Tri-planar mapping scale has changed v13 using older files created in v12?

[edub]

I have updated to Corona 13, and now working in some files that were created in v12, and I'm having some problems with tri-planar mapping scale being different (much smalle tiling of textures)...  is this documented somewhere, or is it a bug? Is there a scale setting that I can change to get the expected scale from my old files?

[TomG]

Can you share some examples? That is, a 12 render then same scene in 13, plus a screengrab of the material set up (or, a version of the scene, can even be stripped down to just a cube with the offending texture applied etc. so long as it shows the difference in 12 and 13).

[edub]

Hi Tom...  I don't have v12 installed anymore, so I don't know how to verify this issue.  I will try to install v12 on another machine and test it out.  thanks

[edub]

-Hi Tom - we are still having the problem.  We even tried uninstalling 13 and reinstalling 12, but that didn't change anything.  Here's a screenshot of partial render done in v12 (intended tiling) and also in v13 (smaller scale tiling).  i have included screenshots of the material setup from v13 as well, but I don't have access to the original v12 install.  I will ask my colleagues if they have a v12 install so we can create a simplified file - this one is too large and highly confidential

[maru]

I've found some older reports and it looks like before Corona 13, the CBitmap + Triplanar combination did not work at all. What you are seeing in V13 should be the correct results.
If you would like to see identical mapping in Corona 13 as it used to be in 12, try removing the Triplanar node from the material setup (just connect CBitmap directly to the material channel). Please let us know if it worked.

[edub]

ok, thanks Maru.  I think I had created this material using a cosmos material as a starting point, and modified some shaders.  I probably don't need the triplanar in this case, but a lot of the cosmos materials are using it, so I can see this being an issue in many cases - or are you saying it's only when combined with the mapping randomizer?  Or was it just a bug in the v12 that was never identified?