Touch Screen Test Request


Vitel

Active Member
Joined
May 16, 2009
Messages
560
Website
vminko.org
Devs please do a similar test on Pandora: http://www.youtube.com/watch?v=_T4S70ygXwY
I thinks it's interesting for lots of users.
 
Holy hell, I haven't finished watching the video yet, but curvature at the edges? Are they serious? I mean... can they really not just extend the size of the touch sensor beyond the visible size of the screen...

Also.

Confirmed: The Motorola Droid touch sensor sucks. I got this impression when stabbing links uselessly in the web browser (the iPhone schools it), but it's nice to see it demonstrated here.

Also, a little more on topic. The Pandora screen is resistive, therefore it doesn't need to do any computation to figure out where the centre point of your finger is, in fact it doesn't do any such computation, which is why you need to generally use the screen with a fingernail or a stylus to get a good level of accuracy.

Of course, in the right conditions resistive screens are far more accurate than the capacitive ones demonstrated in the linked video, and lines will be straight invariably if you can draw that straight. I also believe that resistive touchscreens will not exhibit the curvature towards the edge of the screen, simply because they do not do any computation to figure out the position of your finger.
 
gadgetoid said:
Holy hell, I haven't finished watching the video yet, but curvature at the edges? Are they serious? I mean... can they really not just extend the size of the touch sensor beyond the visible size of the screen...
The curvature at the end is actually intended. basically it's to make it easier to hit something way out at the edge of the screen. Try 25 point calibrating a touch screen and you'll probably see what I mean
 
Last edited by a moderator:
How useful would a test be? Right.. the display/touchscreen won't be changed anyway, your best bet is a software patch. But then again: Who needs something like a test pattern?
If it works in a real world case it doesn't matter how it performs in such (stupid) tests.
 
anyone know is anything like UScrybe is available on linux?

Synaptics released software for windows not too long ago, can be used on most laptops with synaptics touchpads :)
 
JayFoxRox said:
How useful would a test be? Right.. the display/touchscreen won't be changed anyway, your best bet is a software patch. But then again: Who needs something like a test pattern?
If it works in a real world case it doesn't matter how it performs in such (stupid) tests.
Very useful. Currently you are unable to use Pandora in real world and decide whether the touch screen is clear enough for you. Nobody can except devs.
 
Last edited by a moderator:
It has been said before, Pandora's screen is resistive and this kind of test doesn't make sense. A resistive screen works and behaves differently :)
 
GizmoTheGreen said:
anyone know is anything like UScrybe is available on linux?

Synaptics released software for windows not too long ago, can be used on most laptops with synaptics touchpads :)
I don't know how this ended up in here but I'll give you the intricate details. The xf86-input-synaptics driver has supported 2 finger and 3 finger gestures since god knows when. Last I tried on Windows it didn't work... at all... with the preinstalled driver or the update offered to me. Anyway, you can map that to a bunch of different things. My pad only supports 2 fingers which I've in turn mapped to scrolling.

Now get back on topic.

The resistive screen works just fine it would seem. Capacitive screens are extremely imprecise in comparison.
 
Last edited by a moderator:
mali said:
this kind of test doesn't make sense
Why? I have a Zaurus C1000 with a resistive touch screen. This test (using stylus instead of finger, of course) perfectly demonstrates the touch screen jittering.
 
Last edited by a moderator:
Vitel said:
mali said:
this kind of test doesn't make sense
Why?

It's for capacitive screens, which work in a different way to how the resistive screen works on the pandora. Don't ask me how though.
 
Last edited by a moderator:
Because you will just see an accurately drawn line anyway. The reason for that meander effect is the additional calculation that has to take place on a capacitive screen. If a resistive screen is calibrated accurately, the line will appear exactly under the tip of the stylus and not meander.
 
mali said:
Because you will just see an accurately drawn line anyway. The reason for that meander effect is the additional calculation that has to take place on a capacitive screen. If a resistive screen is calibrated accurately, the line will appear exactly under the tip of the stylus and not meander.
You're wrong. Take a look here please:
http://bugzilla.kernel.org/show_bug.cgi?id=14434&action=View
 
Last edited by a moderator:
I can't take those test results at face value, as the human finger is inaccurate itself. They need to have a machine drawing perfectly straight lines (or they need to use a ruler) to really show the inaccuracy of the screen.
 
Vitel said:
You're wrong. Take a look here please:
http://bugzilla.kernel.org/show_bug.cgi?id=14434&action=View

I'm not exactly wrong, because that phenomenon is not the same on the Zaurus ;)
Whatever, Im not against testing the touchscreen, either. I just wanted to say that you won't see that effect you see on the Droid, on Pandora :)
 
Last edited by a moderator:
zhasha said:
gadgetoid said:
Holy hell, I haven't finished watching the video yet, but curvature at the edges? Are they serious? I mean... can they really not just extend the size of the touch sensor beyond the visible size of the screen...
The curvature at the end is actually intended. basically it's to make it easier to hit something way out at the edge of the screen. Try 25 point calibrating a touch screen and you'll probably see what I mean

The curvature is because the finger slides off the touch sensor. As less and less of the finger is touching the sensor you will only have the side of the fingertip within, resulting in those curves. As you can see only the iPhone have a noticeable problem with this, the other two tracks the finger outside of the viewable area to avoid this problem.
 
Last edited by a moderator:
Awakening said:
The curvature is because the finger slides off the touch sensor. As less and less of the finger is touching the sensor you will only have the side of the fingertip within, resulting in those curves. As you can see only the iPhone have a noticeable problem with this, the other two tracks the finger outside of the viewable area to avoid this problem.
Of all the touch screen/Wacom sensors etc. I've tested, this effect has always been well documented and completely intentional. The general problem arises when a screen is sunken into its frame rather than perfectly aligned with it, so you adjust it with edge-resistance, which makes the edge bleed towards the middle. Think of a widescreen TV on panorama setting, and then invert the stretching. Hell, it was an option in the configuration utility that came with my last touch screen.
 
Last edited by a moderator:
zhasha said:
Awakening said:
The curvature is because the finger slides off the touch sensor. As less and less of the finger is touching the sensor you will only have the side of the fingertip within, resulting in those curves. As you can see only the iPhone have a noticeable problem with this, the other two tracks the finger outside of the viewable area to avoid this problem.
Of all the touch screen/Wacom sensors etc. I've tested, this effect has always been well documented and completely intentional. The general problem arises when a screen is sunken into its frame rather than perfectly aligned with it, so you adjust it with edge-resistance, which makes the edge bleed towards the middle. Think of a widescreen TV on panorama setting, and then invert the stretching. Hell, it was an option in the configuration utility that came with my last touch screen.

Irrespective of whether it is intentional, or unintentional, surely extending the size of the touch sensitive area beyond that of the screen would provide a better, although less cost effective solution.

I'll have to grab a ruler to properly test it, but I can draw some pretty straight lines on my N900 using the stylus and a half eaten box of shortbread as a ruler.
 
Last edited by a moderator:
As I just commented in another thread, I'd be more interested in seeing a demo/test of pen pressure detection.

We know basic pressure detection is possible on resistive screens, and it looks like it's already confirmed on the Pandora hardware (according to the FAQ, anyway). I'd like to see a real-world demo though. If the stock driver can report on pressure, it should be easy enough to test; Gimp's already ported, and is designed to use pressure data with brushes.

I expect the pressure reported will be pretty erratic... I wonder if it's possible or practical to implement some sort of smoothing algorithm on the raw data to get a more consistent response from the pen?
 
Last edited by a moderator:
Back
Top