Uoti Urpala [Sun, 18 Dec 2011 14:17:07 +0000 (16:17 +0200)]
tmv decoder: set correct pix_fmt
Previously the decoder only worked if the user had set avctx->pix_fmt
manually. For some reason the libavformat tmv demuxer sets this, so
the problem was not visible in avplay etc.
Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
Mans Rullgard [Sat, 17 Dec 2011 15:05:24 +0000 (15:05 +0000)]
build: link test programs with static libraries
Many of the test programs directly access internal symbols not
exported from the shared libraries. This allows tests to run
when configured with shared libraries.
Ronald S. Bultje [Sat, 17 Dec 2011 19:56:40 +0000 (19:56 +0000)]
swscale: fix overflow in gray16 vertical scaling.
This fixes the same overflow as in the RGB48/16-bit YUV scaling;
some filters can overflow both negatively and positively (e.g.
spline/lanczos), so we bias a signed integer so it's "half signed"
and "half unsigned", and can cover overflows in both directions
while maintaining full 31-bit depth.
Ronald S. Bultje [Fri, 16 Dec 2011 00:03:54 +0000 (00:03 +0000)]
swscale: fix overflows in output of RGB48 pixels.
For certain types of filters where the intermediate sum of coefficients
can go above the fixed-point equivalent of 1.0 in the middle of a filter,
the sum of a 31-bit calculation can overflow in both directions and can
thus not be represented in a 32-bit signed or unsigned integer. To work
around this, we subtract 0x40000000 from a signed integer base, so that
we're halfway signed/unsigned, which makes it fit even if it overflows.
After the filter finishes, we add the scaled bias back after a shift.
We use the same trick for 16-bit bpc YUV output routines.
Mans Rullgard [Sat, 17 Dec 2011 16:19:24 +0000 (16:19 +0000)]
mp3dec: hack: fix decoding with safe bitstream reader
The buffer splicing relies on the bitstream reader over-reading
the end of the buffer as declared in init_get_bits(), although
more data is actually present. Manually moving the bitstream
boundary after init_get_bits() allows this to work as expected.
Mans Rullgard [Sat, 17 Dec 2011 15:48:19 +0000 (15:48 +0000)]
fate: fix eatqi test
The sample has an incomplete last frame. Decoding it is pointless.
The garbage produced was changed by the bitstream reader now
protecting against over-reads.
Ronald S. Bultje [Fri, 16 Dec 2011 21:42:04 +0000 (21:42 +0000)]
get_bits: introduce safe bitreading to prevent overreads.
When turned on, H264/CAVLC gets ~15% (CVPCMNL1_SVA_C.264) slower for
ultra-high-bitrate files, or ~2.5% (CVFI1_SVA_C.264) for lower-bitrate
files. Other codecs are affected to a lesser extent because they are
less optimized; e.g., VC-1 slows down by less than 1% (all on x86).
The patch generated 3 extra instructions (cmp, cmovae and mov) per
call to get_bits().
The performance penalty on ARM is within the error margin for most
files, up to 4% in extreme cases such as CVPCMNL1_SVA_C.264.
Based on work (for GCI) by Aneesh Dogra <lionaneesh@gmail.com>, and
inspired by patch in Chromium by Chris Evans <cevans@chromium.org>.
Mans Rullgard [Fri, 16 Dec 2011 21:19:50 +0000 (21:19 +0000)]
get_bits: remove A32 variant
The A32 bitstream reader variant is only used on ARMv5 and for
Prores due to the larger bit cache this decoder requires.
In benchmarks on ARMv5 (Marvell Sheeva) with gcc 4.6, the only
statistically significant difference between ALT and A32 is
a 4% advantage for ALT in FLAC decoding. There is thus no (longer)
any reason to keep the A32 reader from this point of view.
This patch adds an option to the ALT reader increasing the bit
cache to 32 bits as required by the Prores decoder. Benchmarking
shows no significant change in speed on Intel i7. Again, the
A32 reader fails to justify its existence.
Ronald S. Bultje [Mon, 12 Dec 2011 20:02:58 +0000 (12:02 -0800)]
h264: don't drop B-frames after next keyframe on POC reset.
The keyframe after a POC reset may not be the first to be returned to
the user. Therefore, don't reset the expected next POC once we return
a keyframe to the user, but once we know that the next frame in the
return-queue is a keyframe.
Shitiz Garg [Thu, 15 Dec 2011 09:02:59 +0000 (14:32 +0530)]
westwood: Make sure audio header info is present when parsing audio packets
Audio header information might get scrambled and would not parse,
yet wsqva_read_packet would try to parse audio packets causing
segfaults such as floating point exception.
Martin Storsjö [Wed, 14 Dec 2011 20:32:29 +0000 (22:32 +0200)]
libgsm: Reset the MS mode of GSM in the flush function
The mode is set in libgsm_decode_init, but the decoder
object is simply destroyed and recreated in the flush
function - therefore the mode has to be set again.
This fixes playback using the libgsm_ms decoder in avplay.
Mans Rullgard [Mon, 12 Dec 2011 00:50:08 +0000 (00:50 +0000)]
bswap: make generic implementation more compiler-friendly
With these changes, gcc 4.5 and later recognise it as a bswap
and use the proper instructions on ARM and x86. On x86, the
16-bit bswap is recognised from gcc 4.1.
Mans Rullgard [Sun, 11 Dec 2011 21:41:59 +0000 (21:41 +0000)]
x86: cabac: replace explicit memory references with "m" operands
This replaces the explicit offset(reg) memory references with
"m" operands for the same locations. As a result, one fewer
register operand is needed for these inline asm statements.
Mans Rullgard [Mon, 5 Dec 2011 15:38:10 +0000 (15:38 +0000)]
lavc: avoid invalid memcpy() in avcodec_default_release_buffer()
When the buf and last pointers are equal, the FFSWAP() results
in an invalid call to memcpy() with same source and destination
on some targets. Although assigning a struct to itself is valid
C99, gcc does not check for this before calling memcpy().
See http://gcc.gnu.org/bugzilla/show_bug.cgi?id=32667
Mans Rullgard [Sun, 27 Nov 2011 14:04:16 +0000 (14:04 +0000)]
lavu: replace int/float punning functions
The existing functions defined in intfloat_readwrite.[ch] are
both slow and incorrect (infinities are not handled).
This introduces a new header with fast, inline conversion
functions using direct union punning assuming an IEEE-754
system, an assumption already made throughout the code.
The one use of Intel/Motorola extended 80-bit format is
replaced by simpler code sufficient under the present
constraints (positive normal values).
The old functions are marked deprecated and retained for
compatibility.