| Age | Commit message (Collapse) | Author |
|---|
|
|
|
|
|
|
|
|
|
detaching, etc.). This is supposed to look for the first input/output of every socket type, but was actually taking the first matching link from the link list, regardless of the linked socket's position.
|
|
|
|
|
|
could doo easily blur very high depths and cause artifacts.
|
|
links would often result in unrelated color sockets or otherwise miss updates.
The reason is that the per-node updates used for Reroute node type inheritance are not supposed to be looking at connected nodes, they are purely for "local" updates. For this sort of "global" update which requires depth-first search, the update function on the node tree level must be used instead.
|
|
the socket data structs, leading to crash. This is caused by the node verification procedure, which resets any socket type to the initial type defined in the socket templates. Adding sockets dynamically without templates solves this (the sockets are then ignored by verification).
|
|
http://en.wikipedia.org/wiki/Inpainting
|
|
|
|
function calls (we had a few of these).
|
|
|
|
|
|
required to get consistent ID numbers for particles. The Object ID is not usable since it's a user defined value of the instanced object, which does not vary per instance. Also the random value from the object info node is not consistent over time, since it only depends on the index in the dupli list (so each emitted or dying particle shifts the value).
The particle index is always the same for a specific particle. Randomized values can be generated from this with the use of a noise texture.
|
|
selected.
The image format for the node and sockets were not properly initialized. The file output node only supports image types (not movies), so it needs to check for proper format type after copying from the render settings.
|
|
|
|
|
|
|
|
Issue was caused by threading conflict between compositor output node which
is freeing buffers used by render result image and image draw code which
could use buffers at the same time as compositor frees this buffers.
Solved by adding adding lock around viewer image invalidation and image
drawing.
Use renamed LOCK_PREVIEW mutex for this, which si not called LOCK_DRAW_IMAGE.
With new compositor locking for preview is not needed so it could be removed.
Added the same lock around viewer operation which also frees buffers used
by viewer image. It's actually quite difficult to check whether this is
indeed needed. This code seems to be using acquire/release technique, but
somehow acquiring ImBuf before invalidating it in compositor operation
doesn't resolve the issue, so probably it's not actually locking acquire
and things should be checked deeper.
|
|
- Fixed issues with calculating matte with balance != 0.5
It used to be used concave combination of minimal and maximal
channel values which could be inpredictable.
Use concave combination of two non-major channels sorted
by their index, so such combination would always use the same
coefficients for particular non-major channels.
- Added despill balance slider which defines balance between
non-major channels used for calculating average of two
colors. Difference between average value and pixel value of
major screen channel defines amount of despill. Balance of
0.5 gives the same behavior as it was before this slider
was added.
---
svn merge -r48678:48679 -r48789:48790 ^/branches/soc-2011-tomato
|
|
uninitialized memory. it happened when you delete a node that was being
executed. in the compostor
|
|
--
svn merge -r48088:48089 -r48091:48092 ^/branches/soc-2011-tomato
|
|
|
|
|
|
crashes.
|
|
|
|
normalise->normalize).
|
|
|
|
* WITH_BF_COMPOSITOR option, to disable tile compositor compilation.
* Removed unused SCons file.
|
|
|
|
|
|
Tiles Compositor: Fails without 'Compositor' output node. Regression.
|
|
|
|
masks its useful to be able to disable.
also rename RNA to 'use_antialiasing'
|
|
|
|
- pre calculate the UV dot product
- use image width and height converted to floats in the inner loop.
|
|
|
|
also fix for building ghost test and fix double free in one of the tests
|
|
- get the image width and height once rather then calculating on every access (was doing min/max subtract).
- use unsigned int's - faster for looping.
|
|
|
|
|
|
|
|
This matte could be used to force alpha be at high values in areas where
algorithm detects it as edge or background color.
|
|
|
|
- Re-arrange functions in headers and implementation file to make them
more grouped by entity they're operating with. Also order of functions
in implementation file should match order of functions in header for
easier navigation.
- Rename some functions to match conventions of naming public functions.
- Some code de-duplication, still some room for improvements tho.
- Split main 2D tracking functions into smaller steps to make it more clear.
Accidentally OpenMP was disabled in some of previous commits, re-enable it.
|
|
|
|
This adds garbage matte input to new keying node which is used to
force occluding things which can not be eliminated by color operations.
White areas defines areas which should be removed from final result.
|
|
commit.
|
