Is that true? 15% seems very low to me. Source?
These 2 Artickes may help. The 2nd one on a skim does discuss needing to increase Brightness quite a bit.
“One disadvantage of the Pancake optics is low efficiency which is typically around 10%. Figure 2 shows the assembly schematic and pictures of the working all-plastic Pancake optical lenses.”
oh, I thought pancakes were just a compact lens design, I didn’t know that the light went through a circular polarizer and two half mirror, so 15% is optimistic
Well rbh I wasn’t that familiar either til you encouraged me to explore.
Definitely very interesting worth while read.
Thank you very much for this article,very very good
Pancake is the future Of vr
I thought the rule of thumb was 10%, so 15% is being generous ha.
This is why all professional SLR lens are all glasses.
Photographers complain about the heavy lens a lot, but none of them will use plastic.
The brightness loss could possibly be mitigated by brute forcing higher nits by using different more novel kinds of backlighting.
If an LCD used a laser phosphor backlight, it would be able to use less power and still get HDR brightness.
If it wasn’t just using the laser as a light source akin to an LED or today’s laser-based projectors
but instead used a MEMS scanning laser against a phosphor, combined with the really good local dimming algorithms available, you could probably get very high nits specifically targeted where you need it and only there with less power and heat output for similar performance to an HDR fald backlight.
I think a MEMS system would allow an LCD to finally accurately mimic an emissive display closer than even a full array local dimming back light because of the way it works.
A MEMS laser system is a system that uses mechanical deflection of laser light via a microelectric mirror. So it would allow a laser to essentially act like the electron gun on an old emissive CRT.
You up the scan rate and get tighter control of the beam. Add the really Advanced local dimming algorithms that have been developed all these years for LCDs, and I don’t believe you could get an LCD any closer to an emissive display for performance.
Good news is it’s very low power also, and because it would be used in a backlight, as opposed to say projection, what mems has been used for up to now, like the show WX or Nebra Anybeam the efficiency you would gain in the backlight would overcome the inefficiency of the pancake optic’s poor light transmittance.
What happens if such a HMD’s pancake lens broke or lose? Will all those MEMS laser beams directly hit the poor user’s eyes with 7 times the usual intensity?
You would use the MEMS system as part of a phosphor based backlight system, so the user’s eyes would have no direct contact with ithe lasers?
If the mechanical mirror stops functioning and there is an intense concentration of lasers to the point of burning the phosphor layer, you would have a fail safe in place
to immediately cut power.
Laser phosphor back lights are going to have to be where LCDs generally and probably hmds also will have to go. The EU just put out new power requirements for TVs, and the only way to accommodate HDR within that power envelope is with lasers and a phosphor. Mini LED, micro LED, simply put out too much heat and use too much power to get the kind of brightness.
The MetaHDR Prototype that Norm from tested got to try seemed like it was wall mounted, probably to account for the added electricity needed to feed all those LEDs or whatever bulb it was using.
When you get LEDs even when they’re really tiny packed in to such a small space, heat dissipation and Power requirements start to go way up. That HDR prototype had a metal mesh cage as it’s housing that essentially acted like a giant heatsink.
They also had to stack two LCDs (showing the same image), in order to stop enough of that strong light, that the blacks were somewhat dark… :7
That’s what a MEMS laser system would hopefully help solve outright. You could probably forgo the need for a second light modulating cell layer because it wouldn’t behave like a transmissive light source.
A scanning laser interacting with a phosphor layer would essentially behave just like an emissive display. Finally a backlight nearly capable of per pixel illumination and simpler than stacking two LCDs.
LED is inherently inefficient because it’s transmissive. You can have thousands of LEDs packed in and best case you’re only getting about 3% of the light through the polarizers, and that’s 3% of the light that manages to actually get through. Now if you add polarizers within the lenses as well as in the LCD, you see the problem.
The other 97% of the light either doesn’t get through or is wasted as heat. That prototype Norm tried was basically a 3D printed heatsink with some lenses and a display.
Micro LED is pretty awesome, because it does away with polarizers, but it’s still not there and it has the same issues with heat and Power especially when it comes to something like a head mounted display.
Mems systems are pretty small. The projectors they used to sell fit in your pocket, and when you take them apart, the guts amount to the thickness of about four credit cards stacked on top of each other.
Pancake lenses are glass though, and there are pancake lenses that professionals use on cameras.
But obviously we are not talking about pancake lens used by cameras, but the ones used by VR headsets.
The pancake lens for VR developed by Kopin need to bounce light within the lens themselves, resulting in low light efficiency and more obvious ghosting problems.
Maybe there is a pancake lens system for VR which uses glasses and have high efficiency? If you know, please share with us.
Just checked my pico 4 and they do seem to be plastic actually. I guess those are the all plastic ones announced by Kopin last year, afaik other pancake lenses so far are glass (Arpara, quest pro, vive flow). I could be wrong though, I was surprised that Pico 4 was plastic actually so I was wrong about that! Good to know, thanks.
That said, none have high efficiency it seems.
This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.