I1Profiler : Invalid measurment

[mmel]

Need help, I have I1pro, trying to profile my Epson pro 9800,so I did print  patch set , during measurement of test chat, only one  line keep getting this message Invalid measurement, color do not match measure value, I did reprint same patch set, still same message. any one have same issue, any solution.

Mel

[stalisman]

Sometimes when I get that problem I resort to measuring the offending line patch by patch, then when successful, return to full line scanning.

[Paul2660]

Are you running the most current version of i1profiler?  This problem was very common with the 1st release of the
software (at least on the windows side).  For about 3 months, I had to use a modified version of the code that changed the color lookup table.  Net my 9880 was printing too dark for the software and I would get the same error constantly.  You could fix the problem by backing off the ink density about -10, but I didn't want to do that as a permanent fix. 

With the 1.1.x update of the software, this problem went away. 

When I was having the problem, most times if I did the individual scans of the problem colors, when I generated the profile, it didn't look right on the 3D view, i.e. it had big holes in it.  By backing off the ink density, I got good profiles. 

With the fix from xRite, this problem went away and has stayed fixed with later versions of the software, Win7, 64 bit.

Paul

[mmel]

Hi Paul,

I am running the most current version of i1profiler.

[Paul2660]

Mel,

I would try one other thing, print the same target but back off the ink density by around -5 and see if this will scan
without the errors.   i1profiler seems to be somewhat sensitive to the density of certain patches depending on the
printer and paper or canvas.   If that works, then I would open a case with xrite and send them the sets of targets.

Paul

[Wayne Fox]

Try increasing the patch size of your target a little, I had similar problems and that solved it.

[jemmyell]

Hi,

I had to set MeasurementTolerance field in XRi1G2WorkflowSettings.ini to .7 from .5 to get reliable patch reading.  YMMV.

-James

[Rhossydd]

Try increasing the patch size of your target a little, I had similar problems and that solved it.

I don't understand why that would make any difference with hand read data from the i1pro. Effectively it's the same as moving the i1 more slowly when scanning the chart.
FWIW I find I need to move the i1 very slowly to get reliable results.
I may try James' ini hack soon, but I'd really like to know if changing that will effect the quality of the results.

[Wayne Fox]

Hard to go slow enough some time, a larger patch size helps that.  But you are right, kind of a long shot.

How many patches is your target?  Did you try a different number of patches by chance?  Also have you set the patches to scramble?  I think that would help the i1Pro.

As far as the tolerance, I've heard of others needing to do this as well.  Seems as though the error is because a color is read out outside of tolerance (which doesn't make sense since who knows what an ink/paper combination will do).  Changing the tolerance shouldn't affect the data, the numbers are what they are, and changing the tolerance won't have any impact on the actual numbers being read.

[Rhossydd]

How many patches is your target?  Did you try a different number of patches by chance?  Also have you set the patches to scramble?  I think that would help the i1Pro.

I've used charts from 400 to 3000 patches, usually with scrambled patches.

As you say; after years of smooth steady patch reading with the i1 with PMP5, trying to dramatically slow down the scanning speed doesn't come easy

[digitaldog]

Try increasing the patch size of your target a little, I had similar problems and that solved it.

I suspect the minimum patch size spec is off. At least with an iSis based on some testing I’m doing. Wider patches are better anyway (more averaging of data scanned).

[Wayne Fox]

Have you tested your i1Pro to make sure it is a problem?  Maybe the light source is getting old and not consistent?

[mmel]

Hi Wayne,

By default the patch width is 0.394in and patch height 0.354in, how much increase would you recommend.

Mel

[Ethan_Hansen]

As far as the tolerance, I've heard of others needing to do this as well.  Seems as though the error is because a color is read out outside of tolerance (which doesn't make sense since who knows what an ink/paper combination will do).  Changing the tolerance shouldn't affect the data, the numbers are what they are, and changing the tolerance won't have any impact on the actual numbers being read.

The good news is that the tolerance setting does something in i1Profiler V1.1.1. In the previous version it had no effect. The bad news is that i1P does not indicate which patch or patches were out of tolerance. With the iSis it gets even sillier; the entire target is read in before i1P complains about bad patches although it shows a visual comparison of expected to measured after each pair of rows.

The best way to debug the problem is to measure the same target several times and compare the data. Crank the tolerance up until it reads. I don't know if ColorThink or BabelColor handle i1P's mxf file formats. A free alternative is to save the measurements out in CGATS format and compare in MeasureTool (no dongle needed for this). You need to first edit the i1P CGATS files and do a global replace to delete every instance of SPECTRAL_ before MeasureTool can read the files.

Look to see if there is close agreement between the different readings. If the maximum Delta E 2000 is over 0.8 or so, something is wrong. MeasureTool highlights the 10% of patches with the greatest differences. If two or more patches in a row are highlighted, both having high dE values, this indicates a measurement error in which the i1Pro did not adequately detect the patch boundaries. If the errors are randomly scattered, you are more likely seeing the inherent inaccuracy of the instrument.

Repeating this process with targets having larger patch sizes will nail down whether you are seeing actual measurement errors or just i1Profiler being weird. If the latter, increase the measurement tolerance value until you get consistent chart reads and don't worry about it.

[FMueller]

Hi,

I had to set MeasurementTolerance field in XRi1G2WorkflowSettings.ini to .7 from .5 to get reliable patch reading.  YMMV.

-James

The above is what worked for me. I'm using version 1.2