Can labview create a 3D vision and present it in a 3D monitor?

[lshi]

Hello all,

Here's my situation. Currently I am using labview to create some simple animation (2D picture ctr) that can interact with user, presented on a LCD,. But we recently decided to try it in 3D. I did some research and found that the LG cinema 3D monitor D2342p with passive glasses (polarize) best suit our project. Since the image has too be interlaced, can labview create such 3D vision (3d picture ctr), and how? Can a 3D image has content projected beyond the monitor, say 10m behind the panel? I really appreicate it if someone can answer or direct me to a better place to look for.

Thanks,

Lei

[X.]

Not an expert at all, but I am interested by your project, so I am looking at it from a potential user point of view..

The 3D picture would probably allow you to compute and represent the alternating orientation of the scene needed for the s and p-polarization fields (with the proper lighting). Essentially you need to compute two images for each frame, both with slightly different parallax and alternate them at 60 Hz.

My only experience with projecting a LV generated image on an alternate screen is using a LCOS (liquid crystal on silicon) spatial light modulator. What I do is create a front panel which is a single LV image (not picture, but I am sure it would work just as well with a picture control) and size it to the LCOS "screen" size (800x600 pixels). Then I move the VI panel to my second monitor (my video card supports two), which I set to a 800x600 resolution. Finally, I disconnect my second monitor and connect the cable to the LCOS. That works like a charm. I am using a DVI cable, but I am not sure whether the signal is interlaced or not.

In any case, one thing which I don't do is try to update my LCOS at 60 Hz (which you would need for 1080i output to your monitor). Since there is no way to time things without drifting off synch over time using LV on Windows, my understanding is that you would need some kind of real time system to be able to do that (an RT controller with a video board may be able to do that). When this is done, I suppose you would just have to worry outputting your images at this fixed rate and the video board would generate the properly synchronized stream... But I am not sure and would be also curious to have an authoritative answer.

[X.]

The RT idea is probably over the top and impractical. What you probably want to do is some sort of low-level transfer of your fullscreen images to the video buffer (as a video game would do). The problem is that you would have to be able to keep up with the required 60 fields/s rate, otherwise you will drop frames (which is OK), or worst, go out of phase. I suppose that there is way to tell the video board that a field in the buffer is "upper" or "lower" (s or p in case of a 3D display based on polarization). But all this is not controllable from LV at this point, AFAIK.

[lshi]

Hi X.

Thank you so much for your reply. Yet I'm not going to pretend I fully understand what you are saying. So Labview can not do 3D (interlaced) in 3D display? Is it possible that I create a series of these interlaced 3D images, show them in the 3D picture ctl (or maybe an object, a ball?), and present it on the 3D screen, so the two eyes see them as a 3D image?

Thanks,

Lei

[X.]

From what I understand, what this display does is alternate between two polarizations at 60 Hz. One field is displayed as a s-polarized image and the next as a p-polarized one and so on. What is crucial is that the s-polarized image always shows you the point of view corresponding to the eye which is behind the s-polarizer [let's pretend for the sake of this discussion that it is the lft eye] (and the p-polarized image, the point of view of the eye behind the p-polarizer). That's how your brain is going to reconstruct the 3D perspective.

If, because of desynchronization, you are sending the image corresponding to the left eye at an instant in time where the field is displayed as a p-polarized one, it will be seen by your right eye, and that will confuse you.

This is why I am saying it is crucial that the display always gets left-eye/right-eye images in synch. There is pretty much no chance that by sending images to your front panel, they will be or keep in synch with the display, so the problem I described above WILL happen.

It MAY BE that it is sufficient to provide an image that is comprised (artificially) of two interlaced fields, one corresponding to the left-eye perspective and the other to the right-eye perspective, and the video board will to the proper deinterlacing and send each field to the same eye for you. But I'd check with the manufacturer first. The problem with this approach is that you would have to do the interlacing yourself. That is, generate two images with different perspectives for each frame and interlace them before displaying the result in your picture indicator. That's a little bit more involved than alternatively displaying a left-eye perspective image and a right-eye perspective image, but not that complicated.

Again, I have absolutely no expertise in this domain! I am just interested in the question.

[lshi]

The lg monitor use so called FPR (Film Patterned Retarder, film type polarized glasses) technology. So it is horizontally interleaved image. The monitor will arrive in a few days. We'll see how it works with labview, and what are the other options. Thanks again for your helpful suggestions.

[X.]

Sure, no pb. When and if you obtain some results, it would be nice to report them back here!