hkuang [Fri, 16 Aug 2013 23:36:07 +0000 (16:36 -0700)]
Add neon optimize vp9_short_idct10_8x8_add.
vp9_short_idct10_8x8_add is used to handle the block that only have valid data
at top left 4x4 block. All the other datas are 0. So we could cut several
unnecessary calculations in order to save instructions.
Deb Mukherjee [Fri, 16 Aug 2013 20:51:00 +0000 (13:51 -0700)]
Cleanup/enhancements of switchable filter search
Cleans up the switchable filter search logic. Also adds a
speed feature - a variance threshold - to disable filter search
if source variance is lower than this value.
Jim Bankoski [Tue, 20 Aug 2013 15:14:52 +0000 (08:14 -0700)]
fix the mv_ref_idx issue
The following issue was reported :
https://code.google.com/p/webm/issues/detail?id=601&q=jimbankoski&sort=-id&colspec=ID%20Pri%20mstone%20ReleaseBlock%20Type%20Component%20Status%20Owner%20Summary
This code makes the choice and code cleaner and removes any question
about whether the border needs to be checked.
Yaowu Xu [Sat, 10 Aug 2013 22:04:02 +0000 (15:04 -0700)]
Change to limit the mv search range
As the pixel values beyond image border are duplicates of pixels
on edge, the change limits the mv search range, any mv beyond
the limits no longer produce new/different prediction values
as entire block with pixels used for subpel interpolation are
outside image border.
Jingning Han [Wed, 14 Aug 2013 23:50:45 +0000 (16:50 -0700)]
Enable early termination in uv rd loop
This commit enables early termination in the rate-distortion
optimization search loop for chroma components. When the cumulative
rd cost is above the current best value, skip the rest per-block
transform/quantization/coeff_cost and continue to the next
prediction mode.
For bus_cif at 2000 kbps, the average run-time goes down from
168546ms -> 164678ms, (2% speed-up) at speed 0
36197ms -> 34465ms, (4% speed-up) at speed 1
Dmitry Kovalev [Mon, 19 Aug 2013 20:20:21 +0000 (13:20 -0700)]
Using plane_bsize instead of bsize.
This change set is intermediate. The next one will remove all repetitive
plane_bsize calculations, because it will be passed as argument to
foreach_transformed_block_visitor.
Adrian Grange [Mon, 19 Aug 2013 18:58:52 +0000 (11:58 -0700)]
Further correct bug in loopfilter initialization
The intent was to initialize the deltas for the
segment to the computed value, irrespective of mode
and reference frame if (mode_ref_delta_enabled == 0).
(In response to bug posted by Manjit Hota to codec-devel
and webm-discuss lists)
Jingning Han [Fri, 16 Aug 2013 00:03:14 +0000 (17:03 -0700)]
Fix the returned distortion value in rd_pick_intra
Return the distortion value in vp9_rd_pick_intra_mode_sb as sum of
dist_y and dist_uv. Remove the right shift operation on dist_uv,
and make it consistent with that of vp9_rd_pick_inter_mode_sb.
Deb Mukherjee [Sat, 3 Aug 2013 00:15:38 +0000 (17:15 -0700)]
Speed feature to skip split partition based on var
Adds a speed feature to disable split partition search based on a
given threshold on the source variance. A tighter threshold derived
from the threshold provided is used to also disable horizontal and
vertical partitions.
Results on stdhdraw250:
threshold = 32, psnr = -0.18%, speedup is somewhat more than derf
because of a larger number of smoother blocks at higher resolution.
Based on these results, a threshold of 32 is chosen for speed 1,
and a threshold of 64 is chosen for speeds 2 and above.
Jingning Han [Wed, 14 Aug 2013 01:58:21 +0000 (18:58 -0700)]
Unify luma and chroma rd-cost estimation
This commit unifies the rate-distortion cost calculation process of
luma and chroma components. It allows early termination to be enabled
later in the rd search loop of chroma components, in consistent with
luma pixels.
James Zern [Thu, 15 Aug 2013 01:28:42 +0000 (18:28 -0700)]
vpxenc: open output file after setting pass #
write_ivf_file_header would incorrectly skip writing the file header in
the 2nd pass, causing the initial frame header to be overwritten on
close potential causing an overly large frame header to be read and a
crash.
most likely broken since: 9e50ed7 vpxenc: initial implementation of multistream support
Dmitry Kovalev [Wed, 14 Aug 2013 18:39:31 +0000 (11:39 -0700)]
foreach_transformed_block_in_plane cleanup, explicit tx_size var.
Making foreach_transformed_block_in_plane more clear (it's not finished
yet). Using explicit tx_size variable consistently instead of
(ss_txfrm_size / 2) or (ss_txfrm_size >> 1) expression.
Paul Wilkins [Tue, 13 Aug 2013 16:47:40 +0000 (17:47 +0100)]
Renaming in MB_MODE_INFO
The macro block mode info context originally contained an
entry for each 16x16 macroblock. In VP9 each entry refers
to an 8x8 region not a macro block, so the naming is misleading.
This first stage clean up changes the names of 3 entries in the
structure to remove the mb_ prefix.
TODO clean up the nomenclature more widely in respect of
mbmi and bmi.
Jingning Han [Wed, 7 Aug 2013 21:45:37 +0000 (14:45 -0700)]
SSE2 high precision 32x32 forward DCT
Enable SSE2 implementation of high precision 32x32 forward DCT. The
intermediate stacks are of 32-bits. The run-time goes down from
32126 cycles to 13442 cycles.
projects / libvpx / log