Signed-off-by: Paul B Mahol <onemda@gmail.com>

With complex filters, an output can come from any input,
or several inputs, including inputs of a different type.
Copying the codec parameters from the first input with
the same type does not make any sense.

This does not change anything for simple 1->1 filters,
as source_index is set in that case.

With complex filter graphs, it can end up accessed
without having been set.

This frame rate is more reliable than the one copied
from the input stream, so it is used in priority.

The arguments now use the key=value syntax.

The current flat arguments syntax is not easily extensible
due to sws_param possibly containing commas.
This is also consistent with abuffersrc.

This should fix the fate failure
http://fate.ffmpeg.org/log.cgi?time=20120609002213&log=compile&slot=x86_64-archlinux-gcc-enableshared

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

* cus/stable:
  ffplay: reset AVFrame to defaults before decoding each new frame.
  ffplay: check return code of avcodec_decode_video2()

Merged-by: Michael Niedermayer <michaelni@gmx.at>

* qatar/master:
  float_dsp: ppc: add a separate header for Altivec function prototypes
  ARM: fix float_dsp breakage from d5a7229
  Add a float DSP framework to libavutil
  PPC: Move types_altivec.h and util_altivec.h from libavcodec to libavutil
  ARM: Move asm.S from libavcodec to libavutil
  vc1dsp: mark put/avg_vc1_mspel_mc() always_inline

Merged-by: Michael Niedermayer <michaelni@gmx.at>

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

2.5 times faster than the generic code

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

This fixes: ffplay -f lavfi -i cellauto
This was a regression since factorizing the filter code with ffmpeg.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: Marton Balint <cus@passwd.hu>

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: Marton Balint <cus@passwd.hu>

Also include config.h so that HAVE_ALTIVEC will be defined.
Fixes compilation on PPC with Altivec enabled.

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

Move vector_fmul() from DSPContext to AVFloatDSPContext.

This will allow for easier implementation of Altivec functions in libraries
other than libavcodec.

This will allow for easier implementation of ARM-optimized functions in
libraries other than libavcodec.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

This message is printed whenever the RAWVIDEO codec is used, whether the
output file contains any frames or not. Test command:
./ffmpeg -i infile.y4m outfile.y4m

Signed-off-by: Mike Scheutzow <mjs973@optonline.net>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Reverting "Fix" because it disables amrwb decoder.

This reverts commit e11020a145.

Found and analysed by trac user Jamal.

Fixes the remaining part of ticket #1428.

Found, analysed and tested by trac user Jamal.

Fixes part of ticket #1428.

Found, analysed and tested by trac user Jamal.

Fixes part of ticket #1428.

Found, analysed and tested by trac user Jamal.

Fixes part of ticket #1428.

Found, analysed and tested by trac user Jamal.

Fixes part of ticket #1428.

Fixes #1210.

Signed-off-by: Paul B Mahol <onemda@gmail.com>

Signed-off-by: Paul B Mahol <onemda@gmail.com>

this should fix some fate failures

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

using as_assert() in the table generator would make the generators
have to be linked to libavutil.

This reverts commit c3b6efa4cb.

This ensures that these functions are inlined into the per-position
entry points, allowing constant propagation as needed for proper
optimisation.

18% faster VC1 decoding on Cortex-A9.

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

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Suggested by Roger Pack on ffmpeg-user.

With that change, there is only one array for the inputs,
it will make easier to implement a variable number of inputs.

And reimplement av_get_channel_layout_string() on top of it.

Makes valgrind happy.

Otherwise, setting string will try to free a random address.