Exporting locking up with DTS objects

#1 Alright this is fun. i spent time tearing bits of the model off until i could locate what was causing the lockups.

Can anyone guess?

No?

Alright well here's a DTS file. It has one triangle. And this will not export to dif.

subvoicestudios.com/~ramensama/error.dts

Well here's a better description of the triangle. all three points have same Z and Y coords. So they are basically a strait line.



Granted this picture was taken at an angle for dramatic reasons :p

I rotated this "object" a few degrees and then exported it and all was fine. so it seems to have issues if the vertecies are lined up like this along 1 axis.

i rotated the object 90 degree, and it wouldn't work.


Then again this is the type of polygon that was a byproduct of extruding and other things, wasn't intentionally modeled. But hey, now I know. Perhaps a check like this should be included in the newest version of constructor?

#2 Ok, more research. These types of polygons are called Non-Planar. Lightwave will list these polygons in it's statistics window, so this is something to check for before exporting to DTS and before loading up in constructor.

I've noticed that you can have 4 sided nonplanars, but these aren't the bad guys, it's the 3 sided non-planars. I model as quads, and tripple before exporting as DTS. So after trippling, i'll make sure there are no non-planar 3 sided polys.

Maybe i should have created this post in the "feedback forum"

#3 Although no one replied - rest assured that people are reading and taking note!

I've had similar issues, and I might even go back and look for this as the cause.

Kelly

#4 All shapes, regardless of source, must have a volume (cannot be 2 dimensional) to be part of a DIF structure.

Three points (like a triangle) that exist on a plane (or so close as to be within floating point roundoff/error) will not convert to DIF, period.

#5 But we can make it discard the bad polygons - I will look into it.

#6 As noted in the other "fixes" thread, this is now fixed.

#7 Where/When do we get to see your fixes? Oh yeah and thank you for your work on this Jaimi.

#8 As soon as I get to a good stopping point, I will post a download for everyone. Until then, I don't want to swamp my server (it's currently on my DSL line, and is kind of slow).
I have a couple more items to look into before releasing another test. I'll send you an email with the link when it's ready, and you can download it if you like.

#9 @ Stephen, the leftover polygon there ended up being some mistake from the modeling process, so i wasn't aware of it till i started digging. of course it makes sense that a polygon must have volume, it's just a matter of the lightwave exporter, and possibly others, didn't catch.

#10 Some tools create degenerate triangles on purpose - you can use them to stitch tristrips together for more efficient batching.