[Soc-2011-dev] Weekly progress report #11 - 1th-7th August

Sergey I. Sharybin g.ulairi at gmail.com
Fri Aug 5 16:56:39 CEST 2011

Hi, Blender and libmv Communities!

I'm still wirking on camera tracking integration project 
(soc-2011-tomato branch). I think no large commit would be done this 
evening, so here's some small progress for this week
> 1) what you did this week
(big changes only)

- Fixed occasional crash caused by conflicting access to non-threadsafe 
- A bit nicer feedback from libmv when doing reconstruction.
- If "Zoom To Mouse" is enabled in user preferences then zooming would 
happen to mouse position.
- Separate stabilization settings from display in clip editor.
- Added sliders to 2D stabilization panel which controls intensity of 
translation/scale which applies on shot.
- Added filter type to Stabilize2D compositor node. Supports nearest, 
bilinear and bicubic interpolation.
- Added new node called Transformation.
- Added output sockets to movie clip compositor node. They holds 
stabilization data which can be used by Translate or Transform nodes.
- Highlight failed to reconstruct frames with red in cache line.
- Added group "Failed Tracks" in Select Grouped operator.
- Reconstruction should now deal correct with aspect ratio.
- Fixed bug with resetting marker's flag for non-transforming tracks 
when canceling transformation.
- Fixed bug with keyframing non-selected tracks when initializing 
tracking tools.
- Added operator redo panel to Clip Editor.
- Added operator "Clean Tracks" which currently checks if tracks are 
tracked in bad way and can select bad tracks, delete them or delete 
tracking segments which are bad.
- Added some additional camera presets from Francois request.

- Not really implemented, but there are some ideas how to implement 
dopesheet/graph-related tasks. Need to be discussed with other developers.

> 2) what you will do next week
- Rotation 2d stabilization.
- Improve useability of UI. No real proposals atm but several requests 
to make things more modal (like mode for 2d tracking, mode for 
reconstruction and so)
- Finish things i've planned for this (i've spent ,ore time than i've 
expected on solving proxies-related things like buildbot environment and 
so, also discussions discovers more critical things)
- Make some research on stabilization Francois provided to me. 
Interesting from both of algoritmic and workflow POV.
- Solve sensor-size workflow issuses.
> 3) any problems that will require extra attention and the impact on your proposed schedule
Well, its not like things which can break schedule. Just there is one 
discovered issue in proxies patch which makes them useless for tracking. 
I've sent detailed report to Peter this week.
Also there's thing which i'm not sure how to solve best -- having tracks 
in dopesheet and grapheditor are useful for artists. Ways of bringing 
such functionality should be discussed with other develoeprs.

Project homepage: http://wiki.blender.org/index.php/User:Nazg-gul/GSoC-2011

More info here: 

Videos: there's no really new tools, so there's no video today.

SVN repo: 

With best regards, Sergey I. Sharybin

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.blender.org/pipermail/soc-2011-dev/attachments/20110805/6fa6f5a9/attachment.htm 

More information about the Soc-2011-dev mailing list