[#15135] imbuf DDS support: sync to nvidia texture tools revision 602 + fix for ATI2 compressed normal maps
It just updates the dds stuff.
Provided by Amorilia
Kent
This fixes:
[#16655] Blender crashes when Fraps videos are added to the sequencer
(sorry, color conversion is not supported by swscaler for some reason,
so currently, only the crash is fixed)
Small one: changed the string attribute in multilayer exr files to be:
"2.43 and newer"
To indicate that it's about compatibility of the exr file, not the version
Blender saved it in.
For simple RGB(A) files, the channel names in openexr were supposed
to be simply "R" "G" "B" and "A" too.
Other programs like other names... like lower case, or like "ambient.r"
Tested with file from renderman.
Added serious interlacing to movies opened using ffmpeg.
(Other video decoders to be done)
Rational: deinterlacing, if done seriously _has_ to be done
in YUV-space. Since internal interface first converts data
to RGB we are pretty much lost (and fall back to IMB_filtery
in that case).
This commit fix the following problem in the DPX code:
1) The code always assume a depth of 10 bits
2) The code don't check the file type (Log or Linear)
This adds redcode (the file format of RED one, R3D) support to blender.
Seems to work fine with the footage I found on the web, but keep in
mind, that because of the unoptimized nature of libopenjpeg, frame
decoding isn't that fast.
It is also a rather challenging task, to make 4k-float-footage realtime :)
Secunia Research
SAID: SA29818 (http://secunia.com/advisories/29818/)
Credit: Stefan Cornelius, Secunia Research
The old code trys to do a sscanf %s %d %s %d from a line in the
image file.
Now it copies over that line to a max buffer of size 540 chars before doing
the sscanf.
(I just picked a constant that was siginficatly large)
It also checks to see if it gets all 4 values if not return NULL.
Kent
[#8578] imbuf for DDS textures: minor bugs fixed
(syncing with upstream nvidia texture tools)
[#8727] imbuf for DDS textures: fix for DXT5 alpha channel corruption
Submitted by our DDS person, Amorilia
Kent
Fixed a long standing bug in IMB_dupImBuf: refcounter should be better 0
within the duplicated ImBuf...
Finally removed this strange STRIPELEM_META mumbo jumbo.
On a way to asset management: made METAs behave like true input strips:
* proxying should work
* input filter options should work (including transform and color balance :)
* IPOs now work in replace mode
ray mirror and transparency, was using the same stack for all
recursion depths, now creates new stacks as needed.
Fix for bug #8592: crash saving non-float image as 16bit tiff.
reference black, reference white and gamma.
Added 16 bit TIFF saving.
This needs more work to cleanup code and add 16 bit TIFF reading, but
committing it now so it can be tested.
value that is an offset along the normal when looking for the nearest
face, which allows baking faces further away, e.g. an ID badge onto a
shirt.
Also fixes a bug baking to float images, for things other than
displacement it didn't work sometimes, and a memory leak in the
extend filter.
Press Rkey in compositor for reading back render results and invoke a compo.
This now correctly reads AO (skipped it sometimes) and it makes a correct
composite.
FSA todo:
- hotkey + button for reading all samples back + composite
- solve black border around image
Added a faster linear interpolation scaler to IMB_scaleImBuf.
Only handles common cases when we either
enlarge both, width and height or
shrink both, width and height
but that is pretty fast:
* does only blit once instead of two passes like the old code
(fewer cache misses)
* uses fixed point integer arithmetic for byte buffers
* doesn't branch in tight loops
Should be comparable in speed to the ImBuf ..._fast functions at least
for byte-buffers.
For the rather seldom situations that e.g. width is enlarged and height
is shrunken, the old code is still in use.
-------
Also small fix in sequencer: free float buffers on read now.
This fixes:
[#7989] Sequence editor preview and anim render output broken on Linux PPC
and also optimizes RGBA -> ABGR conversion function a little bit.
(Fixing also a crash, if there is no ibuf->rect available...)
This is actually just the alpha value as currently being calculated
by the mist code. It is in many cases not very useful to have this as
alpha in shading result, also for postprocess and composite.
Note: this pass also works with "Mist" not set in World, of course.
