|
Click the pic for more
info
Heres some more examples of my software at work, rendered in POV-Ray:
Heres how studs-out building compares to studs-up:
Enjoy!
--Bram
|
|
|
In lugnet.announce.moc, Bram Lambrecht wrote:
|
Click the pic for more
info
Heres some more examples of my software at work, rendered in POV-Ray:
Heres how studs-out building compares to studs-up:
Enjoy!
--Bram
|
Hi Bram,
Some excellent algorithmic design there. Are you sure you shouldnt be doing
combinatorics rather than engineering?
A few questions about it:
1) Is the whole thing algorithmic or do you have to specify where your
snot-centers (I see there are at least two in the bunny) will be placed? 2) If
the whole thing is algorithmic how do you determine where the snot-centers are
placed? 3) Will you be sharing your code at any point?
Superb work anyway.
Tim
|
|
|
In lugnet.build.sculpture, Timothy Gould wrote:
|
In lugnet.announce.moc, Bram Lambrecht wrote:
|
Click the pic for more
info
Heres some more examples of my software at work, rendered in POV-Ray:
Heres how studs-out building compares to studs-up:
Enjoy!
--Bram
|
Hi Bram,
Some excellent algorithmic design there. Are you sure you shouldnt be doing
combinatorics rather than engineering?
A few questions about it:
1) Is the whole thing algorithmic or do you have to specify where your
snot-centers (I see there are at least two in the bunny) will be placed? 2)
If the whole thing is algorithmic how do you determine where the snot-centers
are placed? 3) Will you be sharing your code at any point?
Superb work anyway.
Tim
|
And apparently my ability to ask questions far supersedes my ability to follow
links. I see youve written a handy paper to explain the method. Thanks.
Tim
|
|
|
In lugnet.build.sculpture, Timothy Gould wrote:
> In lugnet.announce.moc, Bram Lambrecht wrote:
> > http://lego.bldesign.org/LSculpt/
>
> Some excellent algorithmic design there. Are you sure you shouldn't be doing
> combinatorics rather than engineering?
Building stuff is fun.
> A few questions about it:
> 1) Is the whole thing algorithmic or do you have to specify where your
> snot-centers (I see there are at least two in the bunny) will be placed? 2)
> If the whole thing is algorithmic how do you determine where the snot-centers
> are placed? 3) Will you be sharing your code at any point?
>
> Superb work anyway.
Thanks! Shoot me an email if you're interested in the code.
--Bram
|
|
|
Havent had time to really read into this (will though!) but wanted to post a
quick congrats based only on first impressions. They look great! Well done!
God Bless,
Nathan
|
|
|
Wow... if I were still building witth LEGO bricks with any regularity, this
style is all Id be doing. Beautiful stuff.
Eric
In lugnet.announce.moc, Bram Lambrecht wrote:
|
Click the pic for more
info
Heres some more examples of my software at work, rendered in POV-Ray:
Heres how studs-out building compares to studs-up:
Enjoy!
--Bram
|
|
|
|
In lugnet.announce.moc, Bram Lambrecht wrote:
> <http://lego.bldesign.org/LSculpt/
> Enjoy!¬
> --Bram
I sure did!!! Outstanding...
Three things I found:
- LSculpt seem to prefer "Unix" style PLY (with only CR at the end of text file
instead of CR/LF Windows style). I had to convert Stanford bunny
(bun_zipper.ply) to be able to use it.
- After conversion with 3D Object Converter (from ply to ply), Lsculpt barfs and
generates nothing. It says "output size is -1.#J x -1.#J x -1.#J studs".
- I tried the horse
(http://www-static.cc.gatech.edu/projects/large_models/horse.html) and get a
studs IN sculpture. Wrong mesh winding? Any way to correct that?
Once again, congratulations!
Philo
PS: idea for improvement: include half stud jumper offset ;o)
|
|
|
Awesome work, Bram. Any chance of getting an MPD or some screenshots of the
innards of the cow?
James Wilson
|
|
|
|
lugnet.announce.moc, Bram Lambrecht wrote:
|
|
Heres some more examples of my software at work, rendered in POV-Ray:
|
|
Heres how studs-out building compares to studs-up:
|
Wow!
I have been working on some sculptures, scenerys and a huge globe using these
studs out building technique lately. And this would be a great addition.
I presume I could also use this in making a landscape with the correct 3D
triangle meshes files? This would save me so much time
no more placing all the
plates and then rendering to see if it looks ok, Now I can just make a quick
landscape model and plug it into your new program.
Im so glad to see some programmers still coming up with some cool software for
us AFOLS! Keep it up! Now if I can just get back to work with out thinking of
all the possibilities today.
Mike Gallagher
mikeslego
|
|
|
In lugnet.announce.moc, Bram Lambrecht wrote:
|
Click the pic for more
info
Heres some more examples of my software at work, rendered in POV-Ray:
Heres how studs-out building compares to studs-up:
Enjoy!
--Bram
|
Wow. It amazes me what you can come up with :-)
I love LEGO. It allows people to do such cool things like this.
Spotlighted.
-Anne
|
|
|
Very cool.
So it looks like this was used for a class project (based on the header of the
paper). If so, I hope you got an A. The paper itself looks exactly like papers
Ive looked at in the past that were presented at SIGGraph.
It seems to not like the binary_little_endian format in ply files. When I run
the cow and elephant files found
here it doesnt complain,
but it also doesnt generate any output. Both of those files use
binary_little_endian as their format.
--Travis
|
|
|
In lugnet.build.sculpture, Philippe Hurbain wrote:
> In lugnet.announce.moc, Bram Lambrecht wrote:
> > http://lego.bldesign.org/LSculpt/
> > Enjoy!
> > --Bram
>
> I sure did!!! Outstanding...
>
> Three things I found:
> - LSculpt seem to prefer "Unix" style PLY (with only CR at the end
> of text file instead of CR/LF Windows style). I had to convert
> Stanford bunny (bun_zipper.ply) to be able to use it.
I used someone else's library to read in the PLY files, so I'm not sure what it
prefers. STL is a simpler format, so I wrote that parser myself...and haven't
had any problems with it.
> - After conversion with 3D Object Converter (from ply to ply),
> Lsculpt barfs and generates nothing. It says "output size is
> -1.#J x -1.#J x -1.#J studs".
Try saving as PLY binary from 3D Object Converter. As a plus, it'll make your
file smaller!
> - I tried the horse
> (http://www-static.cc.gatech.edu/projects/large_models/horse.html) and get a
> studs IN sculpture. Wrong mesh winding? Any way to correct that?
Hmm... I suppose a commandline switch for flipped normals is in order. The
Princeton horse worked for me. It's the same model as the GT horse, but I'm not
sure if it's the exact same model.
> Once again, congratulations!
Thanks!
> PS: idea for improvement: include half stud jumper offset ;o)
That doesn't really fit with my algorithm, but it would be cool.
--Bram
|
|
|
In lugnet.build.sculpture, James Wilson wrote:
|
Awesome work, Bram. Any chance of getting an MPD or some screenshots of the
innards of the cow?
|
Hey James, right now there are no inner workings on any of the rendered models.
Its up to you to figure out how to build the shape my program spits out. --Bram
|
|
|
In lugnet.build.sculpture, Mike Gallagher wrote:
|
Wow!
I have been working on some sculptures, scenerys and a huge globe using
these studs out building technique lately. And this would be a great
addition.
I presume I could also use this in making a landscape with the correct 3D
triangle meshes files? This would save me so much time
no more placing all
the plates and then rendering to see if it looks ok, Now I can just make a
quick landscape model and plug it into your new program.
|
Yep, that should work. You might get some funky pieces around the edges of the
mesh, but most of it should be good. If you want a studs up landscape, use -q
-n in your commandline.
|
Im so glad to see some programmers still coming up with some cool software
for us AFOLS! Keep it up! Now if I can just get back to work with out
thinking of all the possibilities today.
|
I look forward to some cool results!
--Bram
|
|
|
In lugnet.build.sculpture, Bram Lambrecht wrote:
> In lugnet.build.sculpture, Philippe Hurbain wrote:
> > - I tried the horse
> > (http://www-static.cc.gatech.edu/projects/large_models/horse.html) and get a
> > studs IN sculpture. Wrong mesh winding? Any way to correct that?
>
> Hmm... I suppose a commandline switch for flipped normals is in order. The
> Princeton horse worked for me. It's the same model as the GT horse, but I'm not
> sure if it's the exact same model.
Another handy option might be one to flip bricks pointing down so that they
point up. So the sculpture would be completely studs out execept for the
bottom, which would always be studs up instead of studs down like it is now.
Not sure how this would look on the heads of animals, though. Maybe a tolerance
so that any studs-down pieces within a certain percentage of the bottom of the
model are made studs up. (Of course, that could also be done with
post-processing of the LDraw file.)
--Travis
|
|
|
In lugnet.announce.moc, Bram Lambrecht wrote:
|
Click the pic for more
info
Heres some more examples of my software at work, rendered in POV-Ray:
Heres how studs-out building compares to studs-up:
Enjoy!
--Bram
|
Amazing work Bram! Loved the paper.
Just a quick note: The 3D converter that you link to can take as import
Renderman files (RIB) and output PLY files. RIB is one of the supported file
formats to E-frontiers Poser software. This allows for posable human shapes.
-dw
|
|
|
In lugnet.build.sculpture, Travis Cobbs wrote:
> In lugnet.build.sculpture, Bram Lambrecht wrote:
> > In lugnet.build.sculpture, Philippe Hurbain wrote:
>
> Another handy option might be one to flip bricks pointing down so that they
> point up. So the sculpture would be completely studs out execept for the
> bottom, which would always be studs up instead of studs down like it is now.
> Not sure how this would look on the heads of animals, though. Maybe a
> tolerance so that any studs-down pieces within a certain percentage of
> the bottom of the model are made studs up. (Of course, that could
> also be done with post-processing of the LDraw file.)
This is already an option! Use the command line "-b 1" to get the bottom layer
of plates studs up instead of studs down. Increase 1 to whatever number of
layers you want studs up (it'll do 5 layers at a time, so the next meaningful
layer command is "-b 6") Also handy is "-d z" if your input mesh happens to use
a Z-up coordinate system instead of a Y-up coordinate system. So if your models
are showing up rotated wrong, try that.
--Bram
|
|
|
In lugnet.announce.moc, Bram Lambrecht wrote:
> <http://lego.bldesign.org/LSculpt/
> <http://lego.bldesign.org/LSculpt/bunny_sides.jpg>>¬ {Click the pic for more
> info}
Hmm. Most of that went over my head. Still, there were a few things that I
think confused me due to more than my incipient senility:
In formula 3, shouldn't that be "Let k\in \{1,2,3\} such that . . ." rather than
"Let k=\{1,2,3\} such that . . ."?
And in formula 10, shouldn't it be H_i that is defined, not H? H is the sum
itself, yes?
Nice bunny.
TWS Garrison
|
|
|
In lugnet.build.sculpture, Thomas Garrison wrote:
> In lugnet.announce.moc, Bram Lambrecht wrote:
> > <http://lego.bldesign.org/LSculpt/
> > <http://lego.bldesign.org/LSculpt/bunny_sides.jpg>>¬ {Click the pic for more
> > info}
>
> Hmm. Most of that went over my head. Still, there were a few things that I
> think confused me due to more than my incipient senility:
>
> In formula 3, shouldn't that be "Let k\in \{1,2,3\} such that . . ." rather than
> "Let k=\{1,2,3\} such that . . ."?
I always thought $k\in\{1,2,3\}$ was pretty much equivalent in this situation to
$k=1,2,3$. By your definition you are assigning k to be equal to the set as
opposed to a value.
> And in formula 10, shouldn't it be H_i that is defined, not H? H is the sum
> itself, yes?
I think you're right about this one.
> Nice bunny.
>
> TWS Garrison
Tim
|
|
|
In lugnet.build.sculpture, Thomas Garrison wrote:
> In lugnet.announce.moc, Bram Lambrecht wrote:
> > http://lego.bldesign.org/LSculpt/
>
> Hmm. Most of that went over my head. Still, there were a few things that I
> think confused me due to more than my incipient senility:
>
> In formula 3, shouldn't that be "Let k\in \{1,2,3\} such that . . ."
> rather than "Let k=\{1,2,3\} such that . . ."?
I'm not a math or a CS person really, so some of my notation could very well be
wrong. What I meant is that k should be 1, 2, or 3, whichever gives you the
biggest answer.
> And in formula 10, shouldn't it be H_i that is defined, not H?
Yes.
> Nice bunny.
Thanks!
--Bram
|
|
|
In lugnet.announce.moc, Bram Lambrecht wrote:
> http://lego.bldesign.org/LSculpt/
Some of you have reported errors that LSculpt was not running on your machine.
It seems that by default, Visual C++ expects some libraries to exist on the
user's computer. I've compiled a static build instead, so hopefully it should
work for everyone now.
The fixed Windows binary is up on the site. Let me know if you still get
errors!
--Bram
|
|
|