This mode has the following differences:
- Video timing is correct for streams with B frames, at least with some
demuxers.
- Video filters can modify frame timestamps and insert new frames, and
removing frames is handled better than before.
- Some things are known to break, it's not usable as the default yet.
Things should work as before when the -correct-pts option is not used.
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@18922 b3059339-0415-0410-9bf9-f77b7e298cf2
note, the code could probably benefit from some finetuning / experimentation in case anyone is bored ...
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@18591 b3059339-0415-0410-9bf9-f77b7e298cf2
patch replaces '()' for the correct '(void)' in function
declarations/prototypes which have no parameters. The '()' syntax tell
thats there is a variable list of arguments, so that the compiler cannot
check this. The extra CFLAG '-Wstrict-declarations' shows those cases.
Comments about a similar patch applied to ffmpeg:
That in C++ these mean the same, but in ANSI C the semantics are
different; function() is an (obsolete) K&R C style forward declaration,
it basically means that the function can have any number and any types
of parameters, effectively completely preventing the compiler from doing
any sort of type checking. -- Erik Slagter
Defining functions with unspecified arguments is allowed but bad.
With arguments unspecified the compiler can't report an error/warning
if the function is called with incorrect arguments. -- Måns Rullgård
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@17567 b3059339-0415-0410-9bf9-f77b7e298cf2
this one does actually compress&decompress the video at various shifts with lavc while the other spp filters are doing optimized intra only filtering
limitations:
mpeg4 is hardcoded, all options too, pretty trivial to change though, even filtering with non dct codecs like snow could be tried ...
the qscale/qp is only taken fron the first MB of each image and then used for the whole image (would needs some small changes to lavc to let the user set the qscales for the mbs themselfs but iam to lazy ...)
this needs ALOT of cpu time and memory especially at uspp=8 ...
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@16947 b3059339-0415-0410-9bf9-f77b7e298cf2
these differences from spp lead to a few nice symmetries which significantly reduce the computational cost
almost not mmx optimized (iam lazy ...)
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@15945 b3059339-0415-0410-9bf9-f77b7e298cf2
- do not uninitialize video encoder between files
- checks for image size & format change moved from mencoder.c to vfilters
by Oded Shimon <ods15@ods15.dyndns.org>
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@14879 b3059339-0415-0410-9bf9-f77b7e298cf2
- Fixed PlanarToNV12Wrapper() and made it handle NV21.
- Added yuv2nv12XinC() to handle software scaling.
- Added NV12/NV21 handling to various places.
- Removed NV12 from vf_hue and vf_spp as they don't look like they can
actually handle it.
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@14716 b3059339-0415-0410-9bf9-f77b7e298cf2
frames that will be skipped, mencoded tells vf_softskip (if present)
that it should drop the next frame. this allows filters that need to
see every input frame (inverse telecine, denoise3d, ...) to see
skipped frames before they get dropped.
in principle, a smarter softskip filter could be written that would
buffer frames and choose to drop the one with least change, rather
than strictly dropping the next one.
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@12338 b3059339-0415-0410-9bf9-f77b7e298cf2
to generate valid mpeg output from avi's that have duplicate frames in
them, or when using inverse telecine filters. to use it, put the
"harddup" filter at the end of your filter chain.
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@12335 b3059339-0415-0410-9bf9-f77b7e298cf2
for removing duplicate frames from telecined video that was
incorrectly deinterlaced. Minor bugfixes added by me.
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@12226 b3059339-0415-0410-9bf9-f77b7e298cf2
colorspace input supported for now. Autocropping is also not implemented.
Example usage: mplayer -vo zr2 -vf zrmjpeg foo.avi.
vf_zrmjpeg and vo_zr2 should obsolete vo_zr and libvo/jpeg_enc.c in the future.
Problem is that it needs some paramters of the zoran card (max resolution), for
now the user needs to tell vf_zrmjpeg those parameters (which is stupid,
because zrmjpeg should be able to query vo_zr2 for that information....) The
filter currently uses code which is also present in libvo/jpeg_enc.c, in the
future the (then enhanced) ffmpeg mjpeg encoder should/will be used.
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@11663 b3059339-0415-0410-9bf9-f77b7e298cf2
(pullup.[ch]) is not mplayer-specific and is designed to work well
with g2; vf_pullup.c is the g1 wrapper. see man page for details, and
keep in mind, this is a work in progress.
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@10665 b3059339-0415-0410-9bf9-f77b7e298cf2
1) cosmetic change, no reason these need to be mpeg2-specific
2) add a flag to tell when fields are ordered, so we don't have to
assume bff (which would usually be wrong) when field flags are not
available.
3) add other flags for future use :)
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@10664 b3059339-0415-0410-9bf9-f77b7e298cf2
