Subject:
|
Re: Crazy, OK Heretical Idea ...
|
Newsgroups:
|
lugnet.cad
|
Date:
|
Thu, 1 May 2003 01:25:02 GMT
|
Viewed:
|
736 times
|
| |
|
|
Steve Bliss wrote:
> In lugnet.cad, Xanthra47 wrote:
>
> > What existing formats come close ?
> > Md2 (Quake II)
> > Milkshape
> > OBJ
> > Md3 (Quake III)
>
>
> How do these formats do with the number of triangles normally found in an LDraw
> model? My impression is that these formats (and their rendering engines) normally
> deal with far fewer triangles/quads than LDraw.
That's my impression also. They seem to be oriented towards
(relatively) lower polygon counts, and increased usage of
textures. They use the textures to achieve the details so
that they don't need to bog down the system with more vertices
and polygons.
The CAD oriented file formats though are probably the other
way around. Optimized to truly represent the shapes and
volumes of objects accurately. While not exactly what we
might need either, these CAD formats are probably closer
than the game formats.
>
> > So what would we want in a new file format ? Here's a short list
> > 1. Easy upgrade path (convert the current parts easily)
> > 2. Fast rendering
> > 3. Cross-Platform portablity
> > 4. Supported by common modellers/renders
> > 5. Human Readable/Editable
> > 6. Capable of animation
> > 7. supports "mating" of parts
> > 8. supports texture/material mapping
>
>
> Here's some features of LDraw I'd like to see retained:
>
> 1. Simple language for people new to 3D graphics.
>
> Hmm, I guess that's the only one. :)
I'm with you on that.
I don't know if there is a format out there that is designed
for the things we are interested in, but I'm not sure that
it's going to be that easy to use (or to read) if there is.
If you want to talk real Heresy, then I think we should
discuss having different (but related) file formats for
primitives and parts vs. models, assemblies, and scenes.
Most of what we want to do doesn't really need the part
files to change much.
A new format could be developed from the level just above
the part level (the model or submodel level,) that was
totally new, and still referenced the current part files.
These types of things need to be explored if we want to
move this subject out of displaying static models, and
into the animation/movie, or simulation areas.
A movie might contain many scenes, Each scene might have
a script. To show hoe the different models (or parts of
models) move as time passes.
A model needs to be able to describe the different collections
(possibly subfiles) of parts that all move as one unit
(think all the gears on an axle) and the way that these
collections relate, and cause the movement of the other
collections (when this gear on this axle turns X degrees
then this gear on this other axle turns that axle X*40/8
degrees - for example)
All of this starts to imply that the parts in models may
be sorted or organized differently, possibly having all
the parts that move together near each other in a model
file. This would be incompatible with the way we currently
tend to sort parts in a model to make creating instructions
easy.
I'm not trying to solve this problem now, just pointing
out that many of the new areas we want to take this LEGO
CAD stuff into are going to mean rethinking much more about
how we work and interact with our creations than just
the file format. On the other hand, Hardly any of it
really needs the *part* and *primitive* file format to
change.
Xanthra47 also mentioned the 'Vertex lists' as an example
of (somewhat) recent 3d technology that we should really
be making use of. There are others.
Actually I think all of the common programs today probably
do keep a hash table of points per part, and reference them
through an index, or passing the tables to the hardware.
Testing coordinates to see if they are new or not is not that
computationally expensive, and only needs to be done at part
load time.
On the other hand though saving vertex lists in the part
files, could help to make them even smaller. And Size is
currently one of the LDraw format's biggest advantages.
(I think anyway.)
The 3D optimization I'd most like to see the parts library
add (After BFC is done.) is to group the polygons in
a part in 'strips'. Having hints in the part files for
which polygons should be in which strip, would increase
performance of many tools significantly (I think) and
figuring out strips automatically at load time isn't
always easy, correct, or fast.
Just more things to think about....
-Kyle
--
_
-------------------------------ooO( )Ooo-------------------------------
Kyle J. McDonald (o o) Systems Support Engineer
Sun Microsystems Inc. |||||
Enterprise Server Products Kyle.McDonald@Sun.COM
1 Network Drive BUR03-4630 \\\// voice: (781) 442-2184
Burlington, MA 01803 (o o) fax: (781) 442-1542
-------------------------------ooO(_)Ooo-------------------------------
|
|
Message has 1 Reply:
 | | Re: Crazy, OK Heretical Idea ...
|
| (...) OK, so that leads to one of my desires in an upgraded (or different) GDL: level-of-detail control. Basically, there should be language elements that allow authors to flag the priority of different elements in a file -- so the main outside body (...) (22 years ago, 1-May-03, to lugnet.cad)
|
Message is in Reply To:
| | Re: Crazy, OK Heretical Idea ...
|
| Quoting out of the original order: (...) Flaming? Well, you should go research LDraw II, now that Jacob's server is back online. (...) How do these formats do with the number of triangles normally found in an LDraw model? My impression is that these (...) (22 years ago, 1-May-03, to lugnet.cad)
|
41 Messages in This Thread:
 
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
This Message and its Replies on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
