Page MenuHome

Fix T37937, smooth brushes are not symmetrical. This is an issue with order of calculation because smooth brush relies on neighboring vertices and it modifies the vertex coordinates directly. This means that for the mirrored part, order of...
AbandonedPublic

Authored by Antonis Ryakiotakis (psy-fi) on Mar 4 2014, 4:28 PM.

Details

Summary

...calculation might vary, creating those
differences. To fix this, we need to use proxies and ensure that the
original coordinates stay intact while a stroke is underway. This is
impossible in the current implementation which uses multiple passes for
higher brush strengths but if we restrict the smooth brush to one pass
then it can work.

Diff Detail

Branch
smooth_symmetry_fix

Event Timeline

The iteration was there for a reason I guess, how does this impact the behavior of the smooth brush?

Implementation seems ok by the way, will leave it up to you or user feedback to decide if leaving out the iterations is a problem.

I am unsure if committing this is a good idea after all. First of all it only solves the mesh case, not the multires case. Second, it is harder to smooth out the mesh if this is to be accepted. Current behaviour is hacky, yes, but it works and I'm not sure if sacrificing the current behaviour is OK for such small differences.