Signed-off-by: Timothy Gu <timothygu99@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Also correctly namespace other functions in vidstabutils, and decrease
difference from Libav.

Initial-patch-by: Vittorio Giovara <vittorio.giovara@gmail.com>
Signed-off-by: Timothy Gu <timothygu99@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

In particular, allow pad_len and whole_len to have value set to 0, which
means that no padding will be added. Previously a value set to 0 meant
that that the filter had to pad forever.

The new semantics is clearer, also simplifies scripting since the option
value might be automatically computed, so that no checks need to be done
in case it is 0.

The old semantics was never documented and the logic was broken (the
filter was always adding samples indefinitely), so this should not break
backward compatibility.

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

Also convert a debug av_log() to av_dlog().

The only remaining floats are in the user interface, they are left as they
should not cause a problem in practice

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

there are some still left for 1 time initialization

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

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

This avoids an unused variable warning with MSVC since the variable is
only used in a debug mode printf statement.

Fixes ticket #3829

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

Signed-off-by: Timothy Gu <timothygu99@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Get rid of the many, slightly differing, implementations
of basically the same thing.

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

Based-on suggestion by JULIAN GARDNER <joolzg@btinternet.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Reviewed-by: Clément Bœsch <u@pkh.me>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Fix a segfault with large window size.

The framework can ensure that each input frame has exactly
the correct number of samples, except the last one.

It is always identical to the last pushed frame.
The samples in the last incomplete window were ignored,
this is unchanged.
Possible enhancement: pad the last incomplete window with
silence.

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

Fixes compilation failures on msvc/icl shared builds

Signed-off-by: James Almer <jamrial@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

It was wrongly being exported and used by libavfilter.

Signed-off-by: James Almer <jamrial@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

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

Signed-off-by: James Almer <jamrial@gmail.com>
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>

No change in the generated asm.

This makes the code about 1.5x faster without any noticeable difference
in the output.

This has no impact on overall performance, since the block DCT taking
most of the time anyway.

Add a tlength option with frequency and timeclamp variable

Add to the volume option support for frequency and timeclamp variable,
   a_weighting, b_weighting and c_weighting functions

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

8x8 is about 5x faster than 16x16 on 1080p input. Since a block size of
8x8 makes the filter almost usable (time wise) and it's not obvious if
8x8 or 16x16 is better from a quality PoV (it really depends on the
input and parameters), the filter now defaults to 8x8, and as a result
libavfilter is micro bumped.

This removes the avcodec dependency and make the code almost twice as
fast. More to come.

The DCT factorization is based on "Fast and numerically stable
algorithms for discrete cosine transforms" from Gerlind Plonkaa &
Manfred Tasche (DOI: 10.1016/j.laa.2004.07.015).