Recently in Multitouch Category

Look at this, isn't it wonderful?!?!

Yessss!!! It looks very very nice!
Anyway, of course it is still to be completed, I need to prepare all the braces for the circuit boards, the bracket for the webcam and, of course the infrared setup and the front mask.

So, let's start with this second part of the construction notes!

After bending all the parts, I had to refine them because, as I said, I abounded with measures because I was not knowing how much material I would loose with the bend. In this phase I discovered how useful is the circular saw for the Dremel instead of the electric hacksaw that I used to cut the parts before. The advantage of the circular saw is that it can make very straight cuts and you will have better result.

So, basically I cut again quite all the sides of all pieces of plexiglas refining them to the right size, then I started to try to put all together and see what was needing to be accomodated.

Once I got the situation where all pieces were complying well together I started with the holes for the screws. As I said in the previous post, I decided for the most difficult assembly solution: screws on the edge of the plexiglas.


There was a lot of difficulties: the first one is to keep the drill parallel to the edge all along the hole just using the hands, the second one may be to drill the hole without melt or break the polystirene ecc. Other than these issues, I had also a lot of oblique sides that means oblique holes!

I have to admid that it was really taugh but, luckily I had good results. I used a 1,5 mm drill bit for 2 mm screws. I made a lot of tests on spare polistyrene pieces to understand how I could drill it without melting it. At the end of the game I decided to don't care about the melting plastic!

I just regulated the Dremel to the slowest speed and made all the holes keeping the Dremel in position just by hand.

As you surely remember, three weeks ago I posted about my first test with an LCD panel for multitouch applications. I saw that it had enough chances to work out but I was so scared about the fact that the bare LCD panel is extremely delicate. So, as I was saying at the end of the post, I started to think, then work, to a case for this multitouch display.

Of course I thought about a case that should also look nicely, what's better than a plexiglas one?

So, after some weekends of work, this is how the case is looking today:

I designed it trying to keep its size as small as possible, so I went for a cut pyramid shape. The pyramid is (or better, it should) be deep enough to hold the webcam in a position that, using the wide angle optics, is far enough to capture the entire screen view.

I can honestly say that probably I choose a very complex shape to build, infact it took me several days to bring it to this stage. For this reason I'll talk about how I proceed and I think that this will fill at least 2-3 posts. Let's start!

First of all, reading also several modding forums, I undestood that plexiglas is quite difficult to work so I choose a similar material, the transparent polystyrene. It seems more tollerant to mistakes even if it is more easy to melt while working with electrical tools like drill or saw.

Anyway, all begins with a good drawing scheme of the whole case and the pieces that needs to be cut. I paid big attention in this trying to consider every factor like angles because there are very few square angles in my design, and exact polystyrene sheet thickness.

In addition, there are some bends in my design and I was not knowing how much material the bend was taking, so, basically, I was abounding with measures (3-5 millimeters more) impacted by bends and I considered to cut again the parts after bending them.

I used the electric hacksaw for cutting the polystyrene sheet but, keep in mind that it tend to melt very easily. So, to avoid that, I used a dimmer to regulate at very low speed the saw. Let's say a 2 saw movements per second under load. This useful also to be very accurate in cutting following a straight line.

I know, every single sencond that you spend working this kind of material you are tempted to take away the protective film but, I can say that it is very important to keep it on until the very end. In bad cases keep it in place with adhesive tape but don't remove it!


Then I cut some pieces of wood to use as mold for bending the pieces of polystyrene. Then I used those pieces to build a complete mold for every piece. It was a little bit of work more but absolutely essential to have a precise bent shape.


Once I had the molds, I started to heat the polystyrene. I used a heat gun, one of that tools that are useful to remove the paint. Basically it is a sorta like of hairdryer with a high temperature airflow.

I used it very gently, moving it very often, slowly but constantly to avoid any break in the plastic sheet. Despite this, one of the pieces broke and it was the last one (of course)!

The polystirene sheet that I used was 5 millimeters thick and it took about 2-3 minutes with the heat gun to soften enough.



When it was ready I gently pressed it on the mold and I used other wood sheet and small morsel to keep it in place. The polystyrene needs another 3-5 minutes to cool down enough to be hard again and, with the morsels it could take all the time it needs.


I paid particular attention to the angles. In order to have clear ones I used pieces of wood to press the polystyrend exactly on the edge.

I can say that the more you heat up evenly the polystyrene sheet, the best bending results you will have.


The last step that I will show in this post is regarding the piece for the back and bottom of the case. It was the most difficult one because it had two bends. Of course I had to create an ad-hoc mold just for it.

I'll continue to illustrate the case building process in the next posts, stay tuned!

Quite a long time has passed since I did something more for my multitouch display. My last post on this topic was on November!

Unfortunately I did not any progress basically because I don't own a projector, so, everytime I should ask to someone to lend me a projector then I have limited time to do some tests with it. This is not exactly comfortable especially because I have very limited spare time and quite never in consecutive days.

So, in the last months, I was thinking more often that it would be nice to achieve the goal to use an LCD instead of a projected screen.....and I started to do it!

To be honest, I thought about using an LCD for a multitouch display since the very first moment. Infact, when I bought the webcam, the Fire-i from Unibrain, I bought also a wide angle, non IR coated lenses for the cam. This was exactly because I thought that, if all things worked out I could try with LCD and use the wide angle lenses to have a short distance beetween the camera and the display itself.

Ok, I was a little bit optimist at the first time but....at the end of the game it was definitively good!

Some days ago, at work, I was speaking with a great new colleague, Pietro, who read this blog and discovered all the works that I've done at home. I was very pleased to talk with him illustrating all the hints and techniques used with the multitouch display. Then I said "yes but I don't have a projector, actually I'm thinking about to buy an old used 14" LCD on eBay to sacrifce for the experiment". He had a great answer: "I have one abandoned in my cellar, I may give it to you"!!!

He was so gentle to give me this great LCD monitor for free!

So, last weekend, I started to disassemble the monitor to get out the bare LCD panel and test it with the infrared camera and the IR LEDs.

I was really surprised of how many pieces there were in the monitor! Kilos of screws, panels, the speakers in the basement, plastic covers and supports...

It took me more than two hours to fully disassemble it, especially because I didn't wanted to break something so I was very gentle trying to have a safe and, hopefully reversible, disassembling. At the end I had the bare LCD monitor panel and a little fear that this was a non-return point because I already forgot how to reassemble the monitor :o)

Of course the first step was to carefully inspect the backlight complex and its filters to understand what needed I needed to remove in order to have a monitor transparent to IR but not so transparent to the visible light (I need the backlight anyway to see the monitor image!).

I think I was really lucky because I needed to remove only the reflective white sheet that was in the back of the backlight, I kept the Fresnel lens and the diffuser. If I will have time I'll post some pictures of those layers for those of you that wants to understand exactly what I'm talking about.

Then I immediately put the LCD panel on the desk (keeping it vertical by two Fonera boxes :o). I reconnected all the electronics, double-checking the connections in order to be sure not to burn out the monitor on the first test. But, very soon, I realized that, even if the panel worked out, I could have a serious problem.

The problem is in the Flat Flex Cable (FFC) that is connecting the board on the right side of the display with the board on the bottom side. It is really too short and requires that both boards sits in the back of the panel, without the possibility to bend the boards out to free the view of the entire display surface from the back.

Anyway, I didn't care at that moment and switched on the monitor....all ok, it was still working!
Then, second test, I put the IR webcam on the back of the LCD and tested the IR remote of the TV: wow, I saw its IR light!

In addition the LCD didn't lost too much brightness with the removal of the reflective white sheet and the image was still clearly visible.

My last test was with a couple of spare Osram SFH485P IR LEDs that I had from my first multitouch: all ok, I was able to see very very clearly them by the webcam through the LCD display and I was able also to see the IR light reflected by my fingers. I took this as the definitive evidence that this setup could work greatly.

So, scared by the risk to break the very thin and delicate LCD panel, I put it apart in a safe place. I started to think about a good box to hold all the things together safely and steadily.

While brainstorming for the box, I found a great site that was selling spare parts for video projector and.....YES! On that magic site I found the right FFC extension! So the problem was solved!

I hope to post some news on this exciting new adventure very soon!

Stay tuned!

In the last month I was very busy with my work so I didn't find enough spare time to dedicate to my multitouch. That's why I didn't posted anything, because there were no news to post about.

In this week finally I arranged to find a little bit of time at evening so, paragraph after paragraph I completed what I can say the "alpha" version of my multitouch display DIY guide!

Compared to the "work in progress" version that hopefully you saw after my last post, I completed the section two and basically wrote down from scratch the section three where I describe the construction of the FTIR multitouch itself.

As usual, I hope that it may be useful to you to face all the issues and doubt you can find along this building path and, why not, I hope that it gives all the common informations that probably were still missing in this site and that some of you were asking me in the comments or by email.

As I was saying in the last post, this guide is not intended as the ultimate reference for the construction of those kind of device. Honestly it is aiming to be at least a complete walkthrough for those who want to understand what a multitouch display is and how an FTIR one may be built at home, with very common materials and tools.

Please feel free to comment this post or the guide page to point out parts that are lacking informations or that are not so understandable so I can improve the guide and, hopefully, your experience in reading it.

I know that my Multitouch Display page has been left alone since I wrote it, with no updates even while I was anyway going on with my multitouch work and tests. In addition, when I write something on the multitouch matter, I'm assuming that everybody knows a lot of things, quite all I've learnt on it until now and I know that is not so useful. On the other side some of you rightly wrote me saying "why you don't give further and detailed informations on the construction, the components used etc.?".

I know that there are already tons of informations on the internet on this subject and there are some very good information depots as well. One of the best is the NUI Group wiki and, of course, the blogs of his users (thanks to DzHiBaS that aggregated them). But, in my opinion, there is still the need for something that may be a more or less complete walkthrough in the theory and the construction of those displays, like it is the how-to guide by Gravano that, unfortunately, was not updated lately with all the great improvements he made.

For all those reasons early this week I started to update the Multitouch Display page...and not only! While writing I reealized  all those reasond and I decided to try to explain more clearly and starting from the beginning the whole story, aiming to reach the goal to put up together a quite comprehensive homemade multitouch guide.

Obviously this is not aiming to be "the ultimate guide" because I played just with FTIR setup and, in addition, my experience is limited just to my project and to the solutions and materials I used. But, anyway, I tried to put together all I know, I experienced and I read on this subject so I can surely say that, hopefully, this will be the best I can do.

Of course this is still a work in progress: there are already three pages to look at but there are some sections that  still need to be written. Anyway, I really hope that at least anybody who is approaching for the first time this fabulous world may enjoy it and find it useful for its multitouch work.

I'll really appreciate any comments and suggestions you would post to enrich this guide and, why not, to correct any mistake in it. It could be a good tools for everybody or at least a good starting point.

Some days ago I was finally able to make a complete test of my new compliant surface made of the silicone rubber layer and the sketch paper. As usual, I can say that this was also a partial success.

The layers of silicone and paper gave great sensibility to my FTIR multitouch setup. Actually a very light touch is enough to trigger the blob recogniction, even the passage of a paper tape roll is detected!

As I was saying, there are anyway something to work on. The silicone rubber is still a little bit sticky and this cause the sketch paper to adhere to it for a little time after the finger touch passage. The result is a visible tray left after the finger movement. The software I used until now, Touchlib from NUI Group, is recognizing those trays as finger touches as well so that moving a finger result to several touches moving in a tray.

This bad effect is not impacting on simple application like Smoke but, unfortunately, it hardly impact on applications like the one used to play with photos. With this app, moving a photo result in enlarging it because of the multiple touches detected in the tray of the real finger touch. Anyway, you'll have a better idea looking at the video I shooted during the last test:



Actually, to be honest, I don't have big ideas on how to overcome this issue. I'm quite convinced that all those problems are coming from the type of silicone rubber I used, the Prochima Crystal Rubber, that is a good silicone rubber but probably not for this kind of (strange) applications. The other guys from NUI Group had infact better results with the silicone rubber layer without this kind of issues. My only workaround, for now, will be to buy and try other kind of papers and, probably later on, to try a different kind of silicone rubber itself.

But, anyway, actually I'm moving my interests also on software for multitouch. What I saw until now is that there's a lot of pieces around, most of them built using libraries or scriptable frameworks, few of them are compiled, mostly commercial and closed source. Unfortunately it seems to me that it is a lot of work done but in a fragmented way: there is no coordination and every app is a monolithic, separate app so that you cannot have a "multitouch system" but rather some multitouch-capable applications launched on a computer.

Actually I'm thinking about how to approach this matter, also because hopefully sometime in the near future I'll have a good multitouch screen and I will like to use it without the need for a mouse, keyboard or other "old" computer equipment.

In this brainstorming phase, every comment and suggestion is really welcome, feel free to contact me also privately.

When I tested for the first time my homemade multitouch display setup, I was too excited at the beginning to complain about the lack of sensibility. I had to press hard on the rear projection film to let the webcam detects the "blobs" of the infrared light bounced out of the acrylic. Later I understood this and I decided that it was the time to get a compliant surface made of silicone rubber as I was reading on the NUI Group forums.

I had a lot of doubts regarding the preparation because, of course, I never used this kind of material so I didn't had a clue of how to work with it. But the need was so strong so I decided to try at least once. Half a Kg of silicone rubber costed me 40 euro so I decided that this was my maximum budget to give a chance. But of course I didn't want to waste it!

Anyway, at the end of the game, I have to say that it was far more simple than I was expecting. I have to say also that I spent some time thinking about all aspects of the preparation process in order to avoid to waste the only chance I had for stupid mistakes.

I had great results regarding the silicone rubber layer even if later I had the surprise that, in my case, this layer alone with the rear projection surface won't work: I had to add an additional layer made of sketch paper to have the great sensibility that I was looking for.

I intentionally filmed all the preparation process thinking to make a video that may be helpful to someone else that is in the same situation I was, so, now that I saw the good effects of the silicone layer and after some (too many) hours of video editing, I can finally say:
HABEMUS VIDEO!



I really hope that you enjoy the video and that it may solve out any doubts you may have about the preparation of the silicone layer. Feel free to comment this post to ask for additional aspects or questions that I forgot to mention in the video or in the past posts. I would like also to know if you're going to prepare such compliant surface and, if you already did it, if it is working out for you even without the additional sketch paper layer.
Enjoy!

After I successfully tested the sketch paper I promoted the silicone rubber layer to the status of "useful". So I proceeded to acquire from the camcorder the video of the silicone layer preparation that I shooted two weeks ago.

Actually the video is nearly two hours long, so I will take some time to me to edit it and shorten to a more reasonable lenght but without loosing important informations. So, if you're going to prepare a silicone rubber layer by your own and you have particular doubts or questions, feel free to comment in this blog and to ask, so I can try to answer your questions and to keep in the video not only the informations that were important to me.

I posted here just those two images for now so you can get a rough idea about how I made it. I think I can post the video in a week, hopefully less. In the meantime I really suggest you to take a look at the compliant surface topic in the NUI Group forum.

Finally I was able to use some spare time to buy and try the tracing paper, to be honest, I din't post the tests done with it because it didn't worked out as expected! Sorry for being away so much time!

Of course there's another reason why I'm writing this post, I could not be so motivated by a tracing paper flop!

Last sunday, after I wrote the last post in which I said that I didn't have the tracing paper to test, a friend of us came home. Before she came, I was speaking with her on the phone and, knowing that she's an artist and she's always experimenting a lot of materials, I was asking her for a small piece of tracing paper. She brought an entire roll of sketch paper saying that she didn't had the tracing paper but this was so similar to it and I could give a try. I really have to say: THANKS CAMILLA!




If you look at the video you'll understand everything. Infact the tracing paper is not working as expected. I think that this is due to the fact that the tracing paper is too smooth and bright and it just sticks on the silicone rubber like the rear projection was doing alone. This partially sets off FTIR on the whole contact surface and finger touches are not so bright and recognizable.

The sketch paper is instead less smooth, more porous and even if it sticks a little bit on the silicone rubber, it detaches from it very quickly. As result it gives very good and usable blobs with just some persistance but anyway really a good behaviour compared to what I got from the tracing paper.

Luckily this is happening also with the projection film on it and not only! With the whole setup that is: silicone rubber, the sketch paper on it and the projection film on top I have constant and bright blobs even when I quickly move the fingers around. With just the bare silicon rubber I had great blobs that become very faded and less recognizable when moving the fingers.

I'm pretty happy of this results, hopefully this is the solution to the silicone rubber failure. Now I feel really more confortable with my multitouch table and I'm ready to go over. Even if the sketch paper roll width is smaller than my plexiglass I already covered the whole display surface with it, using two stripes. I hope that the background remove feature of the multitouch software may overcome the possible issue of the paper junction line. On top of all, I really hope to find some time during this week because I don't have enough time today to test the whole screen setup! But, don't worry I have anyway other news to post so I will not leave this blog alone for another whole week!
Stay tuned!

Thanks to the first great suggestions from Cirrus and Cerupcat from NUI Group, this morning I tried an early workaround to solve my problems with the silicone rubber. Some indications came out also from Gravano and, actually, we're quite sure that the silicone rubber failure is most probably due to its excessive stickiness.

This silicone rubber stikiness is the cause of the high cohesiveness of the rear projection surface on it and this frustrate the whole screen surface setting off the FTIR effect without any finger touch. Most probably this is why any other touch is not detected and why it seems there is low IR light: because it exit the plexiglas on the whole surface and not only in the touch spots.

To confirm those assumptions, as Cirrus suggested, I tried to place a paper sheet on the silicone rubber and tested it with and without the rear projection film on it. Cirrus rightly suggested to use tracing paper but, unfortunately, I don't have it at home and today is sunday, so I gave a try with normal A4 printing paper. Here are some pictures:

The above picture is the IR camera view that shows good blobs of my fingers on the bare silicone rubber layer. This is a little bit more than what I should have also with the rear projection film on it. But unfortunately this is (still) not happening.

The above picture is made placing a paper sheet on the silicone rubber layer. It is sticking anyway but really less than the rear projection film. The stiked paper makes a noisy pattern and just around the three fingers touches.

Last picture is with the rear projection film on top, so the paper is laying in beetween the silicone rubber and the projection film. As you mai notice, the paper is adhering not only around the three fingers touch but in the large area where it is covered by the projection film, this is because the (even small) weight of the projection film is pressing it on the silicone.

This result is really really more relaxing than what I got on the first silicone rubber test. The blobs are slightly visible using a sheet of paper beetween the two layers! This means, of course, that tomorrow I will run to the shop to buy a piece of tracing paper big enough to cover the whole screen. The tracing paper should work better than the normal one, so I hope to get brighter blobs.

I don't want to consider this issue closed and run too further but, if the tracing paper will work, next step will be to build an holding structure for the screen, so I can avoid to use the kitchen chairs to hold it and switch to a pretty stable and fine-tunable setup.