Yunqing Wang [Wed, 3 Jul 2013 21:43:23 +0000 (14:43 -0700)]
Speed up motion estimation using small partitions' result(experiment)
Current partition checking starts from small sizes, and then goes up
to large sizes. This experiment uses the small partitions' motion
estimation result, which is already available, to speed up the
large partition's motion estimation. We can decide to skip some
patition checkings if they are unlikely choices. We could use the
motion vector(MV) result as current partition's prediction MV, limit
the search range and reference frame.
Current result at speed 1:
psnr loss: 1.19% for stdhd, 0.287% for derf.
speed gain: 14% for sunflower(hd), 11% for akiyo.
Jingning Han [Tue, 16 Jul 2013 19:04:07 +0000 (12:04 -0700)]
Skip redundant motion search in 4x4 level rd loop
This commit makes the encoder to perform motion search only once
per reference frame type for each 4x4/4x8/8x4 block. For bus_cif
at 2000 kbps, the runtime goes from 253812ms -> 217817ms
(14% speed-up) for speed 0.
Removing tile_rows and tile_columns from VP9Common, removing redundant
constants MIN_TILE_WIDTH and MAX_TILE_WIDTH, changing signature of
vp9_get_tile_n_bits.
Jingning Han [Mon, 15 Jul 2013 18:05:31 +0000 (11:05 -0700)]
SSE2 16x16 inverse ADST/DCT hybrid transform
This commit enables SSE2 implementation of 16x16 inverse ADST/DCT
hybrid transform. The runtime goes from 5742 cycles -> 1821 cycles.
This provides about 1% encoding speed-up at speed 0.
Making implementation of vp9_set_pred_flag_{seg_id, mbskip} consistent
with vp9_get_segment_id without using confusing sub(a, b) macro. Passing
mi_row and mi_col to functions explicitly instead of replying on
mb_to_right_edge and mb_to_bottom_edge.
John Koleszar [Tue, 16 Jul 2013 17:20:56 +0000 (10:20 -0700)]
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.
Ronald S. Bultje [Fri, 12 Jul 2013 19:59:19 +0000 (12:59 -0700)]
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).
Ronald S. Bultje [Thu, 11 Jul 2013 18:35:13 +0000 (11:35 -0700)]
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.
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.
Jingning Han [Mon, 15 Jul 2013 18:28:46 +0000 (11:28 -0700)]
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.
Jingning Han [Mon, 8 Jul 2013 23:48:47 +0000 (16:48 -0700)]
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.
Jingning Han [Sat, 13 Jul 2013 03:54:14 +0000 (20:54 -0700)]
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.
Deb Mukherjee [Wed, 10 Jul 2013 23:51:07 +0000 (16:51 -0700)]
Some minor cleanups for efficiency
Implements some of the helper functions more efficiently with
lookups rathers than branches. Modeling function is consolidated
to reduce some computations.
Also merged the two enums BLOCK_SIZE_TYPES and BlockSize into
one because there is no need to keep them separate (even though
the semantics are a little different).
projects / libvpx / log