advanced rendering for AutoCAD
Would it be possible to have a setting for pattern repeats on materials? Many times carpets, wall papers and fabrics have the pattern repeat slid one direction or the other some distance. Most of the time this isn't an issue, however I have just suffered through dealing with a very particular client and making a large enough pattern to take this into account slowed the rendering substantially.
Thanks for your consideration.
Tags:
No clue what you mean. You'll need to provide an example.
Your talking about seamless tiles and repeating tile patterns. nXt handles these well. It is up to the operator (you) to to use correctly made material maps to achieve a good result. Google seamless tiles or repeating tiles and you'll get a number of tutorials that will walk you through the process. Here are two examples: http://www.photoshopessentials.com/basics/repeating-patterns/
http://www.myjanee.com/tuts/tilings/tilings.htm
BW
I could not find a good example that is easily seen so I've Tried to do a simple diagram. Hopefully that plus the explanation below is enough.
Many items with patterns do not have a direct pattern match where the pattern simply fits into a square grid. When using a one of these fabrics there is an amount of material that has to be removed in order that the picture continues successfully. This creates something of a brick pattern turned on its side - see the attached diagram.
Since it is a repeating pattern, you can create a square with that pattern in a program like Photoshop, that meets the criteria required to create a seamless tile for material mapping in nXt and every other rendering engine available.
The other option would be to use the tile tool. Create a new 'advanced' material,. Under the 'material tree setting' select "TILE". Set the scale to the size of the image you are useing as a texture map. You'll need to 'unlock' the constraints to enter a dimension which is not square. Set the joint to "0" (zero). Then set the offset to a known dimensional offset or get to it by trial and error.
BW
Almost-- the texture coords. won't respect the offset unless I add that feature. Maybe (it would be a confusing checkbox that would be seldom used.)
I've actually gotten this method to work for me in the past, that's the only reason I suggested it. But as I try it again, prompted by Roy's "almost..." reply, I see that it is problematic. I'm not sure I understand why the coords. don't respect the offset. Why would that not be the default.
None the less, a repeating seamless tile can be created from any image to duplicate or represent carpet or wall coverings.
BW
Roy Hirshkowitz said:
Almost-- the texture coords. won't respect the offset unless I add that feature. Maybe (it would be a confusing checkbox that would be seldom used.)
I'm pretty sure it never worked that way.
OK
I think Kingjin has seen exactly what I had tried & also the result that I was getting. When you try and tile a pattern with the tile or the brick tool you get a slight shadow line. For a low-res image its not a big deal but for high-resolution its a problem.
When I add enough pattern to get a full repeat in the typical square format the file size for the image is a lot larger than the original square pattern, many times 16 times as large. This really slows the rendering down.
I don't have any idea of the inner workings of nxt or how complicated what I'm asking for is. IF the tile/brick tool was a little more precise and the joint thickness could successfully be zero it would do the trick. It would also be helpful to be able to align one face on an object while leaving the others as they currently are.
The tile pattern can be set to 0 without joint lines. It still won't work for your application (like I mentioned above), not until I add a feature that locks the mapping to the tiling-- haven't decided yet and certainly won't do it any time soon. You can probably reduce the resolution of the compound image somewhat in your case-- depends on the scene.
Kingjin is using the bump feature which can sometimes show some odd results. He is not using a joint spacing of 0.