FAQ

1. What is D3D
2. How can I get D3D
3. Who is using D3D
4. Is there a D3D mailing list
5. What is the timetable for having D3D utilize hardware rendering
6. We ran into a problem installing D3D on a Linxu/PC running RedHat 7.2. The program complained about a compatiability issue with a library.
7. How do I bring other models into the D3D Volume Browser
8. When adjusting the opacity of an isosurface the surface is very blotchy, why

What is D3D?

Display 3-Dimensional (D3D) is part of the WFO-Advanced system development aimed at providing an advanced operational workstation to operational forecasters. It allows users to view model output and real-time meteorological data in a three-dimensional interactive display.

How can I get D3D?

D3D is available to NWS Forecast Offices and others.
Note: D3D reads netCDF data files from the AWIPS database, only.

Who is using D3D?

A visual image show the of field deployment locations and status of D3D users.

Is there a D3D mailing list

There are two D3D email addresses for users to interact with the D3D community:
d3d@fsl.noaa.gov: to direct questions to the D3D developers at the Forecast Systems Lab.
d3d-list@fsl.noaa.gov: to communicate with users of D3D, those interested in D3D, D3D developers and onlookers. If you would like to join (or leave) the mailing, read how.

What is the timetable for having D3D utilize hardware rendering?

We are looking into this.

We ran into a problem installing D3D on a Linxu/PC running RedHat 7.2. The program complained about a compatiability issue with a library.

D3D can fail on Red Hat 7.2 because of a shared library file. The 7.2 operating system looks for shared library files ending in ".so" and appears to tack this onto the end of our shared library fui_ipc.sl A system administrator cleverly decided to set up a symbolic link from the existing file to a filename that the system wanted. And, volia -- D3D works.

Change to the tcl/tclD3D/fui_ipc directory:

cd d3d/d3d_demo/app/tcl/tclD3D/fui_ipc

Create a symbolic link to the expected file from the existing file:

ln -s fui_ipc.sl fui_ipc.sl.so

How do I bring other models into the D3D Volume Browser

The answer assumes that D2D can read the model. I'll use an example to help illustrate. Say we wanted to add a regional model called "mesoEta212".

First, make sure you know the model 'key lookup name' in the D2D grid look up tables. (Or, search for it the following way. 'find /awips/fxa/data/ -name browserSourceMenu_3.txt' where /awips/fxa/data/ is your FXA_HOME)

Second, add the keyname to the list of models for the Volume Browser. In this example, the model fits on the Regional Scale so edit d3d/d3d_demo/localization/FSL/config/browserSourceMenu_3.txt and add the keyname.

Third, create a model configuration file. These are located in d3d/d3d_demo/localization/FSL/data and named, by convention, v5d_keyname.fxa. In this example is would be called v5d_mesoEta212.fxa.

To create this file copy a similar file and edit it appropriately.
Line 1 is always FXA
Line 2 is keyname keyname
Line 3 is Nx Ny Nz
Lines 4 to Nz+4 are all the specific Nz pressure levels
Lines (Nz+4)+1 sources the common variables list. If you wanted your own list I can send you more info. Basically, you declare the number of variables and list them, but need to include the a
Line N-1 from the last line is the topo file
Line N the political boundary file. Usually a bcd file found in d3d/d3d_demo/localization/FSL/topo

When adjusting the opacity of an isosurface the surface

What you are seeing are artifacts from the rendering process when a portion of a transparent surface is viewed through another portion of the same surface. When a relatively smooth isosurface (like one of theta) is viewed from above, it will look pretty good. But if it has a significant wave in it and it is viewed obliquely so that one side of the wave is viewed through another side, the artifacts (or "blotches") will appear. These artifacts will also occur with a transparent "blobby" isosurface (like one of wind speed) since this type of isosurface is closed and will almost always have a side viewed through another. These artifacts seem to only occur when viewing a transparent isosurface through itself, but not when viewing one transparent isosurface through another. Consequently, the amount of blotchiness will be a function of the parameter whose isosurface is viewed and the value of the isosurface. Perhaps during training, you were just smart (or lucky) and picked good isosurfaces ;-).

If simulated (or dithered) transparency is used rather than "alpha" transparency, these troublesome artifacts disappear. However, there is less control of transparency values (the only effective values are multiples of 0.25). The type of transparency is toggled with Options->Transparency.

Since supporting hardware rendering for D3D would be difficult with the resources that we have, we have opted to use software rendering only. As a result, we are using the Mesa software rendering library. I _think_ that the source of the artifacts is in the way the Mesa library handles alpha transparency. This as been an annoyance for a while, but I've never had the time that it would take to pull the library apart and try to fix it (if, in fact, this is where the problem lies). The library is a rather hefty piece of code with which I am totally unfamilar, so fixing it will probably take a lot of time. I'll get to it when and if I can.

Last updated: Tue 20 Feb 2002