Will Holcomb wrote:
> I noticed that an overlay library recently was posted. I looked at it and
> it is beyond my current experience but I'm sure some playing would be
> productive.
The overlays that you see are not true "overlays" per-se. They don't use
any of the card's native hardware overlay capabilities (only just
recently making it to PC quality video cards with the GeForce2).
For Java3D there are two basic techniques in use currently - placing a
textured object right in the viewing plane and the overlay is the
texture on the object, and reading the raster from the video card,
drawing to it, and then putting it back (typically done in the preSwap()
method of Canvas3d). Neither of these are highly efficient or use any of
the native resources of the card, but they do work.
> I never have understood though why exactly I can't just draw onto the
> image after it has been rendered.
Well, the approach you outline below is exactly that. The problem is
really in the implementation. As you have discovered, the use of the
Graphics3D is for all intents and purposes useless. That's the reason
for the efforts with the competing overlay implementations. Myself and a
few others have been mainly working with Dave Ys implementation for a
little while now. I'm working on the drawing functionality now so that
you can "scribble" on the screen. All of this code will be soon
available in the j3d.org code repository (http://code.j3d.org/) along
with a lot of instructions on how to use it etc.
> P.S. Is anyone working on bridging the java3d api and the jmf?
[Wave's hand] Nothing that I'm doing fulltime. I've got a bunch of code
from some brazillian students for a movie texture node for Xj3D that I
haven't integrated yet. There is also a good demo for this on the JMF
site. Don't have the link handy, but it is floating around in the Java3D
FAQ (http://www.j3d.org/faq/). Also have other interests in using it for
various things. One known problem is that you can't currently use JMF
for the audio rendering.
> working on something that makes a Canvas3D a jmf PushBufferStream but it
> is still very buggy. I have had to put it off to do this VB stuff but I
> was wondering if anyone else was working on something like this that they
> were planning on releasing under some sort of open licence. The stuff I
> have is at:
>
> http://www.himinbi.org/public-cvsweb/cvsweb.cgi/RecordingCanvas3D
Well, if you want to contribute it to the j3d.org code repository, I'd
be happy to take it and help you along the way. I already have the
basics in place for image capturing from the canvas so if you felt you
could work with that, then feel free to go for it.
> P.P.S. I read the stuff about floating point errors and what not in large
> transformations and I was wondering if this was the cause of the problem
> shown in:
>
> http://www.himinbi.org/jars/Translation_Test.html
>
> When you run the distance up to around a million meters you start seeing a
> jitter that gets worse and worse as you get farther out. It looks like a
> rounding error somewhere, but it would have to do with the Transform3D
> class.
Nope. that's floating point inaccuracy alright. Finally found the MERL
paper about this that I read a long time ago. Take a read of this:
http://www.merl.com/reports/TR95-16a/locales.html
Go down to figure 3 and the section "Precise Positions and Velocities".
--
Justin Couch http://www.vlc.com.au/~justin/
Freelance Java Consultant http://www.yumetech.com/
Author, Java 3D FAQ Maintainer http://www.j3d.org/
-------------------------------------------------------------------
"Humanism is dead. Animals think, feel; so do machines now.
Neither man nor woman is the measure of all things. Every organism
processes data according to its domain, its environment; you, with
all your brains, would be useless in a mouse's universe..."
- Greg Bear, Slant
-------------------------------------------------------------------
===========================================================================
To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
of the message "signoff JAVA3D-INTEREST". For general help, send email to
[EMAIL PROTECTED] and include in the body of the message "help".
