Patch from Guillaume Poirier. I didn't test the functionality of the
filter but at least it fixes compilation.
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@27822 b3059339-0415-0410-9bf9-f77b7e298cf2
We were using an inconsistent mix of the three variants and 'volatile' should
be the most correct and portable variant.
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@27791 b3059339-0415-0410-9bf9-f77b7e298cf2
Neither variant is valid C99 syntax, but __asm__ is the most portable variant.
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@27788 b3059339-0415-0410-9bf9-f77b7e298cf2
This patch comes from Andrew de Quincey <adq_dvb at lidskialf dot net>.
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@27732 b3059339-0415-0410-9bf9-f77b7e298cf2
It has been done this way in libswscale for years without apparent ill effect.
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@27699 b3059339-0415-0410-9bf9-f77b7e298cf2
Conditionalize enabling of some the acceleration because
liba52-0.7.4 doesn't support all that MPlayer's included
copy does.
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@27619 b3059339-0415-0410-9bf9-f77b7e298cf2
Some symbols were dropped or renamed, requiring corresponding changes
in MPlayer.
- Use AVCodecContext->bits_per_coded_sample instead of ->bits_per_sample.
- Use AVCodecContext->trellis instead of ->flags&CODEC_FLAG_TRELLIS_QUANT.
- Don't set AVCodecContext->rtp_mode (already marked unused before).
- Use ff_eval2() instead of ff_eval().
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@27548 b3059339-0415-0410-9bf9-f77b7e298cf2
It is available in other places and needs to be updated continuously.
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@27532 b3059339-0415-0410-9bf9-f77b7e298cf2
libmpcodecs/vd_qtvideo.c:69: warning: 'ImageCodecDrawBand' defined but not used
libmpcodecs/vd_qtvideo.c:71: warning: 'ImageCodecEndBand' defined but not used
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@27464 b3059339-0415-0410-9bf9-f77b7e298cf2
ad_pcm used the old audio timestamp tracking system that calculated
timestamp at end of decoder output as
last_timestamp_in_input_decoder_has_read +
bytes_read_after_that_timestamp / input_bitrate.
For PCM this can be accurate as input bitrate is constant. However it
relies on input bitrate being known and actually set. At least in some
case with .mov input and libavformat demuxer it wasn't set. Instead of
special-casing PCM to make sure input bitrate is set (in general it
may not be known or constant at all) change ad_pcm to explicitly set
the pts information on the decoder output side.
Replace all USE_ prefixes by CONFIG_ prefixes to indicate
options which are configurable.
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@27373 b3059339-0415-0410-9bf9-f77b7e298cf2
Conflicts:
libvo/x11_common.c
libvo/x11_common.h
Rename the vo_gl macro "vo_border()" to "vo_gl_border" as it conflicts
with the global variable "vo_border"; done in the merge commit because
uses of the macro needed changes anyway to resolve conflicts.
The filter copied the chroma planes from every potentially changed row
of the original image to its non-subsampled work area. Change it to
calculate the minimum and maximum x coordinate for each chroma row and
only copy the part between those.
The performance increase is only minor (though observable), but I
think the resulting code is no more complex than the original.
The effect of alpha blending was calculated as
color = orig_color * alpha / 255
where alpha and color range from 0 to 255.
Change this to
color = (orig_color * alpha + 255) / 256
where the "/ 256" can be expressed as a shift whereas the compiler
would probably generate a multiply+shift for the original "/ 255".
This formula gives a result that is too high by 1 for some inputs.
However it gives the exact result if alpha is 0 or 255 which is
probably the case where small errors would matter most.
Seems to be enough to avoid crashes (due to unaligned SSE2) with FFmpeg vorbis decoding for now.
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@27281 b3059339-0415-0410-9bf9-f77b7e298cf2
Uoti Urpala