The Ultimate 3D projection system: A Practical Discussion Thread - Page 6

OK gang! The first round of testing started tonight!

Testing, you say? Didn't you already jump into this thing headfirst, wallet flying, blindfolds superglued???

Kinda... well, yes and no...

I've definitely committed to going this route. Technically I now feel 100% confident I can pull it off. So there are items that I am committed to purchasing, such as the projectors and the 3DXL boxes.

However! Even with the KNOWNS being settled, there are still various unknowns. Unexpected surprises should be fully EXPECTED.

So wherever there is the potential for variables in the quality or performance of components that are available in multiple guises, from multiple vendors, then I feel COMPELLED to experiment with the combinations of those variables, to find the best solutions, before I shell out the remaining moolah.

Some things I cannot change at the moment, such as my Black Diamond II 1.4 screen, which is touted as being compatible with polarised 3D systems (more on THAT a few posts down from this one).

Other things are DEFINITELY easier to experiment with, such as the polarised filters for the projectors, and the glasses used.

I learned a LOT tonight - more than I expected. I also was driven to certain decisions about how the project needs to be completed.

There was REALLY GOOD NEWS, there was some PRETTY BAD NEWS, and there was NEWS THAT MADE THINGS TRICKY.

To Find out which was which ....? read on...!

First up - we have the Circular Polarising filters that were used in tonights show, err tests...

From Sunny California, err polarization.com, we have the AFlash Circular polarising filter "sample" kit. This was acquired for the stunningly low price of just $25.

In the picture below, you can see it is just the basic polarising film, not mounted on glass, nor sandwiched between two sheets of anti-reflective glass as some of the higher-end solutions are.

As a result, it is more likely to have issues with heat degrading the filters and shortening their life. In addition, as you may notice in the picture, they are subject to really annoying "curling" - as in not staying flat.

But we have not yet acquired the high end filters, so we will proceed with these good buddies. Did their low price affect their performance? Not in the least (apart from the curling).

Next we rounded up a set of RealD 3D glasses, just like those in the theaters... actually, they ARE the ones that are in the theaters. I found these on Amazon for $2.50 apiece, in 10-packs. Surprisingly, they turned out to have performance that bested the other early entrant.

That other early entrant was a pair of MicroVision "Joshua" circular polarised glasses, in "shiny black" - also obtained on Amazon, but at a healthier bite of $32.50 - more than ten times the price of the RealD.

Secretly, I was convinced that they would outperform the RealD cheapos and force me to spend big bucks equipping the home theater for the throngs of family, friends and odd, scary neighbors that would now flock to my home for 3D movie nights.

Alas, while certainly much sexier, they were pretty soundly beaten by the RealD specs...

OK - teasers out of the way, let's move on to the testing methodologies.

The testing methodology is very simple - start at the very beginning of the "food chain" and work your way up, trying to understand what is happening at each point, and how it may affect the next step, or the project as a whole.

Tonight was all about the Optics - not the Electronics

(1) STEP ONE - So first we have looked at the projectors, and gotten a feel for how their light is output and what, if anything will need to be looked at. In the case of this first project, I have tested thoroughly the HD550 and the RS40. In both cases, these projectors have all three colors polarised together in unison horizontally, or, at Zero Degrees. Also, there is no circular polarization present.

What does this mean? It means, for various technical reasons, that they are ideally suited for a Circular Polarising filter setup, or perhaps even a Infitec (Dolby3D) color filter system. It also means that they are NOT suitable for a Linear Polarised filter system.

In addition, it means that the light loss will be MUCH less than expected, which is GOOD! Why is this? Because all circular polarising filter have, as part of their basic design, a linear polarising filter as their first layer, to exclude light that is not properly oriented to be converted to cicular polarisation.

For a projector that does not have polarised light output, HALF the output of the projector may be filtered out immediately at the first layer of the circular filter. So for our JVC units here, we expected to lose a lot less light. And MAN, were we RIGHT on this one!!!

However, there IS A CATCH... which would factor significantly in tonight's testing round. You see, the JVC's are a bit of an oddball in having in having horizontal polarisation at zero degrees. It would seem that most 3D projectors in the theaters are set up with vertical polarisation, and the glasses are also configured to perform this way, as became evident in the test. (If your confused, just remember that the circular filters in the glasses ALSO have a layer of linear polariser).

(2) STEP TWO - Next, we play with the filters that are to be applied in front of the projector lens, to see how this works out, and if there is any immediate negative impact. This was an area I was especially concerned about up to tonight, as I have been hearing all kinds of horror stories about how much light I would lose. We proved them wrong, as you'll see further on.

(3) STEP THREE - Next, I wanted a "pure" test of the interaction between the circular filters on the projectors, and the glasses used. This means I HAD TO exclude any interaction from the screen, so that it didn't create any confusion in interpretation. Fortunately, this is much simpler than it sounds.

All we had to do was insert the glasses into the light path between the (filtered) projector and the screen. In theory, and light not polarised with the correct orientation would be immediately blocked in the appropriate glasses lens (left or right) and the resultant shadow would clearly illustrate the performance of the combination.

This means that we could use ANY screen for these tests, no silver-based screen is required at this stage.

(4) STEP FOUR - Finally, when we have run through all of combinations and determined the best ones, and recognized just what we should be able to expect as a best case scenario, THEN it is time to introduce our final bugbear, the SCREEN, into the equation. No where have I had a harder time determining the performance of any component of this system, than with the screen.

To test the screen, we simply took the glasses out of the light path, and put them on our faces. Ideally, at this point, the light reflecting off of the Black Diamond II would still be polarised when it hit our faces, and the glasses would do their thang! Sadly, in the case of the BDII, the results were pretty disastrous. This was the most serious set-back of the night, and sets the stage for a major upgrade that I had not counted on quite yet.

I'm not sure what "generation" of the BDII I have (whether 2nd or 3rd), but I would advise anyone looking to move up to a polarised passive 3D system NOT to make that jump into a BDII just yet, until we have a better understanding of the horror we witnessed tonight! This is covered in more detail below.

OK - so lets move on with some of the observations...

So just how much LIGHT LOSS did we experience, when we slapped our new AFLASH filter over the front of our beloved JVC light factory?

Not much at all - in fact WAAAY less than we were prepared for. This is a HUGELY significant finding, and bodes really well for the final light output of the finished product.

Getting the orientation of the new lense right was very simple - if it was oriented incorrectly, it blocked pretty much 100% of the light from the projector. This is because the JVC's light is already polarised, and the circular polarising filter uses a linear polariser as it's first layer.

On a normal projector, with non-polarised light, almost half the light would have been lost right here at this first step. But not so with our JVC.

In the picture below, the filter is covering the right half of the screen. Look REALLY closely and you can just barely make out the transistion in brightness down the center of the picture.

Can't see it? heh Try tilting your screen back and dropping your head to make your picture dimmer, or just adjust your display's brightness till you can see it. I promise, I HAVE posted the right picture.

The light loss here is negligible - I'm going to guess on the order of 5% or so. But I haven't measured it yet, so this is a subjective guess.

Frankly, I was astounded - I had expected much more light to be sacrificed.

OK - Next, we look at the two glasses candidates, combine them with the filtered output from the projector, and then pit them against each other like Ricky Martin fans...

If you liked the good news in THIS section, then you will positively LOVE the good news in this next section... except maybe for that one small catch we hinted at earlier..

OK - so now that our test projector is equipped with our fancy new polarised filter and is outputting gorgeous circularly polarised light beamz, it's time to throw up the glasses into said light path, and read the tea leaves, err shadows, cast upon the canvas which we calleth our screen!

Again, I was VERY surprised at how much light passed through the setup. Very PLEASANTLY surprised. You are going to mess yourself when you see how transparent these lenses are in the below pictures.

In fact, you may be even more excited when you notice how much light was occluded (blocked) by the appropriate filter in the glasses (particularly with the RealD pair).

After all, this is EXACTLY what we want! LOTS of light passing through one lens, with just about ALL the light blocked from the other lens.

However, after opening a couple of pictures, your going to become confused and come back to read the rest of what I am saying here After all, why the heck has rdjam posted pictures of the glasses in both the "normal" (horizontal) position - AS WELL AS this "wierd-as-hell" sidways (vertical) orientation????

The answer seems pretty self-evident to me - these glasses are manufactured for polarised projector systems that are based on a VERTICALLY polarised light source. As such, their performance, as illustrated below, is severely compromised when held or worn at the "proper" position on your nose. But when oriented sideways, as if for spidermonleys from Mars, their performance is superb.

Right off the bat, that means we are going to have to look a little harder for an off-the-shelf pair of glasses that is set up for a horizontally polarised system, or at worst, may have to custom build our own from sheets of polariser film. Neither of these is a deal-breaker in my view.

Now, moving on - you will also notice that the performance of the RealD glasses is significantly better than the MicroVision glasses, as evidenced by the higher light pass-thru on one lens, and MUCH darker image thru the other lens.

In the images below, the order is as follows: (1) RealD in "normal" (worst) orientation (2) RealD in Vertical (best) orientation (3) MicroVision in "normal (worst) position (4) Microvision in "vertical" (better) position (5) MicroVision in "halfway" (best) position...

I'm very impressed with the RealD glasses - ideally, I'd want to track down the manufacturer of the film used in these. More on that later...

In our next section, we cover the biggest disappointment of the night - the performance of my pride-and-joy BDII screen...

OK - it's late, I'm tired, and there's no easy way to say it, so I'll just get right to the point.

There is NO WAY that the BDII screen I have here in my home will EVER be able to be used for a passive 3D system, based on these tests.

The polarised light appears to expose the manufacturing process and an inconsistent application of the various surface layers in a way that I just wouldn't have imagined possible. Effectively, whether viewed through the RealD lenses, or though the MicroVision lenses, the exact same flaws in the surface structure were blindingly apparent. Trying to watch 3D on this screen, in this particular configuration, would be like trying to watch 2D on a multicolored wall covered in grafitti. Impossible.

Now, to be fair, I have heard rumours that there is a new version of the BDII on sale now, technically, their "3rd Gen" screen. I bought mine mid-year last year, so it's POSSIBLE that I have either Gen 2 or Gen 3, I just don't know. But I would advise anyone CONSIDERING a passive 3D system to hold off on buying a BDII until this is clarified a bit more.

I'd also invite SI to submit their 3rd Gen to me for testing, to see if we can make better music, err video, together.

(PostScript: I have file a warranty replacement request on the screen with SI, as it looks like a manufacturing defect to me, in the application of the surface coatings. The response from SI will elucidate whether this is a one-off with my screen, or a normal condition.) Hopefully this is a problem with just my screen, as SI clearly states that the BDII should work fine with passive 3D systems, in which case I will be able to test again with a replacement screen.

In the meantime, it's clear that at this point I had better get some other 3D silver-based screens in house to continue testing.

The BDII not ONLY displayed terrible corruption of the image, but it had pretty bad retention of the polarisation. This is evident by the "clear" lens showing lots of dark bands, as well as by what was SUPPOSED to be the dark lens showing huge amounts of light and image (although again cluttered by surface structure).

The first two images below are throught the right and left lenses of the RealD glasses. The second two images are through the right and left lenses of the MicroVision glasses.

Ironically, the images below actually bear a resemblance to the BDII going up in flames, at least as a passive 3D screen. I can only hope that rumoured 3rd Gen screen gives a completely different result. If not, I'll be forced to cry "foul" on SI's claim that the BDII is a suitable passive 3D screen.

Thanks David!

I was actually shocked at how little light loss there was. Even after passing through both the PJ filter, and the glasses lenses, the light loss was negligible, and the extinction ratio on the opposing glasses lens was ALSO very good.

You're right, of course, the issue with the horizontal/vertical plane of the glasses is very easily overcome. If one can't find these off the shelf, they are easily cut out from sheets to replace the existing lenses.

This whole thing is going to come down to the screen as the single most important component that makes or breaks it all. Everything in the optical chain was near-perfect until the light hit the BDII... then BOOM!

I will be looking for some proper 3D screens for testing now. I'll go for smaller sizes and bigger variety of choices, until I find the best performer.

Yeah, I was planning to have the whole system running by the end of the month, but the BDII surprise may have put a spanner in that plan. Hopefully I can get some alternatives in quickly and at least have a small screen running the final config soon.

In the meantime, I'll keep testing. Maybe next weekend I'll set up the electronics testbed with the two projectors and the 3DXL boxes. This will allow me to get all of that side tested out and sorted. During the week I will be installing intake fans at the back of the PJ enclosure, to blow fresh air directly at the cooling intakes of the RS40 units. It's going to get crowded in there!

Based on these initial tests, I may have to run these projectors in low-lamp mode! lol

Anytime! This is good fun so far!

Ouch, that makes it tricky. There MAY be two options for you with that projector, both of which will cost you a lot of light.

1) Doing straight linear or circular filters can be tried, but using them so that the linear filtering (also part of circular filters) is actually at 45 degrees to each of the colors. The trouble with this is that it would kill half your light, and the stray light left over (these filters block gradually over a broad angle) may result in poor extinction ratios.

2) Doly/Infitec filters may work, but they are expensive. Until I have tried them out on my setup, I don't know if they also have any linear pre-polarisation like the circular filters do.

I'd have to say that a projector with different angles of polarisation for any color is probably just not suited for a two-projector passive system.

BTW - Could you post pictures of your tests here? I'd love to get everyone to contribute what they see in their testing of ANY projectors here, as it would be a great resource for others considering this.

Thanks BlackShark. Yes, the BDII was a surprise, given how much I've read about its suitability for passive 3D.

However, I'm taking the approach for now that this may be a manufacturing defect on my particular screen, given the obvious inconsistent coatings I'm seeing in the photographs, and their public position that it should work for passive 3D.

So I'm sending them an RMA request on the grounds of a manufacturing defect, and looking to see if their response sheds some light on the matter in one way or the other.

Testing for the electronics will be later this week.

I'm not so concerned about that issue, since all current active 3D displays (such as plasmas, LCDs and active projectors) display Bluray 3D squentially also, in the same way.

The right and left frames will be decoded and displayed in the same order as received in the stream, so it is extremely unlikely that there will be any issue in this regard.

We'll have the answers this week, at any rate, but that is not even one of my concerns, thankfully.

OK - I got word back from a senior level at Screen Innovations today. Very impressed with their quick response.

Sadly, the current BDII does not preserve circular polarisation. He says that their testing was fine for linear polarisation, for some reason, but that when they got into testing a new Sony projector with circular polarisation, it all came undone, with much the same results as I have posted here.

All is not lost though - they are working on a specialised version of the BDII which is designed to retain circular polarisation. This screen will be available very soon, perhaps even within two months. They have identified the coating layer that is breaking the polarisation.

He indicates it will be a higher gain screen than the 1.4, so you can get a brighter picture in 3D. I will keep you posted on any further news on this new screen as it becomes available. They are going to be in Europe over the next week for the big show.

I actually also considered this for a split second or two but two things against

1) picture would be portrait, making it just a wee bit difficult to reformat the image

2) projector is only designed to be tilted off the horizontal plane by a few degrees, otherwise the cooling system cannot function as designed