Scott LaVarnway [Fri, 7 Dec 2012 20:19:52 +0000 (12:19 -0800)]
Moved denoiser frame copy/updates out of loopfilter thread
The loopfilter thread from the previous frame can be running while
starting the current frame. cpi->Source will change during this time causing
the wrong data to be copied. The refresh_x_frame flags also change, which
will cause incorrect updates of the denoised buffers.
Scott LaVarnway [Thu, 6 Dec 2012 03:09:05 +0000 (19:09 -0800)]
added work buffer for denoiser
The denoiser was writing to LAST_FRAME buffer. If LAST_FRAME isn't being
updated, the reference frame buffers were out of sync between the encoder and the
denoised raw buffers. This patch resolves the discrepancy by always writing to a work
buffer (INTRA_FRAME) and then copying from that buffer to any buffers that needs to
be updated.
Martin Storsjo [Wed, 28 Nov 2012 20:05:42 +0000 (22:05 +0200)]
Disable fast unaligned on armv5te
This is enabled by default in the main configure.sh, but apparently
is supposed to be disabled if the hardware doesn't support it.
Unaligned reads is only supported on armv6 and newer.
James Zern [Tue, 27 Nov 2012 22:03:36 +0000 (14:03 -0800)]
avoid redefining DECLSPEC_DEPRECATED
fixes, e.g.:
In file included from ../vpx/internal/../vpx_decoder.h:33:0,
from ../vpx/internal/vpx_codec_internal.h:46,
from ../vp8/common/onyx.h:21,
from ../vp8/encoder/block.h:15,
from ../test/subtract_test.cc:18:
../vpx/internal/../vpx_codec.h:52:0: warning: "DECLSPEC_DEPRECATED"
redefined
/usr/x86_64-w64-mingw32/sys-root/mingw/include/winnt.h:164:0: note: this
is the location of the previous definition
Marco Paniconi [Mon, 26 Nov 2012 22:14:50 +0000 (14:14 -0800)]
Updates to qp-regulate and rate correction factor.
Don't use the switch to gf_rate_correction factor when
temporal layers is used (i.e., cpi->oxcf.number_of_layers > 1).
In temporal layers, we prefer to avoid this as any frame
(e.g., base layer frame at anchor of pattern) may update
both last and golden (and possibly alt-ref), and so we would get
different rate correction factors within the same layer.
This change will make sure one rate correction factor exists for each layer.
Also, made some other code in qp-regulate that depends on
alt/golden update specific to the 1 layer case.
The Android NDK automatically manages the include directories. Trying
to do so manually for the Android GYP files can cause the wrong setjmp.h
to be included.
Marco Paniconi [Wed, 14 Nov 2012 23:05:06 +0000 (15:05 -0800)]
Update to datarate_test.
Exlcude key frame from buffer underrun check, and increase
lowest bitrate in BasicBufferModel.
Both changes are needed because of a known issue (#495).
John Koleszar [Mon, 22 Oct 2012 21:21:59 +0000 (14:21 -0700)]
postproc: allocate enough memory for limits buffer
The vp8_post_proc_down_and_across_mb_row_sse2() needs space for an
even number of macroblocks, as they are read two at a time for the
chroma planes. Round up the width during the allocation of
pp_limits_buffer to support this.
Yunqing Wang [Thu, 4 Oct 2012 19:59:36 +0000 (12:59 -0700)]
Add unit test for decoder test_vector_test
Got 61 test vectors from vp8-test-vectors.git
(http://git.chromium.org/gitweb/?p=webm/vp8-test-vectors.git)
Added decoder test vectors downloading in unit tests. Uploaded
the test vectors and their md5 files to WebM website.
$ gsutil cp *.* gs://downloads.webmproject.org/test_data/libvpx
Added their sha1sum to the test/test-data.sha1 file.
In unit tests, download the test vectors to LIBVPX_TEST_DATA_PATH.
Test_vector_test goes through the test vectors, decodes them, and
compute the md5 checksums. The checksums are compared with the
expected md5 checksums to tell if the decoder decodes correctly.
John Koleszar [Mon, 8 Oct 2012 22:58:04 +0000 (15:58 -0700)]
multi-res: disable intra on forced ref frames
If a reference frame is forced because of low dissimilarity, then
shut off the search of intra modes. This change has mixed results. On
one clip (QVGA), it hurt quality by ~1.5% with negligible speed impact.
On another (VGA) it had negligible affect on quality, but a ~0.2% speed
impact.
John Koleszar [Mon, 8 Oct 2012 22:54:24 +0000 (15:54 -0700)]
multi-res: add parent_ref_valid flag
Rather than overloading the parent_ref_frame value to shut off the
search in some cases, add a new validity flag. This cleans up some
of the duplicated mr_encoder_id && mr_low_res_mv_avail checks as
well, for readability.
Yunqing Wang [Fri, 28 Sep 2012 17:13:07 +0000 (10:13 -0700)]
post-proc: deblock filter optimization
1. Algorithm modification:
Instead of having same filter threshold for a whole frame, now we
allow the thresholds to be adjusted for each macroblock. In current
implementation, to avoid excessive blur on background as reported
in issue480(http://code.google.com/p/webm/issues/detail?id=480), we
reduce the thresholds for skipped macroblocks.
2. SSE2 optimization:
As started in issue479(http://code.google.com/p/webm/issues/detail?id=479),
the filter calculation was adjusted for better performance. The c
code was also modified accordingly. This made the deblock filter
2x faster, and the decoder was 1.2x faster overall.
Next, the demacroblock filter will be modified similarly.
John Koleszar [Mon, 8 Oct 2012 15:39:47 +0000 (08:39 -0700)]
multi-res: work around reference mismatch
In some situations, believed to be an interaction between temporal
scalability and dropped frames, the references available to an
encoder may not be the same references available to its parent.
Previously, the code tried to force the reference frame chosen by
the parent to be used on this frame, even if it was disabled. This
was preventing the pick mode loop from running even once, which led
to a crash.
Attempts to reproduce this bug locally were unsuccessful, so it is
still undetermined what the underlying cause of this issue is. In
the specific case that was failing, the application did not set
any flags which influenced the reference selection on that frame.
ref_frame_flags indicated that the golden frame was disabled,
believed to be because the last frame updated the last and golden
frames, so golden was shut off by default. It's not clear why this
wouldn't have also been true in the lower res encoder, ie, why the
lower res encoder decided to use and/or was allowed to use the
golden frame. We weren't able to debug into the non-crashing
lower res encoder as the crash couldn't be reproduced locally.
John Koleszar [Wed, 3 Oct 2012 21:52:56 +0000 (14:52 -0700)]
fix uninitialized value in multi-res encoding
If a parent mb is available but is intra coded, then parent_ref_mv is
invalid. Check that the parent is inter coded before trying to access
the parent_ref_mv. Previously the parent_ref_mv was being read from
an uninitialized stack allocation, causing potential OOB reads and
other undefined behavior.
Adrian Grange [Tue, 2 Oct 2012 16:36:41 +0000 (09:36 -0700)]
Added Reset method to TwopassStatsStore
The stats buffer needs to be reset between runs of the
encoder. I added a Reset() function to TwopassStatsStore
and called it at the beginning of each encode.
This enables us to run multiple encodes which was
previously not possible since there was no way to reset
the stats between runs.
John Koleszar [Mon, 24 Sep 2012 21:52:18 +0000 (14:52 -0700)]
rtcd/win32: use InitializeCriticalSection explicitly
Protect the call to {Initialize,Delete}CriticalSection() with an
Interlocked{Inc,Dec}rement() pair, rather than the previous static
initialization. This should play better with AppVerifier, and fix issue
Scott LaVarnway [Wed, 26 Sep 2012 20:45:53 +0000 (13:45 -0700)]
Added unit test for subtract functions
Patch Set 1: gain familiarity with unit tests... added simple
4x4 subtract test
Patch Set 2: fixed mistakes, parameterized as suggested
Patch Set 3: randomized the source/predictor data
Change-Id: I33432bdf7c9f2a9b8c2533a37106382c2a8209ee Signed-off-by: Scott LaVarnway <slavarnway@google.com>
projects / libvpx / log