Color RP Systems

From: Albin Hastbacka (Sanders  Design International, Inc.), Peter Gien (Pogo  International, Inc.), Juergen Bauer (University of Bremen), André Dolenc (Helsinki University of Technology), Michael Brindley, Chuck Kirschman (Clemson  University), Stephen Rock (Rensselaer  Polytechnic Institute), C. Rooney (Brock Rooney & Associates  Inc.), Gill Barequet (Tel Aviv  University), Terry Wohlers (Wohlers Associates)
To: RP-ML
Date: Sunday, December 3, 1995
Subject: Color RP Systems
     I have asked several people what their opinion is relative to a color RP
machine.  The concept would be a machine that use a clear "build" material
so that colors placed outside or inside of the model could be seen.  It
would have obvious application in the medical area for showing a tumor in
blue and the blood vessels in red in an organ like the liver.
     MATERIALIZE of Belgium has shown an interesting use of color in RP at the
recent AUTOFACT conference.  At their exhibit, they showed a two color
version of a skull with a tumor shown in a different color than the rest of
the skull.  The extent to which the tumor had spread around the base of the
tumor was clearly visible by using this technique.
     Besides mediical applications, are there some other good uses of color in
the RP business?

Date: 03 Dec 95 23:33:57 EST
     Al, in years past I was involved in a project dealing with a power-plant design. Through the course of this project, I became familiar with specialized CAD tools for plant design. In some phases of the design, there was such a tangled mess of pipes, that a color coded model would have been great. The current state of the art is to employ model builders to create a life-like model of the plant from foam and balsa wood. This would apply to all petro-chemical plants as well.
     But coming back to the old STL format, how are we going to shoe-horn color information into it? As for textures, perhaps Mr. Brock Rooney might wish to enlighten us on how many triangles would be needed to represent realistic surfaces, such as grass and hair?
Peter H. Gien
POGO International, Inc.

Date: Mon, 04 Dec 1995 12:14 ZE1
     Yes !
     When I had my first contact to PR in 1993, it was an article in the  journal "Deutsches Architektenblatt", which means "German Architect  Paper". The issue countained an article from a Mr. Weisgerber of  University of Karlsruhe, architectural department. He explained the SLA  principle and showed some examples of buildings "materialised" by SLA.
    The article ended up with remarks on the role of material and colour  reception in architectural models, and therefore with the inadequacy of  current rp techniques and materials.
     Since then, I am waiting for two things to happen in the rp industry:
1.  Manufacturing machines available which build metal or high-strenght  plastic PARTS.
2.  Low-cost desktop 3D-Printers available which build coloured MODELS.
     Houses (normally) are one-of-a-kind products. Think of the number of  architects and construction engineers in your city, and you get an idea  of a market potential !
Regards, Juergen

Date: Mon, 4 Dec 1995 13:33:05 +0200
BAU writes:
 > Houses (normally) are one-of-a-kind products. Think of the number of 
 > architects and construction engineers in your city, and you get an idea 
 > of a market potential !
     There is a place for RP, and its use will tend to grow as the costs of making models decreases, but the market potential does not seem to be as large as you suggest.
     Several years ago I contacted the architecture Dept. of this University. One person there knew more about RP than many engineers. The problem with using RP in architecture is costs. In general, architecture models do not require high accuracy, they can be large, and it is very very cheap to build them the traditional way.
     IMHO, virtual reality technology is more appropriate for this application area.
Regards, Andre'

Date: Tue, 05 Dec 95 20:59:35 PST
Peter Gien wrote:
> But coming back to the old STL format, how are we going to shoe-horn color
> information into it? As for textures, perhaps Mr. Brock Rooney might wish to
     If we cram color information into an STL file, it will no longer be an STL file!  Of course, if the STL file format had been tagged, it would be quite easy to add additional information (of any type) to  the file/model.  And any application which does not know about or care about some of the data could easily skip it.
     For RP purposes, we do not want different colors for different triangles;  we want different colors/materials for different VOLUMES.  The triangles are just a way of defining the desired volume.  Therefore, it seems that you would want to associate different colors/materials with different tessaleted volumes.  It would be best to keep both the model and  any auxillary information (including color/material) together (i.e., in the same file).  It would also be desireable to keep the various pieces together (in the same file).  Of course, a tagged data format would handle all of this in its sleep, so to speak.
     I hereby publically request that anyone, anywhere, making a data file format which is intended to be used beyond one isolated computer for a limitted time, use a tagged data format.  As a minimum, the beginning of the data file should identify the data which is to follow.  File names are readily and easily changed and so are not adequate for identifying the contents of a file.
     Ok, I'll get down off my soapbox now.  Thanks for your patience.
  --> Mike Brindley   brindley@ece.orst.edu  Corvallis, Oregon, USA, Terra

Date: Wed, 6 Dec 1995 13:08:37 -0500
From: ckirsch@eng.clemson.edu
> Mike Brindley wrote:
> an STL file!  Of course, if the STL file format had been tagged, it
> would be quite easy to add additional information (of any type) to 
     Well, if you look at the Binary STL file, there is that 2 byte pad at the end of each facet.  I have on occasion used that for "special use" type stuff when playing with STL info.  The readers don't care, so it is backward compatable.  Just a thought...
> we want different colors/materials for different VOLUMES.  The triangles
     This is the downfall of what I said above.  Mike is absolutely right. Tagging facet colors will work fine to build a model of Mount Fuji, but to build the Tumor-in-the-Skull you need to color volumes corectly. This poses some interesting file and data manipluation questions.  If I can lay down any of 4 colors, which one(s) do I use in an area??  I figure if we throw the problem at Mr. (Dr.?) Rock, Mr. Rooney and Mr.  Brindley, 3 good answers should fall out.
chuck

Date: Wed, 6 Dec 1995 15:10:18 -0500
From: rocks@cat.rpi.edu (Stephen J Rock)
Chuck Kirschman Wrote: 
> I figure if we throw the problem at Mr. (Dr.?) Rock, Mr. Rooney and Mr. 
> Brindley, 3 good answers should fall out.
     Thanks Chuck...
     You could create multiple boundary reps to represent each unique (homogenous) material or color volume within a part, as Mike suggests, and then tag each  solid accordingly.  Perhaps you can embed the color in the solid name?
     However, I'd say the fun starts when you want to vary the material  composition (color) continuously in 3D.  This issue was raised at the SFF95  Symposium and someone mentioned using a voxel-based appraoch, where the model is discretized into small elements each having a specified composition.
     It seems there must be a more efficient representation and composition    definition method which does not involve voxels.  I think it is worth considering interpolation-based methods.  Specify composition at certain critical locations within a structure, then let the composition in between be some combination of the specified materials (colors). 
     Solving this problem effectively should have utility far beyond making colored hardcopy.
(Mr.) Steve

Date: Thu, 7 Dec 1995 10:54:23 -0500 (EST)
From: BROCK ROONEY <brockrooney@delphi.com>
     Using multiple volumes to indicate color:
     Many programs that read STL files allow concatenated STL files.  Ours do. At least some 3D Systems programs do.  Each one could indicate a different color.
  unix cat file1.stl file2.stl >total.stl
   dos copy /b file1.stl+file2.stl  total.stl
     Many programs which manipultate or slice STL files allow the processing of mulltiple STL files.
     BR&A STH format supports multiple volumes and color info, among other things.
     The 2 byte attribute count in STL is generally ignored, and has been used by some CAD systems to store Layer/Level or Color info.  This can indicate Triangle Color.  Volume Color could be derived by the silce software.
     Intermediate colors/shades could be produced by the slice software by dithering.
     Cubital systems have been used to produce colored regions by trapping the (colored) support wax in voids.
C. Brock Rooney, Pres., Brock Rooney & Associates Inc.  (Brockware)

From: Barequet Gill <barequet@math.tau.ac.il>
Date: Sun, 10 Dec 1995 12:56:16 +0200
BROCK ROONEY <brockrooney@delphi.com> writes:
> Cubital systems have been used to produce colored regions by trapping
> the (colored) support wax in voids.
     Back to the old days, there were several applications of this feature.
1. Molecular models. Identifying different atom types by different shapes and sizes of the fillers (wax voids trapped in the plastic spheres ["atoms"]).
2. Grey level. A grid of voids, uniformly positioned but varying in void sizes was used for generating a grey-level "picture" on the boundary of the model.
3. Business cards. Flat plastic "business cards" contained the owner's name and details written in engraved voids filled with wax. (The "font" however was quite ugly.)
4. Assembly marks. Huge models were broken into several pieces with wax-filled voids serving as assembly marks.
5. Blood. There was a blood-vessels model (data obtained by an MRI scanner), where the (blue) wax indicated the blood and the (light-yellow) plastic formed the vessels.
Etc. Etc.
     Cubital also organized a workshop on heterogeneous models during March 1993. Proceedings are available (so I assume; at least I have a copy) from their p.r. people.
Gill.

Date: 11 Dec 95 19:12:47 EST
From: Terry Wohlers <73417.1465@compuserve.com>
Subject: Heterogeneous Models
> Cubital also organized a workshop on heterogeneous models during March 1993.
> Proceedings are available (so I assume; at least I have a copy) from their
> p.r. people.
     Gill is being modest.  He organized the workshop, which, by the way, was very successful.  The event was the brainchild of Itzchak Pomerantz, inventor of the Solider system and the former president and CEO of Cubial.  Interestingly, the workshop marked the beginning of his departure with the company.  The first day of the workshop was the first day on the job for Yehudah Baron, the current president and CEO of Cubital.