This was actually a regression after color management re-implementation, need
to copy settings from saved image buffer to an original one since they might
be modified during save.
Also noticed image format planes detection didn't work properly from an image
buffer. Made it so save operator works fine now, but also marked a TODO in
BKE_imbuf_to_image_format() which needs to be investigated further.
Summary:
Behaves very much the same as cache for Movie Clip datablock:
- Image now have `MovieCache *cache` field which replaced
legacy `ListBase ibufs`.
This allows image datablock to easily keep of image
buffers which are owned by itself. This field isn't
saved to the file and getting restored on undo steps.
However, cache limit is global for movies, sequences
and image datablocks now. So overall cached image buffers
size will not go above cache limit size in user
preferences.
- Image buffers which are marked as BITMAPDIRTY will never
be freed from the cache.
- Added utility function to iterate over image buffers
saved in movie cache.
- Movie cache cleanup check callback now have ImBuf argument
which can be used in a condition of cleanup.
- Added some utility functions which replaces legacy ibufs
iterations with image cache iteration which happens from
inside a lock.
- Fixed `image_mem_size()` which was only counting one of
the buffers if both float and byte buffer present.
Additional notes:
- `BKE_image_get_first_ibuf()` is rather stupid, but direct
access to ibufs->first was also the same stupid idea.
Would consider avoid this function is another project.
- There are some places which doesn't look threadsafe, but
they already were not so much threadsafe anyway before.
So think not a big deal with solving this later.
Finally solves infinite memory usage by image sequences! :)
Reviewers: brecht, campbellbarton
Reviewed By: brecht
CC: sebastian_k
Differential Revision: http://developer.blender.org/D95
Fixes this problem:
*Go to the Blender (BI) textures tab and open an existing image.
*Hit on the plus sign next to datablock to create a new image.
*Complete the creation process and switch back to the first (loaded) image.
-> The image user count raises... will raise as often as you hit the plus sign
The scene pointer used for looking up the appropriate source of render result images in the image editor was always taken from context. This means that render results for a different scene would never be
displayed in the image editor.
To give feedback on running renders, try to get the running render job's scene pointer in the image editor for render result type images. This only happens during rendering, apart from that the regular
context scene result is displayed.
We now have openimageio building when cycles builds or when it's
manually set to build.
(I reverted the _IMAGE_ in the define name because I think the closer
the cmake flags match the defines in the software the better, and there
is no reason to rename all the existent WITH_OPENIMAGEIO references in
CMakeLists.txt - which would be the alternative)
Currently supports only two modes:
- Show alpha channel of the mask
- Multiply footage by the mask, which will give
you final-looking combined image.
TODO: Currently rasterization happens on every
redraw, need to cache rasterized mask
somewhere to make redraw more realtime.
operator for opening images in the image editor header, it's easily one
of the most common operations, don't see why it has to be only accessed
from menu.
We now support the combined layer of Photoshop files (stored as layer 0
in the file). This way users can keep their files as multilayer PSD and
Blender always handle them as flat images.
For perfect alpha this requires an OpenImageIO update:
342cc2633f
Photoshop sample files:
https://github.com/OpenImageIO/oiio-images
Brecht has some pending fixes to push for OIIO as well, so we may as
well wait to update our libraries.
What works:
===========
* 8bit images (with or without alpha)
* 16bits images (alpha discarded)
* Photoshop files saved with 'Maximum Compatibility'
* Cycles, Blender internal, BGE (and player)
Known limitations
(due to OIIO dependency):
=========================
* Images with less than 4 channels show a wrong thumbnail (bug may be in OIIO)
* Packed images are not supported
* We do not write PSD files.
Note: old Blenders have support for PSD via Quicktime library. But due
to license issues this was discontinued.
Many thanks for Brecht van Lommel for reviewing the patch, suggesting
multiple improvements and to help solving the alpha issue.
The SAMPLELINE flag in histogram was set during the BKE_histogram_update_sample_line function. That function in turn is called during every scope update in area draw function, meaning that during render
it constantly gets set. OTOH the operator tries to disable the flag on invoke, which "cancels" the sample line by default. So during render the operator un-setting of the flag has no effect, because the
render job immediatly triggers a redraw, which updates scopes and sets it again.
Moved the flag out of the actual sample line update function into the operator execute. Now only the operator enables/disables overall sample line drawing, while the rest of the update works as before.
HSV values were calculated from a linear space color, which is
not so much useful and correct. Now RGB(A) buffers will use
color managed color for HSV values.
Still not sure which color to use for HSV when there's only one
channel in a buffer. This part left unchanged for now.
Currently only circle and square, might be easily
extended in the future.
New primitives are creating at cursor location.
This also implied adding 2d cursor to space clip.
Also fix set 2D cursor location which didn't work
in image editor's mask mode since 2.67.
TODO: draw_image_cursor better be moved to some
more generic file, but it's not so much
important for now and might be solved later.
Thanks Campbell for the review!
This mimics the behaviour we have in the Clip Editor.
I personally would prefer if we had no border once in fullscreen
(current border is 5 pixels).
I will consult Sergey Sharybin first to see if we can change that in the clip editor as well (though there I
believe the border is useful - the bottom of the editor is used to indicate 'tracked' frames.
Issue was caused by operator redo saving values for previous
inverted channels, meaning the same channels will be inverted
next time operator runs.
Don't think it's useful to save operator values here, since
you don;t have visual feedback about which channels were
inverted. So marked all this properties as SKIP_SAVE. Gives
much more predictable results.
in fact the problem was caused by own previous fix/improvement for a different case, now this works as follows...
- render uses last-saved name, falls back to 'untitled' in blend file path.
- non render uses id-name always, saves into dir of last-saved image, fall back to blend file path.
of the active object in the 3D view. This was due to sharing a global G.moving
flag to indicate that transform is active, now it's only set per transform data
type so different editors don't influence each other.
node materials.
Area and region listener callbacks now get the screen and area pointers passed, so
they can do more fine grained checks to see if redraw is really needed, for example
depending on the 3D view drawtype.
2 bugs here
- missing NULL check in IMB_colormanagement_display_settings_from_ctx()
- deadlock in draw_image_paint_helpers()
Simple solution is to not draw paint helpers for render/viewer images.
were no properly updating when rendering animations.
The render engine was only updating the image user current frame on images used
by material textures. Now moved the function that updates all from the editors
to blenkernel level and do it on all frame changes.
After the paint refactoring for 2.67, the OpenGL texture was getting updated for
every stroke point, rather than once for every redraw. With a small brush radius
and low spacing the number of stroke points can be quite large, which might have
a big performance impact depending on the graphics card / drivers.
Also for 2D image paint, avoid redrawing the button panels and properties editor
during painting.
There is another possible cause for slowdowns with 3D texture painting which was
not fixed. Projection painting is creating and destroying threads for every stroke
point. Depending on the CPU/OS there might be a lot of overhead in doing that if
the brush size is small.
Was caused by svn rev53181, and it worked before because
image buffer didn't have quality set and in this case
fall back to scene settings happened.
Now or render result quality from scene settings is always
used, image buffer's settings is ignored.
