webm/libvpx.git
42 hours agotemporal_svc encoder: Change default setting for layering_mode = 0. 64/72264/1 master
Marco [Wed, 26 Nov 2014 19:11:34 +0000 (11:11 -0800)]
temporal_svc encoder: Change default setting for layering_mode = 0.

Change-Id: I5f723ada144d0bd82ee700ef28a4916fd64ba5e4

45 hours agoMerge "Separate rate_correction_factor for boosted GFs"
Yaowu Xu [Wed, 26 Nov 2014 15:56:43 +0000 (07:56 -0800)]
Merge "Separate rate_correction_factor for boosted GFs"

2 days agoMerge "tests: enable resize_test for vp8"
James Zern [Tue, 25 Nov 2014 23:12:41 +0000 (15:12 -0800)]
Merge "tests: enable resize_test for vp8"

2 days agoSeparate rate_correction_factor for boosted GFs 33/72233/2
Yaowu Xu [Fri, 21 Nov 2014 19:55:27 +0000 (11:55 -0800)]
Separate rate_correction_factor for boosted GFs

When the golden frame is boosted, the rate correction factor is not
correlated well with other inter frames even in CBR mode. This commit
changes to use GF specific rate_correction_factor when gf_cbr_boost
is greater than 20%.

Change-Id: I6312c1564387bcacc11f4c5e8a9cfdc781b5c3ab

2 days agoRemove unfreed memory in quantize_test.cc 56/72256/2
Johann [Tue, 25 Nov 2014 20:06:56 +0000 (12:06 -0800)]
Remove unfreed memory in quantize_test.cc

vp8_config doesn't need to be on the stack. Just use a local copy.

Change-Id: Ia241b1d87fd9422556d957f3ea43ad71eb5c414a

2 days agoCosmetic change in vp9_pick_inter_mode 49/72249/2
Jingning Han [Mon, 24 Nov 2014 23:54:55 +0000 (15:54 -0800)]
Cosmetic change in vp9_pick_inter_mode

Change-Id: Ic072585ebffdb36982ed7b8b9f875ca6c1c656c4

2 days agoAdaptively adjust mode test kick-off thresholds in RTC coding 47/72247/2
Jingning Han [Mon, 24 Nov 2014 22:40:42 +0000 (14:40 -0800)]
Adaptively adjust mode test kick-off thresholds in RTC coding

This commit allows the encoder to increase the mode test kick-off
thresholds if the previous best mode renders all zero quantized
coefficients, thereby saving motion search runs when possible.
The compression performance of speed -5 and -6 is down by -0.446%
and 0.591%, respectively. The runtime of speed -6 is improved by
10% for many test clips.

vidyo1, 1000 kbps
16578 b/f, 40.316 dB, 7873 ms -> 16575 b/f, 40.262 dB, 7126 ms

nik720p, 1000 kbps
33311 b/f, 38.651 dB, 7263 ms -> 33304 b/f, 38.629 dB, 6865 ms

dark720p, 1000 kbps
33331 b/f, 39.718 dB, 13596 ms -> 33324 b/f, 39.651 dB, 12000 ms

mmoving, 1000 kbps
33263 b/f, 40.983 dB, 7566 ms -> 33259 b/f, 40.978 dB, 7531 ms

Change-Id: I7591617ff113e91125ec32c9b853e257fbc41d90

2 days agoMerge "Rework forward txfm/quantization skip system in RTC coding mode"
Jingning Han [Tue, 25 Nov 2014 17:33:57 +0000 (09:33 -0800)]
Merge "Rework forward txfm/quantization skip system in RTC coding mode"

3 days agoMerge "Remove redundant intra mode penalty from vp9_pick_inter_mode"
Jingning Han [Tue, 25 Nov 2014 06:13:44 +0000 (22:13 -0800)]
Merge "Remove redundant intra mode penalty from vp9_pick_inter_mode"

3 days agovp9_ethread: modify VP9_COMP structure 32/72232/5
Yunqing Wang [Fri, 21 Nov 2014 19:11:06 +0000 (11:11 -0800)]
vp9_ethread: modify VP9_COMP structure

This patch modified struct VP9_COMP. Created a struct ThreadData
to include data that need to be copied for each thread. In
multiple thread case, one thread processes one tile. all threads
share one copy of VP9_COMP,
(refer to VP9_COMP *cpi in the code)
but each thread has its own copy of ThreadData,
(refer to ThreadData *td in the code).
Therefore, within the scope of encode_tiles(), both cpi and td
need to be passed as function parameters.

In single thread case, the FRAME_COUNTS pointer in ThreadData
points to "counts" in VP9_COMMON.

Change-Id: Ib37908b2d8e2c0f4f9c18f38017df5ce60e8b13e

3 days agoMerge "Fix a tautological assert."
Alex Converse [Tue, 25 Nov 2014 00:36:53 +0000 (16:36 -0800)]
Merge "Fix a tautological assert."

3 days agoFix a tautological assert. 48/72248/1
Alex Converse [Mon, 24 Nov 2014 23:00:20 +0000 (15:00 -0800)]
Fix a tautological assert.

Change-Id: I90ad08823e1d038384536fa9f458caadc2c87f38

3 days agoRemove redundant intra mode penalty from vp9_pick_inter_mode 40/72240/3
Jingning Han [Mon, 24 Nov 2014 17:31:10 +0000 (09:31 -0800)]
Remove redundant intra mode penalty from vp9_pick_inter_mode

The intra mode penalty is covered by intra_cost_penalty. This
commit removes the other intra cost threshold, provided that the
constant 50 is negligible in normal rate-distortion cost.

Change-Id: I9b8b7483c43b9a41741622e7057def1f7d51bb72

3 days agoMerge "Key frame non-RD mode decision process"
Jingning Han [Mon, 24 Nov 2014 21:21:56 +0000 (13:21 -0800)]
Merge "Key frame non-RD mode decision process"

3 days agoMerge "Refactored idct routines and headers"
Debargha Mukherjee [Mon, 24 Nov 2014 20:47:03 +0000 (12:47 -0800)]
Merge "Refactored idct routines and headers"

3 days agoMerge "sad_test: initialize bit_depth_ in all cases"
James Zern [Mon, 24 Nov 2014 20:26:15 +0000 (12:26 -0800)]
Merge "sad_test: initialize bit_depth_ in all cases"

3 days agoRefactored idct routines and headers 12/72212/2
Peter de Rivaz [Fri, 24 Oct 2014 07:37:39 +0000 (08:37 +0100)]
Refactored idct routines and headers

This change is made in preparation for a
subsequent patch which adds acceleration
for the highbitdepth transform functions.

The highbitdepth transform functions attempt
to use 16/32bit sse instructions where possible,
but fallback to using the C implementations if
potential overflow is detected.  For this reason
the dct routines are made global so they can be
called from the acceleration functions in the
subsequent patch.

Change-Id: Ia921f191bf6936ccba4f13e8461624b120c1f665
(cherry picked from commit 454342d4e77dbb67f4a3c10f97a57a6fcb46d9a0)

3 days agoKey frame non-RD mode decision process 16/72216/6
Jingning Han [Thu, 20 Nov 2014 19:28:00 +0000 (11:28 -0800)]
Key frame non-RD mode decision process

This commit makes a non-RD coding mode decision process for key
frame coding. It can be optionally turned on in speed -6 and above.

Change-Id: I0847258b392877a0210b4768bef88ebc9ad009b5

3 days agoMerge "Only allow for cyclic refresh (aq=3 mode) for base layer."
Marco [Mon, 24 Nov 2014 15:46:36 +0000 (07:46 -0800)]
Merge "Only allow for cyclic refresh (aq=3 mode) for base layer."

5 days agoMerge "Indent build/make/configure.sh"
Johann [Sat, 22 Nov 2014 21:14:56 +0000 (13:14 -0800)]
Merge "Indent build/make/configure.sh"

5 days agosad_test: initialize bit_depth_ in all cases 37/72237/1
James Zern [Sat, 22 Nov 2014 20:12:59 +0000 (12:12 -0800)]
sad_test: initialize bit_depth_ in all cases

previously 'bit_depth_', which is later used to calculate 'mask_', would
be left uninitialized in non-high-bitdepth builds

Change-Id: Ia72035f4645baf3bb0f191504f491b934cdf1e0e

6 days agoMerge "Fix some minor nits."
Paul Wilkins [Sat, 22 Nov 2014 01:39:43 +0000 (17:39 -0800)]
Merge "Fix some minor nits."

6 days agoIndent build/make/configure.sh 27/72227/9
Johann [Fri, 21 Nov 2014 02:20:06 +0000 (18:20 -0800)]
Indent build/make/configure.sh

Whitespace changes only.

Change-Id: I92eef49db1798bc69ec003fca69995fea560d6fa

6 days agoMerge "Added highbitdepth sse2 acceleration for quantize"
Debargha Mukherjee [Sat, 22 Nov 2014 00:08:47 +0000 (16:08 -0800)]
Merge "Added highbitdepth sse2 acceleration for quantize"

6 days agoMerge changes Ie077edd0,Id31a74fc
Paul Wilkins [Fri, 21 Nov 2014 23:38:32 +0000 (15:38 -0800)]
Merge changes Ie077edd0,Id31a74fc

* changes:
  Remove rate component adjustment for AQ1
  Switch AQ1 segment basis from q ratio to rate ratio.

6 days agoMerge "Add adaptive midpoint for AQ1."
Paul Wilkins [Fri, 21 Nov 2014 23:26:18 +0000 (15:26 -0800)]
Merge "Add adaptive midpoint for AQ1."

6 days agoMerge "Add variance restriction to AQ2."
Paul Wilkins [Fri, 21 Nov 2014 23:25:43 +0000 (15:25 -0800)]
Merge "Add variance restriction to AQ2."

6 days agoOnly allow for cyclic refresh (aq=3 mode) for base layer. 06/72206/4
Marco [Wed, 19 Nov 2014 23:31:29 +0000 (15:31 -0800)]
Only allow for cyclic refresh (aq=3 mode) for base layer.

Condition existed for temporal case, added it for spatial as well.
Issue: https://code.google.com/p/webm/issues/detail?id=878.

Change-Id: I38339207f9a94924f5568a081eabe64f867a686d

6 days agoFix some minor nits. 35/72235/3
Paul Wilkins [Fri, 21 Nov 2014 20:27:52 +0000 (12:27 -0800)]
Fix some minor nits.

Change-Id: Ib8810d431fa20a2c78e0caaa28eb2c99903e60fb

6 days agoMerge "Further AQ1 clean up."
Paul Wilkins [Fri, 21 Nov 2014 20:58:03 +0000 (12:58 -0800)]
Merge "Further AQ1 clean up."

6 days agoRework forward txfm/quantization skip system in RTC coding mode 36/72236/2
Jingning Han [Fri, 21 Nov 2014 20:18:53 +0000 (12:18 -0800)]
Rework forward txfm/quantization skip system in RTC coding mode

This commit allows more aggressive decision to skip forward
transform and quantization for luma component in RTC coding mode.
The chroma components remains going through the normal coding
routine, since they are not included in the non-RD mode search
process.

It reduces the runtime cost by 2% - 10%. In speed -6,
vidyo1 1000 kbps
16576 b/f, 40.281 dB, 8402 ms -> 16576 b/f, 40.323 dB, 7764 ms

nik720p 1000 kbps
33337 b/f, 38.622 dB, 7473 ms -> 33299 b/f, 38.660 dB, 7314 ms

dark720p 1000 kbps
33330 b/f, 39.785 dB, 13505 ms -> 33325 b/f, 39.714 dB, 13105 ms

The compression performance of speed -6 is improved by 0.44% in
PSNR and 1.31% in SSIM.

Change-Id: Iae9e3738de6255babea734e5897f29118bebc6d7

6 days agoMerge "Initial AQ1 restructuring."
Paul Wilkins [Fri, 21 Nov 2014 20:10:03 +0000 (12:10 -0800)]
Merge "Initial AQ1 restructuring."

6 days agoMerge "vp9_thread_test: fix 'had_error' assignment"
James Zern [Fri, 21 Nov 2014 19:35:59 +0000 (11:35 -0800)]
Merge "vp9_thread_test: fix 'had_error' assignment"

6 days agoMerge "Allow DC/H/V/TM on screen content."
Alex Converse [Fri, 21 Nov 2014 17:38:39 +0000 (09:38 -0800)]
Merge "Allow DC/H/V/TM on screen content."

6 days agoMerge "Drop special inter mode selection for screen content."
Alex Converse [Fri, 21 Nov 2014 17:38:02 +0000 (09:38 -0800)]
Merge "Drop special inter mode selection for screen content."

6 days agoMerge "Remove armv5te target"
Johann [Fri, 21 Nov 2014 16:55:22 +0000 (08:55 -0800)]
Merge "Remove armv5te target"

6 days agoRemove rate component adjustment for AQ1 29/72229/2
Paul Wilkins [Fri, 21 Nov 2014 02:32:44 +0000 (18:32 -0800)]
Remove rate component adjustment for AQ1

In AQ1 a rate adjustment was applied for blocks coded with a
deltaq. This tends to skew the partition selection and cause
rate overshoot.

For example, consider a 64x64 super block where some but not all
sub blocks are in a low q segment and some are in a high q segment.
The choice of Q when considering large partition and transform sizes
is defined by the lowest sub block segment id (currently this implies the
lowest Q). If some parts of the larger partition are very hard this will
cause a high rate component.

The correct behavior here is for the rd code to discard the large partition
choice and break down to sub blocks where some have low and some
have high Q.  However the rate correction factor above mask the high
cost of coding at a larger partition size.

Change-Id: Ie077edd0b1b43c094898f481df772ea280b35960

6 days agoSwitch AQ1 segment basis from q ratio to rate ratio. 26/72226/5
Paul Wilkins [Fri, 21 Nov 2014 01:28:21 +0000 (17:28 -0800)]
Switch AQ1 segment basis from q ratio to rate ratio.

In defining the Q deltas for segments in AQ1 use a rate
ratio rather than a q ratio.

Change-Id: Id31a74fcf2b7e55437e42a51c21b3cbcb57028d4

7 days agoAdd adaptive midpoint for AQ1. 18/72218/5
Paul Wilkins [Thu, 20 Nov 2014 23:03:01 +0000 (15:03 -0800)]
Add adaptive midpoint for AQ1.

Make the midpoint variance used in AQ mode 1 segmentation
depend on the overall complexity of the frame in two pass.

Change-Id: I452814ec57f7a32352e41bb250e78066abe952dd

7 days agoCorrectly initialize "ones" value in neon quantize 20/72220/3
Johann [Thu, 20 Nov 2014 21:24:55 +0000 (13:24 -0800)]
Correctly initialize "ones" value in neon quantize

By using 0xff for a short it was not setting the high bits. When
comparing the output with vtst to find non-zero elements it was skipping
vaules which had no low bits set such as -512 / 0xFE00.

Using -8191 as the first element of coeff will generate this condition.

BUG=883

Change-Id: Ia1e10fb809d1e7866f28c56769fe703e6231a657

7 days agoAllow DC/H/V/TM on screen content. 09/72209/5
Alex Converse [Fri, 24 Oct 2014 22:57:48 +0000 (15:57 -0700)]
Allow DC/H/V/TM on screen content.

6.3% better compression
less than 1% compression time increase

Change-Id: Ie83c059436e54c09de9e7c87e06e0a6d40dc38fe

7 days agoDrop special inter mode selection for screen content. 08/72208/5
Alex Converse [Tue, 18 Nov 2014 21:31:49 +0000 (13:31 -0800)]
Drop special inter mode selection for screen content.

Better mode selection was implemented for all content.

Change-Id: I479778ed21d3968892f4dce396c83733583f4f23

7 days agoRemove armv5te target 25/72225/1
Johann [Fri, 21 Nov 2014 01:08:11 +0000 (17:08 -0800)]
Remove armv5te target

All the assembly code has been removed, the tests no longer check for
the target, and android and chrome do not use the targets.

Change-Id: I193993f7b2b0bd6478453402f573ce3606e04e8d

7 days agoMerge "vp9_ethread: move filter_cache out of RD_OPT struct"
Yunqing Wang [Fri, 21 Nov 2014 00:51:31 +0000 (16:51 -0800)]
Merge "vp9_ethread: move filter_cache out of RD_OPT struct"

7 days agoAdd variance restriction to AQ2. 12/72112/7
Paul Wilkins [Fri, 7 Nov 2014 16:32:50 +0000 (16:32 +0000)]
Add variance restriction to AQ2.

Add an additional restriction to bit/complexity based
segmentation based on spatial variance.

Only lower Q when both the number of bits spent
in the initial encoding pass and the spatial complexity are
below a threshold. This will prevent the low Q segments
being used just because there is a surfeit of bits.

Small metrics gains especially opsnr.
derf ~0.2% std-hd ~0.3%

Change-Id: I6a8496d466d673f9b0e2b2ca6304ea7b6d8e1cce

7 days agoFurther AQ1 clean up. 66/72166/4
Paul Wilkins [Thu, 13 Nov 2014 20:35:07 +0000 (12:35 -0800)]
Further AQ1 clean up.

Further patch to restructure AQ mode 1.

Change-Id: I566452a033d047a49a40441a7be24690ea69412d

7 days agoInitial AQ1 restructuring. 51/72151/6
Paul Wilkins [Wed, 12 Nov 2014 20:36:16 +0000 (12:36 -0800)]
Initial AQ1 restructuring.

This is the first of a series of patches to restructure and
improve AQ mode 1 (variance based AQ).

Change-Id: Idcf693131a3ea2459dcfd957a54a65b971fa4a2a

7 days agoMerge "Fix bug in calculating number of mbs with scaling."
Paul Wilkins [Thu, 20 Nov 2014 23:45:41 +0000 (15:45 -0800)]
Merge "Fix bug in calculating number of mbs with scaling."

7 days agotests: enable resize_test for vp8 23/72223/1
James Zern [Thu, 20 Nov 2014 22:34:40 +0000 (14:34 -0800)]
tests: enable resize_test for vp8

this was incorrectly set in test.mk by
93ffd37 Enable and fix resize_test for VP9

the test is now available when using --disable-vp9

Change-Id: I6acf44b0de647b34812ef5e18fd96447cdf9b25d

7 days agoMerge "sad_test: fix vp8-only build"
James Zern [Thu, 20 Nov 2014 22:18:37 +0000 (14:18 -0800)]
Merge "sad_test: fix vp8-only build"

7 days agovp9_thread_test: fix 'had_error' assignment 22/72222/1
James Zern [Thu, 20 Nov 2014 22:16:28 +0000 (14:16 -0800)]
vp9_thread_test: fix 'had_error' assignment

worker hooks return false on error, fix the assignment in Execute() used
in the TestSerialInterface test

Change-Id: I93c2e45f270330ae6d35a3a303411c4ee0f31337

7 days agoMerge "vp9_ethread: move max/min partition size to mb struct"
Yunqing Wang [Thu, 20 Nov 2014 22:00:37 +0000 (14:00 -0800)]
Merge "vp9_ethread: move max/min partition size to mb struct"

7 days agovp9_ethread: move filter_cache out of RD_OPT struct 19/72219/2
Yunqing Wang [Thu, 20 Nov 2014 20:42:36 +0000 (12:42 -0800)]
vp9_ethread: move filter_cache out of RD_OPT struct

Similar to mask_filter, the filter_cache in RD_OPT struct can be
moved out, and declared as a local variable since it is only
used in pick_inter_mode functions.

Change-Id: I412b99cca82bade07ac912064ec03dd1de6b2c17

7 days agoMerge "vp9_ethread: change mask_filter to a local variable"
Yunqing Wang [Thu, 20 Nov 2014 21:02:55 +0000 (13:02 -0800)]
Merge "vp9_ethread: change mask_filter to a local variable"

7 days agoMerge "Revert "vp9_ethread: include a pointer to mb in VP9_COMP""
Yunqing Wang [Thu, 20 Nov 2014 20:27:34 +0000 (12:27 -0800)]
Merge "Revert "vp9_ethread: include a pointer to mb in VP9_COMP""

7 days agoFix bug in calculating number of mbs with scaling. 17/72217/1
Paul Wilkins [Thu, 20 Nov 2014 20:22:53 +0000 (12:22 -0800)]
Fix bug in calculating number of mbs with scaling.

Correct calculation of number of mbs in two pass code when
frame resizing is enabled. Always use initial number of mbs if
scaling is enabled, as this is what was used in the first pass.

Change-Id: I49a4280ab5a8b1000efcc157a449a081cbb6d410

7 days agovp9_ethread: change mask_filter to a local variable 14/72214/1
Yunqing Wang [Thu, 20 Nov 2014 17:41:49 +0000 (09:41 -0800)]
vp9_ethread: change mask_filter to a local variable

The mask_filter in RD_OPT struct is used to record rd result in
filter decision. It is only used in pick_inter_mode functions,
and is removed from the struct and declared as a local variable.

Change-Id: I3c95c8632ba7241591ce00ef2ef5677b5e297d7b

7 days agovp9_ethread: move max/min partition size to mb struct 13/72213/1
Yunqing Wang [Thu, 20 Nov 2014 17:24:50 +0000 (09:24 -0800)]
vp9_ethread: move max/min partition size to mb struct

The max_partition_size and max_partition_size are set at the
beginning while setting speed features, and then adjusted at
SB level. Moving them to mb struct ensures there is a local
copy for each thread.

Change-Id: I7dd08dc918d9f772fcd718bbd6533e0787720ad4

7 days agoMerge "Enable ssse3 version of vp9_fdct8x8_quant"
Jingning Han [Thu, 20 Nov 2014 17:06:39 +0000 (09:06 -0800)]
Merge "Enable ssse3 version of vp9_fdct8x8_quant"

7 days agoRevert "vp9_ethread: include a pointer to mb in VP9_COMP" 11/72211/2
Yunqing Wang [Thu, 20 Nov 2014 03:04:07 +0000 (19:04 -0800)]
Revert "vp9_ethread: include a pointer to mb in VP9_COMP"

This reverts commit 6906d218ddd1af97228a797f4558e402231d94f1.

Another way will be used to handle mb struct.

Change-Id: Ic1111a46b2b1ee00f8f9e3fcd4cf3eb6030b2dc4

8 days agoAdded highbitdepth sse2 acceleration for quantize 69/72169/4
Peter de Rivaz [Thu, 16 Oct 2014 12:38:46 +0000 (13:38 +0100)]
Added highbitdepth sse2 acceleration for quantize

Also includes block error.

(This patch is mostly cherry picked from
commit db7192e0b014a331a1dcb102c8a1148e9f0e1081)

Change-Id: Idef18f90b111a0d0c9546543d3347e551908fd78

8 days agoMerge "Changes SvcContext_t to SvcContext"
Deb Mukherjee [Thu, 20 Nov 2014 07:22:18 +0000 (23:22 -0800)]
Merge "Changes SvcContext_t to SvcContext"

8 days agoEnable ssse3 version of vp9_fdct8x8_quant 95/72195/6
Jingning Han [Tue, 18 Nov 2014 20:37:59 +0000 (12:37 -0800)]
Enable ssse3 version of vp9_fdct8x8_quant

It improves the speed performance of vp9_fdct8x8_quant_sse2 by
about 5%.

Change-Id: I74b093ba4d81df64caf71ac7693f3d917f673097

8 days agosad_test: fix vp8-only build 00/72200/2
James Zern [Wed, 19 Nov 2014 03:27:06 +0000 (19:27 -0800)]
sad_test: fix vp8-only build

ROUND_POWER_OF_TWO() is defined in vp9 headers currently, avoid it in
non-high-bitdepth code

Change-Id: Ic28b8f95ef7964800475ee8b35be5f9cea9afab6

8 days agoAdd a reset to rc tracking for dropped frames 07/72207/1
Yaowu Xu [Wed, 19 Nov 2014 23:32:11 +0000 (15:32 -0800)]
Add a reset to rc tracking for dropped frames

VP9/DatarateTestVP9Large.ChangingDropFrameThresh/[34] fails post the
merge of commit#ffa06b37. This commit adds reset of rc tracking info
when frame is dropped, and fixes the causes of the bad interaction
between the tests and the previous commit.

Change-Id: I848acfd9fcb336359662274325190f94aac76eae

8 days agoMerge "Combine fdct8x8 and quantization process"
Jingning Han [Wed, 19 Nov 2014 19:17:44 +0000 (11:17 -0800)]
Merge "Combine fdct8x8 and quantization process"

8 days agoMerge "Add sse2 version for vp9_quantize_fp"
Jingning Han [Wed, 19 Nov 2014 19:17:36 +0000 (11:17 -0800)]
Merge "Add sse2 version for vp9_quantize_fp"

9 days agoCombine fdct8x8 and quantization process 93/72193/5
Jingning Han [Tue, 18 Nov 2014 19:53:14 +0000 (11:53 -0800)]
Combine fdct8x8 and quantization process

This commit reworks the forward transform and quantization process
for 8x8 block coding. It combines the two operations in a single
function to save a store/load stage of the original transform
coefficients. Overall the speed -6 is slightly faster (around 1%
range). The compression performance of speed -6 is improved by
3.4%.

Change-Id: Id6628daef123f3e4649248735ec2ad7423629387

9 days agoMerge "Prevent severe rate control errors in CBR mode"
Yaowu Xu [Wed, 19 Nov 2014 00:56:06 +0000 (16:56 -0800)]
Merge "Prevent severe rate control errors in CBR mode"

9 days agoChanges SvcContext_t to SvcContext 94/72194/1
Deb Mukherjee [Tue, 18 Nov 2014 20:25:31 +0000 (12:25 -0800)]
Changes SvcContext_t to SvcContext

Fixes a borg build.

Change-Id: I8cb510577e8d8bbc7a2e64f9e1bdfe883f49cb61

9 days agoMerge "Modify active_worst_quality setting for one pass CBR."
Marco [Tue, 18 Nov 2014 19:22:18 +0000 (11:22 -0800)]
Merge "Modify active_worst_quality setting for one pass CBR."

9 days agoMerge "change to call vp9_refining_search_sad() directly"
Yaowu Xu [Tue, 18 Nov 2014 17:14:02 +0000 (09:14 -0800)]
Merge "change to call vp9_refining_search_sad() directly"

9 days agoPrevent severe rate control errors in CBR mode 90/72190/2
Yaowu Xu [Tue, 18 Nov 2014 16:52:21 +0000 (08:52 -0800)]
Prevent severe rate control errors in CBR mode

In rare cases, the interaction between rate correction factor and Q
choices may cause severe oscillating frame sizes that are way off
target bandwidth. This commit adds tracking of rate control results
for last two frames, and use the information to prevent oscillating
Q choices.

Change-Id: I9a6d125a15652b9bcac0e1fec6d7a1aedc4ed97e

9 days agoAdd sse2 version for vp9_quantize_fp 91/72191/1
Jingning Han [Tue, 18 Nov 2014 16:58:09 +0000 (08:58 -0800)]
Add sse2 version for vp9_quantize_fp

vp9_quantize_fp is the quantization process used by rtc coding
mode. This commit adds a sse2 implementation of it. The
implementation is modified based on vp9_quantize_b_sse2. No speed
difference from ssse3 version.

Change-Id: I24949c5b27df160b4f35117d28858d269454e64a

10 days agoMerge "Add empty pointer check to pred buffering in rtc coding mode"
Jingning Han [Tue, 18 Nov 2014 01:40:54 +0000 (17:40 -0800)]
Merge "Add empty pointer check to pred buffering in rtc coding mode"

10 days agoModify active_worst_quality setting for one pass CBR. 34/72134/6
Marco [Tue, 11 Nov 2014 19:01:55 +0000 (11:01 -0800)]
Modify active_worst_quality setting for one pass CBR.

Current setting had active_worst_quality set too high (close to worst_quality)
for first frame(s) following first key frame. This changes that to be somewhat
more aggressive in allowing active_worst_quality to be lower following key frame.

Also remove the 4/5 reduction in active_worst for key frame as
this should be set by the user qp_max setting.

Change-Id: I0530b3ddcc85c00e3eb7568de1b14a31206c4a4c

10 days agochange to call vp9_refining_search_sad() directly 86/72186/2
Yaowu Xu [Mon, 17 Nov 2014 19:25:43 +0000 (11:25 -0800)]
change to call vp9_refining_search_sad() directly

The function pointer in compressor instance does not change, so this
commit changes to call the function directly.

Change-Id: I9c9c460e3475711c384b74c9842f0b4f3d037cc5

10 days agoAdd empty pointer check to pred buffering in rtc coding mode 85/72185/1
Jingning Han [Mon, 17 Nov 2014 19:21:42 +0000 (11:21 -0800)]
Add empty pointer check to pred buffering in rtc coding mode

This commit adds a check condition to the prediction buffering
operation used in the rtc coding mode. This resolves a unit test
warning in example/vpx_tsvc_encoder_vp9_mode_7.

Change-Id: I9fd50d5956948b73b53bd8fc5a16ee66aff61995

10 days agoMerge "Code cleanup: remove unused members in RD_OPT"
Yunqing Wang [Mon, 17 Nov 2014 17:10:28 +0000 (09:10 -0800)]
Merge "Code cleanup: remove unused members in RD_OPT"

10 days agoMerge "vp9_ethread: combine encoder counts in separate struct"
Yunqing Wang [Mon, 17 Nov 2014 16:57:38 +0000 (08:57 -0800)]
Merge "vp9_ethread: combine encoder counts in separate struct"

13 days agoMerge "Added sse2 acceleration for highbitdepth variance"
Debargha Mukherjee [Sat, 15 Nov 2014 05:11:27 +0000 (21:11 -0800)]
Merge "Added sse2 acceleration for highbitdepth variance"

13 days agoMerge "Visual studio build fix using explicit cast"
James Zern [Sat, 15 Nov 2014 04:26:28 +0000 (20:26 -0800)]
Merge "Visual studio build fix using explicit cast"

13 days agoCode cleanup: remove unused members in RD_OPT 78/72178/1
Yunqing Wang [Sat, 15 Nov 2014 00:33:25 +0000 (16:33 -0800)]
Code cleanup: remove unused members in RD_OPT

These 2 members in RD_OPT were moved to TileDataEnc struct
already, and therefore were removed here.

Change-Id: I22fee3b67f96e473a58e194a7edc76dbd48bfa04

13 days agovp9_ethread: combine encoder counts in separate struct 77/72177/1
Yunqing Wang [Sat, 15 Nov 2014 00:04:15 +0000 (16:04 -0800)]
vp9_ethread: combine encoder counts in separate struct

Several frame counters in encoder are updated at SB level. Combine
those counters and put them in a separate struct, which allows us
to allocate one copy for each thread.

Change-Id: I00366296a13c0ada4d8fa12f5e07728388b6cab7

13 days agoAdded sse2 acceleration for highbitdepth variance 21/72121/6
Peter de Rivaz [Thu, 16 Oct 2014 13:00:54 +0000 (14:00 +0100)]
Added sse2 acceleration for highbitdepth variance

Change-Id: I446bdf3a405e4e9d2aa633d6281d66ea0cdfd79f
(cherry picked from commit d7422b2b1eb9f0011a8c379c2be680d6892b16bc)
(cherry picked from commit 6d741e4d76a7d9ece69ca117d1d9e2f9ee48ef8c)

13 days agoVisual studio build fix using explicit cast 74/72174/1
Deb Mukherjee [Fri, 14 Nov 2014 23:12:01 +0000 (15:12 -0800)]
Visual studio build fix using explicit cast

Change-Id: If74510370723e497f4f33d988b8b398124edf69b

13 days agoMerge "vp9_ethread: include a pointer to mb in VP9_COMP"
Yunqing Wang [Fri, 14 Nov 2014 23:11:47 +0000 (15:11 -0800)]
Merge "vp9_ethread: include a pointer to mb in VP9_COMP"

13 days agovp9_ethread: include a pointer to mb in VP9_COMP 73/72173/1
Yunqing Wang [Fri, 14 Nov 2014 20:31:06 +0000 (12:31 -0800)]
vp9_ethread: include a pointer to mb in VP9_COMP

Modified VP9_COMP struct to include MACROBLOCK *mb. This change
makes it feasible in multi-thread case to allocate a mb for each
thread.

Change-Id: I624d6d1aa9c132362200753e5d90b581b1738d6e

13 days agoMerge "Add more rate control stats to vpx_temporal_svc_encoder."
Marco [Fri, 14 Nov 2014 19:36:05 +0000 (11:36 -0800)]
Merge "Add more rate control stats to vpx_temporal_svc_encoder."

13 days agoAdd more rate control stats to vpx_temporal_svc_encoder. 63/72163/5
Marco [Thu, 13 Nov 2014 23:03:45 +0000 (15:03 -0800)]
Add more rate control stats to vpx_temporal_svc_encoder.

Average and variance of actual encoding bitrate over fixed short-time window.

Change-Id: Icc3d6d93792c81e1fb9fd3b56008bf47a503b204

2 weeks agoMerge "vp9_ethread: modify the cyclic refresh struct"
Yunqing Wang [Fri, 14 Nov 2014 02:35:01 +0000 (18:35 -0800)]
Merge "vp9_ethread: modify the cyclic refresh struct"

2 weeks agoMerge "adapt the adjustment limit for rate correction factor in RTC mode"
Yaowu Xu [Thu, 13 Nov 2014 23:50:30 +0000 (15:50 -0800)]
Merge "adapt the adjustment limit for rate correction factor in RTC mode"

2 weeks agovp9_ethread: modify the cyclic refresh struct 59/72159/3
Yunqing Wang [Thu, 13 Nov 2014 21:38:23 +0000 (13:38 -0800)]
vp9_ethread: modify the cyclic refresh struct

Two members in struct CYCLIC_REFRESH
  int64_t projected_rate_sb;
  int64_t projected_dist_sb;
are updated at the superblock level, which makes them shared data
in the multi-thread situation, and requires extra work to handle
them. However, those values are updated and used immediately, and
therefore can be removed. This patch cleaned up the code and
removed the two members.

Change-Id: I2c6ee4552bf49fb63ce590cdb47f9723974fffb1

2 weeks agoMerge "Prepare for dynamic frame resizing in the recode loop"
Adrian Grange [Thu, 13 Nov 2014 23:01:49 +0000 (15:01 -0800)]
Merge "Prepare for dynamic frame resizing in the recode loop"

2 weeks agoMerge "Fix 32 bit build emms problem."
Paul Wilkins [Thu, 13 Nov 2014 23:00:41 +0000 (15:00 -0800)]
Merge "Fix 32 bit build emms problem."

2 weeks agoMerge "Revert "Fix SSSE3 partial_idct_test detection""
Johann [Thu, 13 Nov 2014 22:49:22 +0000 (14:49 -0800)]
Merge "Revert "Fix SSSE3 partial_idct_test detection""

2 weeks agoMerge "Use reconstructed pixels for intra prediction"
Jingning Han [Thu, 13 Nov 2014 21:59:02 +0000 (13:59 -0800)]
Merge "Use reconstructed pixels for intra prediction"

2 weeks agoMerge "Refactor nonrd_use_partition coding process"
Jingning Han [Thu, 13 Nov 2014 21:58:21 +0000 (13:58 -0800)]
Merge "Refactor nonrd_use_partition coding process"

2 weeks agoPrepare for dynamic frame resizing in the recode loop 12/72012/19
Adrian Grange [Mon, 10 Nov 2014 17:51:38 +0000 (09:51 -0800)]
Prepare for dynamic frame resizing in the recode loop

Prepare for the introduction of frame-size change
logic into the recode loop.

Separated the speed dependent features into
separate static and dynamic parts, the latter being
those features that are dependent on the frame size.

Change-Id: Ia693e28c5cf069a1a7bf12e49ecf83e440e1d313

2 weeks agoFix 32 bit build emms problem. 39/72139/5
Paul Wilkins [Wed, 12 Nov 2014 17:36:49 +0000 (09:36 -0800)]
Fix 32 bit build emms problem.

Add extra vp9_clear_system_state() calls to fix
double / mmx issue introduced into first pass
code for 32 bit builds.

Change-Id: I84cd2986b80d83650a091ab25c43755efeb82e03