generally speaking, if multilayer image fails to load for current
frame doesn't mean anything bad. It might be used to make it so
image sequence is being alpha-overed somewhere in the middle of
scene time.
Made it so if the whole file fails to load, image node will
deliver black transparent color, the same what happens for
regular (non-multilayer images).
Also needed to tweak code in load_multilayer_sequwnce to make
sure no cached frames are pointing to a freed memory.
The RenderLayers node would use the "combined" image result for all passes which don't have a valid render result yet. This causes problems when the buffer element size is not actually 4 floats (RGBA) as
with the 3 float normal passes. Also the result is rather meaningless then, so just keep the image buffer at NULL for unavailable passes, which will return plain (0,0,0) color.
After discussion with Brecht decided that automatically updating the sockets of the node based on externally modified data (removed EXR file passes) is not desirable behavior. But at least making sure
the correct passes are assigned to the output sockets of the Image node is possible. Now the passes are matched by name instead of using the faulty index stored in the socket data, which is more
reliable. Still may break if changing pass names externally, but an image reload is highly recommended anyway and will fix that.
This was own error in r60049 which fixed chunk number calculation. This was mixing int and unsigned int values from ExecutionGroup, which leads to huge chunk numbers which are then skipped.
Debug code for graphviz output moved to a dedicated file COM_Debug.h/cpp.
The DebugInfo class has only static functions, which are called from a number of places to keep track of what is happening in the compositor. If debugging is disabled these are just inline stubs, so we
don't need #ifdefs everywhere and don't get any overhead.
The graphviz output is much more useful now. DebugInfo keeps track of node names in a static string map for meaningful names. It uses a number of colors for various special operation classes.
ExecutionGroups are indicated in graphviz with clusters.
Currently the graphviz .dot files are stored in the BLI_temporary_dir() folder. A separate dot file is generated for each stage of the ExecutionGroup scheduling, this is intended to give some idea of the
compositor progress, but could still be improved.
The chunk indices for scheduling chunks based on a given area were calculated incorrectly. This caused chunks at the very border of the render (pixels 256..257) to be omitted, leading to incorrect values
in the Z buffer of the test file, which in turn caused wrong normalization range and the resulting almost-white image.
Also added a dedicated executePixel function for Z buffer to avoid any interpolation of Z values.
- add missing headers from cmake (own omission)
- quiet rna_test.c unused define warnings.
- minor style edits
- spelling corrections and ignore all uppercase words with spell checking script.
The issue with wrapping is that it requires correct interpolation of the border pixels. Since interpolation is done at the far left end of the node tree in buffer/image/etc read operations, the wrapping
setting can not be used directly in those operations (otherwise in-line translate operations would cause conflicts). To make wrapping work correctly we need to add a buffer in front of the translate
operation, which can then be interpolated correctly based on wrapping. The WrapOperation becomes a variant of ReadBufferOperation, which uses its wrapping setting to determine the correct "extend" mode
for interpolation of the buffer.
This merges all mix operations into a COM_MixBaseOperation
(naming could be better, but this way it corresponds to what's
going on with math operations.
Same was done with RenderLayers operations.
Overall this gives 20% of bf_compositor library compilation
time decrease. And it was rather annoying to have tens of
files with just a single-line constructors anyway.
TODO:
- All Convert operations could also be merged into a single file,
but that would require adding some ConvertBaseOperation to
reduce code duplication (ideally). Leaving it unchanged for now.
- Some operations' headers seems to be wrongly including MixOperation
header, they need to include NodeOperation instead it seems.
This is rather harmless, but would be nice to doublecheck on this
eventually.
It was caused by wrong copy-paste thing, which replaced check
"whether alpha channel is enabled" with "whether alpha channel
is not zero" (which is always zero in accumulator).
Compositor always works with RGBA, so no need to do any special
checks here.
TODO: Maybe MapUV ode shall ignore alpha channel?
This disables crazy adaptive sampling happening in diagonal direction.
This still gives some doggyness, but it's much less dramatic now,
and behavior is pretty damn the same as EWA filtering when rendering
textures with Blender Internal.
Problem is that the read/write buffer operations only work with actual
image inputs. If a singular value is used as group input no actual
buffer will be created, the write operation does not schedule any chunks
and the ReadBufferOperation subsequently returns zero
(MemoryBuffer::read).
The fix uses the (0,0) resolution to detect single value input of the
WriteBufferOperation. The actual resolution is then clamped to (1,1) to
ensure we have a single pixel to store the value in. A m_single_value
flag is also set, so we can reliably distinguish this from genuine image
resolutions without having to check m_width/m_height later on.
The ReadBufferOperation copies this flag from the associated
WriteBufferOperation and if set will always return the single value from
pixel (0,0).
In fact, there's no need to get float buffer at all,
conversion could be done in pixel processor level
after interpolation.
It might give slightly worse interpolation results
(which i'm not sure would be visible by eye) but
it gives more than 2x speedup on my laptop on node
setups used for warping image.
--
svn merge -r58988:58989 ^/branches/soc-2011-tomato
This commit includes all the changes made for plane tracker
in tomato branch.
Movie clip editor changes:
- Artist might create a plane track out of multiple point
tracks which belongs to the same track (minimum amount of
point tracks is 4, maximum is not actually limited).
When new plane track is added, it's getting "tracked"
across all point tracks, which makes it stick to the same
plane point tracks belong to.
- After plane track was added, it need to be manually adjusted
in a way it covers feature one might to mask/replace.
General transform tools (G, R, S) or sliding corners with
a mouse could be sued for this. Plane corner which
corresponds to left bottom image corner has got X/Y axis
on it (red is for X axis, green for Y).
- Re-adjusting plane corners makes plane to be "re-tracked"
for the frames sequence between current frame and next
and previous keyframes.
- Kayframes might be removed from the plane, using Shit-X
(Marker Delete) operator. However, currently manual
re-adjustment or "re-track" trigger is needed.
Compositor changes:
- Added new node called Plane Track Deform.
- User selects which plane track to use (for this he need
to select movie clip datablock, object and track names).
- Node gets an image input, which need to be warped into
the plane.
- Node outputs:
* Input image warped into the plane.
* Plane, rasterized to a mask.
Masking changes:
- Mask points might be parented to a plane track, which
makes this point deforming in a way as if it belongs
to the tracked plane.
Some video tutorials are available:
- Coder video: http://www.youtube.com/watch?v=vISEwqNHqe4
- Artist video: https://vimeo.com/71727578
This is mine and Keir's holiday code project :)
- fix thumbnail preview (previously it showed only one input)
- make SplitViewer node update even if the second input is not connected
- now it works when the first socket is connected to a zero-sized node tree (e. g. Color Input node)
- SplitViewer node is now based on 2 operations: SplitOperation and ViewerOperation.
- ViewerBaseOperation was removed as a redundant one. Any future viewer style node can use the same principle and prepare the output before passing to an actual ViewerOperation.
Thanks Lukas Toenne for reviewing this patch and giving me get few pieces of advice.
