Intel Xeon only), doing a lot of mutex locking is really slow. Getting
the image buffer for each texture read then made using more threads
actually slow down the render. Now I've split up the function in two
parts, one parts that checks if the image is available, and another
that does a mutex lock and loading if needed.
Changes quite a lot of code, so hopefully doesn't break stuff, but it
seemed to survive test with rendering a number of frames using all
image types and many threads, though this kind of threading problem
only happens once in a while .. so hard to test for.
was calling screenmain() before executing the python script which meant it was never executed (therefore no crash :) )
Moved screenmain() back to the the end of main() and added a TESTBASELIB_BGMODE which checks for G.vd and uses the scene layer if its not there. Of course python should not be running stuff that uses G.vd :/
Also made python scripts stay attached to screens when LOAD UI is disabled. This means you can load a new blend file and the python console can stay open, has been tested for a while in the apricot branch.
somefile_##.png -> somefile_01.png
somefile_########-image.png -> somefile_00000001-image.png
Before, A hash at the end of the string would be replaced by a number with 4 characters. This is still default if no #'s are in the string, so nothing has changed.
To use this function from the python api use scene.render.getFrameFilename()
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.
- renderwindow didn't show values of previous buffer correct.
- renderwindow only showed floats and no char values like image editor.
- renderwindow didn't show x,y values.
- image editor didn't show z values.
Patch to fix these problems provided by Rob Hausauer, thanks!
Reversion of premul bugfix, as it was apparently not
working all that well.
Note that this brings back the bug where the erase alpha
paint tool won't display correctly, since the UV image
editor just draws images in key alpha now.
commit. I still think this is a good idea, but needs more discussion.
Basically, the way the premul option worked before is it actually
changed the image data to be premul (for each pixel it multipled r,g,b by a) when
loading an image. So if a user wanted his image to be key, yet
still work in the renderer (which expects premul) he'd be stuck.
Also, it was kindof confusing how if you painted something in the image,
then saved it with premul on, when next you loaded the image it'd have
changed (especially since the image editor painting seems to paint in
key, not premul).
basically flagged the image so that on next load/reload, the image
data would be converted to premul. This was very confusing to the
user, as it meant premul wouldn't take effect will the image was
reloaded, and it would also change the image data, which the user
might've been painting.
To fix this, I've removed this behaviour and instead made the premul
option apply at render time. During render while evaluating an image
texture, if the image has the premul flag set then the premul operation
is done on the texture result data, thus not touching the image data
at all.
Also, I've made premul be turned on by default.
Attention! Rather large sequencer rewrite:
* Implemented layer blending using implicit effects. (works like layers
in "The Gimp" or Photoshop.)
* Fixed Space-Bar start-stop in preview windows.
You can start playback using spacebar within a preview-window and it _works_!
* Fixed Flip Y (didn't work for float)
* Fixed premul (didn't work for float)
* Added IPOs to _all_ tracks. In blend-mode REPLACE it drives the
"mul"-parameter in all other blend modes it drives the effect.
* you can meta single tracks.
* moved "mute track" from "M" to "Shift-M"
* added "Shift-L" for "lock track"
* changed inner workings for Metas. Now all ImBufs have to use the
reference counting mechanism. (Only interesting for coders :)
!!! Really important change, that affects current files!
Since you can mute tracks and now there is real layer blending capabilities
in place, I changed the silly behaviour that chose the output track.
Old behaviour: if we have an effect track visible, use the uppermost effect
track. If there is _no_ effect track visible, use the lowest input track.
New behaviour: always use the uppermost track. With blend modes active:
work our way down starting from the uppermost track to the first
"replace"-mode track. This is the way the gimp, photoshop, basically _all_
other applications work...
So if this change ruins your day: please try to fix your files using
"mute". If this doesn't work out, I can still restore the old behaviour,
but I really hope, that this is _not_ necessary!
Rational: most people won't get affected by this change, since you can't
really do anything usefull with the (old) sequencer without at least one
effect track and then you are on the safe side...
-----
Bugfix #7831: fix image filename extensions on saved files. Make FFMPEG save JPEG images by default, allow JPEG images to have either ".jpg" or ".jpeg" extension.
* made stamp filename optional
* renamed weightpaint "Filter" to "Blur"
* made the defailt weightpaint opacity 1.0 rather then 0.2 so when you select 1.0 weight you can paint it with without multiple clicks.
The list of changes (some are fixes):
- Properly horizontally centered tags in all fields (bug?).
- File area does not have trailing space and has leading "File " at
start instead (probably a bug).
- Small separation between to time related fields, space saving.
- Removed colons, for consistency and space saving again.
- Frame field is zero aligned for higher visual stability.
- Marker name shows a rarer name, "<none>" (using <> is typical for cases
in which there is nothing: <none>, <empty>, <blank>, etc).
- Top area for misc info that can be really long (file, note and render date).
- Bottom area for more constantly changing but short ones (marker, SMPTE,
frame, camera and scene).
- Only render date moves a line (when note field is not used), and frame one
moves if no SMPTE (still in same line, so no big jump), for extra visual
stability (marker is fixed, assuming most people would show frame and or
SMPTE).
- ISO 8601 date format for render date, localization independant.
Comparision images:
http://www.infernal-iceberg.com/blender/stamp-original.pnghttp://www.infernal-iceberg.com/blender/stamp-cleanup.png
draws into the
frame.
This patch includes some changes I made...
* use blenders bitmap fonts (rather then own fonts)
* select font size
* user interface layout changes
* Marker as another image stamp option
Also added some new API calls
BMF_GetFontHeight(font);
BMF_DrawStringBuf(...); - so we can draw text into an imbuf's image buffer.
get_frame_marker(frame) - get the last marker from the frame.
IMB_rectfill_area(...) - fill in an image buffer with a rectangle area of color.
TODO - draw stamp info in 3d view, at the moment it just displays in the animation.
using the last selected face was almost good enough however when selecting verts and edges the last selected face would become inactive and the space image would flicker about too much.
The active face is used for getting the space image at the moment and keeps scripts that use this flag working also.
This has 2 commands to get and set, so the variable is not accessed directly.
all "UV Calculate" scripts work now
last commit crashed when in solid draw mode, it seems subsurf modifier is ignoring the displayMask since MTFACE is available. just made it do a null check for now.
uvcalc_follow_active_coords.py - should be done inC and put in the snap menu.
Initial commit of imagebrowser in trunk.
BIG COMMIT!
Main changes:
* completely reworked imasel space
* creation and storage of the preview images for materials, textures, world and lamp
* thumbnails of images and movie files when browsing in the file system
* loading previews from external .blend when linking or appending
* thumbnail caching according to the Thumbnail Managing Standard: http://jens.triq.net/thumbnail-spec/
* for now just kept imasel access mostly as old imgbrowser (CTRL+F4, CTRL+F1) a bit hidden still.
* filtering of file types (images, movies, .blend, py,...)
* preliminary managing of bookmarks ('B' button to add, XKEY while bookmark active to delete)
More detailed info which will be updated here: http://wiki.blender.org/index.php/User:Elubie/PreviewImageBrowser
Places that need special review (and probably fixes):
* BLO_blendhandle_get_previews in readblenentry
* readfile.c: do_version and refactorings of do_library_append
* UI integration
TODO and known issues still:
* Accented characters do not display correctly with international fonts
* Crash was reported when browsing in directory with movie files
* Bookmark management still needs some UI work (second scrollbar?), feedback here is welcome!
Credits:
Samir Bharadwaj (samirbharadwaj@yahoo.com) for the icon images.
Many thanks to everyone who gave feedback and helped so far!
Baked textures not updating in the 3D view. This bug was a side
effect of the image not being marked as loaded after changing
parameters (uv test grid, size) of a generated type image.
- Bugfix #5714
New option "Save changed images" didn't respect relative file names
- Bugfix #5621
New camera (shiftx, shifty) was missing in IpoWindow as animatable option
This header defines function stubs for builds without OpenEXR.
A quote from openexr_multi.h:
/* ugly... but we only use it on pipeline.c, render module, now */
No longer true! Function definitions should never be in header files,
I say.
Please read:
http://www.blender3d.org/cms/Imaging.834.0.html
Or in short:
- adding MultiLayer Image support
- recoded entire Image API
- better integration of movie/sequence Images
Was a whole load of work... went down for a week to do this. So, will need
a lot of testing! Will be in irc all evening.
When linking a referenced (Library) Image to a texture face, it wasn't
tagged to become saved in file. So you lost data.
Also fixed: tooltip for the "Li" icons was wrong.
