From: mbernat@gel.ulaval.ca (Marc Bernatchez) Subject: [ON LINE] HMD-Resolution (third version) Date: 17 Jun 1995 01:29:55 GMT Organization: dep. de Genie electrique Last update: 16/06/95 Hi all, I would like to submit to all of you some calculations I made to compare HMD resolutions regardless of their types and optics. Note that this is a third version of the text previously posted the this group. Due to feedback I got in the previous posting, I thought it would be useful to repost this new version. My reference is a person looking at a 14" PC screen at a distance of about 2 feets (60 cm). The FOV is 25 degrees. Let's assume the display is at 1024x780 pixels. This gives 1.46 arc minutes per pixel horizontally. Some people stand closer to their computer screen so if we take a distance of 1.5 feets (45 cm), the FOV is 33.4 degrees and we have an angular resolution of 1.96 arc min. per pixel for the same resolution on the screen. These two values will give us a good idea of the kind of resolution we will perceive with some HMDs. Further more, the human visual perception system has an acuity of about 1 arc minute per pixel. It can get to 1/2 or 1/3 of arc minute per pixel in the fovea region due to a higher cones density there. I must thank Mr. Joseph Psotka at the U. S. Army Research Institute for is kind help regarding these considerations. You can also look in some good VR books for informations regarding the human physiological relations with virtual environments. I have found useful informations in " Virtual Reality ExCursions with programs in C ", ISBN 0-12-737865-0. Now let's look at the CyberMaxx. The LCDs are about 267x225 (in triads of colors). The FOV is 56 degrees horizontal. So we get 12.58 arc minutes per pixel. The VFX1. The LCD resolution is 278x204 (in triads of colors). The FOV is 46 degrees. With this one, the angular resolution is 10.36 arc minutes per pixel. Finally, the i-glasses have LCD elements showing 263 x 230 color triads per eye. The FOV is 30 degrees. Which gives us an angular resolution of 6.84 arc min. per pixel. Just to get an idea of measures, the CAE electronics fiber optics HMD is about 2.5 arc minutes per pixel in the fovea region as stated in the May 95 issue of VR NEWS. It's about 5 arc minutes per pixel in the peripheral region. So I interpret it this way. The VFX1 will give better sharpness than the CyberMaxx BUT it will give you more of a "tunnel" sensation and less immersion effect. In fact, it would be more appropriate to speak of tunnel vision for the i-glasses which should be considered more as an augmented reality HMD than an immersion HMD. The VFX1 and the CyberMaxx are much alike as long as the FOV is concerned. The CyberMaxx, even if it shows larger pixels than the VFX1, will give a reasonable feeling of immersion. A FOV of 60 degrees is considered to be the minimum to get immersion in a VR environment. What is great about this is that it's a measure of the resolution that is independent of the viewing device, LCD or CRT based. However, don't jump too fast on conclusions. You must also evaluate the FOV which is very important to create the feeling of immersion. I've read some articles from people and they seemed to say that a FOV increase gives more gain than a resolution increase. I have put up a Web page on this topic that shows all these considerations with nice graphics and charts. It is located at: http://www.gel.ulaval.ca:80/~mbernat/analyhmd/analysis.html I would love to hear your comments about all that. Bye. Note: the resolution spec. where taken from the May 95 issue of VR NEWS (vol.4 Issue 4). The calculations are still mine but perfectly correspond to the ones found in the article. Reassuring I guess ;-) Bye all. ========================================================================== == Marc Bernatchez = E-MAIL == == C.O.P.L.(GEL plt-00307) = mbernat@gel.ulaval.ca == == Dep. Genie electrique = or == == Universite LAVAL = marc.bernatchez@gel.ulaval.ca == == Quebec, Canada =============================================== == G1K 7P4 = Virtual Reality is the future of computers == == = http://www.gel.ulaval.ca/~mbernat == ==========================================================================