dyntopo won't collapse unmasked verts while masking is active #36930

New Issue

Danny McGrath · 2013-10-03T14:35:58+02:00

Danny McGrath commented

2013-10-03 14:35:58 +02:00

%%%--- Operating System, Graphics card ---
Ubuntu 12.04 64bit

- Blender version with error, and version that worked ---

r60526

- Short description of error ---

I had some stubborn verts in the attached blend that refused to collapse. Apparently disabling the mask fixed it.

- Steps for others to reproduce the error (preferably based on attached .blend file) ---

Open the attached blend
Enable sculpt/dyntopo
Make sure collapse is activated
switch to shaded to see the mask is active
Try to smooth/draw/etc the greased verts, doesn't work
Alt-m to disable the mask and then you can

%%%

%%%--- Operating System, Graphics card --- Ubuntu 12.04 64bit - Blender version with error, and version that worked --- r60526 - Short description of error --- I had some stubborn verts in the attached blend that refused to collapse. Apparently disabling the mask fixed it. - Steps for others to reproduce the error (preferably based on attached .blend file) --- - Open the attached blend - Enable sculpt/dyntopo - Make sure collapse is activated - switch to shaded to see the mask is active - Try to smooth/draw/etc the greased verts, doesn't work - Alt-m to disable the mask and then you can %%%

Danny McGrath commented

2013-10-03 14:35:58 +02:00

Changed status to: 'Open'

Sergey Sharybin commented

2013-10-04 09:59:01 +02:00

%%%If i understand your issue correct, it's actually caused by the fact vertices are actually masked out. Here's in attachments mask with higher contrast, so you might easily see this. And i think it's expected that masked areas are not affected by topology changes?

Perhaps we'd rather add some different types of mask visualization (i.e. add contrast slider, or use rainbow colors instead of grayscale)?%%%

%%%If i understand your issue correct, it's actually caused by the fact vertices are actually masked out. Here's in attachments mask with higher contrast, so you might easily see this. And i think it's expected that masked areas are not affected by topology changes? Perhaps we'd rather add some different types of mask visualization (i.e. add contrast slider, or use rainbow colors instead of grayscale)?%%%

Brecht Van Lommel commented

2013-10-04 15:24:33 +02:00

%%%I think the problem isn't so much that the mask isn't visible, but the behavior of partially masked vertices. You can still sculpt them, the influence is just less, so it's strange that dynamic topology doesn't work on them.

Currently check_mask_half in pbvh_mesh.c will only do splits if the mask is < 50%. This doesn't make sense to me, why would you not update the topology if the vertices moved, such an arbitrary cutoff is going to give a sharp cutoff in the mesh. If the mask makes the brush influence smaller, then the vertices will be moved less and so edge splits will happen less too, I think that should be enough?

I tried a quick change to that value, but it gives strange results, like a kind of fractal shape around the masked areas, I think the paint mask is not being subdivided correctly.%%%

%%%I think the problem isn't so much that the mask isn't visible, but the behavior of partially masked vertices. You can still sculpt them, the influence is just less, so it's strange that dynamic topology doesn't work on them. Currently check_mask_half in pbvh_mesh.c will only do splits if the mask is < 50%. This doesn't make sense to me, why would you not update the topology if the vertices moved, such an arbitrary cutoff is going to give a sharp cutoff in the mesh. If the mask makes the brush influence smaller, then the vertices will be moved less and so edge splits will happen less too, I think that should be enough? I tried a quick change to that value, but it gives strange results, like a kind of fractal shape around the masked areas, I think the paint mask is not being subdivided correctly.%%%

Brecht Van Lommel commented

2013-10-04 15:42:50 +02:00

%%%Fix committed to svn now in revision 60548, for the dynamic topology behavior.

The mask visualization could be changed, but the fact that slightly influenced vertices show only faintly is pretty much what I would expect. Something like a rainbow color could be used but that gets in the way of seeing the mesh too.%%%

%%%Fix committed to svn now in revision 60548, for the dynamic topology behavior. The mask visualization could be changed, but the fact that slightly influenced vertices show only faintly is pretty much what I would expect. Something like a rainbow color could be used but that gets in the way of seeing the mesh too.%%%

Brecht Van Lommel commented

2013-10-04 15:42:50 +02:00

Changed status from 'Open' to: 'Resolved'

Brecht Van Lommel closed this issue

2013-10-04 15:42:50 +02:00

Sign in to join this conversation.

No Label

Download

What's New

Blender Studio

Manual

Developers Blog

Documentation

Benchmark

Blender Conference

Development Fund

One-time Donations

dyntopo won't collapse unmasked verts while masking is active #36930