Port MPlayer tinterlace filter from MPlayer, with some ideas taken
from the FFmbc/libavfilter port, with the following main differences:

* added support for full-scale YUVJ formats
* added support for YUVA420P
* request_frame() on the filter is forced to return a frame
* some code factorization (related to the copy_picture_fields() function)
* fixed black padding values for mode 3

Rationale: avfilter_copy_frame_props() was already defined in
libavfilter/avcodec.h, and keeping the lavc/lavfi API glue localized in a
specific file should ease maintainance and help the ones which use
libavfilter without depending on libavcodec.

Avoid out-of-buffer data access when nb_channels is 8.

av_samples_alloc() behavior changed in bbb46f3ec, resulting in random
data filling the data[] and linesize[] arrays of the returned
AVFilterBufferRef, which was resulting in wrong behavior in case of code
checking on data[i] nullity.

In particular fixes crash in avfilter_filter_samples():
        for (i = 0; samplesref->data[i]; i++)
            memcpy(link->cur_buf->data[i], samplesref->data[i], samplesref->linesize[0]);

and correctly fills the linesize[] array for planar data.

The traces end up not indented and ugly anyway, but with the additional
spaces are much more painful to grep/parse.

It can be used to directly pass a AVFilterBufferRef to lavfi, avoiding a
memcpy.

avfilter_copy_frame_props() avoids code duplication and increases
robustness.

The sample aspect ratio is a per-frame property, so it makes sense to
define it in AVFrame rather than in the codec/stream context.
Simplify application-level sample aspect ratio information extraction,
and allow further simplifications.

Fix box alpha rendering when a YUV colorspace is selected, in particular
fix trac ticket #763.

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

It's a part of public API.

A list of formats may have been dynamically created by the calling code,
and thus should not be referenced by the sink buffer context.

Avoid possible invalid data reference.

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

This fixes playback with some audio files through filters.

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>

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

Those functions are shared between libs.

Also fix a typo in function names: smtpe → smpte.

It declares API necessary for filtering apps.

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

With this the application can find out how many frames are available.

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

It slipped through the last two iterations.

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>

Address trac ticket #706.

While quality is bad, PAL8 support is needed to allow testing some
encoders that only support PAL8 input.

Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>

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

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

from the fork.

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

Introduce variables "E", "PHI", "PI", "main_w"/"W", "main_h"/"H",
"text_w"/"w", "text_h"/"h", "x", "y", "n" and "t" in line with
vf_overlay and refactor the code accordingly.

Call dtext_prepare_text as early as possible
Do not draw if the memory allocation failed

Split the memory allocation from the actual drawing.

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

It makes more sense for a bit mask to use an unsigned type.
The change should be source and binary compatible on all
supported systems, hence micro version bump.

Fixes a few invalid shifts.

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