#80 eye-one pro does not work

[lt_gustavsen]

There is a problem with the eye-one pro and lprof. I tested this with the qt4 port.
My eye-one pro works with argyllcms. When I use it with argyll it normally prompt for calibration against a white tile every time I use it.
In lprof it dose's not ask. If I run the monitor calibration as an example, it runs through the entity calibration in a 20 seconds. In the message box under the patch window I got warnings like:

Spot read failed because instrument needs calibration.
read_sample returned “instrument needs calibration (mode needs calibration)
Instrument calibration failed.

Since my English skills is limited, here is a video of the problem
http://ltgustavsen.googlepages.com/video.ogg (480K)

I have done the same step with my dtp-94 and then it works as expected.

[Hal Engel]

Logged In: YES
user_id=1052244
Originator: NO

OK I will look into this. Since I don't have an EyeOne Pro I will need your to help with testing things when I put in a fix. I should have something to test in the next few days since I was able to throw together some initial test code this evening. But it will need more work before I put it into CVS for you to test since I need to work some UI issues with this first.

Thanks for reporting this. The code where this is happening is has a FIXME comment but since we didn't have any way to test the fix it had not been worked on. So it is good that we have someone who can do the testing now.

[Hal Engel]

Logged In: YES
user_id=1052244
Originator: NO

There is now an initial check in to CVS of the test fix for this. This new code now checks to see if a calibration is needed as part of the device initialization and should perform all needed calibration steps before beginning the measurements. The old code only ran display related calibrations. The UI around this might still be a little rough but I would like this tested to make sure that it is working before putting too much effort into cleaning up the UI. Since the old code was working with every color meter I had I was able to test this by forcing the code to always go to the new calibration routines. But since my meters only required display type calibrations (IE. monitor white and frequency calibration) only a few paths through the new code have been tested at this point. But it should detect that the EyeOne Pro needs a reflective white calibration and prompt you to do one.

Please check out the Qt4-port branch and test it.

[lt_gustavsen]

Logged In: YES
user_id=1361762
Originator: YES

Thanks for the fix. It works better now, but not good enough.
The problem now is that when I’m finished with the calibration, I can not access the instrument for the profiling part of it.

I have illustrated the problem in the screen cast under but with the calibration report instead, since it quicker.
http://ltgustavsen.googlepages.com/lprof-i1.ogg (sorry to say it’s 2.3mb)
As you can see it asks nicely for the calibration of the instrument. One small issue is that when I click ok on the box “The measurement device has been calibrated ….” It immediately starts measuring without any instruction to place the device on the monitor.

If I try to do this one more time, it just fall back to the previous window when I click “go”. This is the same behavior I have when I create the measurements sheet after calibrating. If I go back to the main menu, and start from scratch it works. This is of course not a big problem for a calibration report, but a show stopper for the profiling.

The screencast is taken on a fedora 8 box. I build it this time with cmake. I used fedora 8 on this test since the policykit rules included in the rpm found on http://koji.fedoraproject.org/koji/packageinfo?packageID=5775 works very well. When I use Argyll with this rules it is very quick to release the device after use. When I use ubuntu with normall udev rules, I often have to wait a few seconds before I run dispcal –r again.

I have tested also tested this on ubuntu with the same results.

For the normal calibration I get the warning about "no lcd support…" after it has measured 5-6 patches. I guess it more intuitive to have this at the start of the measurements?

[Hal Engel]

Logged In: YES
user_id=1052244
Originator: NO

"Thanks for the fix. It works better now, but not good enough."

I would have been surprised if everything worked as well as it needs too since I was not able to test with a device that behaves like yours. So I expected this.

"As you can see it asks nicely for the calibration of the instrument. One
small issue is that when I click ok on the box “The measurement device
has been calibrated ….” It immediately starts measuring without any
instruction to place the device on the monitor."

This is one of the things I wanted to find out. None of my devices need an external calibration like reading a reflective calibration patch so I was not sure exactly how this would work from a UI point of view. I think just adding another line to the Finished Calibrating pop up asking the user to place the instrument back on the measurement patch before clicking Ok would work here.

"If I try to do this one more time, it just fall back to the previous
window when I click “go”. This is the same behavior I have when I
create the measurements sheet after calibrating. If I go back to the main
menu, and start from scratch it works. This is of course not a big problem
for a calibration report, but a show stopper for the profiling."

Yes this is a significant problem. It appears that the first device calibration is leaving things is a strange state and it fails initialization the second time for some reason. It appears this does not happen unless the calibration handler is called. So it is something in the calibration handler that is causing this to happen. I will see what I can do to reproduce and fix this. If I can't reproduce it I may add some debugging code and ask you to run through this again so that you can get the debug messages back to me so that I can get a handle on where this is happening. By the way the screen capture was very helpful.

"For the normal calibration I get the warning about "no lcd support…"
after it has measured 5-6 patches. I guess it more intuitive to have this
at the start of the measurements?"

I modified how some of this works recently as part of putting together a more advance monitor calibration algorithm. The four or five patches are preliminary measurements to set up the main calibration steps. The new algorithm, when it is in place, will work for both LCDs and CRTs. So you are seeing code that is in transition between the old calibration algorithms and the new. In addition, Gerhard is looking into more advanced profiling algorithms that will work for all types of monitors. So at some point the warning message will go away. But you are right that the message should happen before any measurements are taken.