|
|
| |
In lugnet.cad.dev, Travis Cobbs wrote:
| |
Very nice. I don’t author parts, so I won’t be trying it out. However, I
looked at your tutorial, and I have two suggestions:
- Add “Un-Project”, where it takes a projected shape and flattens it automatically on an axis specified by the user (X, Y, or Z).
- Automatically split pattern polygons so that they don’t cross any polygon boundaries of the template part. Use the un-projected geometry that you generated in the above as the “pre-projected” template, and do the splitting while the pattern is still flat.
The first feature above is probably pretty trivial. The second one would
likely be a lot more work, but would make the tool a whole lot more powerful.
--Travis
|
For unproject MLCad does the job. Symply select everything and set the
appropriate matrix row at 0. This is perfect for triangle, quads... If you have
primitives you have to later automatically fix the “all zero matrix row” using
LDDP.
The pattern split program is actually a special case of a tool I consider
writing for a long time. It would take two sheets of tri/quads and cut them at
the intersection between the sheets. I am slowly clearing my mind about the
whole process... so some day I’ll try to write it. But it is far from obvious!!!
Also if you manually build the pattern as an overlay of the original shape, you
optimize it (eg avoid teeny triangles), something an automatic cut wouldn’t do.
Philo
|
|
| |
In lugnet.announce, Philippe Hurbain wrote:
| |
This new tool allows easy creation of non-flat patterned LDraw parts. Create
a flat pattern and stamp it on a 3D former.
Program, documentation, source code are available
here.
|
This is great! Like Tim, I’d been thinking about writing a tool like this for a
long time. Thank you very much for making this available.
Steve
|
|
| |
In lugnet.cad.dev, Timothy Gould wrote:
| |
In lugnet.cad.dev, Philippe Hurbain wrote:
| |
In lugnet.cad.dev, Timothy Gould wrote:
| |
Wow. Awesome. I’ve been meaning to write something like this for ages so
I’m very happy. Would it work on minifig faces too?
|
It should. One interesting aspect is that while photo/scan distorts the
pattern, flattening the shape to put it in quad2dat does the same distortion
so they compensate each other!
Philo
|
I was curious if it was a direct projection or an area preserving one. A nice
variant might be to write an area preserving variant although that would get
quite complex.
Tim
|
It’s a direct parallel projection. I think it’s the best since the photo (with a
long focal lens) or a scan has the same projection.
Philo
|
|
| |
In lugnet.cad.dev, Philippe Hurbain wrote:
| |
In lugnet.cad.dev, Timothy Gould wrote:
| |
Wow. Awesome. I’ve been meaning to write something like this for ages so I’m
very happy. Would it work on minifig faces too?
|
It should. One interesting aspect is that while photo/scan distorts the
pattern, flattening the shape to put it in quad2dat does the same distortion
so they compensate each other!
Philo
|
I was curious if it was a direct projection or an area preserving one. A nice
variant might be to write an area preserving variant although that would get
quite complex.
Tim
|
|
| |
In lugnet.cad.dev, Timothy Gould wrote:
| |
Wow. Awesome. I’ve been meaning to write something like this for ages so I’m
very happy. Would it work on minifig faces too?
|
It should. One interesting aspect is that while photo/scan distorts the pattern,
flattening the shape to put it in quad2dat does the same distortion so they
compensate each other!
Philo
|
|
|
