| Author | Message | ||
     Alexander Oest Rating: N/A |
Hi All I suggested a possible solution to the low max refresh rate of DLP projectors in the 3D News section. I think we'd better continue any discussion here in General 3D Discussion, so here you go: The problem with DLP projectors only being able to run 60 Hz is the heavy flickering. Now, I assume that flickering is caused mainly by the shutterglasses, not by the rapid change of fields in the video signal. This might be solved by having each shutter panel flicker TWICE per field, so the shutter sequence would be RR-LL-RR-LL-RR..... This would obviously result in 120 Hz shuttering - and hopefully eliminate visible flickering. How about that? Do you think it could work? Does anyone have suggestions of how to design a circuit that could go between the video card and the glasses and do this doubling? I use Asus card and glasses (they've got mini-jack connectors), so if you come up with a proposal based on this, I'd go to the electronics shop right away. Alex Btw, if this works, it'd increase stereo viewing comfort on television sets too. | ||
| gb Rating: N/A |
If DLP projectors display the entire field in one go then doubling the shutter speed should reduce perceived flicker. This technique is used in Cinema projection, where there are 24fps but each frame can be blanked upto 3 times. >>>Btw, if this works, it'd increase stereo viewing comfort on television sets too. This wouldn't work for CRT Television sets as the picture is scanned sequentially. What would result is severe strobing effects. Similar to that seen when a video camera is pointed at a TV screen. GB | ||
| Andreas Schulz Rating: N/A |
Hi Alex! I'm afraid doubling the shutter rate won't help at all - for each eye, instead of : XXX___XXX__ you will get X.X._._.X.X., i.e. you will still have the 30 Hz main flicker. The cinema example doesn't apply here, since there the image is interrupted by a gap of the same duration as between two images, i.e. you get _P1_P1_P2_P2_P3_ instead of _P1PP1_P2PP2_P3PP3_. | ||
| Alexander Oest Rating: N/A |
Hi GB and Andreas Hm. I can see the problems. TV is definately out of the question. And after thinking it over, there are some additional obstalces. Andreas, you write the shutter sequence of each panel as X.X._._.X.X .Does this take into account that the shutter panel of one eye must remain shut when the frame for the other eye is on the screen? I didn't take that into account in my original post. I think this would lead to a sequence of X.X.X._.X_.X.X.X - assuming that X means shot and _ means open (you could write it as sssososss. During the frame where the shutter panel originally would be shut, it acutally *is* shut for the whole period, but in the frame where it originally was supposed to be open, it's really open-shut-open. Meaning that it goes as this: shut-shut-shut-open-shut-open-shut-shut-shut-open-shot-open etc. Or put in a more schematic way: RS: sssososssososss... RI: ---+++---+++---... LS: ososssososssoso... LI: +++---+++---+++... Where: RS=Right Shutter RI=Right Image LS=Left Shutter LI=Left Image o=open s=shut +=frame is visible on screen -=frame is not visible on screen With the max 60 Hz refresh rate of the projector, the normal duration of the panel being shot is 1/60 sec. Now, the shutter panel must go through states of open-shut-open in that amount of time meaning that a period openness or shutness in the "oso" interval lasts 1/180 sec. It takes a good pair of shutter glasses. This is getting a bit difficult, because now the shuttering would be with two different intervals: shut for 1/60 sec (3 times 1/180 sec), open for 1/180 sec, shut for 1/180 sec, open for 1/180 sec and over and over again. The only reason that this should work would be if the percieved flickering is due to the duration of the "open" periods (leaving the duration of closed periods without influence on flickering). I don't know if this is actually the case and if reducing open periods from 1/60 to 1/180 would help anything. I admit it's all beginning to sound a bit far fetched. The image would be considerably darker, as the glasses are shut for a longer total part of the time. And the glasses would have to be very quick. And then you'd need a circuit that takes only the open intervals and turns them into three intervals of open-shut-open. That doesn't sound like an off-the-shelf component to me. I probably won't exploit this possiblilty further, unless you insist that it acutally is a brilliant idea ;-) Alex | ||
| Larry Elie Rating: N/A |
Actually, I think there is another solution. Expensive. Cumbersome. But I think workable. Two DLP projectors, wired (for example) so that the left projector gets the odd fields THROUGH a line doubler to give a full frame not field, and similarly the right gets the even fields again THROUGH a line doubler, each at 60 Hz. Instead of field blanking, it is frame blanking. I have seen two LCD's wired to the same source (2D) for additional brightness, and the images can be made to coincide pretty well. | ||
| David C. Qualman Rating: N/A |
If the image is being displayed in a line-scan format, then double-flicking will definitely not work. Assume that you want to flick the shutter in the middle of the image. While it is open, you are seeing the lines that are being drawn. While it is closed, you see no lines. Then, when it opens again, you see lines again. But, you can only see lines that are being drawn when the shutter is open. So, the lines that were drawn when the shutter was briefly closed will be lost. The net effect is that you will see a black band in the middle of the display. The real problem is that the other lens is still closed for a long time. Flicking the displayed image won't fix this - the closed eye still sees a slow flick rate for that one field. Using two projectors, with polarizers and passive polarized glasses should work - I think - unless the projector also polarizes the light, or the screen depolarizes the light. If you have two projectors, then there should be no need to use active glasses. Think of the problem as a multiplexing problem. With a single display, we want tyo take two images, modulate them through a single transmission channel (the monitor), and then demodulate them at the destination back into two channels (the eyes). With two displays, we have have two transmission channels. The problem now becomes preventing cross-talk. That is, by polarizing each channel or shuttering each channel. | ||
| Alexander Oest Rating: N/A |
Hi Larry and David I know about the possibilities of outputting to two projectors. The reason I wanted to explore the one projector plus shutterglasses soulution is: A: it would be cheaper with one projector B: Page flipping drivers exist. But external vga splitters doing the job is expensive (e.g. the CyViz at +10.000€/US$). Larry, do you have any particular line doubler in mind that'd do the trick? What would the price be? David, you speak of drawing lines. I don't think a DLP projector draws lines at all. It has a micromirror chip whit 1024x768 mirrors - meaning it draws the whole picture at once. Still, I expect the long dark period to give so much flicker, that It wouldn't work well no matter the projector type. I did some experiments a while back with Win3D, Michal Husak's video player and VRCaddy, but I was unable to get stereo on separate vga ports to work (on a Nvidia GeForce2MX dual card) with any application. If I give it another try, which version of VRcaddy and Win3d would you recommend? (And which OS?). Alex | ||
| Larry Elie Rating: N/A |
No, I had not picked a line doubler. Yes, DLP does not have to 'draw' in the classic sense, the mirrors could be addressed in that way, but in parallel makes more sense. The main advantages over LCD for 3D are; a. LCD is significantly polarized b. Contrast ratio of DLP is much higher than LCD c. Space between pixels is reduced d. Higher intensity is possible without reducing contrast e. Cooling of a reflected device is easier than a transmitted device DV magazine has listed several companies working not at XGA but at HDTV resolutions. i1080 spec as I recall. These are theatrical systems, 24 fps non-interlaced as I recall. Having said all that; I'm STILL looking for 1 or 2 DLP chips to experiment with. I'm not trying to do this for free. Yes, I want to MAKE a DLP projector. ONE DLP projector that can project TWO images from TWO DLP's at one time through ONE set of optics. Most DLP's are 3/4"; rather small. I intend to put two, either side by side on on top of one another in the optics of a very bright presentation grade reflective optic slide projector; the optics are designed for a 2" slide, so there is plenty of space. Whether I use 2 polarizers in front of them or time them out right/left is secondary. Is this too ambitious? Could be. But right now I need to find a couple of DLP chips. Larry Elie | ||
| Alexander Oest Rating: N/A |
Larry, Some days ago, I asked the company that sells projectors to my workplace, if it was possible to get my hands on a couple of DLP chips for experiments. He sad it would, but I shouldn't expect to get them for free, because he'd have to buy them as replacement parts. Perhaps you can work out something similar with a projector distributor? Have you thought about how to interface the chips with your graphics card? You surely must need something to go in between and convert the signal. Alex | ||
| Larry Elie Rating: N/A |
Yes, I have thought about the interface, but it is dependant on what chips I find. I'm not sure who other than TI even produces DLP's. The harder part is to do the slight angle allignment; I want two images from offset positions to converge at the screen. This requires each to be mounted at a slight angle to the plane of the mirror of the projector. I don't think that is trivial. | ||
| Giorgio Bogoni Rating: N/A |
Maybe this is worth a look: http://www.crystalcanyons.net/3DGuidebook/ViewingMethods_Projection.htm Giorgio. | ||
| Larry Elie Rating: N/A |
Thanks. Looks promising. Now if I can get a couple of DLP's... | ||
| Alexander Oest Rating: N/A |
Some months ago, on avsforum.com, I saw some Davis DLP components for sale. If my memory serves me correctly, they were referred to as "graphics engines" or "light engines". I had the impression that they were integrated units consisting of a lamp, a colour wheel, a DLP chip, and probably some sort of circuit board. Didn't find out what interface they used, neither whether optics were included. They were cheap - I remember thinking that it was just a bit more than the price of a lamp alone, so it must have been below 1000$ per unit. Sorry I can't be more specific, but when I saw it at first, I thought nothing particular about it. And after a couple of weeks, when I had decided that it acutally was pretty interesting, I couldn't find the original posting. The guy probably got them sold and removed the ad. I'll have another look, and see if I can find it. Alternatively, there might be another batch available some day in the future. Alex | ||
| Michael Martin Rating: N/A |
another solution for stereo with one DLP projector would be a 'polarizing Left Right inverse beam splitter'. wearing polarized glasses, you dont' get half the refresh rate, but half the resolution. But how would i build such a beast? use to prisms covered with polarizing foil? Or is it by any chance commercially available? | ||
| Larry Elie Rating: N/A |
Actually, you could do it with a good 'chopper' wheel (used in optics) with a polarizer mounted to the wheel. Look them up from any physical optics company on the web; I have used PAR (Princton Applied Research-VERY expensive) and SRS (Stanford Research Systems-Pretty expensive yet better than the above). You could also make your own. They are used to create 'AC' signals from light sources. You also need either a shutter or accurate control of when the projector/DLP is and is not projecting. The polarizer may do a decent job of shuttering as well. It also has to sync to the frames, but good choppers do this fine. Intensity will suffer. | ||
| FrancoT Rating: N/A |
Hi all, anyone has tried active stereo using the projector NEC 150Z? It's a DLP projector with the new DMD (Digital Micromirror Device) tecnology. It basically means that it should be able to reach true 120Hz of refresh. Stereographics is claiming that this tecnology works on the Christie Digital Systems Mirage projector series, so it might work also on the much cheaper NEC. What do you think? Franco | ||
| Anonymous Rating: N/A |
Cyviz offers you to use DLP or LCD projectors together with their stereo converter. Gives up to 120hz (60 stereo pairs/sec). It's a high-quality solution, but at a modest cost. | ||
| Alexander Oest Rating: N/A |
Anonymous, AFAIK, the Cyviz is not priced modestly. Doesn't it cost more than €/$ 10.000 ? Alex P.s. Franco, you naughty cross-poster ;-) I replied to your post in the other thread. | ||
| Ray Price Rating: N/A |
I have the LT150Z and it does not project in stereo. I am comtemplating 2 LT150Z's with polorizing filters running together from a dual-head NVidia card tho. I will let u know how I get on. | ||
| ItsikW Rating: N/A |
There is a way to combine two 60fps single-panel DLP projectors to get a 120fps projection system. The idea is to use the fact that all modern single-chip projectors make two color wheel revolutions per frame. If you somehow force one projector to show only the first wheel revolutions in each frame, and the second projector to show only the second wheel revolutions, you will get a 120Hz fps projection system. In order to implement this, it is necesary to develop a custom electronic processor. There is no need to modify the projectors. | ||
| Anonymous Rating: N/A |
So how complex would this kind of electronics be? Maybe the people over at Cyviz would be intrested in the idea? http://www.cyviz.com | ||
| ItsikW Rating: N/A |
In my opinion, the electronics is rather straightforward. Everything is purely digital. I presented the scheme to TI engineers, and they said it should work. Thanks for the tip: I will approach Cyviz. | ||
| Ray Price Rating: N/A |
These polarizers claim to have around 90% transmission rate. Can anyone with more optical expertise tell me whether they would be suitable for a polarized DLP setup? http://www.profluxpolarizer.com/pdf/PPL%20MAY%202002.pdf | ||
| ItsikW Rating: N/A |
Ray, You must be careful in interpreting transmission of polarizing material. Normally, the transmission is defined with respect to unpolarized incident light. In this definition, a polarizer cannot have a transmission higher than 50%. Since the light of DLP projector is unpolarized, you will loose at least 50% of the light when you mount a polarizing filter on the projection lens. This is true for the proflux polarizer as well. My impression is that the proflux polarizer is superior to polymer polarizers for this application. Tere are, however, a few precautions. 1. Proflux acts as a polarization beam splitter. This means that the rejected polarization component is reflected rather than absorbed. Therefore, a significant amount of light will be reflected back to the projector. If you do this to an LCD projector, you would kill it immediately. Fortunately, DLP projectors are more immune to back reflections, but I would first try to get an approval from the manufacturer before trying to mount a proflux polarizer on the projection lens. A possible remedy would be to mount the polarizer at an angle (say, 45 deg) but this will increase significantly the size (and, the cost), and also degrade he optical performance. 2. Handling. Read carefully the handling instructions for a proflux polarizer (in the moxtek site) and decide if you can cope with that. | ||
| Ray Price Rating: N/A |
Thanks for the info on the polarizers. With my projector being only 1000 lumens I don't think a polarized solution will work very well then  . I was interested to see your info on the color wheel idea. Did you hear anything back from Cyviz? How would you sync the two projectors? | ||
| ItsikW Rating: N/A |
Ray, On a second thought I decided not to address Cyviz. I know this company, and I believe they will not be interested. The electronic processor that will do all the necessary signal processing for time-multiplexing the wheel revolutions will genrate also a common sync signal for both projectors. | ||
| Ray Price Rating: N/A |
How difficult would this be to build? Would it be using pretty much off-the-shelf components? I have a friend of mine who is an electronics design engineer who could probably knock up a prototype if he knew exactly what was asked for. Regards Ray | ||
| ItsikW Rating: N/A |
Ray, Yes, it can be done from off-the-shelf components. However, I estimate that this is a $200,000 project. Drop me an e-mail to "itsikw at aquanet dot co dot il" and I will tell you more details if you are interested. Itsik | ||
| Alexander Oest Rating: N/A |
ItsikW, A couple of years ago, I asked this board for a VGA splitting layout that took a page flipped VGA signal and send right and left eye's frames to separate VGA outputs. Andreas Schulz made a layout, but had the reservation about it, that every other frame on each VGA out would be blank or black. As the setup was intended for polarized projection, this was a drawback, but for your purposes of a dual projector shutterglasses setup, it might be exactly what is necessary. I never built the circuit, but it was my impression that the price would be nowhere near the $200.000 you mention. Andreas, do you have any opinion on this? Is the layout your intellectual property, or can it be freely used? Alex | ||
| ItsikW Rating: N/A |
Alex, The idea that I mentioned in my post from Aug. 22 (I call it "Rate Doubler") is not a simple stereo demultiplexer. Although it uses two projectors, the idea is to create an inexpensive DLP based 120Hz projection system for flickerless active stereo. To the best of my knowledge, it is impossible to make such a system with a simple page flipping -> parallel signal converter. The Rate Doubler is based on the fact that single panel DLP projectors use two color wheel revolutions per frame. The main expenses in the project are PCB development and fee for TI technical support. Itsik | ||
| Anonymous Rating: N/A |
Just a question that is slight related, but what can you access via the RS232 interface that might be intresting for stereo viewing? Are there some DLP projecters that you can have a lot more control of what is happening? Or are they all bascially remote controls via a PC? | ||
| Alexander Oest Rating: N/A |
The right place to look for ways of "hacking" a projector would be www.avsforum.com - under "Digital Projectors - Under $5000 USD MSRP". There's a guy there who developed control software for Davis projectors, but I don't know about other brands. Alex |
