Re: A 3D mesh and texture.


<stefanbanev@xxxxxxxxx> wrote in message
news:95cbbc2f-bd7d-44db-8ee3-66f83b8d273b@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
On Jan 2, 9:50 am, "jbwest" <jbw...@xxxxxxxxxxx> wrote:
"A n g l e r" <p k o n i u s z @ h o t m a i l . c o m> wrote in
messagenews:flj5kq$kfr$1@xxxxxxxxxxxxxxx





You seem to fail to understand something basic about textures and
texture
cooridnates. No one does one texture per polygon.

The reason why I asked the question in the first place was so to find
out
how to stretch a texture over a certain region - so I don't seem to
misapprehend the concept I hope ;) Anyhow, thanks for the tips.

As to pixel-wise data, erm, perhaps I should ask how to replace quads
with
some convex representation. The goal is to smooth the volumetric data
and
approximate plus remove any arisen discontinuities - though I'm not
sure
again how to address this without fooling around triangle-based mesh.
It's
been like 5 years since I programmed utilizing OpenGL - so please bear
with me.

Cheers.

do you want to create an isosurface? Two basic techniques are marching
cubes
(tetrahedra, etc) that generates triangle data, and isosurface -style
rendering use some sort of directVolume Renderingtechnique. Medical
imaging stuff has pushed the field of texture-basedvolume renderingto
dramatic new displays.

something like this;

http://www.cs.utah.edu/~jmk/sigg_crs_02/courses_0067.html

With a smallish dataset, you can raytrace in realtime using a shader
program. There's tons of stuff onvolume renderingout there, including
open
source & ready-to-use applications.

http://www.volvis.org/- Hide quoted text -

- Show quoted text -

There's tons of stuff on volume renderingout there,


jb>There's tons of stuff on volume rendering out there,

It's quite true however, there's a few who provides an authentic high
quality iso-surfaces by means of direct volume rendering and I'm aware
about only one who may do it interactively for big projection plane
and big datasets. The major issue is that to have authentic high
quality iso-surfaces by means of direct volume rendering the sampling
density along ray has to be around 8...32 samples per cell and for many
cases tri-linear interpolation is not good enough, higher order of
interpolation has to be applied . May be I'm wrong and you may know
who can do that so the reference would be appreciated. Theoretically,
pre-integrated VR should provide an accurate iso-surface
representation however I have not seen any practical implementation
which does better job then straight forward super-sampling ray-casting
VR; besides, pre-integrated VR is not a quite flexible as well.

--sb



Can you better explain your requirements. What do you mean by "big
projection plane" ? How many pixels ?
What do you mean by "big dataset" ? 100's of megabytes? Gigabytes ?
terabytes ?

I guess I will beg to differ re: direct volume rendering (texture based, not
raycasting) not being able to achieve High Quality. Doctors are extremely
picky about visual quality (let's hope so!). There's enormous amounts of
stuff on the subject at IEEE VIS symposia (and/or EUROGRAPHICS). Google
"Visible human", that's considered a good-sized dataset these days ( a few
gigabytes). ALso, you seem to mix raytracing and direct texture-based volume
rendering. There's no explicit ray in the latter.

Your original post was talking about 100x100 -- is that what you consider
big for a volume slice ? Yes, a volume of that order would require skillful
interpolation to avoid clumpy looks. I'd interpolate the volume to a decent
density and then directly render it.

jbw


.

Relevant Pages

  • Re: A 3D mesh and texture.
    ... No one does one texture per polygon. ... rendering use some sort of directVolume Renderingtechnique. ... quality iso-surfaces by means of direct volume rendering and I'm aware ...
    (comp.graphics.api.opengl)
  • Re: Creating Circular Surface in DirectX
    ... rendering your surface by using a textured quad with alpha blending and a ... directly samples your rendered texture. ... draw it onto the screen via a 3D solid quad. ... be able to do the circular cut out yourself... ...
    (microsoft.public.win32.programmer.directx.graphics)
  • Re: frame rate drop after multiple sprites - direct3d
    ... You are rendering a sprite at a time still, as a pair of triangles, ... You are not sorting your sprites by texture. ... > number on top) the frame rate would drop 10fps. ...
    (microsoft.public.win32.programmer.directx.graphics)
  • Re: HARDWARE: Open-Source-Friendly Graphics Cards -- Viable?
    ... more than a fuzzy texture. ... The other more pressing reason is that the host interface ... puts a very low upper-bound on your triangle rendering rate. ...
    (Linux-Kernel)
  • Re: Threading & LockRect?
    ... LockRect() to copy the bits to a Texture surface in "SystemMem". ... rendering with v-sync on to minimize tearing. ... I've tried the "NoSysLock" flag with no help. ...
    (microsoft.public.win32.programmer.directx.graphics)