[weblacky]

Yo,
After losing out on another SGI machine today I suddenly begin thinking about the 3-D features of the SGIs I have. All my machines should have stereo shutter glasses interfaces. I went on eBay and I was very lucky to find a crystaleyes ESGI emitter model that was missing its DB9 cable, not a problem this is all documented, for only $20 with free ship.  I bought it on the spot. There seems to be lots of CE2/3 glasses around so I’m going to come back to those.

Now I remember saying this along with everyone else, you have to have a CRT to do this. But I begin thinking is this still true? I was under the impression that it had to do with the refresh rate. Though I don’t exactly know how the monitor will be commanded to a new rate over VGA.  

This isn’t the biggest deal in the world to me but I thought it was a good time to just ask. Has anyone experimented with the new much faster LCD monitor gaming technology and older shutter glasses? I’ve seen a few remarks online and some people are saying when they tried it either it didn’t work at all or the timing was slightly off. Again I could believe all that, but I thought it prudent to ask the question to see if anyone had tried it or was in the position to know anything new.

Apparently what I’m reading claims that even though new LCDs can transition from one color to the destination color faster than 8 ms that the transition of that is witnessed on the glasses versus on a CRT where the element would go dark or immediately appear as the other color with no transition. So they seem to say that a whole new technique has been used to strobe the backlit in order to hide the transition of the LCD pixel. I’m curious if this is a problem with OLED which isn’t discussed. But perhaps there are no OLED monitors that transition at that speed. I haven’t bothered to look.

[robespierre]

The CrystalEyes glasses are always open on either the right or left side. They switch directly from left to right without any dead time. So no time at all (0 ms) is available for refreshing a screen from one view to the other. This works with a CRT because the phosphor persistence is short. By the time the beam finishes drawing the bottom line, the top half of the screen is black. By the end of the vertical retrace, the whole screen is black. If the glasses are timed to switch at that moment, the eye on the new side will not see any light from the previous eye's view (there is a tiny amount of afterglow from the phosphors, about 1%).

The other important fact about all VESA stereo glasses is that they are synchronized with the video output of the computer system. A CRT has zero latency from when the image data leaves the RAMDAC in the graphics device to when it appears to the eye. By putting the glasses under control of the computer, the switchover time is coincident with the vertical retrace interrupt and the flipping of the active display buffer. LCD displays work differently because they are digital and receive data in packets (over DisplayPort or HDMI). Image data is queued up and processed in chunks, leading to latency from the computer's graphics output. LCD monitors with 3D ability generate and transmit the synchronizing signal from the monitor so it remains locked to what the eye sees.

All of the above is perhaps solvable with enough thinking and work. The stereo timing signal could be adjusted or delayed based on the LCD panel latency. The left and right frames could be interspersed with black video frames to ensure that each eye starts out viewing a black screen. Some fancy-dancy FPGA box could digitize the signal and insert these changes (like the GBS converter does for CGA signals). But a problem is lurking within these anodyne descriptions. LCD displays work by polarizing light. There is a polaroid film behind and in front of the liquid crystal panel. The liquid crystals can be twisted or untwisted to control the angle of polarization, to allow the light to pass or be blocked. Then you see the light that had the proper polarization to pass the front polaroid film. LCD shutter glasses work by polarizing light... See the problem? If the light from the screen is polarized against the front polaroid film of the glasses, they will be forever dark, whether the shutters are on or off. Luckily most LCD screens seem to be vertically polarized, while CrystalEyes are at 45 degrees, so some light does get through. It looks dim and washed out, not nice at all. For a similar reason, 3D movie projection works with DLP projectors, but not with standard LCD projectors.


Here is a fascinating history of stereographic imaging from one of its participants.

Here is a teardown and repair page for CrystalEyes units.
Here are some comparisons of different brands of shutter glasses.

[Raion]

Hey dude, glad to see you made it here. Did you just duck out after Nekochan's death or...?

[robespierre]

Hey dude, glad to see you made it here. Did you just duck out after Nekochan's death or...?

Yes, I found it pretty depressing. I have some projects I'm working on that might be pretty cool, though.
(How about an Indigo2 with 16TB of storage?)

[Raion]

Well I would hope that at the very least eventually irixnet can live up to it. I think it's imperative to keep ourselves around. Not to distract from the topic... just saw you were back and I remembered ya!

[Donjon]

For a similar reason, 3D movie projection works with DLP projectors, but not with standard LCD projectors.

It works fine with LCD! There have been plenty of home systems (both projectors and televisions) that used an LCD display and LCD shutter glasses - all you need to do is to have the same polarisation in the display device and the glasses. This was used in home theatre applications for both 3D projectors as well as 3D televisions.

Passive glasses (which just have vertical polarising filter for one eye and horizontal for the other) are also possible by building a panel which has odd lines in one polarisation and even in the other, albeit at the cost of losing half the resolution. There are also several home systems that use this.

However, the 3D effect adds little to the cinematic experience (plus the panels are always slightly grey, so the image on a projector can be quite dim), so whilst paying a few dollars extra for the gimmick might be entertaining in the cinema it wasn't a huge commercial success in the home and it's essentially dead today.

All of the above is perhaps solvable with enough thinking and work. The stereo timing signal could be adjusted or delayed based on the LCD panel latency. The left and right frames could be interspersed with black video frames to ensure that each eye starts out viewing a black screen. Some fancy-dancy FPGA box could digitize the signal and insert these changes (like the GBS converter does for CGA signals). But a problem is lurking within these anodyne descriptions. LCD displays work by polarizing light. There is a polaroid film behind and in front of the liquid crystal panel.

If you're getting into fancy-dancy FPGA stuff, you might as well digitise the signal and convert it into one of the recognised 3D formats - the top-bottom frame packing 3D format required of 3D displays in HDMI1.4 is quite similar to what an SGI is outputting for  quadbuffered stereo, although only a 720p (50 or 60 Hz) and 1080p (24Hz) are required by the HDMI1.4 standard.

However, just delaying a single, fairly slow digital signal is something that should be quite comfortably handled by a cheap microcontroller, especially as the precision required isn't huge. The same microcontroller should also be able to handle transmission of a variety of IR glasses protocols, meaning scouring eBay for reasonably priced CrystalEyes glasses and transmitters wouldn't be needed.

[adobbins]

One of the posts on this thread links to a page (https://saf.bio.caltech.edu/crystaleyes) that purports to be about Crystal Eyes repair. However the link seems to be defunct.

I have a few pairs of Crystal Eyes that aren't really usable anymore -- applying pressure at points on the plastic frames or on the wings
may bring temporary relief but very flaky. 

Does anyone have experience with restoring the functionality of the glasses?

Help would be very much appreciated.

ACD

[robespierre]

That's really too bad the page is unavailable (and blocked on the Wayback Machine). It had useful information and pictures showing the disassembly process and a few ideas for repairs.

If the problem is cured by pinching the frame, then it seems obvious that there is a connection issue with some of the cables; I believe there is an elaborate flat printed circuit between the two panels, the phototransistor, and the batteries at the sides. Connections between copper and glass are always touchy and there could be possible cures using copper or graphite paint, similar to the products for repairing rubber membrane switches and "zebra strip" LCD connectors in devices like multimeters. But the loss of the web page is really unfortunate because I don't think it will be easy or obvious how to repair these.

addendum:
Here's a thought. Maybe you could contact the webmaster of bio.caltech.edu and ask if they can send you the page and its linked images? It's worth a shot.

[lohmos]

I did test it with a 144 Hz LCD, no problem here.
I am building a cave with modern DLP projectors but not finished the software part yet.

Another thing is the availbity of libs:

Cavelib was available on Irix those days but i never found a copy of it, would be the best choice.

There are opensoure projects as vrjuggler and freevr but i did not test yet if one can get those compiled under irix.

[weblacky]

I did test it with a 144 Hz LCD, no problem here.
I am building a cave with modern DLP projectors but not finished the software part yet.

Another thing is the availbity of libs:

Cavelib was available on Irix those days but i never found a copy of it, would be the best choice.

There are opensoure projects as vrjuggler and freevr but i did not test yet if one can get those compiled under irix.

Can you please describe the 144Hz LCD (projector I assume) you're using in the setup?  Are you doing polarized left/right eye with two projectors or did you get the 3D output format recognized by a single projector?