mjl@ut-emx.UUCP (Maurice LeBrun) (02/25/89)
--- I want to use my Amiga in my research (finally), and am asking for advice on how to address my graphics needs. I develop/run large scale n-body simulations (magnetically confined plasmas, actually), and want to do color fill-type plots on the Amiga to assist in interpreting the results. As most people who've looked at them know, color fill plots are very nice for visualizing functions of two variables. Neither contour plots nor 3-d perspective plots really compare (the latter are really only good for very smooth data sets). In addition, the extra "dimension" of color allows for more complex displays -- if I wanted to plot both contours of temperature and density as a function of (x,y) [on the same plot], I could do one as a color fill and the other as either contours or 3-d perspective. Anyhow, I really don't know where to start with this. The only experience I've had doing this sort of stuff comes from playing with Mike Norman's (of NCSA) ImageTool for the MacII, which takes the easy way out by just using a 256 color palette out of a very large number of possible colors. However, this may be overkill, I don't know. In any case, I want to use the Amiga because (a) I own one and practically live in front of it, (b) it's possibly better at animation, (c) better output options for animations (composite video -> videotape). As you can tell, I'd eventually like to build some animations. But getting back to the central question (s): 1) What graphics modes are likely to be best suited to my problem? Note that my data sets can be delivered to the Amiga in 'binned' form, i.e. as 2-d arrays each with [say] a byte value representing intensity. There may also be significant amounts of short-wavelength information present (sometimes called "noise" :-). Size of the array will be variable, but I'd like to use the full screen size for display (so I'll need to interpolate). 2) Which commercially available tools are likely to help me out? It might be nice to be able to transfer between different graphics modes, to see which looks best. Also, some kind of smoothing function might be nice because of the short-wavelength part. Good control over the palette is a must. 3) How much will I have to write myself? (I could possibly handle it but it sounds like a tremendous time sink since my programming experience with the Amiga is pretty limited.) I'd really like to get something going with this, as it could be a tremendous help to spicing up my output. Everyone loves color, right? If it turns out well, it's likely to be shown at some physics/numerical conferences as well, and I'll be sure to give credit to the Amiga et al. Maybe I'll even write an article for Byte :-). Please post or send email. I will summarize to the net. Maurice LeBrun | "A computer is like a hole in the ground Institute for Fusion Studies | into which you pour money" University of Texas at Austin | Internet: | - some poor, mistaken person, mjl@fusion.ph.utexas.edu | obviously wrong :-)
mjl@ut-emx.UUCP (Maurice LeBrun) (02/25/89)
In article <10767@ut-emx.UUCP>, mjl@ut-emx.UUCP (Maurice LeBrun) writes: > --- > I want to use my Amiga in my research (finally), and am asking for advice on > how to address my graphics needs. [much deleted] After I made my post, I realized I should mention some of the similar research efforts (involving the Amiga) that I know of, to help put my problem in perspective. Jim Crotinger at MIT has put together some nice animations as part of his dissertation work illustrating his simulation results. These were either 3-d perspective plots or contour plots of a function of two variables. Dr. Richard Matzner at UT plots a function of three variables (coordinate space) by plotting three different isosurfaces in three different colors (an isosurface is the 3-d analogue of a contour line, i.e. everywhere the function has a given value, and you "see" the surface via a large number of contour lines covering it [sorry, that's the best I can do at the moment -- everyone's seen them though, in Star Trek or somewhere]). In both cases, graphics software existed on the remote machine to do the actual plotting (both used VLT), and the plot was captured and made into part of an animation. The problems with applying this same approach to my problem are: (1) color-fill type plots in high level graphics packages are still not widespread, and (2) the number of colors supported is often lousy even if color fill routines are present (at NMFECC, the limit is 8 colors, or 7 if you don't count the background). For color fill plots, 7 colors is just not good enough for me. Also I should put in a (3) that color fills are not currently supported in VLT, but I think this is the least concern of the three. So, it looks like I have to download a raw data file and just do it locally. Fortunately, it's *much* simpler to do this than the very complex tasks of building a decent contour plot, or (especially) a 3-d perspective plot with hidden line removal. As I mentioned previously, it's a real simple problem with 8 bit planes, since you can basically just download a raster file and need to have a nice way of selecting the palette. For the Amiga it's a bit harder because I have to decide between graphics modes, and make good use of the available screen area. In any case, your opinions, encouragement, etc. are welcomed. Maurice LeBrun | "A computer is like a hole in the ground Institute for Fusion Studies | into which you pour money" University of Texas at Austin | Internet: | - some poor, mistaken person, mjl@fusion.ph.utexas.edu | obviously wrong :-)
jesup@cbmvax.UUCP (Randell Jesup) (02/27/89)
In article <10767@ut-emx.UUCP> mjl@ut-emx.UUCP (Maurice LeBrun) writes: >I want to use my Amiga in my research (finally), and am asking for advice on >how to address my graphics needs. I develop/run large scale n-body >simulations (magnetically confined plasmas, actually), and want to do color >fill-type plots on the Amiga to assist in interpreting the results. As most >people who've looked at them know, color fill plots are very nice for >visualizing functions of two variables. Unfortunately, I can't make any suggestions since I don't know what you mean by "color-fill" type plots. I suspect others are in the same situation. -- Randell Jesup, Commodore Engineering {uunet|rutgers|allegra}!cbmvax!jesup