D-DCT
  Change eob threshold for partial inverse 8x8 2D-DCT to 12
  SSSE3 8x8 inverse 2D-DCT with first 10 coeffs non-zero
2014-05-09 09:58:39 -07:00
Alex Converse
09d947ecb9 Allow speed 5 lossless
tx_mode supercedes whatever mechanism is used to push for 16x16
allowing for the use of the 4x4 transform.

Change-Id: I6c3f05ab9fe52050e40cc6303de9334653763289
2014-05-09 09:53:46 -07:00
Adrian Grange
8d2fff7cb4 Move frame stats output to after update
Frame-level stats were being output before the
post-encode update had been carried out.

Change-Id: I67405c67c020cde4468f5db3cee76f7de5b624eb
2014-05-09 09:44:47 -07:00
Yaowu Xu
ff3baaef94 Merge "Skip testing large tx sizes when encoding lossless." 2014-05-09 09:28:05 -07:00
James Zern
a8cfbbe33f vp9_dx_iface: subtract ptrs to validate frame_size
Change-Id: Ic5a6a4a2fec802d9c9c7a71dbae59d5b4d3a8b23
2014-05-08 20:20:20 -07:00
James Zern
4aa7691225 vp9_dx_iface: fix implicit ptrdiff_t -> uint32_t conversion
Change-Id: Ib773aac06b22009408cdd29a73190b276a53fcf6
2014-05-08 20:19:49 -07:00
Alex Converse
4bb70ea136 Skip testing large tx sizes when encoding lossless.
Change-Id: Ib7a17893f6dc0f501b1252ae32804ede144dde37
2014-05-08 18:07:14 -07:00
Dmitry Kovalev
0dacecaf20 Removing VP9DecoderConfig.
We only used two members from that struct: max_threads and inv_tile_order.
Moving them directly to VP9Decoder struct.

Change-Id: If696a4e5b5b41868a55f3cc971e1d7c1dd9d5f69
2014-05-08 16:24:36 -07:00
Dmitry Kovalev
55e4b76569 Merge "Simplifying decoder_decode() function." 2014-05-08 15:01:40 -07:00
Alex Converse
b5422fab46 Add an x86inc MMX fwht4x4.
Change-Id: Ib0a73d4863478f9b8a00976379d25d2f6ebbb197
2014-05-08 12:01:27 -07:00
Jingning Han
41a350a83d Change eob threshold for partial inverse 8x8 2D-DCT to 12
The scanning order has the first 12 coefficients of the 8x8 2D-DCT
sitting in the top left 4x4 block. Hence the partial inverse 8x8
2D-DCT allows to handle cases with eob below 12.

The overall runtime of the inverse 8x8 2D-DCT unit is reduced from
166 cycles (using SSE2) to 150 cycles (using SSSE3).

Change-Id: I4514f9748042809ac84df4c14382c00f313f1cd2
2014-05-08 09:48:58 -07:00
Frank Galligan
07f9fa4336 Merge "vp9_pickmode: Fix signed/unsigned mismatch." 2014-05-08 08:49:12 -07:00
Frank Galligan
aa3d59fcf9 Merge "vp9_firstpass: Fix MSVC data loss warning." 2014-05-08 08:49:04 -07:00
Tom Finegan
86a89d740e vp9_svc_layercontext: Clean up function definition mismatch.
vp9_is_upper_layer_key_frame() definition does not match declaration--
it was missing the second const.

Change-Id: I71312579eb443be1924b8b06d8b3177c3dcb40f3
2014-05-07 18:32:27 -07:00
Tom Finegan
7811039d47 vp9_firstpass: Fix MSVC data loss warning.
Change-Id: I5307fe85ee081fbcfdf55a974f7eba1e0070c723
2014-05-07 18:27:29 -07:00
Tom Finegan
0fc027ad33 vp9_pickmode: Fix signed/unsigned mismatch.
Avoids getting reverted on the next libvpx chrome roll attempt.

Change-Id: I2431bd601111a78320af5f211f6413009674f1c8
2014-05-07 18:20:29 -07:00
Jingning Han
9e7b09bc5d SSSE3 8x8 inverse 2D-DCT with first 10 coeffs non-zero
This commit enables ssse3 assembly implementation of the 8x8
inverse 2D-DCT with only first 10 coefficients non-zero. The
average runtime for this unit goes down from 198 cycles to 129
cycles (34.8% faster).

Change-Id: Ie7fa4386f6d3a2fe0d47a2eb26fc2a6bbc592ac7
2014-05-07 17:40:02 -07:00
Jingning Han
7af0e984b8 Merge "Tune rate-distortion modeling to account for frame light change" 2014-05-07 17:36:00 -07:00
Yaowu Xu
5a65790814 Merge "Remove separate arf and gf minq tables." 2014-05-07 17:04:49 -07:00
Yaowu Xu
9b4d0e81a2 Merge "Rate control adjustment." 2014-05-07 17:04:41 -07:00
Dmitry Kovalev
dc554abc59 Merge "Cleaning up vp9_encode_frame() function." 2014-05-07 14:10:07 -07:00
Dmitry Kovalev
e91616a5d4 Merge "Using SPEED_FEATURES instead of VP9_COMP in vp9_init_search_range()." 2014-05-07 14:07:33 -07:00
Paul Wilkins
ed30c8eff2 Remove separate arf and gf minq tables.
Merged minq tables for arf and gf cases.
These tables were almost the same and for
VBR the arf table was not used at all.

Change-Id: Ie3c87e91dab613cf06f6945ac1ace0e0e4213d34
2014-05-07 14:00:15 -07:00
Paul Wilkins
38ff3e328f Rate control adjustment.
Small adjustment to the active Q range calculations.
These changes should slightly extend the available Q range
for KF/GF/ARF and narrow it for other frames.

The results for this change in isolation are broadly positive
for SSIM and average PSNR and slightly up but mixed for opsnr.

derf +0.293% opsnr, +1.286% SSIM
std-hd + 0.528% opsnr, + 1.746% SSIM
yt +0.056% opsnr, +0.457% SSIM
yt-hd -0.147% opsnr, + 0.226% SSIM

Change-Id: If065280342027ecc5d44b49fc1d440dfef041002
2014-05-07 13:56:05 -07:00