From mconway@microsoft.com Mon May 25 15:37:13 1998 Received: from burdell.cc.gatech.edu (root@burdell.cc.gatech.edu [130.207.3.207]) by lennon.cc.gatech.edu (8.8.4/8.6.9) with ESMTP id PAA00834 for ; Mon, 25 May 1998 15:37:11 -0400 (EDT) Received: from wheaten.hitl.washington.edu (iSr5IIkgusjPpnlLBJ0CZ+s2fJz0sVAv@[128.95.73.60]) by burdell.cc.gatech.edu (8.8.4/8.6.9) with ESMTP id PAA09323 for ; Mon, 25 May 1998 15:37:10 -0400 (EDT) Received: from mail4.microsoft.com (mail4.microsoft.com [131.107.3.29]) by wheaten.hitl.washington.edu (8.8.8/8.6.12) with ESMTP id MAA20472 for <3d-ui@hitl.washington.edu>; Mon, 25 May 1998 12:37:00 -0700 (PDT) Received: by INET-04-IMC with Internet Mail Service (5.5.1960.3) id ; Mon, 25 May 1998 12:38:13 -0700 Message-ID: <4FD6422BE942D111908D00805F3158DF05B2673D@red-msg-52.dns.microsoft.com> From: Matt Conway To: "'cdshaw@cs.URegina.ca'" , 3d-ui@hitl.washington.edu Subject: RE: Osmose Date: Mon, 25 May 1998 12:38:04 -0700 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.1960.3) Content-Type: text/plain; charset="ISO-8859-1" Status: RO > -----Original Message----- > From: Chris Shaw [mailto:cdshaw@umbilicus.artsci.washington.edu] > Sent: Saturday, May 23, 1998 12:22 PM > To: 3d-ui@hitl.washington.edu > Subject: Re: Osmose > > > Matt, > > > It's the visual/interactive language that we're lacking in VR. > > no doubt there are implementation challenges, but those go away with > > time > > First, I'd like to point out that the quest for a visual language for > VR has been around since 1990 or before. In the Banff/U of Alberta > Art & Virtual Environments Project, it was mentioned in the funding > request. Yep. Randy's lab at UVA was riding that bandwagon 'round about that time too. > Film is the medium that people usually point to as a useful source of > techniques, especially for scene transitions. However, as we all know, > film does not work the same as VR. The fundamental difference is the > 30-100ms time allotment. Another fundamental difference is > interactivity. Film and VR are different of course. It is still a valuable place to look for inspiration. I'm not talking about lifting techniques verbatim, but in lifting them in spirit, adapting them to the new space. Of course it would be wrong to think that the vocabulary of film is a complete inventory of visual expression, but it would also be wrong to think that there's nothing there to learn. Film's visual lexicon is incomplete in two ways, first, film is constantly reinventing itself (okay, Film as Art outide Hollywood is reinventing itself, but that's a different rant...). To convince yourself of this, go see Greenaway's surreal telling of the Tempest, "Prospero's Books". The second and more important reason that film isn't an exhaustive source for visual techniques is, of course, that film isn't interactive. (Is this what you mean by the "time allotment", Chris?) To be clear, interactivity isn't about the speed of the rendering, it is about the fact that the *Direction* of the rendering (direction in the theatrical sense) isn't known until runtime. Pacing, framing, blocking, lighting, etc...all these things become severely distorted or even break down with VR. So here's my message: while the past isn't complete, it would also be wrong to think that you can't learn from the past. We cannot ignore the wisdom of people who have worried about these issues, even if those issues lived in a more pre-processed environment. Take, for exmaple, a success story: the virtual cinematographer work presented at SIGGRAPH 96. I think this was brilliant work if only because it encoded the idea that you may need to move, and even *delete* characters and obstacles from the scene in order to frame a shot correctly for a given user. Guess what? Stage directors would look at this and yawn, "we do this all the time" yet we as technologists publish these things as "discoveries." There's stuff to learn out there. Note that the authors of this work did their film research and were clearly inspired by the past. ==== === ==== > Potentially, anything is possible in a VR world. Pardon me, gotta get on a soap box here: You're right, but beware: while anything is possible, not everything is good. In my experience, most of the things we can do in VR are confusing, nauseating, or if you're very, very lucky, merely boring. We need to know the RIGHT things to do in VR; film (and other media) can help inform our decisions as to what is and isn't right. so perhaps we can agree on an obvious summary of all this blahblahblah. Learn from the past, but Just becuase the old media does it doesn't mean it's right. Exploit the magic of the new medium, but.. Just because you can do it doesn't mean you should. For the record, I think this is the core tension in all UI design. So, if that frames the question, what's the answer? Here's a shot: Sometimes the old solutions work just fine, but always ask yourself: "am I limiting myself too much?" Sometimes new solutions are better, but always ask yourself, "is the increased power worth the extra cognitive cost I'm putting on the user to learn this new element? Can I amortize this cost somehow?" This also ties in to our previous discussion of "when to use 2D elements in a 3D world." Both issues are asking the same thing: when is it appropriate for an new medium to borrow elements from an old medium? > One way to sidestep the interactivity challenge is to keep the user > busy with threats, and another is to put the user in a car. > On the other hand, people want to be told stories, and the narrative > agenda makes some tricks necessary, or at least allowable. > > Perhaps the desire for narrative will create a dominant style > by virtue > of it being what the users want. > > On a semi-technical point, my experience with artists working in VR is > that they are extremely hungry for polygons and texels. In my experience, the best artists understand the constraints of the medium that they're working in, whether they work in watercolor, oil, polygons, or pixels. It is certainly an difficult skill to acquire but not an impossible one. Perhaps the > current top machines can deliver the desired performance for some > projects, but I doubt it. My point is that steady-state scene > rendering > hogs all the resources, and transitions cause a doubling in resource > use. Another law of VR: VR always runs at 10 frames per second. As soon as a bottleneck is removed so as to raise the frame rate above 10 fps (a faster machine, better graphics card, etc.) authors will raise the amount of content (more characters, more polygons, more textures, more animations....) so as to lower the frame rate to the breaking point again. Transitions do not necessarily cause a doubling in screen resources. Some transistions do, and those have to be managed carefully. Matt