From: mbernat@gel.ulaval.ca (Marc Bernatchez) Subject: Re: TECH: HMD quality Date: Fri, 06 Oct 95 21:03:23 GMT Organization: dep. de Genie electrique From: mbernat@gel.ulaval.ca (Marc Bernatchez) Hi all, To continue the discussion regarding some of the content of my article HMD resolution analysis , here are some results from tests I made. For those who fear numbers, stop right here. After Mr. Robin Hollands showed me his concern about some possible flaws in the way I computed some values regarding the global quality factor, I decided to do some tests on numbers to see what would happen. In particular, I wanted to see if indeed the numbers obtained were staying in the permitted range. I found that, indeed, for some unrealistic values, the Fqt result could be negatives. To solve that I applied a step function (all negatives values are made zero) to the final Fqt column. Note that values in the partial global quality factor column such as Fprice, Ffov and Fhar can be negative. This will prevent that an hypothetical HMD with a FOV of 180 degrees and a LCD raw pixel count of 1 pixel could rank around 25% ! What follows is a table of all the cases I tried with the equations I use. note: FOV = field of view (in degrees) raw LCD pixel = number of horizontal color pixels on the LCD har = horizontal angular resolution (in arc minutes per pixel) Fqt = global quality factor (in %) FOV | raw LCD pixel | har | $ | Fqt | error case ------------------------------------------------------------------------------ Limits test cases: 0 1 0 1 64% X 1 1 60 1 0% 180 1 10800 1 0% 180 1000 10.8 1 82% 180 1E+10 near 0 1 near 100% Possible cases: 180 9000 1.2 10,000 75% 180 1024 10.54 10,000 60% 100 1024 5.9 3,000 55% 100 1024 5.9 800 58% 60 1024 3.5 800 54% 60 640 5.6 800 51% 30 250 7.2 800 42% As you can see, the Fqt does give the right indication as global quality of an HMD goes. To get the maximum from these information, you must consider the Fqt rank in the context where it will always be valid (for several years to come at least). So a 75% rank doesn't mean that this HMD is at 75% of the maximum quality currently achievable technologically speaking. It rather means that this HMD match the human maximal specifications for FOV and har and, all this, considering the price. I admit price is a bit of a problem. Currently, all the three considered points counts for 33.3% of the overall evaluation. Robin also showed is concern about the fact that har and FOV are not independent variables, and thus compromising the validity of the calculations. I admit it's true but it is the only way to put a percentage level for pixel density and FOV. We can't predict what will be the raw pixel count in five or ten years from here, so it's impossible to chose what would correspond to 100% as raw pixels count goes. To summarize, the difference between Fqt percents doesn't really matter. What do matter is that one HMD as a higher Fqt rank than an other one, regardless of the amount. It's up to the reader to chose what is the most important factor for him (i.e. price, FOV or har) then consider the Fqt of those HMDs who are in the desired range. I will certainly continue to add more factor in the calculation in the future but I still think that, despite it is hard to evaluate human factors, such an evaluation is justified. I also think that it gives a good but incomplete evaluation of HMDs strength. As long as I'm concerned, the rank in the evaluation on my web page do represent the subjective rank I had attributed to most of these HMDs prior to the day I made the global factor calculations. I would like to know if other people feel the same way or have different views. So as I always say, please feel free to send me your comments (negative or positive) to the E-mail bellow. Regards, == 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 == == fax: (418) 656-3159 = http://www.gel.ulaval.ca/~mbernat ==