A Pair Of CRTs Drive This Virtual Reality Headset

With the benefit of decades of advances in miniaturization, looking back at the devices of yore can be entertaining. Take camcorders; did we really walk around with these massive devices resting on our shoulders just to record the family trip to Disneyworld? We did, but even if those days are long gone, the hardware remains for the picking in closets and at thrift stores.

Those camcorders can be turned into cool things such as this CRT-based virtual reality headset. [Andy West] removed the viewfinders from a pair of defunct Panasonic camcorders from slightly after the “Reggievision” era, leaving their housings and optics as intact as possible. He reverse-engineered the connections and hooked up the composite video inputs to HDMI-to-composite converters, which connect to the dual HDMI ports on a Raspberry Pi 4. An LM303DLHC accelerometer provides head tracking, and everything is mounted to a bodged headset designed to use a phone for VR. The final build is surprisingly neat for the number of thick cables and large components used, and it bears a passing resemblance to one of those targeting helmets attack helicopter pilots use.

The software is an amalgam of whatever works – Three.js for browser-based 3D animation, some off-the-shelf drivers for the accelerometers, and Python and shell scripts to glue it all together. The video below shows the build and a demo; we don’t get the benefit of seeing what [Andy] is seeing in glorious monochrome SD, but he seems suitably impressed. As are we.

We’ve seen an uptick in projects using CRT viewfinders lately, including this tiny vector display. Time to scour those thrift stores before all the old camcorders are snapped up.

12 thoughts on “A Pair Of CRTs Drive This Virtual Reality Headset

  1. You were double lucky, a matching pair of devices and composite input. Over the years I looked into these viewfinders, all but one were loaded with wires and power being easy no composite input was found. The hum thing may not work because video is normally low Z. I used a camera. I have never seen a hum signal touching a video input on a VCR. Separate sync both H and V were seen on a scope and I just gave up. I have a few and might try the vector scope trick for a stereo music visualizer though.

    1. Well, try to get a game like Half Life (even the original one from 1998!) working on something like RPi4 and you will understand rather quickly why is the expensive hardware needed.

      Also squinting at a low res PAL/NTSC image (those viewfinders have usually only about 400 lines or so) blown up across your field of vision with a lens will maybe make you understand how stupid your comment is.

      But yeah, it is “greedy-ass Valve” and not actually customers who won’t put up with crappy resolution, laggy software and no tracking, right? Sheesh, dude.

      I applaud the author of that CRT hack, though. It is pretty much an exact reproduction of how the first HMDs worked – e.g. Ivan Sutherland’s one from 1968:

      1. >But yeah, it is “greedy-ass Valve” and not actually customers who won’t put up with crappy resolution, laggy software and no tracking, right? Sheesh, dude.

        I’m pretty sure that they’re complaining there’s no way to play it without a VR headset, rather than requiring an expensive headset (Alyx is officially compatible with more or less every PC headset on the market, including the cheaper WMR gear). Some people seem to be mortally offended that a game designed as a pioneering immersive VR story game doesn’t have a non-VR version.

        Anyway, the CRT hack is neat, but I wonder how heavy it is hanging off your face like that. In my head CRTs wiegh a ton but those are tiny CRTs.

    2. I played Half-Life 2, HL2 EP1 and HL2 EP2 at lowest settings using a GeForce 2 PCI card with 32 MB RAM. That video card had a PCI to PCIe bridge so I said correct: PCI. CPU was some AMD XP 1600+ at 1.1GHz, and the PC RAM was 324MB. It worked like crap but I could enjoy a game everyone was talking about. I had to build water cooling system using an old Soviet Union-era refrigerator from the 1960s. There were four ATX power supplies I got from city trash dump, I fixed them. Electric power was failing all the time at the electric sockets, so each of the other three PSUs were connected to each of the three electric phases inside the electric panel on my floor – I had big trouble with the campus management for that, but later they abandoned all the charges for recovering all their lost finance data. No money for UPS available. I could not even afford to buy a mouse – with those money I could survive for a whole week. I used an old track ball, also recovered from trash. Monitor was a RUBIN-102 CRT, made in the Soviet Union in the 1950s. Sound amplifier was an old Magnetophone, made in the 90s, also recovered from trash. Stupid, but it got the work done.

      Regarding my stupid comment, it is from the poor middle-2000s Eastern Europe student point of view – something very hard for you to understand, and impossible to comprehend. Back in the 2000s I had to work hard just to get the money to pay a taxi driver to help me get stuff from trash dump. And that bad experience helped me learn more and more about hardware hacking – to squeeze everything possible from every available piece of hardware. The experience gained during those times was extremely welcome. I saved many petroleum drilling contracts in Sahara desert by hacking impossible malfunctions – some times even building op-amps using components from machines buried in the sands, abandoned by the Wehrmacht more than 70 years ago – search google for RV12P2000 and RV12P4000. The helicopter was scheduled to arrive every two months with supplies, and petroleum drilling costs hundreds of $ per MINUTE. That’s why my employers started to pay me serious $ just to be there, do whatever I want while watching all the operations from rig-up to rig-down. Only at rig-move (once every two months) I was allowed one day off do to whatever I wish (most of the times – search for stinky ‘recycling’ yards). And I learned never to throw money on useless stuff, but to always keep my mind in a opened-state and find alternative ways to do things – and of course never to cross the limits of the legislation. Nobody wants to f*ck with Law Firms and prosecutors.

      About the game: somebody will find a way for sure to bypass the VR restrictions. Not because the lack of money, but to prove it is possible.

      This CRT head set is and amazing hack and deserves the highest respect.

    3. This guy could get this setup to register as a pc headset if he wanted to do a lot more work, and from there it would just be a matter of getting some kind of hand controls and a kinect for hand/head tracking. It just becomes a question of money vs time, considering that half life game is gonna work on sub-$200 windows headsets.

  2. This could be cool and maybe even usable if specific games were developed for it that aren’t reliant on hi res graphics. Like going very simple designing things with geometric shapes pieced together in 3d… I think the word “voxel” applies to what I’m describing. Like an early version of mine craft type graphics.

    I don’t know enough about development etc so it might not work… But I know modern vr games will not work even if you could get a raspberry pi to run (or relay) them on this thing…

    I just realized what I’m describing is probably very similar to the virtual boy (I think that’s what it was called), Nintendos early, headache inducing venture into vr (sorta).

  3. I wonder how long it will take for him to realize how bad an idea it is to have the CRTs oriented in opposite directions like that.

    Ambient magnetic field (even Earth’s field) shifts and rotates the images on the CRTs in opposite directions when they are pointing in opposite directions like that, so it’s impossible to keep the two images in alignment as you move the headset around. It causes eyestrain and headaches and generally drives you bonkers.

    Arranging the optics to have the two CRTubes parallel (e.g., pointing vertically) solves the issue.

    Folding the optical path so the tubes sit along the side of the head fixes it too, and also helps the weight distribution, but complicates the optics: the optical path becomes so long the field of view is restricted, or requires more complicated relay optics.

      1. It is trivial to rotate either the deflection yoke or the whole tube to change the aspect ratio, actually, and engineering the vertical-oriented assembly would be easier and more compact than the ungainly moose-antlers build described here.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.