[REQ] How to reduce ghosting with dual projectors

[Tril]

EDITED BY SHARKY: I make a request out of it.. maybe somebody has another idea.. continues with post 6!
-----------------------------------------------------------------------------------

EDITED:
Does not work, see post 5. You can't make mistakes if you don't try. At least I tried something.

I'm building a passive projection system with two dlp projectors.

I have a tip to improve ghosting reduction at the cost of some reduction in brigtness. Use two filters for each projector.

Let's say you need to place the filter for the left eye at 45 degrees. Use two filters and place them in front of each other at angles of 23 degrees and 67 degrees. Play with the angle between the two filters until you are satisfied with the brightness and the ghosting reduction.

I got filters from Edmund optics (NT45-668). Even when crossed, they don't filter all visible light. To experiment, I crossed two filters and placed them in front of a projector projecting on the wall. I still see some light on the wall. That means that even with a perfect polarization preserving screen, it won't be perfectly without ghosting. I plan on testing other filters later.

I got a tip from experience to give to people experimenting with filters in front of a projector. Don't try to use cheap theme park styles glasses as a filter in front of a projector. It will only take about 30 seconds to 2 minutes before you burn the filter in the glasses, rendering them unusable.

[sharky]

i created a new thread.. its shurely worth!


its interesting what you found.. how much light do you loose? could you explain the thing with the angle a bit better please?

thank you

sharky

[Tril]

In summary :
Place the first filter at the angle to block light + chosen angle.
Place the second filter in front of the first filter at the angle to block light - chosen angle.

Experiment with the chosen angle. As it grows bigger, the brightness will decrease and the ghosting will also decrease.

It's possible that the reduced ghosting is partially caused by the decrease in brightness making the ghosting less noticeable but I don't think that's the only reason.

The way I see it is that a linear filter let pass light at its polarization angle but also at other angles. You can see that by putting two linear filters on top of each other. If they were totally perfect, they would only let light pass when they are in the same direction and as soon as you changed the angle between them, they would turn black. That's not what happens with real linear filters. They only turn completely black when crossed. By using two filters with angle between them, you derease the amount of light polarized at the angle of your glasses (you don't want this as it reduces brightness) and you also decrease the amount of light polarized at other angles (you want this). Since doing that blocks more light not at the angle of your glasses than light at the angle of the glasses, it reduces ghosting.

I'll try to make an ascii drawing. The 1 represent the area under a curve of the amount of light polarized at the angle of the glasses as a function of the angle.

X : angle of the glasses
Y: amount of light polarised

Perfect filter

000000000010000000000
000000000010000000000
000000000010000000000
000000000010000000000
000000000010000000000
000000000010000000000
000000000010000000000


One filter

0000000000001000000000000
0000000001111111000000000
0000000111111111110000000
0000001111111111111000000
0000111111111111111110000
0111111111111111111111110
1111111111111111111111111


Two filter with an angle between them

0000000000000000000000000
0000000000001000000000000
0000000000111110000000000
0000000011111111100000000
0000001111111111110000000
0000111111111111111100000
0001111111111111111110000

With two filters like that, you are closer to the perfect filter than with one filter. As you increase the angle between the two filters, you reduce the width of the area above but you also decrease the height.

I haven't done any measurements with instruments, I only used my eyes. It's cheap to try out as if you already have a passive projection system because you already have two filters and you can test this out with one projector.

[sharky]

good explanation... have you tested circular filters?

[Tril]

Forget everything I said. It does not work. I must have been too tired when I first did my tests.

What I was doing is place the first filter at an angle that's not the angle of polarisation of the glasses and putting the second filter (the one through which the light passes after the first filter) at the angle of the glasses. The combination of the first and second filter was decreasing the brightness and the second filter was polarizing the light for the glasses. The reduced brightness decreased the apparence of ghosting because the picture was darker but there's still as much ghosting.

I'm currently testing 5 samples I got from 4 different companies and one diy sample I made. They all mostly preserve polarization about equally and I still see a lot of ghosting so I want to try other filters. The main difference between the samples is how much they sparkle. The only sample that does not sparkle at all is made from acrylic glass and costs a lot.

[sharky]

what happen if you put 2 filters in the SAME angle one over the other and sear two pairs of glasses (it sounds crazy but just to try..

[Tril]

I tried to put two filters on top of each other. I also tried wiith three.
I tried to wear two pairs of glasses.

All these show no noticeable difference in ghosting apart from reduced brightness.

[Likay]

Hi there!

I have a polarized projector rig with two lcd-projectors and in my experience the ghosting rejection is superb! The thing in my rig that causes most ghosting is the silverscreen itself (mine is silverfabric). It's not a bad one though (look in another thread with posts from Jahun who did tests with different types of silverscreens regarding polarized rigs).
Since i use lcd-projectors i believe that the ghosting is minimized by the fact that an lcd-beamer is not entirely blacked out though. That light seems to drown the ghosting quite well. If i can see any ghosting there has to be very big dark/light contrasts.

A tip: You can try adjusting the display driver by turning up the gamma factor and at the same time reduce the contrast. This should minimize ghosting using dlp's. (DLP have way more capability to shut down the light compared to LCD ones where some light always slips through). I used this technique when using shutterglasses for ghost minimizing but also for being able to see better in dark games. A kinda two flies in one stroke...

When using circular polarizers you should theoretically be able to turn your head more compared to using linear ones. Used circular technique and tried to turn my head and the difference (if any) compared with linear is very small. The experienced ghostrejection circular/linear also gives setpoint to linear technique. Ghostrejection using linear filters is better with linear filters that with circular ones. So in my experience there are no reasons to use circular ones. Circular technique is also a little more expensive, not a big deal though.

Good luck!

edit: Try this link to Jahuns post: http://www.mtbs3d.com/phpBB/viewtopic.php?t=509

[Jahun]

There is just 1 way to decrease ghosting much in passive polarisation..
It's by improving the screen. Just that since polarizers are commonly alot better anyway.

But, there is a limit on how good screens get, you can't increase the polarisation too much, it will get to be a mirror. So it is a trade off. Most companies settled with +- 50:1 to 100:1 polarisation ratios. Still alot worse than polarizers but there is not much to do about it.


For normal viewing, it is just fine though.