[Soc-2016-dev] Weekly Report # 5

[Nathan Vollmer]

This week I was able to:
*Fix a bug with PBVH brush strength.
*I modified what last week I was calling a "blur" brush to instead be
called an "average" brush, since it averages all the colors under the brush
region instead of averaging neighboring colors together. The average brush
can be chosen from the blending mode menu, although I haven't quite figured
out how to get the average brush to appear under the default vertex paint
brush options in the square box on the top of the tools menu.
*I recreated the older "blur" brush using the PBVH. Unlike average, the
blur brush calculates a unique average for each vertex under the brush
region by averaging all neighboring loops to that vertex. Although the
actual blur effect is no different than the older technique, the new blur
is magnitudes faster.
*I am now creating and deleting vert to poly and vert to loop maps while
initializing the "sculpt session". Before, I was unintentionally creating
and deleting the maps every paint stroke cache, causing a performance hit
when initially painting.
*I looked at and am still looking at what all needs to change to get the
"nearest to ray" raycast working in the PBVH. I'm planning on first
creating a new PBVH search and gather method. That method will chose the
nearest pbvh nodes within a certain distance instead of selecting only pbvh
nodes actually being intersected by the ray. From there, the logic should
be similar to a method in BLI_kdopbvh.c, which only supports BVH trees.

Next week:
*get mirrored painting working
*continue working on "nearest to ray" selection.
*fix a seg fault when "undo"ing a paint stroke.

Here's a gif of the difference between blur and average:
http://i.giphy.com/3o6gDYANVxC5xgeKn6.gif

-Nate
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.blender.org/pipermail/soc-2016-dev/attachments/20160625/9eb56ba8/attachment.htm