Re: [guppi-list] Re: 3D plotting library

[stevenj gil-galad mit edu]

On Mon, 24 Jan 2000, Jon Trowbridge wrote:
> On Mon, Jan 24, 2000 at 03:37:10PM +0000, Aaron Lehmann wrote:
> > OpenGL sounds really hard, and I'm not sure how usable it would be to
> > people without 3d accelerator cards... [...]
> 
> Well, OpenGL certainly adds another layer of complexity.  However, I
> am aiming high with Guppi, and I want to eventually be able to support
> things like complicated volumetric rendering, the kind of imaging that
> is most commonly associated with medical applications (MRIs, etc.).
> That level of 3d is way beyond what we'd want to reinvent from
> scratch.  So if OpenGL is going to enter Guppi anyway, there is no
> harm in it happening sooner rather than later.

For an example of a (GPLed) volumetric rendering program that uses OpenGL,
see Vis5D:

	http://www.ssec.wisc.edu/~billh/vis5d.html

I use this myself, and the performance is quite acceptable with
software-only Mesa OpenGL.  And this is volumetric rendering...simple
things like surface plots will be quite fast in Mesa.

(Note that you use this kind of volumetric rendering for a lot of other
applications besides medical MRI images, e.g. weather simulations,
computational fluid dynamics, computational electromagnetism...)

I strongly agree that OpenGL is the best solution for 3d plotting.  Not
only do you get a lot of sophisticated rendering and live-interaction
features for free, but you also take advantage of hardware acceleration if
it is available.

Cordially,
Steven G. Johnson