From: wilson@cs.utexas.edu (Paul Wilson) Subject: Re: TECH: LCD yields, configurations, cost Date: 29 Nov 1995 23:37:47 -0600 Message-ID: <49jfvb$1aj@jive.cs.utexas.edu> Organization: CS Dept, University of Texas at Austin From: wilson@cs.utexas.edu (Paul Wilson) In article <199511290058.LAA56832@mozart.ph.unimelb.EDU.AU>, >From: John Costella >>Hmmm... I'd guess that the problem is that the pixels are on the wrong >>side of the optics, and give the effect of being projected out onto >>the virtual world. If they were on the near side of the optics, they'd >>be defocused and much less objectionable---like specks on specs, which >>millions of people live with every day, even if they don't like them. I'm not sure I expressed myself clearly enough on this, and I'm sure I didn't think it through very clearly, but I'm not sure it's a bad idea. There are some human factors issues as well as optics issues. Certainly, a speck right in front of the eye will either be defocused enough that it won't entriely mask the stuck pixel, or it will have to be so big that it's somewhat obtrusive. I goofed grossly on that. (Dogh!) On the other hand, a speck at about 10" apparent distance does a much better job. (Judging from a quick test using a pinhead held in front of the period at the end of the previous sentence, while sitting back from the monitor.) So suppose instead of an eyeglasses metaphor we use a windscreen metaphor, or more correctly, a big eyeglasses metaphor, where the windscreen tracks your head motion. We put the masking speck in the "middle" of the optics, so that its apparent distance is further than glasses but nearer than what you're usually looking at in your virtual world. How objectionable is this? I'm unclear on why a dead (or even stuck) pixel is really objectionable in the first place, so I really have no idea. My intuition at the moment is that it has something to do with the defect being projected onto the world, rather than in a separate plane nearer the viewer. After all, I can drive a car with the occasional dot of birdshit on the windscreen without being driven nuts, so if I move my LCD defect onto a virtual windscreen, I shouldn't go much crazier. (I don't have much experience with windshields that head-track, but I do have experience with motorcycle face shields---with various defects on the surface---which haven't bothered me excessively. I don't know if that's close enough to the relevant apparent distance for this purpose.) I have also been going on the assumption that having a one-eye defect is not much of a problem because everything is visible in one eye or the other, and as long as you're focused on the world, you can see everyting all the time. But maybe that's exactly wrong. (A previous poster said that single-pixel flaws were surprisingly objectionable and, well, I'm surprised.) Maybe a one-eyed defect gives you mixed cues that freak out your visual system. From one eye, you get a blip (and focus accomodation cue) that there's something there (e.g., 10" ahead, slightly left of center), but the other eye can't see it, and the brain can't figure out why. If *that's* what's objectionable, you could just put a spot in front of the other eye, too, in the right place to properly locate the spot in space. No problem. Adding defects is cheaper than avoiding them. Has anybody done any human factors studies relevant to this stuff? It just seems intuitively obvious to me that a flaw or two should not be so objectionable that people would forgo a major increase in resolution. (I hate low-res, and I don't mind a little dirt, though maybe that says more about me than about most people.) Hmmm... maybe tomorrow I'll put some test spots on my face shield and go out on my bike... I could even put a flat edge-on thingy on there that's white on one side and black on the other to see if the asymmetrical presentation would confuse my visual system... :-) I also wonder whether this is a resolution-dependent phenomenon. Does a small stuck pixel bother you exponentially less than a large one? (i.e., is it more attractive for higher-res situations?) > [...] > >Your best bet would be to simply mask the speck on the LCD itself. >Paint it, maybe. Liquid paper? It would stop it flashing, but it >would still be a fixed colour and intensity. Yeah. Now if only people would start selling the defective HMD's cheap. I'm sure we could figure out some acceptable fix. How about we put a reticle in the field of view, which just happens to overlap the dead pixels. Then we can call it a feature! (But seriously, on that point---surely people have put reticles and other heads-up info the field of view, e.g., for controlling fighter weapons systems. That doesn't bug people much, as long as it's not a lot of stuff, right? So a stuck pixel can't be too much of a problem if we dress it up appropriately.) -- | Paul R. Wilson, Comp. Sci. Dept., U of Texas @ Austin (wilson@cs.utexas.edu) | Papers on memory allocators, garbage collection, memory hierarchies, | persistence and Scheme interpreters and compilers available via ftp from | ftp.cs.utexas.edu, in pub/garbage (or http://www.cs.utexas.edu/users/wilson/)