Surface 2 Sensitivity problems

Question

It's not the ambient light issue here....I think it's sensitivity problem:

the surface 2 will raise many noisy touch spots when I touch the screen with only one finger, although the other parts of may hand is not very far from the surface. But we know such kind of issue seldom happens on the old surface!

what could I do now? my application is just not usable at all...

Answer 1

If you are sure there is no lighting issue and your screen is green when running the light optimizer, then nothing much can be done except ignoring from code all touch wich are not fingers.

Then you are at this time in a kind of normal touch screen environement.

regarsd

serge

---

Your knowledge is enhanced by that of others.

Answer 2

Hello,

I agree to Serge, I had almost the same problem with one of my applications. After adding a check to only accept Finger and Tag placed on the table the stability increased signaficantly.

protected override void OnPreviewTouchDown(TouchEventArgs e)
 {
 bool isFinger = e.TouchDevice.GetIsFingerRecognized();
 bool isTag = e.TouchDevice.GetIsTagRecognized();
 if (isFingerRecognized == false && isTagRecognized == false)
 {
 e.Handled = true;
 return;
 }
 base.OnPreviewTouchDown(e);
 }

With the above code you can quickly determine if the input is a finger or tag.

Hope this helps you.

---

Roy Janssen - Semper IT Services (Please mark an answer as "answered" if it does help you!)

Answer 3

Thank you all.

I solved the problem by your suggestion...

Answer 4

We're using this work-around to improve the touch input on our own controls, but it does not solve the problem for 3rd party controls we're hosting inside of our application (ie the Browser).  Also, the touch input is still getting processed by the driver, so we still get the animated touch spots all over the place from users' hands hovering over the table.  Touch is basically useless in Windows mode as well.  I think a lower level solution is going to be needed to disable blob input, or at least adjust its sensitivity so it only registers actual touches.  Is this something calibration might fix?  (We still don't have the calibration board).

Answer 5

@Robert, calibratin will not solve that. This is part of the technologie.

---

Your knowledge is enhanced by that of others.

Answer 6

Apparently the technology is really bad then.  The Surface 1 did not have this problem, from my recollections of using them at TechEd and PDC in 2008.  We bought two SUR40s for R&D, and have found them to be nearly unusable, particularly for users with light skin tone.

Answer 7

For the code snippet provided, does this have to be added once per application, or at every touch handler?

We definitely noticed different experiences depending on what shirts people are wearing. Horrible noise from forearms. The only consistent way to use was with awkward "stabbing" with fingers.

Answer 8

It only has to be added once, on the main window.

BTW, calibration did dramatically improve the problem for me.  The SUR40 is now pretty usable, and I can even double click icons on the Windows desktop.  I still get excessive touch events from wrists and shirtsleeves but it is much better than before.

Answer 9

We also use the above code to filter out Blobs and all other noise. Unfortunately we get some strange behaviour. The bottom 30-40 pixel rows do not respond to normal touch anymore. So without using IsFingerRecognized the upper part of the screen is more or less unusable, with IsFingerRecognized the bottom part is unusable.

We already tried calibration, but our SUR40 got even more sensitive to the environment light we had here. We have a pretty dark office, and now are only able to use the SUR40 when turning off ALL lightning equipment. It appears like the calibration mechanism does not any environmental influeces into account, as it only scans calibrates with board.

Due to all the questions asked here and in other threads about tag recognition, 'blob' performance and performance in little light environment I am wondering can we expect soms updates to the firmware/algoritm of the PixelSense system in the near future that might improve the way it works?