The Ultimate 3D projection system: A Practical Discussion Thread - Page 5

Good you found it because I cannot post links thanks to all these dear spammers :-)
As usual wait and see....

OK, an update to this thread. We talked before about the feasibility to use lenses from a pair of Dolby 3D glasses to filter the light from two projectors and in this way create a 3D setup at home. There was a discussion whether the glasses themselves only let through specific wavelengths or if they only blocked specific wavelengths. Do you need to buy a special projector filter or can you use the glasses as filters?

Darin has lend me the lenses from his glasses and I have measured them at work. The instrument used was a Varian Cary 50 scientific spectrophotometer scanning between 800 and 280 nm. I measured absorbance which is defined as the tenth logarithm of [the incident light intensity divided by the remaining light intensity]. An absorbance of one means you let 10% of the light through at the specific wavelength, two - 1% and so on. The scale on the instrument goes to 10 but it's not very accurate over 1, i.e you wont get a linear response anymore. I usually dilute my samples to get the absorbance below 0.5 to be sure to have a linear response. However that is a bit difficult with glasses...

As you can see from the graph the gaps in the spectrum's are overlapping a bit at an absorbance of 1.2 and higher (which corresponds to about 6% transmittance). This could potentially cause some crosstalk. However, if a narrow wavelength light source is used this could be minimized or eliminated by staying away from the edges of the gaps where they overlap. Parts of the violet spectrum is also visible for both glasses and the left seem to be letting through some light at the blue gap for the right lens.

I had a look at different UHP projector spectrums and it seems they differ somewhat which makes the choice of projector important. For instance the VW85 it really is low on light between 500 and 510 nm which is smack in the middle of the green gap for the right eye. The Epson TW5500 looks a bit higher even though it still is a bit low. The Xenon powered VW200 looks significantly better for green even if this also would become neutered, but is instead very low at 600 nm - which corresponds to red for the right eye. Therefore it might be difficult to just put these glasses in front of a normal projector without loosing too much light. This since you would need to tame all the other five colors to match the lowest one. Perhaps if one was able to change the internal filtering to allow for a wider spectrum (vivid mode?), otherwise some colors would be very dim and consequently the calibrated light output really, really low. I don't know if there are other possibilities? Replace the internal filter with a customized one that allow wider wavelengths? I know cine4home has played with different filters.

On the flip side, at their specific wavelength intervals the glasses let through almost all of the light. If an LED or laser projector could be designed with the 6 correct native wavelengths the light efficiency would be very high - about 90% for each eye - or 45% of the light in 2D mode.

Is this the future? An LED with six LEDs instead of three wouldn't seem that much harder to make than todays machines. Cheaper, thinner, simpler glasses, no batteries needed, any screen works, high light efficiency. More expensive projector design though, but I gather the LEDs will go down quickly in price....

Thanks Drexler. Very interesting results. I'm going to have to think about what that first graph means with respect to the measurements I did and posted about earlier:


For instance, even with some overlap on some wavelengths and using sunlight off a white piece of paper the 2 filters on top of each other resulted in very little light. I was assuming the white piece of paper would reflect basically the whole spectrum, but I'm not really positive that is the case.

--Darin

Now, I don't remember which was green and red and I have already packed them again... However, it does't really matter for the discussion that much. Unless you have the spectrum for the 9500UB? In which case it would be interesting to see how the gaps in the spectrum fits with the projector output.

The light you get to your meter is the combined light from all three wavelengths. So even if you get 8.8 and 33.9 respectively out of the original 51.8, doesn't mean you have equal amounts of all colors. It could easily be that you from one of the glasses get quite a lot of for instance green and blue, but almost no red. Then you'd have to strangle the green and blue to get the correct color balance again - loosing a lot of light in the process.

I looked through the glasses on top of each other, at the sky and at my computer screen, and could easily see through them with a violet tint. This kinda makes sense from the graph as there's leakage between 350 to about 410 nm for both glasses - corresponding to the violet part of the spectrum. I'm guessing the leakage from the overlap at the gaps will roughly become white light as there's a slight overlap for both red, green and blue, making the violet leakage dominant in color.

#
Art Feierman Says:
November 8th, 2010 at 10:38 pm

Yep, that’s it. As I mentioned, the XP8000 that arrived I inadvertently assumed that was the 3D adapter, but it was the external battery for the Optoma Pk301. I’ve been so swamped with the move (the whole house is under remodel, and I’ve just gotten internet up (first time since Wed). I’ve been meaning to correct that mention. I’ve spoken with Optoma, they will ship me one, when they have them. They did not, as of last week. -a

So all this effort because nobody sells a box with a set of shutter glasses that will can take a 3D HDMI signal (say 1080p/60fps e.g. from a 3D blu-ray player), split it into 2 HDMI outputs L and R you can run into 2 of the same model projector focused at a typical 2D screen (output at 120fps)? Forgive me if I'm way off base here, but wouldn't this set-up work fine to allow you to see a bright 3D image without all the polarizing this, that and the other thing?

That wouldn't work. Shutter glasses are useless with Dual projectors. Framerates do not add up, two projectors at 60fps don't make 120fps.
The entire point of dual projectors is to get rid of the shutter glasses, use passive ones and have the left and right eye views shown simultaneously for a better picture, free of any motion artefacts caused by shutter glasses.

Brilliant news..looking forward to your updates!.i`m thinking of going the same way..when`s the second RS40 coming?

Hey, that's good news !

What type of screen are you using ? Did you get the BlackDiamond or did you opt for more traditional silverscreens ?

You read my mind: passive 3D without a special screen.

I've also wondered with LED or wide-spectrum light source if a single-chassis projector could work if the spectrum filtering could be done in the digital domain? Perhaps that's asking too much from electronics? But my mind wonders if similar to the way that DLP driven RealD 3D is actually sequential left/right projection switching polarization instantly on the fly, I wonder if something similar could be done for Dolby's color spectrum? A mechanically moving filter or "color wheel" might work... just imagine if the usual color wheel was replaced by one designed with Dolby's 3D light spectrum in mind... or if an LED light source without a color wheel, perhaps 3D could throw a color wheel back in for 3D that would rotate in front of the DMD chip to altnernate left/right with the shift in color spectrum?

Hmmm, now that I think about it. - It should be possible to use a 6 segment color wheel that produces 2 slightly different versions of red, green and blue in a normal 1-chip DLP. In this case they wouldn't need to change much of the design, just a more expensive color wheel and new processing.

The downside would be that the light output in 2D would be very low (only 2X that of 3D) since you filter out a lot of the light - so in this case maybe only 10-20% of what you would get in 2D with a normal color wheel. (Assuming Dolby 3D looses about 90-95% of the light, which is what I've heard).

Where are you getting these from?

Has it been confirmed the Optoma's split the output?

I was thinking about trying the mini radiances but that would get a wee bit expensive.

If the color wheel actually was the Dolby/Infitec filters, separated into the 6 bandpasses, for 2D you take off your glasses and combine the images for the left and right into one for both and almost the whole spectrum will come through. At least enough for any decent CMS to fix (you'll need one for the 3D anyway!

At the ~ OD 1.6 crossovers in your earlier spectrum, you will have ~ 3% projected and of that 3% allowed into the wrong eye, sound like quite a low crosstalk, but LEDs and lasers could eliminate even that overlap.

Not really. At any point in time only one part of each color will get through - remember the image is built on the colors being shown sequentially. This means that the accumulated numbers of photons on the screen per unit time will be lower since some are absorbed in the filter. I.e. the light output will go down.

To make an easy example. Say red is divided in two parts with slightly different wavelengths. Each constitute 50% of red. When the colorfilter is on red1 - red2 will be filtered out and vice versa. So in any point in time only 50% of the red light will get through. And that's the theoretical maximum. So with this technique it wouldn't be possible to achieve more than 50% of what's possible with a normal color wheel. And I'm guessing this figure will be even lower in reality.

It's actually exactly the same things as a 3-chip DLP vs a 1-chip. A 1-chip only shows 1 of 3 colors at a time whereas the 3-chipper shows all 3. Therefore it will only produce 1/3 the light and thus is less bright. Now, with the infitec filters you have 6 colors instead of 3 - so only 1/6 of the light from the original 3-chipper will get through.

You're right - I was thinking of the passive system for non-wheel systems, when the same image is fed to both bandpass sets and you take the glasses off - that's when the 2D gets brighter - both eyes getting both sets instead of only one in the 3D mode. I'll go back to sleep now and wait for LED/laser systems...

The light losses aren't that terrible with Dolby 3D.
The less than 10% remaining light relative to 2D is a value I remember reading on an old Barco spreadsheet, but I've read other professional sources (Harkness-screens) estimating Dolby 3D around 15% remaining light relative to 2D, a value comparable with shutter glasses and RealD Z-screen based polarised presentation. And this is calculated using a traditional single DLP projector already equipped with standard colour filters, to which the Dolby/Infitec filter disc is added.

If you were to replace the original colour filters with infitec filters, you would definitely get higher light outputs.
A purpose built dual projector based Dolby/Infitec system should yield just as much light as a polarised system.

That sounds a lot better! The figure I've heard before I think comes from Mark Haflich, 93% if my memory serves me correct. But I'll bet it depends on the setup, which projector that is used and so forth.

15% X2 gives 30% in 2D which isn't that bad if you start off with a lot (1500 lumens?) and have a higher gain screen like the HP. In addition the brightness difference between 2D and 3D wouldn't be night and day anymore - as it is today.

An infitec system would definitely be my first choice. It's the only one I've seen without screwed up colors. Both shutter and polarized glasses have added a tint (greenish or purplish) to whites that is quite objectionable in my opinion. Not so in the Dolby 3D theater I've visited. In addition it's passive with cheaper, simpler glasses (vs shutter) without batteries and sync issues and most importantly - any screen will work.

I didn't notice the screwed up colour with polarised displays. And with shutter glasses, when it happens it's always due to a hardware fault : badly designed glasses that do have a tint where they should not have any, or on my system, the internal polarisation of the projector that interferes with the filter.
Nothing that can't be corrected with a purpose built system though.

Dolby/Infitec however does also screw up colour by design, and is a much worse than the other solutions out there will achieve accidentally.
The difference is that Infitec does provide an optional colour corrector that balances the colour back to normal, but it's expensive. Most people building Infitec systems just use the filters from the glasses which they put in front of their projectors.

I think the tint is inherent in the shutter glasses. My guess is that the tiny lines from the polarization filter causes interference effects increasing/decreasing the amplitude of certain wavelengths, which gives rise to the observed tint.

It might be curable with a color calibration through the glasses though.

I find the whole thread interesting. It appears many in the thread are AV specialists trying to set up 3D systems. I neednt have to remind professionals of the patents protecting this form of 3D projection.

What are you talking about?

It doesn't split the outputs per sae....

You need two on them.
Feed them the same HDMI 3D content
there is a 3 position switch for video output on the back so one unit can be set L and the other be set R.