http://www.reddit.com/r/Cyberpunk/comments/2r9pmj/activewirehead_an_excellent_article_from_our/

“Whoever the author of this article is, he obviously was not involved in early VR experimentation.
One glaring example – he writes:
At the high end of the spectrum there were VR pods using early LCD displays and run on Amiga 1000s.
The original Virtuality pods did not use LCD displays, nor an Amiga 1000; rather, they used a customized Amiga 3000, with a CD-ROM drive – which wasn’t a standard peripheral at the time, as well as a custom graphics card using what was – at the time – fairly state of the art 3D graphics hardware. The HMDs used narrow and small color CRT displays – not LCDs.
The later Virtuality pods used customized 486 hardware, combined again with the high-end 3D graphics hardware – and higher-resolution (640 x 480 per eye) displays in the HMD.
Honestly, though – the Virtuality systems, while great entertainment machines for the time, weren’t anywhere close to “high end” when it came to VR. That was all in the labs and research institutions. You could – for the right amount of money – easily get HMD hardware that, while it might not rival the weight – certainly rivaled and surpassed the Oculus hardware, as far as a display was concerned.
Indeed – the LEEP optics system is something that the Rift system only approximates, because those optics were custom fabricated and ground from a single piece of glass, and didn’t consist of multiple stacked lenses. It had the same effect (un-warping an image pre-warped by hardware or software for a larger FOV), but ultimately resulted in a greater quality display. However, that quality came at a very hefty price – a set of lenses alone clocked in somewhere around $25k – and they weren’t exactly lightweight, either.
On the homebrew end, though, I will say things were pretty lacking. For consumer level hardware, the best one could buy was a toss-up between the Forte VFX-1 and the VictorMaxx CyberMaxx (not the Stuntmaster – that was a toy, but a fun one to experiment with!). The CyberMaxx had a fairly large field of view compared to the VFX-1 (and was one of the largest at the time), but the VFX-1 had a rockin’ flip-up visor (great for VR world building and testing), awesome over-the-ear KOSS headphones, and great balance with little-to-none nose-weight.
Most homebrew HMDs concentrated on using small LCDs from handheld televisions; it wasn’t the greatest display, but it was fairly expensive. With the right lenses, you could build a fairly nice HMD – easily equivalent to the consumer options (if not packaged as well).
Tracking was done (usually) via mechanical means (attached potentiometer gimbal systems), electronic means (electronic compass and electrolytic tilt sensors), and sometimes via modding PowerGlove hardware (put the emitters on your head, arrange the L-bar overhead, and voila – an ultrasonic 6DOF head tracker is born).
Control was typically done with a mouse (I particularly liked this “ring mouse” trackball thing that you could grip like a gun, move the ball with your thumb, and click left/right buttons with your first two fingers), a joy-pad of some sort, or sometimes (if you had the money) one of the few 6DOF 3D controllers that existed back then (the “holy grail” was the 3Dconnexion SpaceBall controller – not cheap by any means). Logitech and others also made lower-cost 3D controllers.
Software on the homebrew end was generally home-grown, or more generally was based around Rend386, VR386, or the later Avril software. All of them had support for a wide range of input and output hardware, and worked on 386 or better machines.
Finally – one of the ways people shared info (besides usenet of course) was via the wonderful, if short-lived publication “PCVR” – I keep saying I’m going to scan my copies, but maybe somebody’s already done it and put it up on archive.org? Dunno.
Researchers, though – had some really great hardware and software; besides the HMDs, they had access to wonders like the VPL DataGlove, trackers from Polhemus and Ascension, and machines from Silicon Graphics (SGI) – hardware that, back in the early-mid 1990s, basically could give you the speed (if not the quality) of hardware that wouldn’t appear for the consumer for 10 years (heck, Carmack had to develop Quake on such hardware in advance of the PC – figuring rightly that by the time the game was completed, the PC hardware would have caught up). SGI eventually folded, and the some of the engineers went on to form a new company that we know today as NVidia.
I think it’s funny – as someone who played around on the low-end homebrew side of VR back in the 1990s, and watched the internet take over everyone’s attention, and VR to fade into the background noise – that so much of what was known by the DIY crowd of the time has been virtually lost. At the same time, that might be a good thing, because it has allowed a new crowd to flourish forth without many pre-conceived notions of the past to hold them back. In that way, these guys have been able to explore new ways of doing things and coming up with unique solutions to some problems that either weren’t possible back then, or were too expensive to even contemplate from a hobbyist perspective.”