- define __BIG_ENDIAN__ or __LITTLE_ENDIAN__ with cmake & scons.
- ENDIAN_ORDER is now a define rather than a global short.
- replace checks like this with single ifdef: #if defined(__sgi) || defined (__sparc) || defined (__sparc__) || defined (__PPC__) || defined (__ppc__) || defined (__hppa__) || defined (__BIG_ENDIAN__)
- remove BKE_endian.h which isn't used
- omit render code from this warning (cmake only), until render branch is merged.
- moved -Wunused-parameter warning to apply to all C code in blender (not just ./source/blender), (cmake only).
- removed deprecated bitmap arg from IMB_allocImBuf (plugins will need updating).
- mostly tagged UNUSED() since some of these functions look like they may need to have the arguments used later.
Only open AVI files with at least one supported video track type.
(This is the real fix for DV-Type1-AVIs, since the FCC of iads, that
is used there, wasn't detected. But the code will happily open AVI-files
with only audio tracks... So the real fix is: only try to open things, that
contains something we know how to handle :) )
Updated scons to work with the ffmpeg binaries in
lib/windows/gcc again. Also fixed a bug reported by
Debolaz on irc. Blender's internal avi reading (which can
only read simple formats like motion jpeg) was telling the
video code it could read all avis, without checking their
codec.
To fix this, I copied AVI_open_movie(), and modified it to
to replace AVI_is_avi(). Now it properly checks the codec,
and validates the header.
This adds fractional FPS support to blender and should finally
make NTSC work correctly.
NTSC has an FPS of 30.0/1.001 which is approximately 29.97 FPS.
Therefore, it is not enough to simply make frs_sec a float, since
you can't represent this accurately enough.
I added a seperate variable frs_sec_base and FPS is now
frs_sec / frs_sec_base.
I changed all the places, where frs_sec was used to my best knowledge.
For convenience sake, I added several macros, that should make life
easier in the future:
FRA2TIME(a) : convert frame number to a double precision time in seconds
TIME2FRA(a) : the same in the opposite direction
FPS : return current FPS as a double precision number
(last resort)
This closes bug #6715
Standard framerates not supported / breaks sync -- 23.967 29.967 etc.
https://projects.blender.org/tracker/?func=detail&aid=6715&group_id=9&atid=125
Please give this heavy testing with NTSC files, quicktime in/export
and the python interface.
Errors are most probably only spotted on longer timelines, so that is
also important.
The patch was tested by Troy Sobotka and me, so it most probably should
work out of the box, but wider testing is important, since errors are
very subtle.
Enjoy!
http://bugzilla.mplayerhq.hu/show_bug.cgi?id=642
Help provided by Reimar Doeffing.
It removes two special cases which were creating empty huffman tables.
So videos would play all black in players that were working correctly
and looking at the tables if they exist.
I also deleted commented out malloc statements that seem to be working
just fine(they were replaced with MEM_mallocN statements).
If people need them they can just look at cvs to see how they were done before.
I also removed an empty else statement.
(This fix does not fix the black bar in the sequencer with blender made avi's but
I hope to have a fix for that later today)
Kent
some avi's have 0 sized chunks that we were not handling properly.
(Need to basically rewind the avi to previous frame in this case)
http://www.cs.umn.edu/~mein/blender/avi/examp.avi has 25 frames,
4,7 and 9 are "dropped frames"
blender use to crash on these now it displays the frame before it assuming
there is one.
Kent
