audio_resample can not reduce the number of channels

This avoids warnings from the overflow checker and simplifies the code.
Signed-off-by: Mans Rullgard <mans@mansr.com>

During slice threading only MpegEncContext is passed to
mpeg_decode_slice, remove a wrong cast and change the function
definition to take MpegEncContext pointer.

This reverts commit da22ba7d since it
broke slice threading. Slice threading just duplicates MpegEncContext
so every value used during mpeg_decode_slice has to be in it.
A second patch will fix the illusion that Mpeg1Context is available
in mpeg_decode_slice.

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

This fixes assembling using "nasm".

for example:
Attempting to set ac3 dsur_mode to "on" results in:
[AC-3 Encoder @ 0x7f508f0a3280] Value 2.000000 for parameter 'dsur_mode' out of range
[AC-3 Encoder @ 0x7f508f0a3280] Error setting option dsur_mode to value on.
Signed-off-by: Justin Ruggles <justin.ruggles@gmail.com>

It's not used anywhere else.

It's MPEG-1/2 specific.

It's MPEG-1 specific.

It will be used in vdpau code.

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

The code was mistakenly removed in cad16562.
It stored some motion vector data for future use in B-pictures.

This fixes Bugzilla bug #57.
Signed-off-by: Diego Biurrun <diego@biurrun.de>

Also implement MMX/MMX2 versions and SSE4 versions.

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>

lavc: replace references to deprecated AVCodecContext.error_recognition to use AVCodecContext.err_recognition
Signed-off-by: Anton Khirnov <anton@khirnov.net>

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

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

Right now it's not actually initialized on 32-bit, leading to crashes
on win32.

Signed-off-by: Janne Grunau <janne-libav@jannau.net>

The blue channel went missing when toggling fullscreen on Mac OS X.

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

This allows for knowing the output size before decoding even when there is no
header (e.g. FLV). Otherwise we would have to do a preliminary full frame
decode to determine the number of frames-per-packet.