Ronald S. Bultje
9b46e30494
Merge "Don't use BLOCKD in vp9_invtrans.c." into experimental
2013-04-09 21:36:09 -07:00
Ronald S. Bultje
a3874850dd
Make SB coding size-independent.
...
Merge sb32x32 and sb64x64 functions; allow for rectangular sizes. Code
gives identical encoder results before and after. There are a few
macros for rectangular block sizes under the sbsegment experiment; this
experiment is not yet functional and should not yet be used.
Change-Id: I71f93b5d2a1596e99a6f01f29c3f0a456694d728
2013-04-09 21:28:27 -07:00
Dmitry Kovalev
f370db0cf4
Adding setup_quantization function.
...
Change-Id: I8fe25a905717a3cd2da5f87ba0403357536183cf
2013-04-09 18:24:08 -07:00
Yunqing Wang
d3c526fbda
Merge "Fix an issue in set_refs()" into experimental
2013-04-09 14:31:51 -07:00
John Koleszar
a3ec4cbd33
Merge "detokenize: use consistent structure for all block sizes" into experimental
2013-04-09 14:18:59 -07:00
Yunqing Wang
335114debe
Configure: make --enable-gprof work
...
Modified for option "--enable-gprof".
Change-Id: I07f78f28c1b789c0b3e9e7db3561463551ae5460
2013-04-09 13:24:59 -07:00
Yunqing Wang
01a3bd67d8
Fix an issue in set_refs()
...
Scale factor for second ref frame wasn't assigned in the code.
Change-Id: I6ef3f3f71bd652a879ad847369c54c744782ea37
2013-04-09 12:33:28 -07:00
Dmitry Kovalev
02349561b6
Renaming inverse hybrid transform functions.
...
Renaming vp9_ht_dequant_idct_add* functions to vp9_dequant_iht_add*.
Change-Id: Ie427b322b1cc7c8f39d1155f5df91dedfbd944af
2013-04-09 11:09:23 -07:00
Dmitry Kovalev
c34f6fcb54
Fixing upper case names.
...
Renaming Y1dequant to y_dequant, UVdequant to uv_dequant, QIndex to qindex.
Change-Id: I1c356e5f886deb3f8807dc212de9799b55b09d58
2013-04-09 10:46:57 -07:00
Dmitry Kovalev
df76a617b4
Cleanup of set_offsets function.
...
Adding ALLOWED_REFS_PER_FRAME constant instead of hard coded number 3.
Change-Id: I46146aa837896936f920c748c7d4aa4c27f026e4
2013-04-09 10:17:22 -07:00
Dmitry Kovalev
2a6e09d8fe
Merge "Simplification of decoder's code." into experimental
2013-04-09 10:10:29 -07:00
Jingning Han
b3935e8348
Merge "Clamp inferred motion vectors only" into experimental
2013-04-09 09:24:08 -07:00
Dmitry Kovalev
d1cff2deb1
Code cleanup in bitstream code.
...
Lower case variable names, less code.
Change-Id: I1abc8f592ad2343ab5c76fe2d16262741a4a894a
2013-04-08 19:07:29 -07:00
John Koleszar
e6deea4e60
detokenize: use consistent structure for all block sizes
...
Restructure the code to avoid the majority of per-block-size
switches, code duplication, etc. All block types (mb/sb32/sb64)
can be handled by the same code.
Change-Id: I4022718d66e31a15a7074e43f3b98cd0a5124ea7
2013-04-08 13:11:40 -07:00
Dmitry Kovalev
5811d7e865
Simplification of decoder's code.
...
Removing several commented code blocks, using uint32_t and uint8_t types,
removing redundant code.
Change-Id: Ifc5cc9863897925ea2a7cab4f7309ccf28d80bfe
2013-04-08 12:14:40 -07:00
Ronald S. Bultje
f42bee7edf
Don't use BLOCKD in vp9_invtrans.c.
...
Change-Id: I40524170334109e2864b06e3c73c8b34e5aa8b0f
2013-04-08 11:37:29 -07:00
Jingning Han
12bf0796e6
Clamp inferred motion vectors only
...
Clamp only the motion vectors inferred from neighboring reference
macroblocks. The motion vectors obtained through motion search in
NEWMV mode are constrained during the search process, which allows
a relatively larger referencing region than the inferred mvs.
Hence further clamping the best mv provided by the motion search may
affect the efficacy of NEWMV mode.
Synchronized the decoding process. The decoded mvs in NEWMV modes
should be guaranteed to fit in the effective range. Put a mv range
clamping function there for security purpose.
This improves the coding performance of high motion sequences, e.g.,
derf set:
foreman 0.233%
husky 0.175%
icd 0.135%
mother_daughter 0.337%
pamphlet 0.561%
stdhd set:
blue_sky 0.408%
city 0.455%
also saw sunflower goes down by -0.469%.
Change-Id: I3fcbba669e56dab779857a8126a91b926e899cb5
2013-04-08 11:37:03 -07:00
Ronald S. Bultje
aeefa6e194
Fix typo which breaks 4x4 splitmv compound prediction RD code.
...
0.15% quality increase on derf, particularly noticeable on hard clips
at the higher bitrate end.
Change-Id: I02415a96eb9bbc361cba923069625fae71844bc9
2013-04-08 09:17:52 -07:00
John Koleszar
0e7b7e47c2
Merge "Small cleanup inside setup_loopfilter function." into experimental
2013-04-05 16:13:46 -07:00
John Koleszar
8bbabbea70
Merge "Segmentation code cleanup." into experimental
2013-04-05 16:03:25 -07:00
John Koleszar
fa135d7b9e
Merge changes Ibbfa68d6,Idb76a0e2 into experimental
...
* changes:
Move EOB to per-plane data
Move qcoeff, dqcoeff from BLOCKD to per-plane data
2013-04-05 15:56:50 -07:00
Ronald S. Bultje
9161127ee9
Merge "Remove full-pixel-related code." into experimental
2013-04-05 13:46:07 -07:00
Ronald S. Bultje
fd2a747038
Merge "Remove some unused macros." into experimental
2013-04-05 13:46:02 -07:00
Ronald S. Bultje
c6c07d7013
Merge "Remove struct POS." into experimental
2013-04-05 13:45:58 -07:00
Ronald S. Bultje
a9688dfdfb
Merge "Remove unused vpx_log() function prototype." into experimental
2013-04-05 13:45:51 -07:00
Ronald S. Bultje
ac28c3169a
Merge "Remove "tx_type" member from union b_mode_info." into experimental
2013-04-05 13:45:48 -07:00
Yaowu Xu
2e23c74794
Merge "Removed a speed feature no longer used" into experimental
2013-04-05 13:34:57 -07:00
Yaowu Xu
3dca0d44d2
Merge "make one_shot_q an experiment" into experimental
2013-04-05 13:34:45 -07:00
Ronald S. Bultje
36c3a67c20
Remove full-pixel-related code.
...
This is a VP8-only feature (part of profile 3) that is unsupported in
VP9.
Change-Id: I78016eede8d9c834d44d4c517f3e8b8fc2a378b1
2013-04-05 12:50:19 -07:00
Dmitry Kovalev
421baef49e
Small cleanup inside setup_loopfilter function.
...
Change-Id: If7fa8aea02f26c2c2bb5daf4e65c3e661d7031ca
2013-04-05 12:48:48 -07:00
Ronald S. Bultje
61834f7325
Remove some unused macros.
...
Change-Id: Ic219e7878428128e4bb1b3995e8151f92b6bd9c3
2013-04-05 12:40:56 -07:00
Ronald S. Bultje
0732a61c37
Remove struct POS.
...
It is never used.
Change-Id: If7462357c0498ed05af2645f0c272124381d3aab
2013-04-05 12:38:40 -07:00
Ronald S. Bultje
1cb34c32ed
Remove unused vpx_log() function prototype.
...
Change-Id: Icd6b4322841fefcc86f06645e6aaf1ea42fdfabd
2013-04-05 12:37:45 -07:00
Ronald S. Bultje
5cd235c6cd
Remove "tx_type" member from union b_mode_info.
...
It is never used.
Change-Id: Ibae898c52c766aabf65868611060f9c38fb85b35
2013-04-05 12:36:15 -07:00
Dmitry Kovalev
2c42499513
Segmentation code cleanup.
...
Cleaning up the code, removing unused vp9_check_segref_inter function and
useless comments.
Change-Id: Ia0e1a3878dc0f9789cba84aeb507a83d9dccd26b
2013-04-05 11:55:52 -07:00
Yaowu Xu
e79a3ff5f3
Removed a speed feature no longer used
...
Change-Id: Id0c2e44daa936f1d6fb76469fd1bd72a4d7c19fd
2013-04-05 10:43:20 -07:00
John Koleszar
98466e8962
Merge "Simplifying get_delta_q function." into experimental
2013-04-05 09:16:15 -07:00
John Koleszar
05a79f2fbf
Move EOB to per-plane data
...
Continue migrating data from BLOCKD/MACROBLOCKD to the per-plane
structures.
Change-Id: Ibbfa68d6da438d32dcbe8df68245ee28b0a2fa2c
2013-04-04 21:30:23 -07:00
John Koleszar
4c05a051ab
Move qcoeff, dqcoeff from BLOCKD to per-plane data
...
Start grouping data per-plane, as part of refactoring to support
additional planes, and chroma planes with other-than 4:2:0
subsampling.
Change-Id: Idb76a0e23ab239180c818025bae1f36f1608bb23
2013-04-04 16:30:57 -07:00
Yaowu Xu
9780d58e94
make one_shot_q an experiment
...
so it is configurable to faciliate testings
Change-Id: I247b62736c3a08ec2934793959d1ae605a05efa3
2013-04-04 14:14:51 -07:00
Deb Mukherjee
ffc92da4c2
Fixing the newbintramodes experiment
...
Adds back special casing B_PRED mode decoding but protected
within the experimental macro.
Change-Id: If98dc8e56b0ecfb1202540c2b7dfdd070cb81ca0
2013-04-04 12:40:55 -07:00
Dmitry Kovalev
52128c5894
Simplifying get_delta_q function.
...
Change-Id: I3a1e9cc5c3ed5be01ff75a84a6c82ec02c75af9c
2013-04-04 12:10:39 -07:00
Deb Mukherjee
a9e94301f7
Merge "Bugfix in encode_inter_mb_segment_8x8" into experimental
2013-04-04 11:17:48 -07:00
Deb Mukherjee
73031aaa7d
Bugfix in encode_inter_mb_segment_8x8
...
Fixes an indexing bug. Looks like the bug has been there for a while.
Change-Id: I9fc04b0c30754bcb47366ad94a08112925600c4d
2013-04-04 11:07:19 -07:00
Dmitry Kovalev
f857e074d7
Fixing bug introduced by previous commit.
...
Inside decode_sb_4x4 it should be
"get_tx_type_4x4(mb, y_idx * y_size + x_idx)"
but it was
"get_tx_type_4x4(mb, y_idx * (2 * y_size) + x_idx)".
Also making code of decode_sb_4x4, decode_sb_8x8, and decode_sb_16x16
formatted in the same way.
Change-Id: I15c7bef4fb575f7e9da19f953912324cb35d24dd
2013-04-04 10:49:17 -07:00
John Koleszar
ccc0577ab2
Merge "Remove special case vp9_decode_coefs_4x4" into experimental
2013-04-04 07:11:31 -07:00
Paul Wilkins
9b9136f8a2
Fixed incorrect use of compute_qdelta()
...
This function expects real Q values as inputs
not index values.
The use-age her impacts the Q chosen for force key
frames. Though this is a bug fix I have not yet verified
whether following the bug fix the q multiplier value used is
correct.
Change-Id: I49f6da894d90baeb1e86c820c335f02dc80d3b66
2013-04-04 10:19:16 +01:00
John Koleszar
74e8bd11c2
Merge "Adding decode_sb_16x16 function." into experimental
2013-04-03 21:07:53 -07:00
John Koleszar
4d9dbb2ae8
Merge "Reimplementation of setup_frame_size." into experimental
2013-04-03 21:04:29 -07:00
John Koleszar
0520833591
Merge "Adding setup_pred_probs and read_txfm_mode functions." into experimental
2013-04-03 21:02:59 -07:00