I compared various surfaces to make the diffuser: frosted glass, vellum, tracing paper and frosted acrylic. I have a genius VideoCAM NB (a cheap camera) and 10 OP298 leds (100 mA). Here is my setup.
Great post camilosw! I think camera settings, especially when I tested the SPC900nc and unibrain fire-i, play a factor and can help get a nice image before it ever hits the tracking program.
My personal preference, and the tests above may back this up, is that the diffuser produces more defined blobs when is nearest the hand side.
If possible camilosw, since you have all these diffusers, can you also test for us how symbols look through these diffusers? For example, place something an image, card, printed symbol (something flat) and take a picture with each diffuser to see how clear it shows through the different diffuser setups. Some of use will eventually like to use fiducial symbols, and a diffuser needs to be able to pick up blobs as well as clear images (symbols).
My personal preference, and the tests above may back this up, is that the diffuser produces more defined blobs when is nearest the hand side.
I do the sandwich glass-vellum-glass test thinking in that. The glass on the hand side could be a protective covering for the vellum and is the thinner glass so the vellum is near the hand. The problem is the hotspot, but I think that the best choice could be to use more leds and a diffuser or reflector for them (as discused previously). In this way we get more defined blobs.
cerupcat - 17 July 2007 09:26 PM
If possible camilosw, since you have all these diffusers, can you also test for us how symbols look through these diffusers? For example, place something an image, card, printed symbol (something flat) and take a picture with each diffuser to see how clear it shows through the different diffuser setups.
I will test in the next days with black and white simbols as the used in the reactable.
Here are my results, feel free to add them to your post if you’d like to keep all the images at the top (pictures taken with unibrain fire-i):
Tracing Paper on Top of Acrylic (6mm): As you can see the fingers and symbol are fairly clear. Keep in mind the symbol I used wasn’t completley black to begin with (my printer was running out). Hotspots are a problem, but can be fixed by angling or reflecting the light.
Tracing Paper on Bottom of Acrylic (6mm): Hot spots are gone, the fingers still work (less bright) but the symbol isn’t clear enough.
camilosw, did you get to test frosted glass with the frosted part on the hand side (not camera side)? I like the frosted acrylic of yours best, and I’m thinking frosted glass (one sided with frosted part on hand side) will be what I will end up using (and using angled LEDs or reflectors to avoid hotspots).
I don’t tested the frosted part on the hand side because I think it’s not nice because this feels rough, but I will do the test so you can compare. The frosted acrylic is 1mm thickness so we can use them with the frosted side facing the camera (as I tested). I cut them from a cd case. I will search a similar material.
Hey Cerupcat,
I did some more testing tonight as I really want to get things working with my diffuser on top so I can use markers and stuff. That was one of the things that got me really excited about using DI in the first place over FTIR. Anyways, I had my wife stand up and be my touch surface tester while I looked at the webcam image and adjusted the position of my leds. I found a good angle and position that seemed to work well… and without a hotspot. I’ve attached a crude drawing. Basically, what I did was put my bank of leds just off to the side of the acrylic angled towards the opposite side. I believe that if you did this on both sides, with each side lighting the opposite side, you would get consistent, hotspot-free lighting. I only tested lighting from one side, but it seems suggestive that this would work when placing another bank of leds for the other side.
Awesome Miketavius! That’s actually very similar to my current setup (not the one on my test picture above). Should we assume the camera is centered in the middle? How big is your screen that you tested and how many LEDs did you try?
Maybe as we all get tests results we should list our table size, number of LEDs used, LED distance from surface, and placement of diffuser.
Yeah the camera is straight in the middle. My test setup is 24” wide and on the one side I have a bank of 8 leds in a row, fairly close together. I’m sure that you could either space them out a little bit, or just add more to the bank to accommodate deeper screens. The setup height was about 32” and the leds were about 10” from the bottom of the screen. I don’t know why I didn’t try this before tonight… oh well I guess hindsight really is 20/20.
Yeah the camera is straight in the middle. My test setup is 24” wide and on the one side I have a bank of 8 leds in a row, fairly close together. I’m sure that you could either space them out a little bit, or just add more to the bank to accommodate deeper screens. The setup height was about 32” and the leds were about 10” from the bottom of the screen. I don’t know why I didn’t try this before tonight… oh well I guess hindsight really is 20/20.
Hah yeah. I think nima in the other post may have said it best when he mentioned putting LEDs in each of the 4 corners. It’ll be similar to what you and I have, but will create even lighting especially in the corner area (which tend to loose light in my setup when angled towards the center).
Yeah I guess that’s the same basic idea. I don’t shine my light at the center though, I aim for the opposite sides. The far ends each get light this way and the center picks up enough as well. I could do the same thing with a corner setup, but I’m trying to keep the size of my tabletop as close to the screen size as possible. If I light from the sides, I only need a few extra inches on the left and right, as opposed to all the way around if I used 4 corners.
I have done some tests with a very crude setup. I used a normal IR-Lamp (100W) rather far away, as we used a mirror anyway. The surface is normal frosted glass (no acrylic), frosted side to the camera. We have good experience with this Material for front-DI , rather clear picture from the projector.
Well it’s my first post here but I figured I’d share what I believe might be a good resource for DI surfaces. The company I work for has been buying acrylic from this company: http://www.astraproducts.com They carry an acrylic called Clarex and to my knowledge, they are the only US distributer of this material. The stuff we use is the DR-IIIC which comes in a range of transmission from 45% to 85%. This material is made for back light displays apparently but we use it for a xenon flash diffuser. It’s available in various thicknesses I believe. Another product that they carry which I do not have any experience with is the Clarex Blue Ocean Screen. http://blueoceanscreen.com/ This may actually be better suited towards a DI surface as it appears to be made specifically for front and rear projection!
I just stumbled upon this site yesterday after being fairly amazed with the Microsoft Surface videos i watched this weekend but am hopefully planning on building a table in the near future. If and when I do get a chance to start experimenting, I will do some tests with these materials.
So, ¿is it good to use tracing paper (in the hand side) both as a projection screen and a diffuser? After quite a reading I reached to that (probably wrong) conclusion for my future DI table. Thx, Gonzalo.
