Wow you’re actually serious… I’ll be honest, most of us thought you were joking. The reason for that is because you seem to be going about this completely the wrong way and throwing far too much money at random things. I’d suggest you hire someone from this group and get them to help you build these displays because if you carry on like you have been doing you’ll waste a massive amount of money and get no-where. So my advice would be to slow down and research this a bit more. You can’t just buy a $30k enterprise server rack and expect it to all work smoothly, you’ll need some experts to set it up and you’re also adding a lot of work to your project.

Make sure to get some sort of USB or Firewire array to link in all these computers to this beast!  1 / display and also a video out / display as well GL

Gimme an Apple iWall…

The computer you ordered should be great… I would rec finding a bunch of newest nvidia graphics hards and put em in there (for general gpu use)

As for the sensor:

Thinsight or electronic field sensing (capsense) is not really an option for large scale… do you have general dimensions of unit you wish to accomplish? Your best bet is DI or FTIR with multiple cameras and projectors.

There is several methods you can choose to start… each with pros and cons such as: “object recognition"… or “hand tracking” can only be done with DI (Diffused Illumination) compared to FTIR, LLP or DSI.

Check out: http://nuigroup.com/forums/viewthread/1982/ to begin.

Also it helps to make a mini prototype to learn the general concepts of how the vision tracking systems work.

Try watching this video:

Then read:
http://nuigroup.com/forums/viewthread/1731/

Vlado - 04 September 2008 04:25 PM

Don’t spend more than 6000$ for one table , for that amout you can make a verry verry high quality table/display.

For the camera use the Firefly MV , it does 120fps@320x240 , and is curently the best bet for something like this.
Chose what setup you wanna build DI/FTIR/LLP/DSI.

You may want to try the Sony PS3 EYETOY.  AlexP has made a driver for it and it can get 640 x 480 at 60 frames/second 320 x 240 at 120 frames/second but for far less only 39.99 at most retailers.

I think you are going about building this the wrong way.

From your posts it seems like you want to learn and expereince things for yourself rather than getting a pre made solution.  I think building a cheap prototype first would be the best solution as you will learn so much and your design ideas will change during this prototype build phase.  After you have built a cheap prototype you will be much more efficient in building a more expensive version.  Throwing a ton of money will not produce the best multitouch screen, for example it’s probably better to design a modular system with mid-spec PCs that can be linked up rather than buying one super computer to power several screens…

Really the biggest hurdle at the moment to overcome in MultiTouch technology is the software, rather than the hardware.  You can build a multitouch table in a couple weekends but good software takes months and months to develop.

Thank you all for the replies.  Thanks for the concern Joobs.  In regards to the software, I plan on hiring a few programmers (independent) to create applications software after I have completed my first MT so they have something to test it on, but I estimated 1 month would be enough time to create few MT interface Apps like I mentioned before.  I figure $3,500 - $5,000/month for each depending on experience, complexity, and creativity.  I have to research what the average pay is for a programmer to get a more accurate figure.  I’m also not going to sell the software once it’s made (for my private use).  I’ll have the programmers sign a NDA and also have them create the programs at my office building on a closed network so they can’t carry or copy the work and sell it themselves.

iloveanime213, I love your enthusiasm! But, like Joobs said, I think you’re also going about this the wrong way.

Your expectations are a little high for something you’ve just recently gotten into. Considering Microsoft, Apple, Jeff Han, and several other companies have spent tousands/millions and years (not days, weeks, or months) in building their hardware and software, it’s not very realistic to expect to be able to meet your standards in such a short amount of time. The average programmer needs to understand the hardware aspects of the design in order to properly configure something as complex as you’re expecting. A central server may not even be the best approach. That is something that needs careful understanding and outlining beforehand in order to know.

We’re not trying to discourage you or knock you down, but just letting you know the reality behind building such devices. Money does not mean better nor faster. Experience is really the key factor in all of this and that’s something that can’t be gained in such a short amount of time.

I wish you luck!

Thank you for your input and quick reply everyone.  I don’t have any expectations cerupcat.  I’m just guessing the figures from what I know so far.  It takes as long as it takes and I’m one of those people who will never be satisfied unless I see and do it myself.  Changes and unknown factors will come up, but I’m going to go as far as I can with the budget and time limit I gave myself.  I’ll make refinements here and there which will be ongoing, but nowhere near the time I’ll spend during this project.  Microsoft, Apple, Jeff Jan, and all the other companies must have spent millions and years, but because they did (short pause*) I don’t.  Most of what they’ve done that took years of research and development is accessible now as a tested and working product/application.  Regardless of what I or anyone else believes in completing this project within the time frame I gave myself, it all comes down to doing it.

In regards to the standards and time factor you mentioned, all I see that needs to get done is coding, testing, and refining (repeat cycle until complete).

Ultimately, I’ll understand more in depth about the fundamentals, structures, methods, and all around understanding of MTs.  This is my primary goal.  Giving me new data to play with and incorporating it to benefit my life.

This will be an enjoyable experience for me.

I’m going to do what most of you have done for past/completed projects posted on this forum so I get better mass and understanding.  I’ll gain a progressive perspective which most of you have acquired during your trial and errors.

Giving me new data to play with and incorporating it to benefit my life.

This is why we have community We are always open to new questions you may have.

Next step tho is clearly state your goals being as concrete as possible, for example:

I want 10 foot x 4 foot DI touch wall driven by one workstation with 2xXeons @2.6 Ghz and 8GB RAM… 4 cameras/ 2 projectors… running Touchlib and Native C++ applications…

If possible get Google’s Sketchup and experiment with some 3D sketches… or maybe even someone in community could draw something up for you if you describe your needs fully.

Thanks for the support Nuiman.

amazing idea to use a multitouch interface to control all systems around your house! if this works, in some months or years ^^, it unbelievable, bill gates will be envious :D

Im an industrial designer, post your needs and ill try to come up with a concept on how can it be done.

iloveanime213 - 04 September 2008 04:54 PM

Ok. I wasn’t thinking I just ordered the Dell poweredge 2410 for the rack. I paid $27,739.00 for hardware I don’t need and for performance I could match with $5000.

I corrected your typos for you…

Hello again.  It’s near time to place the order for the dell server.  There was no real objections in regards to the configurations I will be ordering so I’ll be placing the order in a few hours online.  I’ve done some searches regarding the camera.  I went to Circuit City and bought the PS3 camera thanks to the post Alex.  I was about to order the Dragonfly express and the Firefly MV today, but I read on a post here on NUI that the Dragonfly express IR filter is glued on the lens. 

I’m asking if the filter can still be removed without damaging the Camera.
Will this camera work well with the 200FPS?
If so I’ll make the order today.

I’ll be setting up the acrylic and the ir led ribbon today.

I have a 1/2 inch clear 37x21.5 42inch diagonal acrylic.  I’ll use this to mess around and test the ftir and silicone rubber today with the the ps3 camera.

The finished product acrylic will be 4 feet x 8 feet 1.5 or 1.85 inchs thick.

I went and returned the VPL-VW200 and got a VPL-VW60 instead for the projector.

See you all soon.  I’ll record the results and upload it on youtube.

I wouldn’t recommend a server, since you want a lot of graphical performance a BOXX workstation is better for you.

APEXX 8 32 cores 128GB ram
http://boxxtech.com/Products/APEXX/apexx8_overview.asp

APEXX 4 16 cores
http://boxxtech.com/Products/APEXX/apexx8_overview.asp

price is undisclosed, youll have to call, ill guess something like 30k for the 16 core model.  but you can find very good systems at BOXX ( 8 cores ) for about 5 - 9k