(11-09-2018, 10:32 AM)josemendez Wrote: A finger is basically a circular (or ellipsoidal) shape. Using a radial support function you can displace particles to the border of the circle/ellipsoid, to make it look like the user is "pressing down" on the slime.
Keep in mind that while the user is dragging, it is very unlikely he will see what's happening under his finger. Something I'd think about would be splatting a fingerprint normalmap on top of the slime, and vanish it over time (a few seconds). This would allow the user to see his "fingerprints" for a while after raising the finger off the screen.
Using more substeps basically means re-doing the entire simulation more times per frame, including collision detection. So yes, if all you're looking for is relatively simple user interaction, substepping is complete overkill (specially for mobile)
Looking slimy, or behaving slimy?
If the first, use a shader with high smoothness/glossiness and a normal map.
If the second, increase distance constraints slack and decrease stiffness a bit. Also consider increasing the solver's "damping" for a thicker/viscous behavior.
I notice you have a ObiClothDragger.cs in the asset. It appears to be an appropriate script close to the purpose. So I tried applying this script to the slime object, with the y-position vector force reversed in script, to show a "press" instead of a drag.
This is the result.
It appears to push the ground zero particle quite deep so a proper finger press effect appears.
But the mouse position does not translate correctly onto the screen. It seems that it does not respect the mouse's world positions.
Also, why does the script use vector4?