Adrian Grange [Tue, 8 Nov 2011 01:15:28 +0000 (17:15 -0800)]
Third set of checks of buffer level against maximum buffer size
Additional check of buffer level to ensure it doesn't exceed the
maximum buffer size.
Change-Id: I1ba4f8b09bbec89646885040ff47470196af521e
Adrian Grange [Tue, 8 Nov 2011 00:54:40 +0000 (16:54 -0800)]
Additional clipping of buffer level to maximum buffer size
Added additional check of buffer level against maximum
buffer size.
Change-Id: Iaf1fbaf008601161e402b43ce82c3dbc129bf740
Adrian Grange [Tue, 8 Nov 2011 00:28:13 +0000 (16:28 -0800)]
Added check to make sure maximum buffer size not exceeded
Added code to clip the buffer level to the maximum buffer
size. Without this the buffer level would increase
unchecked.
This bug was found when encoding an essentially static
scene at 2Mb/s. The encoder is unable to generate frames
consistent with the high data-rate because Q bottoms out
at Qmin.
As frames generated are consistently undersized the buffer
level increases and does not get checked against the
maximum size specified by the user (or default).
Change-Id: Id8a3c6323d3246da50f7cb53ddbf78b5528032c6
James Zern [Sat, 5 Nov 2011 01:50:35 +0000 (18:50 -0700)]
fix file permissions
all of googletest import (
0ab00a22) was marked executable
Change-Id: Id7b7ee03efc21ab998bb03349bd91644e8af25da
James Berry [Fri, 4 Nov 2011 15:48:30 +0000 (11:48 -0400)]
Add unit test support via google test
Change-Id: I0ab00a22fbea1f38c96ef92ef7eaeda782c0c8d3
Yunqing Wang [Fri, 4 Nov 2011 18:52:27 +0000 (11:52 -0700)]
Merge "Add checks in MB quantizer initialization"
Scott LaVarnway [Fri, 4 Nov 2011 18:31:17 +0000 (11:31 -0700)]
Merge "Fix issue 374: eob read incorrectly"
John Koleszar [Fri, 4 Nov 2011 16:36:27 +0000 (09:36 -0700)]
Merge "Changing decoder input partition API to input fragments."
Tero Rintaluoma [Fri, 4 Nov 2011 07:36:49 +0000 (09:36 +0200)]
Fix issue 374: eob read incorrectly
Updated eob changes to check_reset_2nd_coeffs function.
Change-Id: Id1b21c91c7f0fd286640b487ffe47867009b717d
Scott LaVarnway [Thu, 3 Nov 2011 15:06:06 +0000 (08:06 -0700)]
Merge "Change use of eob in the encoder"
Tero Rintaluoma [Tue, 25 Oct 2011 11:25:11 +0000 (14:25 +0300)]
Change use of eob in the encoder
Changed 'int eob' to 'char *eob' in BLOCKD so that both encoder and
decoder will use eobs[25] array from MACROBLOCKD structure. In future,
this will enable use of the decoder side IDCT in the encoder.
Change-Id: I6e1c011628cb8864fd4a0b80f0279ce16a5ca978
Yaowu Xu [Wed, 2 Nov 2011 15:22:58 +0000 (08:22 -0700)]
Merge "added code to clear 2nd order block when appropriate"
John Koleszar [Tue, 1 Nov 2011 22:25:47 +0000 (15:25 -0700)]
Merge "Fix: Increase default cx_data_size"
John Koleszar [Tue, 1 Nov 2011 21:49:23 +0000 (14:49 -0700)]
Merge "vpxenc: fix crash w/empty input file"
Stefan Holmer [Thu, 29 Sep 2011 07:17:09 +0000 (09:17 +0200)]
Changing decoder input partition API to input fragments.
Adding support for several partitions within one input fragment.
This is necessary to fully support all possible packetization
combinations in the VP8 RTP profile. Several partitions can
be transmitted in the same packet, and they can only be split
by reading the partition lengths from the bitstream.
Change-Id: If7d7ea331cc78cb7efd74c4a976b720c9a655463
Yunqing Wang [Tue, 1 Nov 2011 20:20:00 +0000 (16:20 -0400)]
Add checks in MB quantizer initialization
In some situations (f.g. error-resilient is turned on), vp8cx_mb
_init_quantizer() was called once per macroblock. Added checks
to avoid calculations when there is no change.
Change-Id: Ie4f0a5ade2202041254990a4e9d5b03bd1ac5aea
James Zern [Tue, 1 Nov 2011 20:44:00 +0000 (13:44 -0700)]
vpxenc: fix crash w/empty input file
Fixes issue #362.
Change-Id: Iba6d7e49ed610b44c8a4c0f3d6330a93feb0373d
John Koleszar [Mon, 31 Oct 2011 21:42:51 +0000 (14:42 -0700)]
Correct SPLITMV clamping
Prior to this fix, the clamping state of the last subblock partition
dominated, whereas the correct behavior is to clamp if any partition
needs clamping. This bug was introduced by v0.9.6-232-g6b25501
See also:
[1]: http://code.google.com/p/webm/issues/detail?id=371
[2]: https://bugzilla.mozilla.org/show_bug.cgi?id=696390
Change-Id: I444db492b4c4f05f039c7da6f4216da8207dc138
Yaowu Xu [Thu, 20 Oct 2011 19:32:34 +0000 (15:32 -0400)]
added code to clear 2nd order block when appropriate
It is discovered that in rare situations the 2nd order block may
produce a few small magnitude coefficients that has no effect on
reconstruction. The situations are a combination of low quantizer
values (high quality) and low energy in residual signals (content
dependent). This commit added code to detect such cases and reset
the 2nd order block to all 0.
Patch 1 to 4 used code to do all-zero-check on idct result buffer,
and tests on derf set showed a consistent gain of .12%-.14% on all
metrics.But due to a recent change Ie31d90b, the idct result buffer
is not longer populated. So patch 5&6 use an alternative method to
detect the situations. Tests on derf set now shows a consistent
quality gain of .16%-.20%.
As suggested by Jim, Patch 7&8 removed the condition of all first
order block not having any coefficient, instead we reset 2nd order
coefficients to all 0 if sum of absolute value of the coefficients
is small. So it does slightly more than just detecting the oddity
as discussed above, but tests on derf set now show a consistent
gain of .20%-.23% on all metrics.
It is worth noting here that this change does not have any effect
on mid/high quantizer range, it only affects the quantizer value
18 or blow. Within this range, the change helps compression by up
to 2.5% on clips in the derf set.
Change-Id: I718e19cf59a4fc2462cb7070832759beb9f7e7dd
Scott LaVarnway [Fri, 28 Oct 2011 16:27:17 +0000 (09:27 -0700)]
Merge "Improved decode_split_mv()"
Johann [Thu, 27 Oct 2011 18:17:54 +0000 (11:17 -0700)]
Merge "Fix ARM build problem introduced by CL I3fab6f2b"
Scott LaVarnway [Thu, 27 Oct 2011 15:26:30 +0000 (11:26 -0400)]
Improved decode_split_mv()
Tests showed ~1.2% performance boost on the HD clip used.
Performance will vary based on material.
Change-Id: Icbcf1a828750d5b4ae5252bf596b3ef594042e8a
Scott LaVarnway [Thu, 27 Oct 2011 13:14:00 +0000 (06:14 -0700)]
Merge "Improved mv_bias"
Attila Nagy [Thu, 27 Oct 2011 06:06:45 +0000 (09:06 +0300)]
Fix ARM build problem introduced by CL I3fab6f2b
Update ARM asm implementation of vp8_start_encode to new definition.
Change-Id: Ic44791c969e351082331ba6146c3384c01a0dfad
Johann [Wed, 26 Oct 2011 18:33:14 +0000 (11:33 -0700)]
Merge "Reduce partial frame copy in encoder's pick_filter_level_fast"
Scott LaVarnway [Wed, 26 Oct 2011 15:46:10 +0000 (11:46 -0400)]
Improved mv_bias
Small performance gains.
Change-Id: I709b9390a8a27a70f5f23574313b8db85ac7f23d
Scott LaVarnway [Wed, 26 Oct 2011 15:26:30 +0000 (08:26 -0700)]
Merge "Improved read_mb_modes_mv()"
Scott LaVarnway [Wed, 26 Oct 2011 14:46:36 +0000 (10:46 -0400)]
Improved read_mb_modes_mv()
Interleaved vp8_find_near_mvs and vp8_mv_ref_probs.
2.5% to 4% performance improvement for the HD clips used.
Change-Id: Id888b667cf5ae2f0e19da18743140f055ff7de8d
Attila Nagy [Tue, 18 Oct 2011 06:48:50 +0000 (09:48 +0300)]
Reduce partial frame copy in encoder's pick_filter_level_fast
The partial frame copy function used to copy an extra 8 lines above
and below. The partial frame filtering can only modify 3 pixel rows
above the partial frame. Reduce copy to bare minimum needed, which is
4 lines, so that partial filtering on copied frame is possible.
Define the "magic" fraction number for partial filtering in
loopfilter.h .
Change-Id: I4791ffc541b6884b12759a0d0714a8faf16147ec
Johann [Tue, 25 Oct 2011 21:42:21 +0000 (14:42 -0700)]
Merge "remove uninitialized variable warning"
Scott LaVarnway [Tue, 25 Oct 2011 17:04:11 +0000 (10:04 -0700)]
Merge "Improved token decoder"
Johann [Mon, 24 Oct 2011 23:56:55 +0000 (16:56 -0700)]
remove uninitialized variable warning
Restructure if statement to clarify the error condition. Trigger the
error before clobbering pc-> variables.
Change-Id: Id01cab798a341ce9899078fdcec265a0e942a0b7
Scott LaVarnway [Tue, 25 Oct 2011 15:01:32 +0000 (08:01 -0700)]
Merge "Removed read_mv_ref"
James Zern [Tue, 25 Oct 2011 01:21:12 +0000 (18:21 -0700)]
Merge "remove unused variable warning"
Johann [Mon, 24 Oct 2011 23:33:45 +0000 (16:33 -0700)]
remove unused variable warning
Change-Id: I4fcd6e4656d9823aead941616cd63501aecbd6e2
Johann [Mon, 24 Oct 2011 22:24:07 +0000 (15:24 -0700)]
Fix incorrect pointer definition
Global function pointers can not be defined in header files. Restructure
vpx_scale pointer configuration.
Change-Id: I6f568a263ad770d32f530abad6007f990fd1003a
Scott LaVarnway [Mon, 24 Oct 2011 20:16:08 +0000 (16:16 -0400)]
Removed read_mv_ref
Decode the mv mode with if-then-elses instead of traversing
the vp8_mv_ref_tree data structure. This will make it
easier to interleave vp8_find_near_mvs and vp8_mv_ref_probs.
Change-Id: I1e798d6ec40fcaeeff06ccc82f81201978d12f74
Scott LaVarnway [Mon, 24 Oct 2011 17:52:15 +0000 (13:52 -0400)]
Moved the split motion vector decode
into a function.
Change-Id: Ia023a0587100a52cb084f5d9d5512efa6198dad3
Scott LaVarnway [Mon, 24 Oct 2011 17:27:53 +0000 (10:27 -0700)]
Merge "Removed redundant mv clamps for nearmv and nearestmv"
James Berry [Mon, 24 Oct 2011 15:50:27 +0000 (11:50 -0400)]
Fix: Increase default cx_data_size
Prior size could be too small in some instances
resulting in an error.
Change-Id: Ic601e49cbae92c98a0e7fb51ba8c186b352ffba6
Scott LaVarnway [Mon, 24 Oct 2011 15:37:52 +0000 (11:37 -0400)]
Removed redundant mv clamps for nearmv and nearestmv
Did some cleanup as well.
Patchset 2: Fixed bug. Will revisit the segmentation logic.
Change-Id: Idf9fbcff9aaf467bdace9fbd58ef2cea6c602049
Scott LaVarnway [Fri, 21 Oct 2011 13:30:57 +0000 (06:30 -0700)]
Merge "Remove unused DETOK structure"
Tero Rintaluoma [Thu, 20 Oct 2011 10:22:18 +0000 (13:22 +0300)]
Remove unused DETOK structure
DETOK structure is not used anymore.
Change-Id: Id22e1af78fb85d4bb151237a60290d9364faf217
John Koleszar [Fri, 21 Oct 2011 00:44:26 +0000 (17:44 -0700)]
Merge "Add license text to configure generated files."
John Koleszar [Fri, 21 Oct 2011 00:36:40 +0000 (17:36 -0700)]
Merge "Fix: check cx_data buffer prior to write"
Tom Finegan [Thu, 20 Oct 2011 02:47:02 +0000 (22:47 -0400)]
Add license text to configure generated files.
Applies to generated files with c, h, and mk extensions.
Change-Id: Id82c46673c7aad43e95a9de5058ddcfc6fd72f14
James Berry [Wed, 12 Oct 2011 15:18:50 +0000 (11:18 -0400)]
Fix: check cx_data buffer prior to write
check to make sure that cx_data buffer has enough room before
writting to it, prior behavior did not which could result in a crash.
Change-Id: I3fab6f2bc4a96d7c675ea81acd39ece121738b28
Johann [Tue, 18 Oct 2011 01:21:10 +0000 (18:21 -0700)]
Don't copy borders for loop_filter_pick
During the _pick only the Y plane is examined. In addition, data beyond
the borders of the frame is not read.
Change-Id: Ic549adfca70fc6e0b55f8aab0efe81f0afac89f9
Johann [Wed, 19 Oct 2011 21:37:48 +0000 (14:37 -0700)]
Merge "Fix: NEON copy/extend frame for small sizes"
Johann [Wed, 19 Oct 2011 21:36:29 +0000 (14:36 -0700)]
Merge "enc: save entropy probs only when needed for refresh"
Scott LaVarnway [Wed, 19 Oct 2011 17:38:35 +0000 (13:38 -0400)]
Improved token decoder
Tests showed over 2% improvement on various HD clips.
Change-Id: I94a30d209c92cbd5fef285122f9fc570688635fe
Scott LaVarnway [Wed, 19 Oct 2011 17:24:48 +0000 (10:24 -0700)]
Merge "Remove usage of predict buffer for decode"
James Zern [Wed, 19 Oct 2011 00:34:10 +0000 (17:34 -0700)]
vpxenc: fix rollover in status output
sizeof(unsigned long)=4 in 32-bit builds
Change-Id: I81c9d698c80ffaa332214e5b43e98b4e30cf9e88
Scott LaVarnway [Tue, 18 Oct 2011 16:06:50 +0000 (12:06 -0400)]
Remove usage of predict buffer for decode
Instead of using the predict buffer, the decoder now writes
the predictor into the recon buffer. For blocks with eob=0,
unnecessary idcts can be eliminated. This gave a performance
boost of ~1.8% for the HD clips used.
Tero: Added needed changes to ARM side and scheduled some
assembly code to prevent interlocks.
Patch Set 6: Merged (I1bcdca7a95aacc3a181b9faa6b10e3a71ee24df3)
into this commit because of similarities in the idct
functions.
Patch Set 7: EC bug fix.
Change-Id: Ie31d90b5d3522e1108163f2ac491e455e3f955e6
Johann [Mon, 17 Oct 2011 16:21:31 +0000 (09:21 -0700)]
Merge "Fix: vp8cx_pack_tokens_into_partitions_armv5 crash"
Attila Nagy [Fri, 14 Oct 2011 11:17:24 +0000 (14:17 +0300)]
Fix: NEON copy/extend frame for small sizes
NEON version of copyframeyonly, extendframeborders, copy_frame_func were
not working for plane stride < 128 and/or y_width < 128.
Change-Id: Id6c2e6c795274da0c90134b15c0d5f62d1b17a6c
Johann [Fri, 14 Oct 2011 19:06:24 +0000 (12:06 -0700)]
add 32bit darwin10 (10.6) target
Change-Id: Id1c189350d54919be37f864dae91dee37584945a
Johann [Fri, 14 Oct 2011 19:03:32 +0000 (12:03 -0700)]
allow building for older platforms
Change-Id: Ibbd05e981debee12c16ebcd274150cd75a94a69d
John Koleszar [Fri, 14 Oct 2011 14:56:12 +0000 (07:56 -0700)]
Merge "vpxdec updated to use !feof() instead of *buf_sz in readframe()"
Attila Nagy [Fri, 14 Oct 2011 06:19:03 +0000 (09:19 +0300)]
Fix: vp8cx_pack_tokens_into_partitions_armv5 crash
It was crashing when number of partitions was bigger than the number
of MB rows (ex. 128x96 with 8 partitions).
Start point was not checked against mb_rows, plus extra
"empty" partitions were not written out.
Change-Id: I9c2f013b9ec022354b658fab4ef799ff8b1de93d
John Koleszar [Tue, 11 Oct 2011 19:57:17 +0000 (12:57 -0700)]
Bump ABI version number for temporal scalability
Commit 217591f modified the encoder ABI without incrementing the version number.
Change-Id: I74de01597dadcdcd96f6b817e4ec69d9ab535e4c
Adrian Grange [Tue, 11 Oct 2011 19:54:52 +0000 (12:54 -0700)]
Merge "Added rate-targeted temporal scalability"
Adrian Grange [Thu, 6 Oct 2011 22:49:11 +0000 (15:49 -0700)]
Added rate-targeted temporal scalability
Added the ability to create rate-targeted, temporally
scalable, VP8 compatible bitstreams.
The application vp8_scalable_patterns.c demonstrates how
to use this capability. Users can create output bitstreams
containing upto 5 temporally separable streams encoded
as a single VP8 bitstream.
(previously abandoned as:
I92d1483e887adb274d07ce9e567e4d0314881b0a)
Change-Id: I156250a3fe930be57c069d508c41b6a7a4ea8d6a
James Berry [Fri, 7 Oct 2011 19:42:23 +0000 (15:42 -0400)]
bug fix - starting/optimal/max and buffer_level changed from int to int64_t
buffer_level in VP8_COMP and starting_buffer_level, optimal_buffer_level
and maximum_buffer_size in VP8_CONFIG changed from int to int64_t
to avoid potential crash issues for larger target bit rates.
Change-Id: I0d5ab6c8a44c2fef51f30cd8df4bb4b739c5df26
Attila Nagy [Mon, 10 Oct 2011 08:14:07 +0000 (11:14 +0300)]
enc: save entropy probs only when needed for refresh
Previous entropy probs need to be saved (and restored) only when
current updates are not propagated.
Change-Id: Ie6ee0543066e30874e56258be0a6b7d2dd2fdb2b
Scott LaVarnway [Tue, 4 Oct 2011 16:57:42 +0000 (09:57 -0700)]
Merge "Improved tokenize"
John Koleszar [Tue, 4 Oct 2011 14:40:49 +0000 (07:40 -0700)]
Merge "Fix uninitialized new_mv_count in first pass file"
Yunqing Wang [Tue, 4 Oct 2011 14:04:30 +0000 (07:04 -0700)]
Merge "Multithreaded encoder, late sync loopfilter"
John Koleszar [Tue, 4 Oct 2011 13:50:40 +0000 (09:50 -0400)]
Fix uninitialized new_mv_count in first pass file
Uninitialized data could be written to the first pass file when no
motion vectors are present in the frame.
Also fix a number of compiler warnings.
Change-Id: Icc9f53b6d33da9de4563d86d9fd591910473ea90
John Koleszar [Mon, 3 Oct 2011 15:31:13 +0000 (08:31 -0700)]
Merge "makefile: fix target 'all'"
Johann [Fri, 30 Sep 2011 23:17:56 +0000 (16:17 -0700)]
Merge "Reduce computational complexity of generic C loop filter."
Johann [Fri, 30 Sep 2011 22:47:56 +0000 (15:47 -0700)]
Merge "combine loopfilter data access"
Scott LaVarnway [Fri, 30 Sep 2011 16:49:46 +0000 (12:49 -0400)]
Improved tokenize
For a realtime HD encodings, up to 1.6% gains seen.
Change-Id: If45028e23db95124da63f9d38ffe06e05596cc6e
Johann [Tue, 27 Sep 2011 00:17:20 +0000 (17:17 -0700)]
combine loopfilter data access
The data processed by the loopfilter overlaps. At the block level, this
results in some redundant transforms. Grouping the filtering allows for
a single 16x16 transpose (and inversion) instead of three 16x8 transposes
(and three more inversions).
This implementation is x86_64 only. We retain the previous
implementation for x86.
Improvements are obviously material dependant, but it seems to be ~%1 in
tests here.
Change-Id: I467b7ec3655be98fb5f1a94b5d145e5e5a660007
Alpha Lam [Thu, 1 Sep 2011 16:11:23 +0000 (17:11 +0100)]
Call vp8_find_near_mvs lazily
vp8_find_near_mvs() is being called on all possible reference frames
but the data computed may be used if the loop exits early, which can
be due to x->skip beign set to 1.
Optimize this by call vp8_find_near_mvs() laziy only if it is going
to be used and not computed yet.
Change-Id: Iccdbd4c962a670c9f2c99b8aca8096042ca5dc98
Paul Wilkins [Fri, 30 Sep 2011 09:57:54 +0000 (02:57 -0700)]
Merge "CQ and two pass rate control."
Paul Wilkins [Thu, 15 Sep 2011 13:36:24 +0000 (14:36 +0100)]
CQ and two pass rate control.
Changes to the selection of Q limits for two pass
and two pass CQ mode.
Allowance made for Mode and motion vector costs.
Some refactoring of common code.
For Derf and YT sets CQ mode average improvement
circa 1% (SSIM and Global PSNR).
Some increased tendency to undershoot even when
user CQ not reached.
Patch2: Removed some test code accidentally merged.
Change-Id: Icf74d13af77437c08602571dc7a97e747cce5066
Aaron Watry [Thu, 29 Sep 2011 22:25:48 +0000 (17:25 -0500)]
Reduce computational complexity of generic C loop filter.
Change-Id: I1e7f9ed3cd907844a495b9e0073bc140b87e5c06
John Koleszar [Thu, 29 Sep 2011 13:14:37 +0000 (09:14 -0400)]
makefile: fix target 'all'
'all' is the conventional target for building everything in the
makefile, but the child make was expecting all-$(target), for debugging
reasons that I don't recall exactly. Restore the expected behavior.
Change-Id: Ifbb03610b55be679ce7c5e210b7a69a156bb76b9
Attila Nagy [Fri, 16 Sep 2011 10:54:06 +0000 (13:54 +0300)]
Multithreaded encoder, late sync loopfilter
Sync with loopfilter thread just at the beginning of next frame encoding.
This returns control to application faster and allows a better multicore scaling.
When PSNR packets are generated the final filtered frame is needed imediatly
so we cannot delay the sync.
Change-Id: I288d97b5e331d41d6f5bb49d97986fa12ac6f066
James Berry [Thu, 22 Sep 2011 19:03:28 +0000 (15:03 -0400)]
vpxdec updated to use !feof() instead of *buf_sz in readframe()
For partial droped frames using *buf_sz could incorrectly terminate
a decode.
Change-Id: Id4a1166fa9ae6c0aa7e9f214bfa4c0be0ea82c1c
John Koleszar [Thu, 22 Sep 2011 18:54:15 +0000 (11:54 -0700)]
Merge "clamp_mvs() using the wrong motion vector information"
John Koleszar [Thu, 22 Sep 2011 18:18:00 +0000 (11:18 -0700)]
Merge changes Ie650e9b8,I2427e494
* changes:
vpxenc: get version string programatically
Install missing default_coef_probs.h
John Koleszar [Thu, 22 Sep 2011 17:24:33 +0000 (13:24 -0400)]
vpxenc: get version string programatically
To avoid a dependency on vpx_version.h, call the vpx_codec_version_str()
function and build up the string manually.
Change-Id: Ie650e9b8f2aaaffaa31da5e9ef3b566b972321b4
Johann [Thu, 22 Sep 2011 16:30:18 +0000 (09:30 -0700)]
Merge "Replace vpx_ports/config.h with vpx_config.h"
John Koleszar [Thu, 22 Sep 2011 15:08:21 +0000 (11:08 -0400)]
Install missing default_coef_probs.h
Make sure that this header is listed as one of the sources, so that it
will be installed if necessary.
Change-Id: I2427e494488126b179151dc21043c1e2c8ba5991
Attila Nagy [Thu, 15 Sep 2011 12:34:12 +0000 (15:34 +0300)]
Replace vpx_ports/config.h with vpx_config.h
Just a clean-up.
Change-Id: Iea5b6dc925dcfa7db548bc1ab1a13d26ed5a2c9a
Fritz Koenig [Tue, 20 Sep 2011 22:36:44 +0000 (15:36 -0700)]
Reduce grep match when generating offset files.
Search for the word EQU so that extraneous
symbols are not matched.
Change-Id: Ice6c9ca886211e2ca8a2f5174bdd4103db5c4989
Fritz Koenig [Tue, 20 Sep 2011 17:51:06 +0000 (10:51 -0700)]
Move neon only arm functions under arm/neon.
These files don't contain generic arm code, so should
only be compiled by neon.
Change-Id: Ie712823aa04d4235e7cfe7a3b725e73ee4c3e564
Johann [Tue, 20 Sep 2011 16:51:05 +0000 (09:51 -0700)]
Merge "NEON FDCT updated to match current C code"
Johann [Tue, 20 Sep 2011 16:50:42 +0000 (09:50 -0700)]
Merge "NEON walsh transform updated to match C"
Johann [Tue, 20 Sep 2011 16:50:36 +0000 (09:50 -0700)]
Merge "Updated ARMv6 forward transforms to match C"
Johann [Tue, 20 Sep 2011 16:50:19 +0000 (09:50 -0700)]
Merge "Fixed armv5te multiplications"
Tero Rintaluoma [Mon, 19 Sep 2011 07:10:49 +0000 (10:10 +0300)]
NEON FDCT updated to match current C code
- Removed fast_fdct4x4_neon and fast_fdct8x4_neon
- Uses now short_fdct4x4 and short_fdct8x4
- Gives ~1-2% speed-up on Cortex-A8/A9
Change-Id: Ib62f2cb2080ae719f8fa1d518a3a5e71278a41ec
Tero Rintaluoma [Mon, 19 Sep 2011 07:59:52 +0000 (10:59 +0300)]
Fixed armv5te multiplications
Rd and Rm registers should be different in 'mul'. This register
combination results in unpredictable behaviour. GCC will give
a warning and RVCT an error in this case.
Restriction applies only to armv5 targets and not for armv6 and above.
Change-Id: I378d17c51e1f16a6820814fbed43e115aaabb03e
Stefan Holmer [Tue, 6 Sep 2011 12:34:36 +0000 (14:34 +0200)]
Fix necessary for input partitions iface to match the RTP profile
These changes fixes a glitch between the RTP profile and the input
partitions interface. Since there's no way for the user to know the
actual number of partitions, the decoder have to read the
multi_token_paritition bits also when input partitions mode is
enabled.
Included are also a couple of fixes for issues with independent
partitions and uninitialized memory reads.
Change-Id: I6f93b15287d291169ed681898ed3fbcc5dc81837
Tero Rintaluoma [Mon, 19 Sep 2011 07:24:02 +0000 (10:24 +0300)]
Updated ARMv6 forward transforms to match C
- Updated walsh transform to match C
(based on Change Id24f3392)
- Changed fast_fdct4x4 and 8x4 to short_fdct4x4 and 8x4
correspondingly
Change-Id: I704e862f40e315b0a79997633c7bd9c347166a8e
Tero Rintaluoma [Mon, 19 Sep 2011 07:15:33 +0000 (10:15 +0300)]
NEON walsh transform updated to match C
Modified original patch If2f07220885c4c3a0cae0dace34ea0e36124f001
according to comments. Scheduled code a little bit to prevent some
interlocks.
Change-Id: I338f02b881098782f82af63d97f042b85e63e902
John Koleszar [Fri, 16 Sep 2011 15:09:44 +0000 (08:09 -0700)]
Merge "Fixes the boundary checks for extrapolated and interpolated MVs."
Scott LaVarnway [Fri, 16 Sep 2011 15:03:53 +0000 (11:03 -0400)]
clamp_mvs() using the wrong motion vector information
In the "Removed bmi copy to/from BLOCKD" commit, the copy
to the bmi in BLOCKD was eliminated. The clamp_mvs() used
the bmi in BLOCKD, which now contains incorrect values. This
patch fixes this problem.
Change-Id: I8eca1eaf4015052b0b63e90876f7ad321aba7cff