Re: Re: Visualization site updates
- To: mathgroup at smc.vnet.net
- Subject: [mg61012] Re: [mg60979] Re: Visualization site updates
- From: Chris Chiasson <chris.chiasson at gmail.com>
- Date: Fri, 7 Oct 2005 03:37:59 -0400 (EDT)
- References: <acbec1a40510060535o62e1b2e7jcf8ace36a2e089d1@mail.gmail.com> <20051006150815.15147.qmail@web33609.mail.mud.yahoo.com>
- Reply-to: Chris Chiasson <chris.chiasson at gmail.com>
- Sender: owner-wri-mathgroup at wolfram.com
:-] PS: I think OpenGL has 2D components as well. On 10/6/05, Martin Kraus <martin_kraus_germany at yahoo.com> wrote: > Chris, > > you get it as soon as you've implemented it. ;-) > > Actually, I realized that the Java implementation of > many web browsers (and therefore LiveGraphics3D) > can display a lot of special (unicode) characters. > On the other hand, it is not straightforward to > create something like subscripts or superscripts > with LiveGraphics3D because of the restrictions of > the first argument of the Text primitive within > LiveGraphics3D to either an identifier, a string, > a number or a StyleForm with one of these.) > > But how sophisticated should the type setting for > interactive 3D rendering be? Do you really need > full MathML? Here are some thought: > > 1) you don't want these renderings for static > publications > (because you don't want a bitmap when you can > have the PostScript output of Mathematica). > Since you only get a bitmap of the picture anyways, > the quality of the fonts is not extremely important. > I would even argue that the fonts don't have to > be exactly the same as long as attributes such as > boldface, italics, sanserif etc. are preserved. > Thus, IMHO there really is hardly any need for > interactive 3d rendering with embedded fonts. > > 2) usually you don't need equations in a > visualization, > because they are likely to either clutter your image > or to be unreadable small. Rather you need symbols, > which are usually constructed out of sequences, > subscripts, superscripts, underscripts and overscripts > of unicode characters, i.e., a rather limited subset > of > mathematical type setting. Thus, IMHO there is > hardly any need for full MathML rendering in > interactive 3d rendering. > > 3) a small subset in terms of Mathematica commands and options > for the kind of mathematical expressions which would > be > most useful could look like this: > > expr ::= string | identifier | StyleForm[expr, opts, > ...] | > SequenceForm[expr, ...] | HoldForm[expr] | > Subscript[expr, expr] | Superscript[expr, expr] | > Subsuperscript[expr, expr, expr] | > Underscript[expr, expr] | Overscript[expr, expr] | > Underoverscript[expr, expr, expr] > > opts ::= FontFamily->... | FontSubstitutions->... | > FontSize->... | FontWeight->... | FontSlant->... | > FontColor->... | Background->... | > ScriptSizeMultipliers->... | ScriptMinSize->... | > ScriptBaselineShifts->... > > Obviously, I have already thought about this. And > actually > I've started implementing this for LiveGraphics3D. > Thus, I would appreciate any input what else would be > of > general interest for mathematical annotations in 3d > graphics. > > Cheers > > Martin > > --- Chris Chiasson <chris.chiasson at gmail.com> wrote: > > > What I want to know is: > > (When) Are we going to get OpenGL rendering and/or > > SVG with embedded > > fonts and embedded MathML export? > > :] > > > > On 10/6/05, Martin Kraus > > <martin_kraus_germany at yahoo.com> wrote: > > > Jeff, > > > > > > I definitely agree that the resolution of most of > > the excellent > > > visualizations on your site is too fine for > > LiveGraphics3D, > > > i.e., there are too many primitives for > > interactive rendering > > > with LiveGraphics3D. Rendering at a lower > > resolution is > > > therefore often unavoidable, and I'm usually much > > more > > > aggressive in this respect when adapting > > > Mathematica graphics for LiveGraphics3D in order > > to ensure > > > interactive rendering also on less powerful > > machines. > > > Without the possibility to render graphics > > interactively, it is usually > > > preferable to offer a static picture for a well > > chosen view point > > > and/or a precomputed animation for a rotating > > object (as you > > > did for the earthquake example). > > > > > > The 3d surface offers an example for which the > > interactive version > > > doesn't offer a lot of benefits: there are too > > many polygons for > > > interactive rendering with LiveGraphics3D and the > > geometry of the > > > surface (a height field) appears to be simple > > enough to be understood > > > with a 2d image (apart from some hidden parts). > > > > > > I think the ability to interactively rotate and > > zoom objects is > > > particularly useful to examine the earthquake and > > molecule > > > examples because the 3d point positions cannot be > > understood > > > from a 2d image. In these cases the stereo > > viewing (press "s" to > > > toggle the applet between single view, stereo view > > for > > > divergent fusing and stereo view for cross fusing) > > is also useful. > > > > > > In order to separate points of the same color (in > > particular for > > > molecules) you could try the applet parameter > > > <PARAM NAME="POINT_EDGE_COLOR" VALUE=#000000> > > > to render each point with a thin black edge. > > > > > > Without further coarsening, any of the animations > > would not render > > > interactively with LiveGraphics3D. However, using > > LiveGraphics3D > > > to view animations offers the benefit of more > > control over choosing > > > the time step (ALT/META/COMMAND + left button or > > right button > > > and dragging horizontally). > > > > > > BTW a great site for visualization with > > Mathematica! > > > > > > > > > Martin Kraus > > > > > > > > > Jeff Bryant ha escrito: > > > > > > > I've begun adding some new functionality to my > > visualization site. A > > > > few people have requested the ability to be able > > to interact with the 3D > > > > graphics in real-time. I decided to try using > > LiveGraphics3D by Martin > > > > Krauss on several of my examples. Some of the > > examples are not > > > > appropriate for real-time interaction as the > > render time was too long. > > > > I picked a few of my 3D visualizations that were > > more reasonable. For > > > > those examples that are interactive, they are > > often rendered at lower > > > > resolutions to keep loading time down. I'd > > would be interested in > > > > hearing your thoughts on whether this is useful > > where appropriate. Here > > > > are the examples I currently have: > > > > > > > > > > > http://members.wri.com/jeffb/visualization/3dsurf-interact.shtml > > > > > > > http://members.wri.com/jeffb/visualization/potentials-interact.shtml > > > > > > > http://members.wri.com/jeffb/visualization/galaxies-interact.shtml > > > > > > > http://members.wri.com/jeffb/visualization/quake3d-interact.shtml > > > > > > > http://members.wri.com/jeffb/visualization/algebraic-interact.shtml > > > > > > > http://members.wri.com/jeffb/visualization/prehistoric-interact.shtml > > > > > > > > I have also added a new visualization > > (non-interactive) that uses > > > > Mathematica to visualize molecular structures in > > the protein data bank > > > > format. This particular animation shows an > > insulin molecule with 14,940 > > > > atoms in it. To keep the render time down, I > > used 3D points instead of > > > > spheres which unfortunately means no lighting > > effects: > > > > > > > http://members.wri.com/jeffb/visualization/molecule.shtml > > > > > > > > -Jeff > > > > > > > > > > > > -- > > Chris Chiasson > > http://chrischiasson.com/contact/chris_chiasson > > > > > > > __________________________________ > Yahoo! Mail - PC Magazine Editors' Choice 2005 > http://mail.yahoo.com > -- Chris Chiasson http://chrischiasson.com/contact/chris_chiasson