Граф коммитов

8711 Коммитов

Автор SHA1 Сообщение Дата
Jingning Han e211244b02 Merge "Use dedicated variables to store the best mode" 2014-02-26 09:20:13 -08:00
Dmitry Kovalev 55a1269bd9 Merge "Removing redundant vp9_clear_system_state() call." 2014-02-26 09:19:54 -08:00
Dmitry Kovalev b13ee81d18 Merge "Removing vp9_optimize_b() function." 2014-02-26 09:19:28 -08:00
Jim Bankoski d496cd3074 remove ioc problem with rdcost calculation
Change-Id: Idd800fae21508699e9d973134817493763a65042
2014-02-26 07:46:56 -08:00
James Zern adbb881497 dct16x16_test: add NEON functions
note not all functions have NEON implementations:
- fdct/fht/iht

Change-Id: I6bb7ffe89b8cc23b642e19caf22f04f2d5e39087
2014-02-25 23:11:49 -08:00
Marco Paniconi f61b962c1a Add consective frame loss to error_resilience test.
Modify existing test to also check the case of dropping
(i.e., skip decoding) a consecutive list of frames.

Change-Id: Ia8c1195559f952e86e6697996931d3a920c05ae3
2014-02-25 19:08:30 -08:00
James Yu fb5d281bb6 VP8 for ARMv8 by using NEON intrinsics 05
Add dequantizeb_neon.c
- vp8_dequantize_b_loop_neon

vpxdec  --summary --noblit ../videos/tears_of_steel_1080p.webm
Before => After, 13.25 => 13.23 (fps)

Change-Id: Iebe3b0c6ed2359c778b0570763c5681ae25fef0c
Signed-off-by: James Yu <james.yu@linaro.org>
2014-02-26 10:16:00 +08:00
Dmitry Kovalev abe21e42e7 Removing redundant vp9_clear_system_state() call.
This function is already called in vp9_initialize_rd_consts().

Change-Id: I88c0484a7274660287fa25fdf5cdb721d3284e01
2014-02-25 18:11:42 -08:00
Dmitry Kovalev 8d80c262d1 Using clamp() for active_{best, worst}_quality.
Change-Id: Ia79f975f54829b8ae6b568ca5c05f4f1f541706f
2014-02-25 18:01:49 -08:00
James Yu 28b2f82f97 VP8 for ARMv8 by using NEON intrinsics 04
Add dequant_idct_neon.c
- vp8_dequant_idct_add_neon

vpxdec  --summary --noblit ../videos/tears_of_steel_1080p.webm
Before => After, 13.25 => 13.22 (fps)

Change-Id: Id48f39e1da58dd3d8d37658e94989411997f4f7c
Signed-off-by: James Yu <james.yu@linaro.org>
2014-02-26 09:59:23 +08:00
hkuang 1c4e449133 Fix a bug in intra prediction due to change in
25e5552630.

Change-Id: I17ac67c3ced91ad4f057b296f7e8dc86a3389f26
2014-02-25 17:54:33 -08:00
Dmitry Kovalev 7d121ea86e Cleaning up vp9_rdopt.c.
Change-Id: Ic7553a037493107c33cfb9815178a54095a289ec
2014-02-25 17:46:27 -08:00
James Yu d749ab6221 VP8 for ARMv8 by using NEON intrinsics 03
Add dc_only_idct_add_neon.c
- vp8_dc_only_idct_add_neon

vpxdec  --summary --noblit ../videos/tears_of_steel_1080p.webm
Before => After, 13.25 => 13.24 (fps)

Change-Id: I5e9e277ec3a3ca67e13c8cc4c324a6fbe8a897fc
Signed-off-by: James Yu <james.yu@linaro.org>
2014-02-26 09:28:29 +08:00
Dmitry Kovalev 0444deb8d1 Removing vp9_optimize_b() function.
Change-Id: If859c9d58a8fe3f478d409b9e9ca59bcee8fd35d
2014-02-25 17:23:01 -08:00
Jingning Han c0e99f84f9 Use sub-pixel accuracy prediction non-RD mode
This commit builds the actual prediction block in sub-pixel accuracy
and uses which to calculate SAD for non-RD mode decision. In the trail
run on pedestrian_area at 1080p, rtc speed -7 runtime goes from
23495 ms -> 25107 ms (7% slower). The compression performance is
improved by 20.57% for rtc test set.

Change-Id: I438589cd103fe99f1b50c2d1939ac6ca43fa0157
2014-02-25 16:29:20 -08:00
Johann 0eca2cd3c8 Merge "VP8 for ARMv8 by using NEON intrinsics 02" 2014-02-25 14:53:55 -08:00
Alex Converse cfd0eeeebb Merge "Shrink dct_value_cost[] to int16_t." 2014-02-25 14:30:33 -08:00
Jingning Han fd49e9678e Use dedicated variables to store the best mode
Use a set of dedicated variables to buffer the current best mode
in non-RD mode decision. This allows to use mode_info for more
complicated test in the non-RD process.

Change-Id: I6024c9feb0662afd3eb29f7017f7b5a5446f303f
2014-02-25 13:56:12 -08:00
Deb Mukherjee 276e568997 Merge "Adds variance based fixed size partitioning" 2014-02-25 13:46:21 -08:00
Alex Converse 1075c49a3f Shrink dct_value_cost[] to int16_t.
The largest value is 13358.

Change-Id: I7a6b024a92b6250933d9ebc0cad066b966c96bd4
2014-02-25 13:07:18 -08:00
Deb Mukherjee 10bae82510 Adds variance based fixed size partitioning
Adds a method for determining a fixed size partition based on
variance of a 64x64 SB. This method is added to rtc speed 6.
Also fixes a bug in rtc_use_partition() and includes some
refactoring related to partitioning search, and some cosmetics.

Currently compared to speed 5, the coding efficiency of speed 6
is -19% and that of speed 7 is -55%, in cbr mode.

Change-Id: I057e04125a8b765906bb7d4bf7a36d1e575de7c6
2014-02-25 12:20:59 -08:00
Dmitry Kovalev 4632a96d97 Merge "Using vp9_subtract_plane instead of vp9_subtract_{sb, sby, sbuv}." 2014-02-25 11:06:05 -08:00
Dmitry Kovalev 8fc8583a4c Merge "Consistent names for reference_mode functions." 2014-02-25 11:04:37 -08:00
Dmitry Kovalev 7bca32a6a3 Merge "Changing vp9_full_search_sad{, x3, x8} signatures." 2014-02-25 10:51:17 -08:00
Dmitry Kovalev 55214924ac Merge "Cleaning up vp9_refining_search_sad() function." 2014-02-25 10:44:19 -08:00
Tom Finegan 008ef8dd0d Merge "Change dx_time data type in vpxdec.c" 2014-02-25 09:44:31 -08:00
Adrian Grange da6b224e3c Merge "Trap request to display non-existant frame" 2014-02-25 08:48:14 -08:00
Yunqing Wang da47b3415b Merge "Disable adaptive pred filter for non-split mode" 2014-02-25 08:28:03 -08:00
Yaowu Xu b67bd637e0 Merge changes I7e96d619,I9a7631d5
* changes:
  normalize int64_t high value to INT64_MAX
  resolve issue with arm code failing unit test
2014-02-25 08:22:49 -08:00
Jim Bankoski ec55dd1007 normalize int64_t high value to INT64_MAX
Change-Id: I7e96d6199f882d35357f6a9a08c04ad0af5dc26e
2014-02-25 07:29:48 -08:00
Jim Bankoski df9c8d5bd9 resolve issue with arm code failing unit test
The optimizer did something funny with the code around
line 1412.  Before the call to encode_sb split_dist was
set properly but after it was adjusted and converted to
a negative.

https://code.google.com/p/webm/issues/detail?id=714

Change-Id: I9a7631d5325ade2dc28c1030653a23eecec8721b
2014-02-24 17:34:27 -08:00
James Yu 6d15132742 Change dx_time data type in vpxdec.c
Change dx_time data type to int64_t to prevent
test time overflow when decoding long video.

Change-Id: I3dd5e324a246843e07e635fd25c50e71e385ed70
Signed-off-by: James Yu <james.yu@linaro.org>
2014-02-25 09:17:06 +08:00
Marco Paniconi c30cc9f2e4 vpx_temporal_scalable_patterns: fix to encoding_bitrate computation.
Change-Id: I56fcb82c4334c7591b765b55b3c2912778af6db5
2014-02-24 15:37:22 -08:00
Yunqing Wang fee045d13a Disable adaptive pred filter for non-split mode
If sf->disable_split_mask is DISABLE_ALL_SPLIT, disable
sf->adaptive_pred_interp_filter to avoid unnecessary operations.

Change-Id: Icb59174b2f4e9a3c3c16a696deb8018e5bd999eb
2014-02-24 14:05:06 -08:00
Deb Mukherjee 8cc54d576f Merge "Adds an intermediate speed level for rtc" 2014-02-24 13:35:56 -08:00
Deb Mukherjee 5f542408b4 Adds an intermediate speed level for rtc
Moves the existing speed 6 to speed 7 and adds an
intermediate level 6 which is roughly in between
speeds 6 and 7 in both speed and coding efficiency.
Also includes some minor fixes/adjustments.

Change-Id: I98befc4d82d750e79fe426c457c4a2571f6b6cc7
2014-02-24 12:15:33 -08:00
Adrian Grange 709fecac83 Trap request to display non-existant frame
If show_existing_frame indicates that the decoder should
display an existing (previously decoded) frame, add a
check to make sure that the signaled buffer does contain
a valid decoded frame.

Change-Id: Iac8c686b321827414d69a3f2d0467566911bcba2
2014-02-24 10:19:31 -08:00
Yaowu Xu 05e850cb9e added clamp of segment loop filter level
for ABSDATA mode, so segment loop filter level always fall in valid
range for both Absolute and delta modes.

Change-Id: If90df3411479533dbdab63f8ae088d2f5dd174a9
2014-02-24 09:56:48 -08:00
Yaowu Xu bfaf415ea7 Merge "Added clamp of qindex to valid range" 2014-02-24 08:28:07 -08:00
James Zern 6859d0f0d0 fix vs9 build after obj_int_extract changes
broken since:
1d3ca08 obj_int_extract.bat: Pass obj_int_extract.exe path to obj_int_extract.bat.

Change-Id: Ic3de9f8c2b604e3aa9b905e1bc0b578963e01a4a
2014-02-23 16:08:00 -08:00
Dmitry Kovalev 092f32cf78 Merge "Combining if-else branches from different conditions." 2014-02-23 08:28:43 -08:00
Dmitry Kovalev 2aacc66b66 Merge "Cleaning up vp9_mvref_common.{h, c}." 2014-02-23 08:25:40 -08:00
James Yu 300a3bfc73 VP8 for ARMv8 by using NEON intrinsics 02
Add copymem_neon.c
- vp8_copy_mem16x16_neon
- vp8_copy_mem8x8_neon
- vp8_copy_mem8x4_neon

vpxdec  --summary --noblit ../videos/tears_of_steel_1080p.webm
Before => After, 13.25 => 13.25 (fps)

Change-Id: Ib956b5a20522ff57dc8a580bf0aef7b252bddba6
Signed-off-by: James Yu <james.yu@linaro.org>
2014-02-23 22:56:53 +08:00
Yaowu Xu e22b12e304 Added clamp of qindex to valid range
The qindex for a segment was not clamped in ABSDATA mode, which may
cause invalid memory access if an ill-formed stream has a negative
value in ABSDATA mode. This commit added clamp to make sure qindex
for a segment always fall into valid range.

Change-Id: I0a74d00f4ef40aec7edaeca1d03c8645e23ab08c
2014-02-22 12:30:18 -08:00
Jingning Han 4b3e44f91d Merge "Periodically update mode and mv costs" 2014-02-22 11:54:20 -08:00
James Zern 8e3ef6c0e1 Merge "move nestegg to third_party" 2014-02-22 11:43:37 -08:00
James Zern dd18536563 Merge "gen_msvs_vcxproj.sh: Set TargetName instead of OutputFile." 2014-02-22 11:08:39 -08:00
James Zern 8ec2d6eb5a Merge "tests: enable datarate_test based on CONFIG_ENCODERS" 2014-02-22 10:16:25 -08:00
Yaowu Xu 5a7ac03b9e Changed a constant array to static const type
Change-Id: I5b3918441f63ceb7889dcf546e9a2d781be4a3cb
2014-02-21 22:44:13 -08:00
Yaowu Xu f1633e5844 Merge "Remove an unused variable" 2014-02-21 22:44:05 -08:00