Touch screen calibration

QNX Newsgroups (Archives) comp.os.qnx
1

I’m having problems with ‘acalib’. I’m following the directions
as prompted by ‘acalib’. They work at one point. I tried it
again, and then, as I go down, it goes right. As I go right,
it goes down. As x should increase, y increases. As x should
decrease, y decreases. As y should increase, x increases.
As y should decrease, x decreases instead.

At one point months ago, the connector was wrong. Out of probably
30 attempts at calibration to see what was wrong, 3-4 of them
resulted in a good calibration. Ideas?

I’d also like to know if anyone knows the format of the
/etc/config/absf/absf.x (x is 1 in my case) file. I’ll
write my own calibration routine if anyone know’s what
the format is. It looks easy enough, but might require lots
of trial and error.

Anyone else seen weird behavior with ‘acalib’?

Thanks for your time,
-Barry

2

Barry Robertson <brobertson@softwareremodeling.com> wrote:

I’m having problems with ‘acalib’. I’m following the directions
as prompted by ‘acalib’. They work at one point. I tried it
again, and then, as I go down, it goes right. As I go right,
it goes down. As x should increase, y increases. As x should
decrease, y decreases. As y should increase, x increases.
As y should decrease, x decreases instead.

sounds like you have x and y swapped.
after calibration, try slaying and restarting Input.
does that ‘fix’ it ?

At one point months ago, the connector was wrong. Out of probably
30 attempts at calibration to see what was wrong, 3-4 of them
resulted in a good calibration. Ideas?

changing the connector, may have changed the x/y swapping…

I’d also like to know if anyone knows the format of the
/etc/config/absf/absf.x (x is 1 in my case) file. I’ll
write my own calibration routine if anyone know’s what
the format is. It looks easy enough, but might require lots
of trial and error.

basically, its:
“%dx%d:%dx%d:%d %d %d %d %d”,
screen_upper_left_x, screen_upper_left_y, // usually 0x0
screen_lower_right_x, screen_lower_right_y // usually graphics mode (639x479)
touch_upper_left_x, touch_upper_left_y // acalib values
touch_lower_right_x, touch_lower_right_y // more acalib values
swap_x_and_y // 0 means no swap, 1 means swap

Anyone else seen weird behavior with ‘acalib’?

Thanks for your time,
-Barry

3

My latest calibration attempt:

Values reflected by acalib when I pressed them:
(62,103) (910,102)

(516,524)

(71,944) (929,935)

The absf.1 file contains:
0x0:639x479:66 919 102 939 0

I understand the 0x0:639x479 bit. But the
“66 919 102 939” bit has me confused if it’s
supposed to be where I pressed (top,left)
(bottom,right) x,y coordinates.

After calibration, x and y were crazy (swapped).
I exited photon and restart it, and x & y were
back to normal.

So, I’m still hung up on the last section of the
calibration file.

Thanks,
-Barry


Mike Taillon wrote:

Barry Robertson <> brobertson@softwareremodeling.com> > wrote:

I’m having problems with ‘acalib’. I’m following the directions
as prompted by ‘acalib’. They work at one point. I tried it
again, and then, as I go down, it goes right. As I go right,
it goes down. As x should increase, y increases. As x should
decrease, y decreases. As y should increase, x increases.
As y should decrease, x decreases instead.

sounds like you have x and y swapped.
after calibration, try slaying and restarting Input.
does that ‘fix’ it ?

At one point months ago, the connector was wrong. Out of probably
30 attempts at calibration to see what was wrong, 3-4 of them
resulted in a good calibration. Ideas?

changing the connector, may have changed the x/y swapping…

I’d also like to know if anyone knows the format of the
/etc/config/absf/absf.x (x is 1 in my case) file. I’ll
write my own calibration routine if anyone know’s what
the format is. It looks easy enough, but might require lots
of trial and error.

basically, its:
“%dx%d:%dx%d:%d %d %d %d %d”,
screen_upper_left_x, screen_upper_left_y, // usually 0x0
screen_lower_right_x, screen_lower_right_y // usually graphics mode (639x479)
touch_upper_left_x, touch_upper_left_y // acalib values
touch_lower_right_x, touch_lower_right_y // more acalib values
swap_x_and_y // 0 means no swap, 1 means swap

Anyone else seen weird behavior with ‘acalib’?

Thanks for your time,
-Barry

–

__| _ \ _ | Barry Robertson
_ \ | | | Software Consultant Phone: 972-758-9349
| __ < | Software Remodeling, Inc. Fax: 972-964-7524
___/ _| __| brobertson@SoftwareRemodeling.com

4

Forgot to ask: if the screen resolution is set
to 640x480 why are there values in the 900’s?

-Barry

Barry Robertson wrote:

My latest calibration attempt:

Values reflected by acalib when I pressed them:
(62,103) (910,102)

(516,524)

(71,944) (929,935)

The absf.1 file contains:
0x0:639x479:66 919 102 939 0

I understand the 0x0:639x479 bit. But the
“66 919 102 939” bit has me confused if it’s
supposed to be where I pressed (top,left)
(bottom,right) x,y coordinates.

After calibration, x and y were crazy (swapped).
I exited photon and restart it, and x & y were
back to normal.

So, I’m still hung up on the last section of the
calibration file.

Thanks,
-Barry

Mike Taillon wrote:

Barry Robertson <> brobertson@softwareremodeling.com> > wrote:

I’m having problems with ‘acalib’. I’m following the directions
as prompted by ‘acalib’. They work at one point. I tried it
again, and then, as I go down, it goes right. As I go right,
it goes down. As x should increase, y increases. As x should
decrease, y decreases. As y should increase, x increases.
As y should decrease, x decreases instead.

sounds like you have x and y swapped.
after calibration, try slaying and restarting Input.
does that ‘fix’ it ?

At one point months ago, the connector was wrong. Out of probably
30 attempts at calibration to see what was wrong, 3-4 of them
resulted in a good calibration. Ideas?

changing the connector, may have changed the x/y swapping…

I’d also like to know if anyone knows the format of the
/etc/config/absf/absf.x (x is 1 in my case) file. I’ll
write my own calibration routine if anyone know’s what
the format is. It looks easy enough, but might require lots
of trial and error.

basically, its:
“%dx%d:%dx%d:%d %d %d %d %d”,
screen_upper_left_x, screen_upper_left_y, // usually 0x0
screen_lower_right_x, screen_lower_right_y // usually graphics mode (639x479)
touch_upper_left_x, touch_upper_left_y // acalib values
touch_lower_right_x, touch_lower_right_y // more acalib values
swap_x_and_y // 0 means no swap, 1 means swap

Anyone else seen weird behavior with ‘acalib’?

Thanks for your time,
-Barry

–

__| _ \ _ | Barry Robertson
_ \ | | | Software Consultant Phone: 972-758-9349
| __ < | Software Remodeling, Inc. Fax: 972-964-7524
___/ _| __| > brobertson@SoftwareRemodeling.com

–

__| _ \ _ | Barry Robertson
_ \ | | | Software Consultant Phone: 972-758-9349
| __ < | Software Remodeling, Inc. Fax: 972-964-7524
___/ _| __| brobertson@SoftwareRemodeling.com

5

Don’t know if this applies to your case, but try changing the
last 0 in the absf.1 file to a 1. This is a “swap” bit needed
in some cases that I don’t remember.

Previously, Barry Robertson wrote in comp.os.qnx:

My latest calibration attempt:

Values reflected by acalib when I pressed them:
(62,103) (910,102)

(516,524)

(71,944) (929,935)

The absf.1 file contains:
0x0:639x479:66 919 102 939 0

I understand the 0x0:639x479 bit. But the
“66 919 102 939” bit has me confused if it’s
supposed to be where I pressed (top,left)
(bottom,right) x,y coordinates.

After calibration, x and y were crazy (swapped).
I exited photon and restart it, and x & y were
back to normal.

So, I’m still hung up on the last section of the
calibration file.

Thanks,
-Barry


Mike Taillon wrote:

Barry Robertson <> brobertson@softwareremodeling.com> > wrote:

I’m having problems with ‘acalib’. I’m following the directions
as prompted by ‘acalib’. They work at one point. I tried it
again, and then, as I go down, it goes right. As I go right,
it goes down. As x should increase, y increases. As x should
decrease, y decreases. As y should increase, x increases.
As y should decrease, x decreases instead.

sounds like you have x and y swapped.
after calibration, try slaying and restarting Input.
does that ‘fix’ it ?

At one point months ago, the connector was wrong. Out of probably
30 attempts at calibration to see what was wrong, 3-4 of them
resulted in a good calibration. Ideas?

changing the connector, may have changed the x/y swapping…

I’d also like to know if anyone knows the format of the
/etc/config/absf/absf.x (x is 1 in my case) file. I’ll
write my own calibration routine if anyone know’s what
the format is. It looks easy enough, but might require lots
of trial and error.

basically, its:
“%dx%d:%dx%d:%d %d %d %d %d”,
screen_upper_left_x, screen_upper_left_y, // usually 0x0
screen_lower_right_x, screen_lower_right_y // usually graphics mode (639x479)
touch_upper_left_x, touch_upper_left_y // acalib values
touch_lower_right_x, touch_lower_right_y // more acalib values
swap_x_and_y // 0 means no swap, 1 means swap

Anyone else seen weird behavior with ‘acalib’?

Thanks for your time,
-Barry

–

__| _ \ _ | Barry Robertson
_ \ | | | Software Consultant Phone: 972-758-9349
| __ < | Software Remodeling, Inc. Fax: 972-964-7524
___/ _| __| > brobertson@SoftwareRemodeling.com

6

Previously, Barry Robertson wrote in comp.os.qnx:

Forgot to ask: if the screen resolution is set
to 640x480 why are there values in the 900’s?

The screen resolution is unrelated to the resolution
of the touch screen (the video doesn’t know the touch
screen exists and vice versa). If the touch screen uses
a 10 bit A/D converter, full range values might be 0 to
1023, so the middle of the screen might have a value of
~512.

[…]

7

This seems like a nightmare returned. I had a similar
problem using an Elo smartset controller. We had
working systems, and non working systems. I think
that it came down to an acalib version, but I can’t
be sure.

Barry Robertson <brobertson@softwareremodeling.com> wrote:

I’m having problems with ‘acalib’. I’m following the directions
as prompted by ‘acalib’. They work at one point. I tried it
again, and then, as I go down, it goes right. As I go right,
it goes down. As x should increase, y increases. As x should
decrease, y decreases. As y should increase, x increases.
As y should decrease, x decreases instead.

At one point months ago, the connector was wrong. Out of probably
30 attempts at calibration to see what was wrong, 3-4 of them
resulted in a good calibration. Ideas?

I’d also like to know if anyone knows the format of the
/etc/config/absf/absf.x (x is 1 in my case) file. I’ll
write my own calibration routine if anyone know’s what
the format is. It looks easy enough, but might require lots
of trial and error.

Anyone else seen weird behavior with ‘acalib’?

Thanks for your time,
-Barry

–
Mitchell Schoenbrun --------- maschoen@pobox.com

8

The only problem with tweaking the “swap” bit, is that this (or
something similar) is going into an embedded system with no keyboard.
A user could decide to calibrate (or the system), and if the “swap” bit
is wrong, when the user’s done, they have no way to change it.

We’ll probably end up writing our own, but I wanted to get this
figured out so that I’d be starting from known good behavior.

Thanks for your help,
Barry

Jason Lee wrote:

That’s the correct way to do it. At least it worked for me when I am using a
Elotouch screen with the smartset protocol.

By changing the value, it causes a swap in the co-ordinates and it should
work just fine.

Regards,

Jason

“Ken Schumm” <> kwschumm@qsolv.com> > wrote in message
news:Voyager.000808162706.29440A@dilbert…
Don’t know if this applies to your case, but try changing the
last 0 in the absf.1 file to a 1. This is a “swap” bit needed
in some cases that I don’t remember.

Previously, Barry Robertson wrote in comp.os.qnx:
My latest calibration attempt:

Values reflected by acalib when I pressed them:
(62,103) (910,102)

(516,524)

(71,944) (929,935)

The absf.1 file contains:
0x0:639x479:66 919 102 939 0

I understand the 0x0:639x479 bit. But the
“66 919 102 939” bit has me confused if it’s
supposed to be where I pressed (top,left)
(bottom,right) x,y coordinates.

After calibration, x and y were crazy (swapped).
I exited photon and restart it, and x & y were
back to normal.

So, I’m still hung up on the last section of the
calibration file.

Thanks,
-Barry


Mike Taillon wrote:

Barry Robertson <> brobertson@softwareremodeling.com> > wrote:

I’m having problems with ‘acalib’. I’m following the directions
as prompted by ‘acalib’. They work at one point. I tried it
again, and then, as I go down, it goes right. As I go right,
it goes down. As x should increase, y increases. As x should
decrease, y decreases. As y should increase, x increases.
As y should decrease, x decreases instead.

sounds like you have x and y swapped.
after calibration, try slaying and restarting Input.
does that ‘fix’ it ?

At one point months ago, the connector was wrong. Out of probably
30 attempts at calibration to see what was wrong, 3-4 of them
resulted in a good calibration. Ideas?

changing the connector, may have changed the x/y swapping…

I’d also like to know if anyone knows the format of the
/etc/config/absf/absf.x (x is 1 in my case) file. I’ll
write my own calibration routine if anyone know’s what
the format is. It looks easy enough, but might require lots
of trial and error.

basically, its:
“%dx%d:%dx%d:%d %d %d %d %d”,
screen_upper_left_x, screen_upper_left_y, //
usually 0x0
screen_lower_right_x, screen_lower_right_y //
usually graphics mode (639x479)
touch_upper_left_x, touch_upper_left_y //
acalib values
touch_lower_right_x, touch_lower_right_y //
more acalib values
swap_x_and_y
// 0 means no swap, 1 means swap

Anyone else seen weird behavior with ‘acalib’?

Thanks for your time,
-Barry

9

That’s the correct way to do it. At least it worked for me when I am using a
Elotouch screen with the smartset protocol.

By changing the value, it causes a swap in the co-ordinates and it should
work just fine.

Regards,

Jason




“Ken Schumm” <kwschumm@qsolv.com> wrote in message
news:Voyager.000808162706.29440A@dilbert…

Don’t know if this applies to your case, but try changing the
last 0 in the absf.1 file to a 1. This is a “swap” bit needed
in some cases that I don’t remember.

Previously, Barry Robertson wrote in comp.os.qnx:
My latest calibration attempt:

Values reflected by acalib when I pressed them:
(62,103) (910,102)

(516,524)

(71,944) (929,935)

The absf.1 file contains:
0x0:639x479:66 919 102 939 0

I understand the 0x0:639x479 bit. But the
“66 919 102 939” bit has me confused if it’s
supposed to be where I pressed (top,left)
(bottom,right) x,y coordinates.

After calibration, x and y were crazy (swapped).
I exited photon and restart it, and x & y were
back to normal.

So, I’m still hung up on the last section of the
calibration file.

Thanks,
-Barry


Mike Taillon wrote:

Barry Robertson <> brobertson@softwareremodeling.com> > wrote:

I’m having problems with ‘acalib’. I’m following the directions
as prompted by ‘acalib’. They work at one point. I tried it
again, and then, as I go down, it goes right. As I go right,
it goes down. As x should increase, y increases. As x should
decrease, y decreases. As y should increase, x increases.
As y should decrease, x decreases instead.

sounds like you have x and y swapped.
after calibration, try slaying and restarting Input.
does that ‘fix’ it ?

At one point months ago, the connector was wrong. Out of probably
30 attempts at calibration to see what was wrong, 3-4 of them
resulted in a good calibration. Ideas?

changing the connector, may have changed the x/y swapping…

I’d also like to know if anyone knows the format of the
/etc/config/absf/absf.x (x is 1 in my case) file. I’ll
write my own calibration routine if anyone know’s what
the format is. It looks easy enough, but might require lots
of trial and error.

basically, its:
“%dx%d:%dx%d:%d %d %d %d %d”,
screen_upper_left_x, screen_upper_left_y, //
usually 0x0
screen_lower_right_x, screen_lower_right_y //
usually graphics mode (639x479)
touch_upper_left_x, touch_upper_left_y //
acalib values
touch_lower_right_x, touch_lower_right_y //
more acalib values
swap_x_and_y
// 0 means no swap, 1 means swap

Anyone else seen weird behavior with ‘acalib’?

Thanks for your time,
-Barry

–

__| _ \ _ | Barry Robertson
_ \ | | | Software Consultant Phone: 972-758-9349
| __ < | Software Remodeling, Inc. Fax: 972-964-7524
___/ _| __| > brobertson@SoftwareRemodeling.com
\

10

We have similar problem with the Elo smartset. I believe it is because Elo
does not guarantee
the orientation of the touch membrane. We have some that work and some that
do not work.
The calibration routine should be smart enough to figure out the
orientation.


Mitchell Schoenbrun <maschoen@tsoft.com> wrote in message
news:sp28496b63a37@corp.supernews.com…

This seems like a nightmare returned. I had a similar
problem using an Elo smartset controller. We had
working systems, and non working systems. I think
that it came down to an acalib version, but I can’t
be sure.

Barry Robertson <> brobertson@softwareremodeling.com> > wrote:

I’m having problems with ‘acalib’. I’m following the directions
as prompted by ‘acalib’. They work at one point. I tried it
again, and then, as I go down, it goes right. As I go right,
it goes down. As x should increase, y increases. As x should
decrease, y decreases. As y should increase, x increases.
As y should decrease, x decreases instead.

At one point months ago, the connector was wrong. Out of probably
30 attempts at calibration to see what was wrong, 3-4 of them
resulted in a good calibration. Ideas?

I’d also like to know if anyone knows the format of the
/etc/config/absf/absf.x (x is 1 in my case) file. I’ll
write my own calibration routine if anyone know’s what
the format is. It looks easy enough, but might require lots
of trial and error.

Anyone else seen weird behavior with ‘acalib’?

Thanks for your time,
-Barry


\

Mitchell Schoenbrun --------- > maschoen@pobox.com