separate dsp.resample to dsp.resample_common and dsp.resample_linear
and choose to call faster resample_common even when linear_interp=on
when c->frac and c->dst_incr_mod are both zero
speed up resampling when exact_rational and linear_interp are both
enabled because exact_rational force c->frac and c->dst_incr_mod to
be zero when soft compensation does not happen
benchmark on exact_rational=on:linear_interp=on
old new
real 8.432s 5.097s
user 7.679s 4.989s
sys 0.125s 0.107s
Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Muhammad Faiz <mfcc64@gmail.com>
Fixes make checkheaders on systems without the Cuda Toolkit, which
was broken after the dynlink changes.
Signed-off-by: James Almer <jamrial@gmail.com>
Move global thread variables to better place.
Use correct variable for simple and complex filtergraphs.
This makes number of threads set per filter work again.
Signed-off-by: Paul B Mahol <onemda@gmail.com>
Rects with positive w/h/linesize but no data are invalid.
Reviewed-by: Petri Hintukainen <phintuka@gmail.com>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
Expands the parser to also accept the separator '-' in addition to
'+', and take the negative sign into consideration.
The optional sign for the first factor in the expression is already
covered by parsing for an integer.
Signed-off-by: Moritz Barsnick <barsnick@gmx.net>
Reviewed-by: Nicolas George <george@nsup.org>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
scaling list is already transfered to raster scan during head parsing,
so no need to transfer it again.
And after this fix, FATE test SLIST_A_Sony_4/SLIST_B_Sony_8/
SLIST_C_Sony_3/SLIST_D_Sony_9 will pass in i965/Skylake.
Signed-off-by: Wang, Yi A <yi.a.wamg@intel.com>
Signed-off-by: Jun Zhao <jun.zhao@intel.com>
Signed-off-by: Mark Thompson <sw@jkqxz.net>
Remove the |checked| variable because the invalid value of -1 for
|flags| can be used to indicate the same condition. Also rename |flags|
to |cpu_flags| because there are a local variable and a function
parameter named |flags| in the same file.
Co-author: Dmitry Vyukov of Google
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
I moved this into the handle_video_sequence callback because that's
the earliest time you can make an accurate decision as to what the
format should be.
However, transcoding requires that the decision between using
the accelerated PIX_FMT_CUDA vs a normal pix format happen at init()
time. There is enough information available to make that decision
and things work out with the underlying format only being discovered
in the sequence callback.
This patch moves the av_frame_make_writable() call from fill_yuv_image
to get_video_frame so that its argument can be the actual frame that
will be sent to the encoder.
This fixes data corruption issues in codecs that keep references on
one or several previous frames.
Signed-off-by: Sam Hocevar <sam@hocevar.net>
Reviewed-by: wm4
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
Requested-by: wm4 ([FFmpeg-devel] [PATCH] avutil/opt: Support max > INT64_MAX in write_number() with AV_OPT_TYPE_INT64)
Requested-by: ronald ([FFmpeg-devel] [PATCH] avutil/opt: Support max > INT64_MAX in write_number() with AV_OPT_TYPE_INT64)
Reviewed-by: Andreas Cadhalpun <andreas.cadhalpun@googlemail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
This fixes division by zero crashes.
Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
This fixes segmentation faults due to stack-overflow caused by too deep
recursion.
Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
Metadata streams have priv_data set to NULL.
Reviewed-by: Josh de Kock <josh@itanimul.li>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
The nvidia 375.xx driver introduces support for P016 output surfaces,
for 10bit and 12bit HEVC content (it's also the first driver to support
hardware decoding of 12bit content).
The cuvid api, as far as I can tell, only declares one output format
that they appear to refer to as P016 in the driver strings. Of course,
10bit content in P016 is identical to P010, and it is useful for
compatibility purposes to declare the format to be P010 to work with
other components that only know how to consume P010 (and to avoid
triggering swscale conversions that are lossy when they shouldn't be).
For simplicity, this change does not maintain the previous ability
to output dithered NV12 for 10/12 bit input video - the user will need
to update their driver to decode such videos.
P016 is the 16-bit variant of NV12 (planar luma, packed chroma), using
two bytes per component.
It may, and in fact is most likely to, be used in situations where
there are less than 16 bits of data. It is the responsibility of
the writer to zero out any unused LSBs.
Currently, it forces IDR frames for both true and false.
Not entirely sure what the original idea behind the tri-state bool
option is.
Reviewed-by: Derek Buitenhuis <derek.buitenhuis@gmail.com>
User selectable surfaces are not working correctly, if you set number of
surfaces on cmdline, it will always use minimum 32 or 48 depends on
selected resolution, but in nvenc it is not necessary to use so many
surfaces.
So from now you can define as low as 1 surface and nvenc will still
work, it will ofcourse lower GPU memory usage by 95% and async_delay to zero
That was the easy part, now littlebit more...
Next part of this patch is to always prefer rc_lookahead to be more
important for number of surfaces, than user defined surfaces value.
Maximum rc_lookahead from nvidia documentation is 32, but could increase
in future generations so there is no limit for this yet. Value
async_depth is still accepted and prefered over rc_lookahead.
There were also bug when you request more than rc_lookahead > 31, it
will always set maximum 31, because surface numbers recalculation was
after setting lookahead, which is now fixed.
Results:
If you set -rc_lookahead 32 and -bf 3 it will now use only 40 surfaces
and lower GPU memory usage by 20%, also it will now increase PSNR by 0.012dB
Two more comments:
1. from my internal test, i don't understand addition of 4 more surfaces
when lookahead is calculated, i didn't used this and everything works as
with those 4 more extra surfaces, does anybody know what is going on
there? I looks like it was used for B frames which are calculated
separately, because B frames maximum is 4.
2. rc_lookahead is defined default to -1, but in test condition if
(ctx->rc_lookahead) which sets lookahead it will be always true, i don't
know if this is intended behavior, so in default behavior is lookahead
always on!
This is default condition when rc_lokkahead is -1 (not defined on
cmdline), whis is maybe something that is not intended:
ctx->encode_config.rcParams.enableLookahead = 1;
ctx->encode_config.rcParams.lookaheadDepth = 0;
ctx->encode_config.rcParams.disableIadapt = 0;
ctx->encode_config.rcParams.disableBadapt = 0;
Signed-off-by: Timo Rothenpieler <timo@rothenpieler.org>
when meeting IDR frame, vaapi_encode_h264 poc number don't reset, now fix
this issue based on h264 spec. Some decoder don't care this case, but this
fix will enhance the encoder action. Before this fix, poc number is
negative in some case.
Reviewed-by: Jun Zhao <jun.zhao@intel.com>
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
Signed-off-by: Mark Thompson <sw@jkqxz.net>
This was not observed earlier because the only syntax element which
it normally misses with the current setup is slice_qp_delta, but that
is always going to be zero (in IDR frames QP isn't varied on the
slice) which will always exp-golomb code as a single 1 bit. The
immediately following part is the byte alignment, which is always a 1
bit followed by 0s which are ignored, so as long as the bitstream is
never aligned at that point we will never notice because the only
difference is that an ignored bit is a 1 instead of a 0.
(cherry picked from commit fc30a90898)
While outwardly bizarre, this change makes the behaviour consistent
with other VAAPI encoders which sync to the encode /input/ picture in
order to wait for /output/ from the encoder. It is not harmful on
i965 (because synchronisation already happens in vaRenderPicture(),
so it has no effect there), and it allows the encoder to work on
mesa/gallium which assumes this behaviour.
(cherry picked from commit 086e4b58b5)
