Commit Graph

5622 Commits

Author SHA1 Message Date
James Zern
9581eb6e8a use consistent framerate naming
s/frame_rate/framerate/g

Change-Id: I6fc3e088e419c5f46e3a9390dd8a2cad2677a2fc
2013-07-16 14:12:47 -07:00
James Zern
c0562d08f6 Merge "VP[89]_COMMON: remove unused near_boffset" 2013-07-16 12:17:04 -07:00
James Zern
63e914bde4 Merge "VP9_COMMON: remove unused framerate/bitrate" 2013-07-16 12:16:37 -07:00
James Zern
3a7c2665d0 Merge "yv12config: remove YUV_TYPE" 2013-07-16 12:16:04 -07:00
Ronald S. Bultje
58a2005367 Merge "Replace generated quant tables with static lookup tables." 2013-07-16 12:07:17 -07:00
Ronald S. Bultje
e965cccce5 Replace generated quant tables with static lookup tables.
This prevents possible float rounding issues between architectures.

Change-Id: I6ed260aebd49feb4cfb5596a5370c44be5f72167
2013-07-16 12:06:26 -07:00
John Koleszar
cc1aac1b3c Merge "Fix above context pointers" 2013-07-16 11:23:38 -07:00
Jingning Han
5851904744 Merge "SSE2 8x8 inverse ADST/DCT transform" 2013-07-16 11:00:11 -07:00
John Koleszar
5efd9609e3 Fix above context pointers
In the prior code, the above context pointers used for entropy
decoding were initialized on the first frame, and not updated when
the frame size changed. The per-frame code which initializes the
contexts assumes that the contexts are contiguous, leading to an
incomplete initialization when the frame is smaller. This commit
updates the pointers so that the context is contigous whenever
the frame size changes.

Change-Id: I08b53e3a30c8289491212311682ff1b8028cff6c
2013-07-16 10:26:56 -07:00
Johann
90ebfe621f Merge "vp9_convolve8_[horiz|vert]_avg" 2013-07-16 09:42:52 -07:00
Jingning Han
dd97c62ab8 Merge "Skip inter-coded block reconstruction in rd loop" 2013-07-16 09:03:38 -07:00
Dmitry Kovalev
e8e7620a1f Merge "Removing and moving around constant definitions." 2013-07-16 00:52:53 -07:00
Yaowu Xu
c5b0cd8405 Merge "Change to extend full border only when needed" 2013-07-15 21:35:32 -07:00
Yaowu Xu
5b915ebd92 Change to extend full border only when needed
This is a short term optimization till we work out a decoder
implementation requiring no frame border extension.

Change-Id: I02d15bfde4d926b50a4e58b393d8c4062d1be70f
2013-07-15 20:52:13 -07:00
Dmitry Kovalev
ca75f1255f Removing and moving around constant definitions.
Removing unused and duplicated constants, moving them from *.h to *.c
if possible.

Change-Id: Ief4d6b984a3ca2e9b38504f0d855ed072cf7133f
2013-07-15 19:26:30 -07:00
Dmitry Kovalev
65762849d1 Merge "Consistent naming for loop-filter filters." 2013-07-15 19:21:32 -07:00
Johann
6eae37f45c Merge "Remove print_nmvcounts" 2013-07-15 18:43:41 -07:00
Ronald S. Bultje
b02c4d364f Increase border size from 96 to 160.
This is required because upon downscaling, if a motion vector points
partially into the UMV (e.g. all minus 1 of 64+7 pixels, i.e. 70),
then we can point up to 140 pixels into the larger-resolution (2x)
reference buffer UMV, which means the UMV for reference buffers in
downscaling needs to be 140 rounded up to the nearest multiple of 32,
i.e. 160.

Longer-term, we should probably handle the UMV differently by detecting
edge coverage on-the-fly and using a temporary buffer for edge extensions
instead of adding 160 pixels on all sides of the image (which means a
CIF image uses 3x its own area size for borders).

Change-Id: I5184443e6731cd6721fc6a5d430a53e7d91b4f7e
2013-07-15 17:30:57 -07:00
Ronald S. Bultje
1ff94fea56 Inline vp9_quantize() in xform_quant().
Cycle times:
4x4:    151 to  131 cycles (15% faster)
8x8:    334 to  306 cycles (9% faster)
16x16: 1401 to 1368 cycles (2.5% faster)
32x32: 7403 to 7367 cycles (0.5% faster)

Total encode time of first 50 frames of bus @ 1500kbps (speed 0)
goes from 1min39.2 to 1min38.6, i.e. a 0.67% overall speedup.

Change-Id: I799a49460e5e3fcab01725564dd49c629bfe935f
2013-07-15 17:30:57 -07:00
Ronald S. Bultje
7e684e2009 Merge "Inline xform_quant() in encode_block_intra()." 2013-07-15 17:29:39 -07:00
Frank Galligan
ce1d69aed9 Merge "Neon: Update mbfilter if all vectors follow one branch." 2013-07-15 17:11:55 -07:00
Dmitry Kovalev
e973b4e2d9 Consistent naming for loop-filter filters.
Renaming flatmask4 to flat_mask4, flatmask5 to flat_mask5, hevmask to
hev_mask, filter to filter4, mbfilter to filter8, wide_mbfilter to
filter16.

Change-Id: Ic61c73e59c2eee505257584867aafac99833cea1
2013-07-15 16:01:31 -07:00
Ronald S. Bultje
6fb418741f Inline xform_quant() in encode_block_intra().
Also inline some of the block calculations to assist the compiler to
not do silly things like calculating the same offset (or converting
between raster/transform block offset or block, mi and pixel unit)
many, many, many times.

Cycle times:
4x4:     584 ->   505 cycles (16% faster)
8x8:    1651 ->  1560 cycles (6% faster)
16x16:  7897 ->  7704 cycles (2.5% faster)
32x32: 16096 -> 15852 cycles (1.5% faster)

Overall, this saves about 0.5 seconds (1min49.8 -> 1min49.3) on the
first 50 frames of bus (speed 0) @ 1500kbps, i.e. 0.5% overall.

Change-Id: If3dd62453f8e2ab9d4ee616bc4ea956fb8874b80
2013-07-15 16:00:42 -07:00
Dmitry Kovalev
2c31729839 Code cleanup inside vp9_decodeframe.c.
Removing unused DEC_DEBUG define and dec_debug variable. Changing function
signatures to eliminate code duplication, renaming function
mb_init_dequantizer to init_dequantizer. Also removing redundant curly
braces, and comments.


Change-Id: Ia56ee1b0be5f24abb0e878581845be8a4773c298
2013-07-15 14:47:25 -07:00
Frank Galligan
f4f60f6005 Neon: Update mbfilter if all vectors follow one branch.
Change the mbfilter Neon code from executing both branches if all
vectors follow only one branch.

The code is about 5% faster when executing only one branch and about
1% slower when executing both branches.

-PS5: Remove local stack space from mbfilter.

Change-Id: I6a23f9b318a9f4568a2718b4c9348db988fe2182
2013-07-15 13:08:28 -07:00
Jingning Han
043e0f9dad Skip inter-coded block reconstruction in rd loop
Skip the inverse transform and reconstruction of inter-mode coded
blocks in the rate-distortion optimization loop, when skip_encode_sb
feature is turned on. This provides about 1% speed-up at speed 0,
and 1.5% speed-up at speed 1. No performance change in both settings.

Change-Id: I2932718bf4d007163702b61b16b6ff100cf9d007
2013-07-15 11:32:14 -07:00
Jingning Han
faff6ed0fb Skip duplicate block encoding in the rd loop
This speed feature allows the encoder to largely remove the spatial
dependency between blocks inside a 64x64 superblock, thereby removing
the need to repeatedly encode superblocks per partition type in the
rate-distortion optimization loop.

A major challenge lies in the intra modes tested in the rate-distortion
optimization loop. The subsequent blocks do not have access to the
reconstructed boundary pixels without the intermediate coding steps.
This was resolved by using the original pixels for intra prediction
in the rd loop, followed by an appropriately designed distortion
modeling on the quantization parameters. Experiments also suggested
that the performance impact is more discernible at lower bit-rate/psnr
settings. Hence a quantizer dependent threshold is applied to deactivate
skip of block coding.

For bus_cif at 2000 kbps,
speed 0: runtime 269854ms -> 237774ms (12% speed-up) at 0.05dB
         performance loss.

speed 1: runtime 65312ms  -> 61536ms, (7% speed-up) at 0.04dB
         performance loss.

This operation is currently turned on in settings of speed 1.

Change-Id: Ib689741dfff8dd38365d8c1b92860a3e176f56ec
2013-07-15 11:08:58 -07:00
Dmitry Kovalev
1f14bbb624 Merge "Fixing vp9_get_pred_context_comp_ref_p function." 2013-07-15 10:51:42 -07:00
James Zern
04092764f7 VP9_COMMON: remove unused framerate/bitrate
+ VP8_COMMON: place them under CONFIG_POSTPROC_VISUALIZER

Change-Id: I2702d5a3e1134b9c5f7ddc14b4173955a400f2cf
2013-07-12 21:43:23 -07:00
Jingning Han
91365addf8 SSE2 8x8 inverse ADST/DCT transform
This commit enables SSE2 implementation of 8x8 inverse ADST/DCT
transform. The runtime goes from 1216 cycles -> 266 cycles.
For bus_cif at 2000 kbps, the overall runtime reduces from
253707ms -> 248430ms, i.e., 2% speed-up at speed 0.

Change-Id: Ib0372e17e9162d7b11a10d653b1c8be547c878fb
2013-07-12 21:03:16 -07:00
James Zern
ce0324d8dd VP[89]_COMMON: remove unused near_boffset
Change-Id: If9b9ca703b997312df85241a0758d414cfdc5228
2013-07-12 19:41:27 -07:00
Dmitry Kovalev
429070987a Using vp9_copy and vp9_zero instead of custom code.
Change-Id: Id9b6ceeddca3f9b34bfada5c499b1e7a2f42c30b
2013-07-12 18:07:43 -07:00
Dmitry Kovalev
31a68bcdff Fixing vp9_get_pred_context_comp_ref_p function.
Adding missed parenthesis around boolean expressions. Bitstream is changed.
Regenerating test vectors.

Change-Id: I4cc00b761e9473f92f180a9fc3a0c607f0aaae56
2013-07-12 17:46:02 -07:00
Dmitry Kovalev
31403080ff Merge "Removing redundant call to set_mi_row_col." 2013-07-12 17:08:23 -07:00
Dmitry Kovalev
3c94fffdb0 Removing redundant call to set_mi_row_col.
This function is actually called from set_offsets which is called right
before vp9_read_mode_info.

Change-Id: Ibb9d5ad606194bc80eab264fad85b31c9dfd8f77
2013-07-12 16:25:23 -07:00
Johann
a15bebfc0a vp9_convolve8_[horiz|vert]_avg
Super basic conversion from the other implementations. Any changes to
one should be trivial to copy over keep in sync.

Change-Id: I1720b4128e0aba4b2779e3761f6494f8a09d3ea8
2013-07-12 16:21:33 -07:00
Yaowu Xu
cdea4a7c66 Merge "Fix a build issue" 2013-07-12 16:17:22 -07:00
Dmitry Kovalev
aa518af8c7 Merge "Adding struct tx_probs and struct tx_counts to cleanup the code." 2013-07-12 16:02:09 -07:00
Dmitry Kovalev
444c8d4c53 Merge "Making functions read_{inter, intra}_segment_id more similar." 2013-07-12 15:50:02 -07:00
James Zern
c9a2a06c20 Merge "vp9_postproc: remove useless self-assign" 2013-07-12 15:41:41 -07:00
James Zern
4fc6c88e9c yv12config: remove YUV_TYPE
this was never fleshed out in the context of VP8, for which it was
added. for VP9 it has no meaning.

Change-Id: Iba2ecc026d9e947067b96690245d337e51e26eff
2013-07-12 15:25:48 -07:00
Dmitry Kovalev
cc662dd768 Adding struct tx_probs and struct tx_counts to cleanup the code.
Also removing unused declarations from vp9_entropymode.h file.

Change-Id: Ib9c5826db3584a32f6bb3297a76c522b99d83402
2013-07-12 15:22:38 -07:00
Dmitry Kovalev
60969da5cb Merge "Code cleanup in vp9_pred_common.c" 2013-07-12 15:04:07 -07:00
Dmitry Kovalev
db0d603b1c Making functions read_{inter, intra}_segment_id more similar.
Change-Id: I51f9ac910834f2d7aba2be4f7ffbce597e61a144
2013-07-12 14:50:33 -07:00
James Zern
cca973a1ab vp9_postproc: remove useless self-assign
Change-Id: I0bc5d2d8c9fec8be18263b0dc2528886bb5b7b61
2013-07-12 14:17:15 -07:00
Dmitry Kovalev
3ab86adb1e Code cleanup in vp9_pred_common.c
No bitstream changes. Using MB_MODE_INFO temp variables instead of
MODE_INFO variables. Removing redundant curly braces.

Change-Id: Ib9d1bedfbd8af97ecc722ccf697ea8177bbe287c
2013-07-12 14:11:48 -07:00
Yaowu Xu
fb754b182f Fix a build issue
Change-Id: I23a75c495ed7ea917d7f312bef0990e20a6b53d9
2013-07-12 11:38:44 -07:00
James Zern
0195fb53cb vp9: consistent 'log2' variable naming
lg2 -> log2

Change-Id: I0602ddff49e42c9c40c29c084d04b7592b9f8edf
2013-07-12 11:37:43 -07:00
James Zern
37c0a1a8d0 Merge changes I33e76c42,I24aeac1e,If4192b40
* changes:
  vp9_dx_iface: s/vp8/vp9/ where possible
  vp[89]_dx_iface: delete unused function
  vp[89]_dx_iface: factorize vp8_mmap_*()
2013-07-12 11:10:18 -07:00
James Zern
563b4b20a8 vp9_dx_iface: s/vp8/vp9/ where possible
drop 'vp9_' from most static functions unrelated to the codec interface
itself.

Change-Id: I33e76c425bb7373570a57a61662a56d65ab4bdf3
2013-07-12 11:05:39 -07:00