The filter has been natively integrated into libavfilter.

The native filter is equivalent from the point of view of the features
and of performances.

Increase performance, to match mp=unsharp.

Found by: durandal_1707
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

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

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

Finish what 99162f8 started

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

No need to recompute it for each new frame.

This improves the performances just enough to match mp=kerndeint.

Signed-off-by: Stefano Sabatini <stefasab@gmail.com>

Also update the documentation accordingly, and suggest a meaningful range.

In particular, avoid out-of-buffer access and crashes with too big
values, and rework documentation accordingly.

Fixes FATE in MSVC DLL builds.

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

Correct the advancement of a pointer to skip over LFE samples as
intended in the loudness computation.

This suggested patch is 3 of 3 for Ticket #2144 "libavfilter ebur128
loudness inaccuracy, irregular time interval, LFE interference".
This suggested patch supersedes the contribution of January 15, 2013
in the thread “[PATCH] lavfi: EBU R.128 irreg time, ch weights, skip
LFE” which combined multiple fixes in one batch.

Correct the recognition of channel layouts for good channel weight
in the loudness computation.
This suggested patch is 2 of 3 for Ticket #2144 "libavfilter ebur128
loudness inaccuracy, irregular time interval, LFE interference".

Signed-off-by: David A. Sedacca <sedacca at comcast.net>

Some changes in the border pixels, visually indistinguishable.

Do it all in config_props().

Remove possibly pointless inconsistency with the ported code.

Also specify parameter value ranges consistent with those of the ported
filter.

Binary object unchanged.

A broken packet with invalid data in the middle of the stream
should not prevent from decoding the rest of the file.

Work around trac ticket #2264.

It is supposed to be used with mencoder only and does
nothing when used with ffmpeg.

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

Filter does nothing, and supposed funcionality
is already available with native filter(s).

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

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

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

Respect order of call during the filter lifecycle, improve readability.

Rely on generic utilities for computing each plane linesize.

In particular, add support to NV12/21 formats and avoid use of
PIX_FMT_PLANAR pixdesc flag, whose semantics is questionable.

It also fixes various crashes.

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

Native il video filter is now available.

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

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

Now it displays all color components (not just luma)
either in parade or overlay and also works with RGB colorspace.

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

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

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

In particular, fix trac ticket #2248.

17% faster
Note the speedup is smaller if its done before
storing in the context (gcc 4.6.3 core i7)

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

This makes the code about 7% faster

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

clang says:
libavfilter/vf_yadif.c:192:28: warning: incompatible pointer types assigning to
'void (*)(uint8_t *, uint8_t *, uint8_t *, uint8_t *, int, int, int, int, int)'
from 'void (uint16_t *, uint16_t *, uint16_t *, uint16_t *, int, int, int, int, int)'

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