./ffmpeg -nostats -f lavfi -i testsrc2=4k:d=2 -vf bench=start,scale=1024x1024,bench=stop -f null -

    before: t:0.489726 avg:0.489883 max:0.491852 min:0.489482
    after:  t:0.256515 avg:0.256458 max:0.256999 min:0.253755

Fix uninitialized access of minsf in short windows
Fix potential invocation of coef2minsf(0)

Reviewed-by: Ronald S. Bultje <rsbultje@gmail.com>
Signed-off-by: James Almer <jamrial@gmail.com>

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

Clamp scalefactors by coef2minsf to avoid undefined behavior
caused by signed integer overflow. It also avoids clipping of
coefficients so it should avoid artifacts as well, on very
rare corner cases.

This avoids unneeded computations
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

Remove false claim that processing in freq domain is
worse quality than processing in time domain.
Signed-off-by: Paul B Mahol <onemda@gmail.com>

This reverts commit c0918613, since it
triggered a new assertion with gcc.
This bug definitely needs a true proper fix.

This reverts commit bf1495d9.

Seems like clang might be miscompiling it and causing a signed integer overflow,
making a FATE test fail.
Doesn't seem to affect performance, it only runs on the ESC codebook.
Reviewed-by: Claudio Freire <klaussfreire@gmail.com>
Signed-off-by: Rostislav Pehlivanov <atomnuker@gmail.com>

Remove false claim that processing in freq domain is
worse quality than processing in time domain.
Signed-off-by: Paul B Mahol <onemda@gmail.com>

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

FFT expanded to 2^17.
Signed-off-by: Jovan Zelincevic <jovan.zelincevic@imgtec.com>
Reviewed-by: Nedeljko Babic <Nedeljko.Babic@imgtec.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

Otherwise probing and stream analisys will report a correct coded size
but an empty visible size.

Approved by: kieran
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

Fixes ticket #5386
Signed-off-by: James Almer <jamrial@gmail.com>

Signed-off-by: Lou Logan <lou@lrcd.com>

Reviewed-by: Hendrik Leppkes <h.leppkes@gmail.com>
Signed-off-by: Marton Balint <cus@passwd.hu>

Signed-off-by: Lou Logan <lou@lrcd.com>

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>

Fixes segfault
Fixes Ticket5333

Regression since bfc8a4daSigned-off-by: Michael Niedermayer <michael@niedermayer.cc>

Without this patch, configure still passes and enables CUDA, no matter
if it was actually found, breaking the build in case it was not.

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

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

Fixes Ticket5371
Fixes null pointer dereference
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

Fixes ticket #5318.
Reviewed-by: Stefano Sabatini <stefasab@gmail.com>
Signed-off-by: Marton Balint <cus@passwd.hu>

Note some tests need vsync drop to produce exact timestamps, these seem not to
need it. quite likely many more dont need it either, ive not checked beyond finding
one that needs it and the ones which have it removed
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

Fixes error "libavcodec/audiotoolboxenc.c:294:50: error: use of undeclared
identifier 'kAudioCodecPropertyPacketSizeLimitForVBR'."

That was added to 10.9:
https://developer.apple.com/library/mac/releasenotes/General/APIDiffsMacOSX10_9/AudioUnit.htmlSigned-off-by: Michael Niedermayer <michael@niedermayer.cc>

Reviewed-by: Reimar

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

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

This silences missing EOC marker when decoding j2k files with Psot set to 0.
Signed-off-by: Paul B Mahol <onemda@gmail.com>