LDraw.org Discussion Forums

Full Version: New LDraw tools for parts authors: Unrectifier and Flipper
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
I started these programs a few months ago, but it took me some time to iron out some bugs... and ages to be motivated enough to write documentation! Neither of them will revolutionize the art of LDraw parts authoring, but they can be helpful nonetheless!

As usual, documentation, source code and Windows executables are available (links below).



[Image: ur1.gif]
Rectifier is a useful tool to consolidate triangles into quads and to keep file size down, but doing so make further improvements to the file more difficult. Unrectifier does the opposite, it expands Rect primitives into quads + lines, and split quads into triangle pairs. By default Unrectifier operates on the whole file, but it can also limit itself to key-colored elements. With some limitations, it is also able to update condlines around modified areas.

[Image: fl1.gif]
Flipper allows to flip a pair of triangles (change abc + SPAM CONTENT into abd + dca). This helps in some areas of "organic" meshes (eg. hair pieces) to make them smoother (or on the contrary to create a crease to add an edge line). Flipper operates on colored pairs of triangles, thus allows the user to choose the triangle pairs to swap. With some limitations, it is also able to update condlines around modified areas.
Thank you so much, this is great! I like the idea of Flipper, it is very annoying to manually flip areas of LDD meshes. There are a lot of such wrong areas.
I'll try to update LETGUI quickly.
Well, it is not always so easy to use, improving an area may degrade nearby areas, that you then try to improve, etc.... But at least with Flipper it is easy to experiment!
I too ran into the flipper issue when generating the deformed meshes in LDCad.

Do you now of a way to 'predict' the best version of a split or is manual inspection the only way?
I'm afraid that the eye is the best judge... I considered something like minimizing overall sum of facet angles, but I doubt it would really improve things (and it's probably over my poor programming skills anyway!)
This paper implies that it is indeed possible to optimize triangulation using "discrete curvature"...
This seems very interesting, but I'm not sure it's worth the cost (post processing on the whole set) as we just need to make the choice for a single quad.
LETGUI is now again uptodate.
Thanks, Mike!