| Author | Message | ||
     ItsikW Rating: N/A |
I have developed a cost-effective method to display passive stereo with two LCD projectors. I am rotating the polarizations of the projectors beams by 45 deg by using a simple half-wavelength retarding filter. The advantage of this method is 100% optical efficiency (as opposed to 50% or less in other methods). Therefore, it is possible to get large and stunningly bright images with simple 1000 lumen projectors. There are, however, two problems. 1. It is necessary to switch the green content of the two images before display, 2. The stereo rejection ratio is ~13db only (depending on projector being used), because this method does not allow the use of "cleanup" filters. (In regular passive stereo one gets approximately 20db.) Nevertheless, I succeeded to get very pleasing images with the system. I saw that there was a lot of discussion in this board on stereo display with LCD projectors about a year ago, but it seems that nobody has actually tried this methd. I am inviting interested members of the community to try it. I will be glad to help anybody who is ready to invest a little effort in this. | ||
| Anonymous Rating: N/A |
Where can one find some half-wavelength retarding filters and which ones have you used? Switch the green content? How? What projectors have you tested with? | ||
| ItsikW Rating: N/A |
Half wavelength retarding filters are not available commercially. I can make such filters for you to experiment with if you send me a pair of transparent filters that fit the thread of your projection lens. I am using 46mm UV filters. Green content can be switched with image processing programs like photoshop or paint shop pro. Another approach would be to cross the green VGA cables (pin no. 2) but this turns out to be a little tricky. I am working on this right now. I am working with Eiki LC-XM2 projectors. Itsik | ||
| Anonymous Rating: N/A |
Hello Itsik! Great idea! Half wavelength retarding filters are commercially available at: http://www.edmundoptics.com Is the rejection ratio 13db only because the retarders rotate only the specified wavelength perfectly 45 degs and the other wavelengths more/less then 45 degs? Or is it because the projector output is not perfectly polarized? What about some internal modification of the projector, e.g. applying 3 retarders directly in front of the 3 lcd panels? Do you think this would be possible? Any solution for changing the L/R green signal by cable/hardware yet? In http://www.curtin.edu.au/cmst/publicat/2001-01.pdf it is stated that there are some LCD projectors which output all colors polarized in one direction. Do you know any? Best regards, Andreas http://www.stereofoto.de | ||
| ItsikW Rating: N/A |
Thank you, I am happy to hear that you like it. Yes, I know that retarding material is available from Edmund Optics. Their material is low quality. I use much better stuff. What I meant is that there are no mounted filters available. This is because retarding filters are useless for photography. My measurements indicate that the main contribution to the low rejection ratio comes from the projector beam itself. The retarders are compensated for wavelength variation, and work quite well. Incidence angle is not compensated, though, and I can see some variations in polarization contrast in the image. However, this seems to be of secondary importance, and can be improved by various means. Each color has a different rejection ratio, 13db being an average. Blue has best ratio, red has worst. I do not recall the exact numbers right now, they are in my lab. Also, I expect the numbers to be different for different projectors. I am going to check a few different models next week. I cannot tell right now what is the major mechanism for polarization contrast reduction in the LCD projection beam. I believe that the light leaves the panels highly polarized. This leaves two suspects: the X-cube or the projection lens itself. In either case, insertion of polarizers before or after the panels will not do any good. I believe that I will have some findings regarding the hardware green crossing within the next two weeks. To my best knowledge, the common LCD projectors are all cross polarized. This was considered to be a great breakthrough in LCD technology a few years ago, and now all manufacturers adopted this technology. Cross-polarized X-cubes function much better optically than their straight polarized versions. Thus, there is no way that a straight polarized LCD projector will survive in the modern market. There may be, though, some specialized projectors that have straight polarization. I have a recent testimony from a colleague working with a JVC D-ILA projector that his projector is straight polarized (circularly). Itsik | ||
| John Billingham Rating: N/A |
Great!!!, Itsik, I too have demonstrated this concept of projection,although in a bit less refined manner. I got a sample of Edmund's 1/2 wave material and was NOT impressed, especially since some common samples of oriented polypropylene from packaging were just as good if not even better ( the best extinction they could achieve was to a dark blue or violet) My particular interest is to 1/2 wave retard ONE HALF of an LCD projection panel to give me a "video-projection device' with both images on the same panel to avoid getting into a "dual-head" drive.Is it possible you could supply a useable amount of your material? Is it a sheet material, or do you "coat" it onto a glass filter base? Thank You Very Much, John Billingham | ||
| John Billingham Rating: N/A |
On a similar topic: I have also just finished making a simple "ruling-engine" that alows me to cut or draw very accurately spaced lines, accurately enough, in fact that I can exactly match the pixel pitch of an LCD screen, When a piece of 1/2 wave retarder is cemented to a clear substrate, and lines are cut at the screen pitch, and then the alternate lines of 1/2 wave retarder are removed via a very fine tweezer (not as hard as it would seem) Voila, I have a polarized/interlaced display viewable with passive pola glasses!( providing the displayed image is interlaced) So, I am most interested in pursuing better 1/2 wave retarding materials, Itsik, how might I contact you? Best Wishes, John Billingham | ||
| Ray Price Rating: N/A |
Does this work only with LCD projectors, or can a DLP projector be used to produce the same result? Regards Ray | ||
| ItsikW Rating: N/A |
Ray, The method I described takes advantage of the fact that the LCD projection beam is polarized. The projection beam of DLP projectors is unpolarized, so the method will not work on these projectors. The easiest way to project passive stereo with DLP projectors is perhaps to mount polarizers on the projection lenses. However, by doing this you loose more than 50% of the light. Itsik | ||
| John Billingham Rating: N/A |
Most LCD "Projection Panels" which are to be used with "overhead" projectors are "straight polarized" This is what I have used (one with a "low-grade" 1/2 wave retarder) to project with. they ARE aligned such that common "3D polarized glasses" work with them.Yes, I could also have "flipped" one of the units instead of using the 1/2 wave material, but they are designed to protect the LCD layer from heat, and this may negate that design in some models. Unlike other LCD or DLP projectors, I CAN afford to play with these! | ||
| Alexander Oest Rating: N/A |
Itsik, Any news on the hardware crossing of the green signal yet? What is the tricky part of it, btw? Alex | ||
| ItsikW Rating: N/A |
Alexander, In order to cross the green content of the two channels by crossing the wires, the two VGA signals must be perfectly in sync. Unfortunately, I have not found a graphics card with this property. In my Matrox G450 the two outputs are not even the same frequency. Michal Husak (private correspondence) testifies that in both Wildcat VP and Quadro 900XL the frequencies are locked, but there is a phase difference. In view of this, I gave up my efforts for hardware green-crossing. It appears that the software approach is more practical. | ||
| Anonymous Rating: N/A |
Hello Guys: This method can work with a the JVC DLA projectors, they have a new model with impressive specs, the SX21, but only with 1500 lumnens, is enough light coming from the projector to get a very good image after the polarizers?. Thank you, JB | ||
| ItsikW Rating: N/A |
JB, I heard that the DLA projectors have the same polarization in all three colors. I did not have a chance to check personally if this is true. However, if this is the case, then stereo display can be easily achieved by a simple lossless manipulation of the output beams polarizations (no need for green switch). All 1500 lumens will be available for the display. Itsik | ||
| Anonymous Rating: N/A |
Hi Itsik: How can achive the the manipulation of the output beams, with a polarizer rortating filter, retarder?, what I have to use to rotate to -45 and +45 degrees?. Thanks JB | ||
| ItsikW Rating: N/A |
JB, You need a half-wavelength retarder to rotate the polarization of the output beam to +/- 45 deg. Such material can be bought from www.edmundoptics.com. Itsik | ||
| hobbes Rating: N/A |
Itsik, I`m really interesting in your method, can you send me more info to this email pls. oscar.boude@unisabana.edu.co or could you pls give me you email, to make you some questions. oscar | ||
| Anonymous Rating: N/A |
Hi Itsik: I found more specific info about the PBS (Polarizer Beam Splitter) on the D-ILA element at: http://www.jvcdig.com/hayakawa_paper.pdf I saw that the light get rotated but never specify which angle, this does matter with your method?, or is just to place the half-wavelength retarder in front of the lenses? Thanks Itsik JB | ||
| Andreas Petersik Rating: N/A |
Has anybody tried D-ILA projectors yet? Maybe the LCOS technology might also be interesting? Could anybody find out the polarization characteristics for D-ILA and LCOS projectors, please? Itsik: What kind of 1/2 wavelength retarders are you using? You said edmundoptics' material is bad. Best regards, Andreas | ||
| ItsikW Rating: N/A |
JB, I had a look at the article. It leads me to an assumption that the polarization properties of D-ILA projectors are the same as those of conventional LCD projectors. If this assumption is correct, then StereoBright (this is how I call this method) should work on D-ILA projectors in the same manner as it works on LCD projectors. The absolute orientation of the polarization states is immaterial; they just have to be mutually orthogonal. But you have to match the final polarization direction to your glasses (which are normally at +/- 45 deg). I have a testimony of a colleague of mine regarding JVC DLA-QX1G projector, a $225,000 device. He claims that all colors in this projector are polarized indentically. This is of course ideal for stereoscopic display. But maybe this super-expensive device has custom optics. Andreas: I am using Nitto Denko retarders (www.nitto.com). This material is better than edmundoptics', but is difficult to get in small quantities. There are other vendors too, but I did not test any others (hard to get samples). To all trying to set up a StereoBright system: this method relies on the native polarization propertis of the projectors. These vary significantly in different projectors. For good display it is necessary to use projectors with high polarization contrast ratio. In most projectors this ratio is approximately 20. The best projector that I found for this purpose is Sharp PG-C20X. It has a polarization contrast ratio of approximately 50. Itsik | ||
| Anonymous Rating: N/A |
Hey Guys tak a look at this page: http://www.jvcdig.com/3d_kits.htm DLA-SX21 Rocks, 1400x1050pixels, 1500 lumens, 800:1 contrast and under $9000 (Street price), This is the way to go, I know that they are not cheap, but: LCD at 2200 lumens like the sanyo PLV-70 plus polarizers, very low light and no more than 1300 x720 pixels. DLP, the same thing, try to get the highest resolution from a DLP (1300x720 aprox) with the new mustang HD2 MDM. http://www.projectorcentral.com/cedia_2002.htm JVC is the way to, I hope that they can get a higher lumens + more contrast projector soon. JB | ||
| Anonymous Rating: N/A |
Itsk: What about circular polarization?, how you can make it, linear is Ok if you have the people looking directly to screen whitout tilting their heads, then is a problem, Barco suggests that Circular polarization is the way to go, if you need to tilt your head. How can achieve this with LCD's?. I saw the option for this stereo setup: http://www.christiedigital.com/News/PressKits/itsec/montagelx33lr_nov02.pdf Totally dark room, head titl, LCD?, (2) Canon LV7555 (4000 Lumens - 1024*768)or (2) Sanyo PLV70 (2200 Lumens 1366*768) Both with 900:1 contrast, how much light you loose with circular polarization and if it's enough the Sanyo's model 2200 lumens?. Thank you, JB | ||
| itsikw Rating: N/A |
John, In principle one can make circular polarization setup with the same ease as linear polarization setup. With StereoBright, you simply have to use L/4 retarders instead of L/2, and eyewear with circular filters. You are right, with a circular setup one can move his head freely. However, circular polarizers are more expensive, and their extinction is inferior compared to linear polarizers. Itsik | ||
| Ray Price Rating: N/A |
Hi Itsik, I am interested in your stereobright setup. Since a software crossing of the green signals is pretty impossible (for me), do you know if any of the Quadro cards are locked to enable crossing of the green signals wires? Thanks Ray | ||
| Ray Price Rating: N/A |
Oh, also... how do you get information of the polarization quality of LCD projectors? Thanks | ||
| John Billingham Rating: N/A |
Don't forget that Peter Wimmer very kindly included "Stereobright" as a format in his recent "Stereoscopic Player" software. Best Wishes, John Billingham | ||
| Anonymous Rating: N/A |
I tried a little experiment this weekend, and made a VGA adapter that crosses the 'green' and 'green gnd' terminals of two vga leads. I then used my Geforce FX 5900 and two LCD monitors to test the results. What I was expecting (or hoping) was that in horizontal span mode, the window displayed on the right monitor would be lacking the green component (as it was), and that IN THE SAME POSITION on the left monitor would be the green component of the right window. If was, but it seemed shifted by about 1-2 inches to the right. In fact the entire left two inches of the right monitor were displayed (as a green channel) on the left monitor. What does this mean? That the signals are out of sync? Or have I made my lead incorrectly? Any information would be appreciated. Thanks | ||
| M.H. Rating: N/A |
Anonymous: Yes. The output signals from dual head graphic cards are out of sync (non genlocked). For genlocked sygnal you must use Wildcat , Quadro 2000 FX or Quadro 3000 FX ... | ||
| Anonymous Rating: N/A |
Thanks. Those cards are VERY expensive. Is there a seperate device that can be purchased to genlock two video signals? | ||
| Peter Wimmer Rating: N/A |
> Is there a seperate device that can be > purchased to genlock two video signals? It would be much better to swap the left and right green in software. For videos, my Stereoscopic Player (http://mitglied.lycos.de/stereo3d) can do this. Although not officially supported, this player can also be used for images (jpg, bmp). For gaming I don't know any solution. | ||
| Anonymous Rating: N/A |
Indeed, it was gaming I was wanting it for. Well, I hear NVidia are about to release an update for their stereo drivers that allow left and right images to be split on dual-head cards, so I dropped them a line suggesting green swapping as a new feature for their following release. | ||
| itsikw Rating: N/A |
Ray, I absolutely agree with Peter that crossing of the green signals by software is much more practical than crossing by hardware. There is no published data on the native polarization ratio of projectors. The only way to know is to measure. For those who want to build a StereoBright setup: the retarder should be specified at 550nm. In most projectors this is the peak wavelength of the green emmission. Itsik | ||
| Anonymous Rating: N/A |
Hey does that mean you can receive 3d on two monitors? .. | ||
| Ray Price Rating: N/A |
Yes, it does, anonymous. BTW, Itsik... you say you are using stuart screen material, are you using the FilmScreen 200. I tested a sample of this for polarization and it seemed very good. Problem is that Stuart is VERY expensive when compared to Da-Lite. My Da-Lite screen which is 54" x 72" cost me $420 (including frame). For a similar size in Filmscreen 200 Stuart want $1652 (including frame). Do you mind if I ask what size screen you have and how much you paid? Thanks Ray | ||
| itsikw Rating: N/A |
I did not pay a single penny for my Stewart screen... Anyway, there is now a new screen SilverFabric. It is excellent quality and much less expensive than Stewart. If you are a DIY guy, you may try the silver gray plotter paper by Oracal. You will get an excellent screen for $10... Itsik | ||
| Ray Price Rating: N/A |
Hmmm, I really want rear projection though, so I can stand directly in front of the screen. I tried some frosted glass which worked, but when compared to the stuart material the difference in quality is HUGE. So is there a trick to not paying for your Stewart screen? :P | ||
| itsikw Rating: N/A |
Ray, The trick is to work for a big company and ask for a sample... For rear projection, the only screen that I know is working is Stewart. Frosted glass is no good. Itsik | ||
| John Billingham Rating: N/A |
Ray, and others, Please be careful with Itsik's advice, he has stated that he only posts his "bad" ideas, not the good ones. (which I would love to see, since the "bad" ones look pretty good to me!) John Billingham  ! | ||
| Anonymous Rating: N/A |
--what is an wavelength retarding filter, what is it made of, what does it do to light so that i can split the images to each eye with only a simple polariser? --moving head makes projection come along with it?(head tilt) does this tan projection as anithing to do with it?: http://www.stereovision.net/articles/cebit2001/part3/cebit3.htm see jpegs: http://www.stereovision.net/articles/cebit2001/part3/tan1.jpg -to--tan4.jpg | ||
| GianCarlo Rating: N/A |
Two questions: 1. Has anyone suggested that NVidia add "flip and rotate second output" to their dual-head stereo mode? Flipping the image horizontally would allow anyone to view stereo with a single-mirror stereoscope and two monitors (very cheap and easy). Rotating the image 45 degrees would elimintate the need for rotating filters with the use of cheap LCD projectors. It would be awesome!!! I hope nVidia understands that dual-head stereo is the future. Because of ghosting and sync problems, page-flipping and anaglyph just don't cut it! 2. Are there LCD projectors out there that are polarized in the horizonal rather than the vertical (or vice versa?). Is it possible to crack open a projector and flip the polarization? Thanks! | ||
| StereoGamer Rating: N/A |
I understand the flip idea but I don't get the rotate idea at all. How do you rotate an image 45 degrees and still show it in a non-rotated rectangle? You'd get all kinds of clipping of the corners. I guess I'm missing some pertinent piece of info. | ||
| GianCarlo Rating: N/A |
Yes, Stereogamer, you're right, flipping horizontally or a "mirror" mode is pretty simple, but to rotate, it's more of a problem. It would be like letterbox on the second projector. So the resolution and zoom of the second projector would have to be adjusted either manually or with drivers to closely match the first projector's image. So yeah, it's a long shot idea, but not impossible. Personally, I will be overjoyed when nVIDIA gets the dual-head mode stable for all kinds of games. If they take the options further with it after that, so much the better. But I'm really interested in an answer to my second question. If it is possible to buy two good, cheap LCD projectors with one polarized in the vertical and one in the horizontal, no filters or rotation would be necessary. Does anyone know if two such projectors exist? Thanks! | ||
| GianCarlo Rating: N/A |
I've started searching the net to try and answer my own question: "Is it possible to buy two good, cheap LCD projectors with one polarized in the vertical and one in the horizontal?" I know Barco sells an opposingly polarized pair of LCD projectors at a high price, but I've seen a few web-pages that suggest that some standard LCD projectors are polarized in the horizontal and some in the vertical! If this is true it would be awesome to compile a table on this site listing which way each projector is polarized. Perhaps then we would all be able to buy an opposingly polarized set of projectors and set up a passive projection system for cheap!!! So please help! Thanks! | ||
| Dr Brian Rating: N/A |
Just wanted to jump in and quickly correct a few misunderstandings.... The references to "Rotating 45 degrees" I believe are refering to rotating the Polarization (not rotating the image). For example, on your cardboard 3-D glasses the plastic Polarized Filters in one eye are rotated -45 degrees, and +45 degrees in the other eye. The polarization of the light from the projector needs to match, so the L/R images will each be visible in the proper eye. Certain simple plastics can be used to rotate the polarization of light, by just sticking it infront of the projector (or mounting it as a lens filter). I think you can do a cheap experiment by holding some cellophanes or plastic package materials in front of your polarized projector, and watch the polarization rotate (you'll need to rotate your head while wearing polarized glasses to see the effect). These are poor/uneven retarding filters, but should work well enough for a playful experiment. In the Vertical & Horizontal orientation discussion, this is somewhat a mute issue. If your glasses are polarized -/+45 degrees, but your projectors were 0/+90 degrees, then you'd still need to use retarder filters to rotate the polarizations to match your glasses, anyway. Unless you are ok with non-sandard glasses, but then you'll get glare issues (light naturally reflecting in the room will often polarize at 0/90 degree angles, as it bounces off surfaces). The other issue (for vertical/horizontal) is that most projectors are cross polarized, not just polarized in horizontal OR vertical (but both). That is, the Red, Green, & Blue color components of one eye are not all polarized in the same direction to begin with. The Green is roughly 90 degrees from the Red & Blue. Since the other eye/projector is already polarized 90 degrees apart, you could achieve a quick & dirty solution, by just swapping the green color channels to the other eye. Hence all the discussion about Green and wire swapping vs software swapping Green channels. ItsikW has a VERY interesting approach to getting Brighter 3-D images by USING the natural polarization of the LCDs in the projectors. Many other people thought of the LCD polarization as an obstacle, while ItsikW thought of it as a usable feature. I'm impressed. I hope I didn't massacre the facts too much, or write this posting too confusingly. (I'm new to posting). -- Brian | ||
| Anonymous Rating: N/A |
"moot' | ||
| M.H. Rating: N/A |
Let me add something to the discusion. The key question si probably is the brightnes of the final image the key for measuring the quality (related to ANSI brightnes of the projector)? After seeing several projection setup, I can say NO. The key factor is maximal contrast ratio, with is aprox 1/200 for LCD and 1/1000 for DLP. The images from dual DLP were looking beter at higer contrast than brighter images from dual LCD. This is my personal expereinces based on comparison of 2 2500 ANSI LCD and 2 1500 ANSI DLP projectors stereoscopic passive setup. | ||
| itsikw Rating: N/A |
Yes, Michal, you are right, contrast is very important. The contrast value of a projector is a design parameter. The problem is that there is a tradeoff between the light output and the contrast. When you increase one you necessarily decrease the other. This is true for both DLP and LCD. Up to recently the major performance parameter of a projector was considered to be the light output (lumens). This caused manufacturers to design high-output and low contrast projectors. Nowadays, as customers got more educated, the focus shifts towards the contrast. The Internet is loaded with articles comparing DLP and LCD technologies. Having experience with both, I can testify that each technology has its advantages and disadvantages. This is always so in life. I have just searched projectorcentral.com for LCD projectors with contrast above 1200:1. There were nine results, with MSRPs ranging between $3,000 and $27,000. The $27,000 projector is by Sony, and designed for high quality HDTV projection. It has a contrast of 3000:1. | ||
| M.H. Rating: N/A |
itiskw: Is sound like hi contrast ratio LCD with intensity enhanced by half wave retarders will be in the future better than DLP with hi contrast but low intensity. But what about the image sharpnes on LCD in comparison to DLP ? Another question - does the retarders + polarization have influence on the contrast ratio ? | ||
| Anonymous Rating: N/A |
The new DLP BenQ7200 has 2000:1 contrast ratio, 2200 lumens, and is 1024x768 native resolution. Not bad for $1500. Infocus really shook up the market with the introduction of the X1. We'll be seeing some very interesting tech advances in the sub $2000 range over the next several months. John | ||
| Anonymous Rating: N/A |
Here's a sellers link http://www.notebookforless.com/products/PB7200 Projectorcentral.com and projectorpeople.com are both GROSSLY overpriced. Hell, Bestbuy even beats their prices and that's saying something. John | ||
| itsikw Rating: N/A |
Michal, Generally LCD has less image sharpness than single chip DLP, becuase of registration errors of the three LCD panels ("color convergence error"). Video looks different on LCD versus DLP, because LCD has greater latency. Some people say video on LCD is more smooth, other say it is blurred. Polarizer actually increases the contrast in LCD projector. This is because the black light is less polarized than image light. I was able to measure 15% contrast increase in my projectors. Itsikw |
