Darb Dabney’s OpenSim Blog

I took a bit of a rest after the ESRI International User Conference (actually, it was more like catching up with real work). Sadly, I missed out on the call for images by Adam, and I can’t even blame it on there being so many time zones between here and Perth.;^)

So in the interest of sharing stuff in bulk, please accept the following pile of shots. All of them were made from the 40-region OpenBerkurodam (OB40) model that has been taking shape over the past few months. All of them are from the 1.024:1 scale model of the UC Berkeley campus and adjacent downtown environs that have been built (precisely 2.5 square km. worth).

Unlike the attractively detailed SketchUp models one finds for selected UC Berkeley buildings in Google Earth, the OB40 sim has every building, every moderately large tree, a lot of light poles, and even a construction crane imaged in 3D. This was because it was built wholesale with reflective LiDAR data. Lots of data, and very little artistic craft!

The resulting mismatch between the LiDAR bump-map surfaces and the 10-cm natural color orthoimagery that I have draped over them create an effect that is quick, dirty, and very complete. At first glance, one might think that we can’t decide where we are–on the continuum between representational and surrealistic art, or that perhaps the trees have not lichen, but a rather different kind of fungus affecting them. Hey, I’m just saying…

The OB40 model was demonstrated live on two and three higher-end laptops running the standard Second Life client; they had NVIDIA Quadro graphics cards and they did OK. The sculpties imaged a little bit differently than they do with fully approved graphics cards but the client never crashed outright.

Some senior ESRI system folks got a look and a see of what OpenSim was about with GIS data loaded into it. Several public safety people expressed some interest in the possibilities. The presentation was not at a booth, but rather in a corner of the showroom floor given to the “User Applications Fair” that was a spot for about 32 non-commercial folks to show their stuff. Strictly speaking, the ESRI software was not the application on display, but without the ESRI (and ERDAS) software, I wouldn’t have been able to get my GIS data loaded into OpenSim in time for the conference.

What sort of shocked me in terms of response was a huge non-linearity in acceptance based on the age of the person viewing the demonstration. At one point, I was describing some obscure details with an experienced GIS person, and within 15 seconds, a group of three teen-aged 4H Club members (I’d seen them in another part of the conference) sat themselves down without questions or introductions and began going all over the place 3X. They had no questions about the SL client interface, the purpose of the OB40 sim, or any of that. They just sat down and started exploring.

For me, the experience of seeing the 4H kids using OB40 intuitively provided great hope that some day not too far off, people will just accept a Multi-User Virtual Environment (MUVE) as readily as I would read a map from the American Automobile Association (AAA). I mean, for me there’s some effort involved in using the SL client, although at this point it is about as familiar to my hands as the ‘vi’ editor is—I just use it, kind of like reading a book without mouthing the words. But for the younger people who interacted with OpenSim, the interface did not seem hardly present for them, they focused at once on the content and enjoyed it for just the fun.

OK, enough blather - I’ll try and share all the shots, including some that did not make it to San Diego. The actual date for all of the shots was 20080731.

This is downtown Berkeley, the BART station, same area that has been modeled at 1:3 scale in Second Life Agni grid, Gualala region. In Gualala, everything has been built in detail by hand, with custom real-world texture shots. In OB40, the scale is nominally 1:1, but at the moment only a LiDAR drape fills the region (and 39 adjacent regions, too.) There is an avatar above the Power Bar building, the tower on the left.

This is synthetic “MS Virtual Earth” or Pictometry high-angle view of the Martin Luther King Jr. Civic Center building, Berkeley’s city hall. There is an avatar on the near-left side of the roof, enjoying a brown-bag lunch.

This is Shattuck Ave and Hearst St, view SWly. Oscar’s hamburger grill is on the right with all the ducting on the roof.

View Wly across Shattuck Ave toward the Berkeley Arts Magnet school campus.

View toward Oxford St, near sunrise. Strawberry fields in foreground.

Farms in Berkeley? Indeed, this strawberry field was imaged on 2006 07 01 just a couple of blocks from the UC campus. View SEly near sunrise.

View uphill on Hearst St towards Euclid, northerly side of UC Berkeley campus. TECHNICAL DETAIL: in the far distance toward sunrise, there are huge eggs floating above the ground, but textured with the orthoimagery. These are the LiDAR megaprims after they have received their photo texture, but before they have rezzed with their bump map. Depending on bandwidth, how much of the model the client may have already visited and cached, and the phase of the (virtual) moon, it might take anywhere from 15 seconds to a minute or two for the bump maps to fully rez out when one arrives near a region. When shooting these pictures, and typically in OB40, I keep the SL client viewing out 512 meters with “ultra” quality graphcs.

Above the top of Hearst, the Lawrence Berkeley National Laboratory (LBNL) sychotron and nearby buildings, view Ely, including some really large Blue Gum (eucalyptus) trees.

Below LBNL, the Foothill student residences, with Sather Tower in view, far right

The UC Berkeley Greek Theater, site of a great many fine performances over many decades, view Ely, and just Sly of the Foothill residences.

View NEly, of UC Berkeley’s International House, with California Memorial Stadium in background. Avatar is hovering over the cupola of the I-House, sneaking a free look at the football scrimmage (or is it cheerleader camp?) 2006 07 01

View Nly up Piedmont Ave, in the Greek housing section of campus. View toward International House with California Memorial Stadium in background. Horizontal scale 1.024:1, vertical scale 1:1; those trees have scaled height and bulk thanks to LiDAR first return gridding.

UC Berkeley main campus near Bancroft and Piedmont. Large red-roofed building in mid left frame is Boalt Hall, lighter building in right mid-far range is Wurster hall, home of the Urban Planning folks. Here’s looking at you, kids!

Finally, UC Berkeley campus toward LBNL, long shot near Sather Tower. All these shots were from the OB40 sim, sometimes running on osim.bargc.org

Wow it has been a lot of effort to get the 40 regions built out with their LiDAR surface sculpties. Someday when the process gets automated I’ll look back and likely feel like a fool for not loading the BLOBs in MySQL directly, and coding the XML to load these directly. But I built them by hand, using 64-prim linked sets that covered each region. Region-wide linking works fine when the Admin settings are used to “go to God”. Also there’s a new SL client 1.20.15.92456 that works OK and helped a wee bit with the build.

My priority is to create good graphics for large-format display, but I’ve posted a rush of the first end-to end plod by the Ruth named “UC08 Visitor2″. At the moment, I’m still a bit shaky about just how solid the sim will be for demonstration purposes. None of he LiDAR surface sculpties are physical; I’ve been able to turn on ODE. I can’t bear the time to let the sculpties get meshed so there’s none of the cool walking on them, but I have tried boosting the detail settings to 64, in hopes that perhaps the sim will put out 4096 points per sculptie [my bump-maps to define each one are 16K so there’s plenty of info behind the detail boost.

As usual, if YouTube is blocked or you don’t have flash in the browser that you read this with, the video is here: http://www.youtube.com/v/srwRPgjFjnQ

OK, not everyone in the immersive modeling universe has been holding their breath on this one, but hey, I’m happy to say that the production line is fired up and creating 2560 sculptie bumpmaps to inflate the Open Berkurodam sim using LiDAR reflective digital elevation model (DEM). The registration with orthoimagery is not perfect, and small offsets are very distracting, but the first two have been placed, and should illustrate the concept. (Two rezzed, 2558 to go…)

The reflective DEM sculpties have 16 times greater resolution (that’s resolution as NURB point density) than the underlying terrain megaprims. This means that for the 40-region Open Berkurodam sim, there are 160 terrain megaprims, and 160 tiles of 10-cm orthoimagery. The reflective DEM sculpties number 2560 and will be textured using the same orthoimagery.

Reflective DEM surfaces ride over the tops of trees, rooftops, or any structure. They are defined by the first return of the LiDAR reflected signal. By contrast, the terrain megaprims are based on a model of the last return (in these data up to the seventh return signal) that represents the ground under and around all structures and trees.

Things are busy and the presentation deadline approaches for the server. Just to show that indeed, stuff has been going on, take a look at how the BARGC server puts out terrain megaprims. What’s seemingly more performant with dual Xeons than the Core2 Duo server is seeing three regions rezzing terrain at once, and getting three regions’ worth of terrain megaprim sculpties hatching all at once.

I’m focusing on the reflective DEM surface with smaller 33-meter megaprims, which will have 16 times the density of sculpty surface. Meanwhile, this evening’s early video is here at http://www.youtube.com/v/1zMdVAeXeI4

The embed is below (in case you find it blocked)