Possible Touchscreen bug?

[PabloMartin]

Please put atention on this.
I stay 99% secure, this no is a HPTM2 BUG!

I think is a bug on windows 7 multitouch core:

https://connect.microsoft.com/VisualStu ... plications#





Please repeat the test if you want, you can do this:
Just need please make this test on your multitouch tablet or device and confirm if this happen on your machine.
This test is only for multitouch devices !!

Basically is a simple test, and consist in open MICROSOFT PAINT , you just should and make very faster SINGLE touchs.

Here a video i created where you can see how i perform the test and how ghost false touchs are generated creating lines.

YouTube - BUG VIDEO 3
http://www.youtube.com/watch?v=ctLLXHq1NZ0

The bug are that lines.

This is a big problematic situation due this make windows 7 multitouch useless.

Here a simple example showing how this bug can cause problems to aplication.
In this video you can see 3 multitouch buttons, you can se when make faster taps on the side rectangles, the rectangle in the middle is activated.

YouTube - BUG-VIDEO-1
http://www.youtube.com/watch?v=jLIGYS0YfUw

What i showed in this video is a multitouch aplication, simple example created by me, if you want test just download from here:
https://connect.microsoft.com/Visual...me-aplications
Expand the post, and download the zip, inside you have the source code and the exe inside bin folder.
Need net framework 4 installed for run this simple appz, if you not have isntalled no problem framework start download in automatic way.
Please repeat this test if possible, if you see same bug and can upload video to youtube i apreciate that.

Imagine now that aplication is a game, or just a simple virtual piano..........this false touchs make appz non reliable.


I really apreciate if somebody can repeat this and confirm the results, i ake all this test over my HP TM2
If this only happen on my tm2, is a bug of wacom system on tm2,, if anyother can reproduce this on any other multitouch device, that indcate and confirm is a windows 7 bug make multitouch useless, and unreliable.

Really apreciate your help, for now, just please make that simple ms paint test, and if can the secodn test.
Remember make faster SINGLE taps, like i showed on the video.
Is this is confirmed like a bug, please Oliver suscribe to microsoft connect, and start push with me for request fix, and update that post please.
https://connect.microsoft.com/VisualStu ... plications#

Best regards.

Pablo Martin

[nay-seven]

first , i really need to make very fast touch to reproduce your bug ( remember it's also depends of your laptop or screen latency ) and I'm not sure we need to use multitouch this way here..

you say " windows 7 bug make multitouch useless, and unreliable." , false too.. cause many people use multitouch screen with Usine and are happy with..

and last you ask about subscribe to microsoft connect....and I've never heard that microsoft take care and listen to his users since now..so i don't think it will began now...

microsoft multitouch will be improved , but you just have to be patient..

[23fx23]

those who can should try to debug if it's a wacom pb or W7 api.. hope it's wacom specific...

[PabloMartin]

nay-seven what machine and what multitouch device you use for reproduce this bug?
Please can go to connect, download teh attached simple appaz, and reproduce the error of the second video?


This bug exist, and affect specifically to music appz, i just builded a multitouch pad, like MPD AKAI, i use that for percusion.
The 2 bugs affectet multitoiuch aplication making the appz non usable., same vor virtual keyboard piano and other.
About MS connect realy work,......... windows first public release is released arround 1 year 6 months ago........is mothe thatn enought time for wait.
If we no start push togheter this problem take long long time to be fixed.
This bug cant be minimized......

[PabloMartin]

waiting form nay-seven repply.
Need knwo what machine he use for test.

[PabloMartin]

Oliver, im working now with multitouch appz, i have a big project incoming : )

I just no want come here to make noise, i back to your forum due your aplications is multitouch and probably can find some guys with multtouch devices interestedin test this.

Please if you have TX2 tablet, or any tablet multitouch repeat the test, just with ms paint, or download the multitouch example i post on https://connect.microsoft.com/VisualStu ... plications#

At the end of post you can fin source code and inside, you canisnide that zip exe aplication.
Anyway this test can be reproduced just usning MS PAINT.

MS paint support multitouch, just open and make very faster cosnecutive single taps.
If you see only pints, the bug not happen on your tablet, if you see lines, the bug happen on your tablet and that confirm happen at windows 7 level.

YouTube - BUG VIDEO 3
http://www.youtube.com/watch?v=ctLLXHq1NZ0

I make all my test on HP TM2 tablet, need see if this happen on my tablet only, or f happen with all tablets.............is a bug on windows 7 multitouch core API.

Aditional to this, on MS connect i showed second bug.

Please can repeat the test on your tablet, inform what model is , and tellme the results.

Really apreciated !

[PabloMartin]

Oh about second bug, if not understadn what i comment on MS connect letmeknow.

Hoppe you can colaborate with this nad for be clear, my appz is multtouch, and MS PAINT support multitouch.
I see teh bug on MS PAINT and on my aplication, so this is a bug of hpm tm2 or is windows 7.
I have soem guys makeing some test right now, i back here to make public and share this info soon for see if the problem is windows 7 or is hp tm2.

Please oliver if you have some tablet multitouch (no hp tm2, just any other model) tellmewhat is your result.

[PabloMartin]

Update:
I have more people conffirmating this happen under under HP TM2 tablet.

Oliver can you repeat this 2 test, here the exe needed:
http://www.pablomartin.com/BugTest.zip (need installed netframework4) if you not have installed when you run appz, should start download automatically from web.

INSIDE YOU HAVE THE 2 MULTITOUCH APPZ, i used for do the test and report the troubles i already rpeorted on
https://connect.microsoft.com/VisualStu ... plications#

Please if you have different tablet of model i have can repeat the test and tellme your results?
Please dont do this test on single touch devices, and do only on windows 7.
Single touch devices fail n this test and create confusion, only for single devices.

Best regards.


Really apreciated, know the results.

[senso]

This only happen with work with touchenter events, no with touchdown.
If i put more buttons or objets in the middle, and make faster touch in the rectangles at the side, all objects in the middle act like if somebody touch.
Please this need some inmediatly fix !!
I suspect is a trouble in how wpf handle multitouch events.

this pb is well known : ghost points describe in the following image:



also see the video http://sensomusic.com/forums/uploads.php?file=gohst.wmv
It's a touch screen issue, not W7. Most of cheap touch-screen have this pb.

After a test of your apps : you're right, they don't work properly, especially, don't receive the TOUCHEVENTF_DOWN event.
but If I make a small patch with Usine like



two play buttons connected to 2 samplers, it works very well, except the overall latency which is almost like the mouse.

So I think that your pb comes from your layers 1 and 3?

Maybe you can make the test yourself with Usine to see where it comes from?

PS:Personally I spent 6 month to create a good layer 3 in Usine and I know it's an hard job.

[runagate]

It's probably Wacom-specific. I have a Wacom Bamboo Touch tablet. I basically don't bother using it.
Not least of which because there's 3 entirely separate configuration applets in my installation under Win7x64.
Spending time configuring it to your particular needs is worthwhile, though.

Also, a Wacom tablet is, unfortunately for me, seen as a mouse-type HCI and therefore can't really be used as a MIDI
controller without intervening software to hijack and translate its input.

[PabloMartin]

Oliver what coputer you use, i remember you have tx2z, you have last driver for 4 points?
You have HP TM2?

i already start talk about this using email, thsi is really technical issue and now ant create confusion to your users.


One more thing, you wrote:


After a test of your apps : you're right, they don't work properly, especially, don't receive the TOUCHEVENTF_DOWN event.
but If I make a small patch with Usine like


I have a confirmation form MS team, nothing is wrong on the code, you already run my examples and see in the 2 exampels the bug?
Again, what machien use for test?

Please can go to microsoft connect and update post, and explain you can reproduce the bug using the code i provide?
This probably cna be a bug on netframework 4, if have soem time please update on connecta nd explain youc an reproduce this bug and put the model of multitouch device,
really apreciate that.


Best regards!

[senso]

hello Pablo,
Thanks for your interesting post.

From my point of view, your bug is not confirmed.
The Multitouch under Usine is separated into 3 layers:

1) Hardware and driver of your touch screen,
2) Windows7
3) Usine multitouch management and dispatcher

To be more precise,

1) the hardware sends basic information about touch on the screen to the driver. The driver has to clean and convert them into proper TOUCH messages to Windows 7. Most of drivers and hardwares I have tested have a very poor multitouch implementation. The generic W7 driver is not fantastic but functional... The ntrig driver works fine.

2) Windows 7 get messages from the touch screen driver and dispatch them to the proper window or application. It's probably the easiest part of the chain?

3) Usine Get WMTOUCH messages from W7 and dispatch them to the proper control, via an invisible panel which covers all the screen. This part is very, very, very, complex and you have to know at any time which point is pressed, moved or released.
I think that Usine has an unique technology, probably in advance compare to other W7 softs (sorry to congrats myself...)

So in you case it's impossible to know where is the problem. It can be on any layer and personally, as the W7 layer is the simplest, I'm not sure that the pb comes from this one.
Also, it doesn't happen with Usine and all the (poor) hardwares I have tested.

Moreover, don't forget that if your application doesn't have the layer N°3, it will not be multitouch but only mono.So by default apps are NOT multitouch but single.

so?

But you're right, the overall touch management could be faster to handle quick & repeated pressures.

[PabloMartin]

Oliver, not understand why paste agian same repply after my last coment.
Can please know what is the multitouch tablet you used for test?
Please can confirm if you already test my sample example, and if you can reproduce the bug of my appz in your tablet?
Best regards.

Pablo

[senso]

Hello pablo,
Your post has been renamed into something smoother; Sorry for that but it was ambiguous for Usiner's.


(our democratic model)

I'm not sure I can help you anymore.
Without a complex post processing your apps will not be multitouch capable. And as I use the same W7 API and it works fine with Usine something can be optimized in your code?

I have a packard-bell viseo touch screen, and I use to have an hp tx2; Usine works fine on both.

I hope you'll find a solution;
If not, why don't use Usine? (French humor)

senso+++

[PabloMartin]

Yes, is ok, rename post is the correct, probably im wrong and the title of my post can cause some of your users think this problem affect usine.
My specific question is:
you test my appz example, on tx2 or in a packard bell?
about what i showed on this video:

http://www.youtube.com/watch?v=ctLLXHq1NZ0
You can reprodyuce thissame situation in your multtouch dveices?


If just can confirm this, you helpme a lot !! : ) and i really apreicate that.
I have reports from guys with dell xt and tx2z reporting tehy cna t reproduce that MS paint situation, and they downlaoded the aplication examples, and cant reproduce the situation i showed on the videos, all that makeme think this is a problem on HP TM2, your confirmation really helpme : )
Thankyou so much for your time.
Best regards.

Pablo !

[senso]

On my viseo T200 it's almost the same, especially with hi speed contacts.
But normal? If the contact speed is faster than the touch latency it's logical to have unpredictable results?

The overall latency could be improved, but it's probably more an hardware pb than W7?

From my point of view it's not a specific HP TM2 issue.

senso+++

[jean-marc duchenne]

Hi,
Just to say that the two fingers fast test works perfectly on my Dell XT2, it makes only small dots in Paint.
On the other hand, in Usine the objects are not able to follow the same speed.
If I have a dual XY pad and I do the same movements with two fingers, I can see the little crosshair pointer that follows my fingers, but the two circles go to the good position one time for five.
(juste pour dire que le test d'alternance rapide de deux doigts fonctionne parfaitement sur mon Dell XT2.
Par contre, dans Usine les objets ne peuvent suivre une telle vitesse. Dans le même test avec un pad à deux canaux, les deux pastilles sautent à la bonne position une fois sur cinq. Mais peut-être est-ce uniquement un problème de raffraichissement des graphismes ?).

[senso]

yes probably a graphic refresh pb?

[PabloMartin]

Dell XT2, AND HP TX, not generate the toruble i explained.
Have confirmation form multiple users now.

Aprently HP tm2 digitizer no is really super good like i think, probably new drivers improve the situation.

[PabloMartin]

Today i have confirmation from DELL XT, ANDHP TX2z users, around 12 users confirm they cna reprodude this problem on that devices.
This confirm TM2, have a innacurate response to faster taps and other troubles.
Thankyou for your help Oliver.