Text archives Help
- From: tom fogal <tfogal@sci.utah.edu>
- To: iv3d-users@sci.utah.edu
- Subject: [IV3D-USERS] Re: odd color volume bug?
- Date: Tue, 08 Mar 2011 10:31:36 -0700
Hi Collin!
Colin Poczatek <jpoczatek@rics.bwh.harvard.edu> writes:
>
Ok, I've convinced myself that this is a bug but as usual it's
>
possible I'm wrong. It's possible that I broke our QVis export code,
>
but I kinda doubt it.
>
>
Background: So we have multichannel data (16bit) and commonly view
>
it in 2D by taking (Channel-A / Channel-B) (now a float) and mapping
>
that to a hue scale and mapping Channel-B to alpha. We've been doing
>
the same thing making color volumes (QVis format) and rendering in
>
iv3d.
>
>
In the attached screenshot on the left there's a file I made a while
>
back, and on the right is the same data with a different hue scale
>
isosurface and 1dtf rendered.
How long ago is "a while back"? A year ago? 2?
>
When I change the bounds on the hue scale so more voxels are
>
saturated (ie close to max hue) those voxels aren't rendered as
>
bright but appear very dark. This is apparent if you look at the
>
"pink ball".
Can you turn off lighting and tell me if that fixes the problem?
>
This isn't an isosurface issue as you can see in the 1dtf view, with
>
a clipping plane to show the interior of the "pink ball".
>
>
To double check that I was writing values that made sense, I ran the
>
.raw part of the QVis file through unu and grabbed a little 3x3x3
>
nrrd from the center of the "pink ball" and looked at the bytes in
>
a hex editor. RGBA values are all (212, 55, 212, 155) (+/- 1 or 2)
>
which seem to me to be sensible values.
I have a vague memory of messing with these shaders so that they'd be
easier to use with 3 (not 4) channel data. One thing to note is that
you really want to set alpha to 1 (i.e. the top of the TFqn) across
your entire dataset if you have 4-channel data. Otherwise you get some
really strange results.
We should really detect 4-channel data and change the default transfer
function, to be honest... patches welcome ;)
If the alpha in the TFqn isn't the issue, could you send me your data
(preferably in QVis format so I can make sure conversion isn't messed
up) + transfer function?
Best,
-tom
Archive powered by MHonArc 2.6.16.