3D Models

Greg has been in love with 3D modeling and virtual worlds since the Philips days, but it's taken him more than two decades to put together a web page gathering them into one place. Better late than never, though, eh?

The original ones were in VRML, and once upon a time, browser plugins to view such things were common. Alas, that is no longer the case. However, there is a silver lining: virtually all modern browsers have WebGL accelerated 3D support built in, and the three.js library makes invoking that support fairly easy...though, of course, there's a learning curve to convert from the VRML mindset to that of three.js.

Greg's favorite is a model of the local universe, which slowly revolves and zooms in and out (use the controls in the corner to stop that if you wish). Unlike its VRML predecessor, it has no functioning viewpoints (can't seem to get them to work right), which means the view-from-the-origin, find-me-a-constellation feature is missing:

This is a less-complete but more-accurate variant, (mostly) limited to the galaxies within 11 Mpc or so of the Milky Way, all of which have "reasonably accurate" distance measurements (as well as labels), though the phrase "reasonably accurate" doesn't mean the same thing to astronomers as it does to normal people. The blue ones are the more distant galaxies, not (yet?) recognized as being part of any particular galaxy cluster. The pink and orange ones in the center-left and center are members of the Local Group (Andromeda satellites and Milky Way satellites, respectively), which extends only to about 1.4 Mpc; the bright green ones at lower left are members of the Centaurus A cluster; the yellow ones are members of the M83 cluster; and other clusters (either official or visually identified) may or may not have still other colors. As with the model above, this one rotates and slowly zooms in and out, which can be relaxing to watch, but feel free to use the controls at upper right to stop both movements and then play with the model using a mouse or finger:

Below is the same model at roughly the same viewing angle but about fifty times further out. Particularly striking is the void to the upper left of both images, which literally contains no more than about half a dozen galaxies.

This one is yet another variation on the theme: just the Milky Way's (dwarf) satellite galaxies and globular clusters, along with the Milky Way itself and a "background" sphere representing the constellations we would see if the viewpoint were on the Earth itself (represented by a smallish blue sphere in the disc of the galaxy at very center of the model). Of course, the constellations are actually composed of stars also embedded in the disc of the galaxy, all of which are very close to Earth and likely would fit within the (outsized) blue sphere if everything were to scale. (In fact, only the relative positions are to scale; the size of the spheres was chosen purely for visualization.)

This is actually the first (and so far only) model using reasonably modern Javascript and three.js. Among other things, the bug preventing constellation labels from being visible in the previous two models is fixed in this one. On the other hand, the new version of the control panel at upper right is not fully compatible with the old one, so the "H to hide" function is broken. Sigh.

At the other end of the complexity spectrum, this simple one was by request of mathematician John Baez:

Quite a few of the models below are very simple conversions of GPS latitude/longitude/elevation data to "extrusions" (VRML terminology), a.k.a. TubeGeometry (three.js nomenclature). Greg hasn't yet entirely automated the conversion, but he's getting close. This pair is based on GPS traces in the Bay Area. The first was a hike along the northwest coast of San Francisco (10x vertical exaggeration), and the second is a set of vertically offset bike rides through Mountain View, Sunnyvale, and Alviso (virtually flat terrain):

    

These are a bunch of Tahoe-area ski tracks; they're all to the same scale (the vertical is exaggerated 2x in each), and like the SF hike model above, time always progresses from green to purple:

(The white segment in the Alpine Meadows/Squaw Valley one is a portion of the bus trip home; there was an unfortunate incident with the GPS's "start/stop" button near the end of the day. :-/ ) Alas, snow in the Tahoe region has been uninspiring for quite a few years now, with the exception of 2017...which, sadly, was the year Greg got in exactly 2.5 hours of skiing. :-( Many of the resorts (Kirkwood, Sierra Ski Ranch, even Heavenly and Squaw) are much bigger than they appear here, but there simply hasn't been enough snow coverage, and in some cases high winds have closed large areas of the mountain. And back when there was plenty of snow, Greg didn't have a tracking GPS. Le sigh...

more to come ...