Signed-off-by: Diego Biurrun <diego@biurrun.de>

The spec states:

 *  Only the lowest 12 spectral coefficients of any LFE may be non-zero

We were using the 12 lowest *bands*.

10l: Forgot about TYPE_LFE being after TYPE_CPE. Which causes a
segfault when encoding 5.1

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>

This caused failure of a few fate tests.

This allows using it in swscale also.

This allows using it in libswscale/ also.

The output of the x264 commandline tool contains the most uptodate
description of the possible values for the -preset, -tune and -profile
options.

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>

Use 'auto' as the default value.

This removes all valgrind-reported invalid writes with one
specific test file.

Fixes http://www.ocert.org/advisories/ocert-2011-002.html

Signed-off-by: Mans Rullgard <mans@mansr.com>

This is faster than using the clip_coefficients() function. Also,
clipping the coordinates before determining reuse is better since it
will compare coordinates that are closer to their final encoded values.

bits_per_raw_sample is used in video as well, where sample_fmt is not used.

Signed-off-by: Anton Khirnov <anton@khirnov.net>

It's called from avcodec_register() anyway, so there's no reason for it
to be public.

channel.

Move bit counting to the bit allocation function. Move exponent grouping to
after bit allocation. This will allow for adjustment of bandwidth parameters
during bit allocation without having to do exponent grouping multiple times.

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>

Signed-off-by: Diego Biurrun <diego@biurrun.de>

Old version was based on lots of guesswork which had different hacks based on
32-bit version field instead of splitting it into major/minor/micro version.
RV2 picture header parser also had a few places where it deviated from binary
decoder.

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>

Avoids uninitialized reads.

Signed-off-by: Anton Khirnov <anton@khirnov.net>

Signed-off-by: Anton Khirnov <anton@khirnov.net>

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>

This fixes building for ARM11 without Thumb2.

Signed-off-by: Martin Storsjö <martin@martin.st>

This allows using the same arm assembler offsets for both EABI
and the mach-o ABI.

Signed-off-by: Martin Storsjö <martin@martin.st>

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>

Old version divided it wrong, which resulted in chroma drift (visible on FATE
sample too as dirty trails left by clouds).

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>

Use AV_PKT_FLAG_CORRUPT flag to mark packets that might be incomplete
or have errors. Formats that have a mean to validate their packets
should use this flag to deliver such information instead of silently
dropping the packets.

Signed-off-by: Diego Biurrun <diego@biurrun.de>

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>

Signed-off-by: Anton Khirnov <anton@khirnov.net>

It's more readable and less prone to breakage.

This fixes build with clang.

Signed-off-by: Mans Rullgard <mans@mansr.com>

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>

Inspection of compiled code shows gcc handles these fine on its own.
Benchmarking also shows no measurable speed difference.

Removing the remaining cases in get_cabac_bypass_sign_x86() does
cause more substantial changes to the compiled code with uncertain
impact.

Signed-off-by: Mans Rullgard <mans@mansr.com>