chart reading in 1i Profiler vs. Colorport vs. Profilemaker5

[smilem]

Hello, I'm looking for best way to read targets with i1iO.

Target generation:

i1Profiler uses all patches in a target to generate a profile, uses thin white and normal black separators
Colorport uses all patches in a target to generate a profile, uses multicolor separators
Profilemaker Measuretool5 uses some patches to detect errors (how many etc, I don't know), uses normal blak and white separators.

Target reading:

i1Profiler has so far unreliable target reading, frequent pauses, re-reads some strips that are fine.
Colorport has normal reading, with occasional slower reading certain rows.
Profilemaker Measuretool5 has so far best reading ability, with errors detection refuses to read bad targets at all, like cmyk target that was printer in a wrong way.

So the questions is:

Is optimal to use Profilemaker Measuretool5 to read i1Profiler targets? Perhaps it is better to use colorport, but then there is no error detection at all. It is possible to end up building a profile for wrongly printed target.

Isn't using Profilemaker Measuretool5 to measure targets built with other software means there is no error detection as these targets will not contain normal "error detection (blue color?)" patches to begin with?

Thank you.

[digitaldog]

What I can tell you is I’ve built targets in i1Profiler and can read them fine there. I can then use MeasureTool to load the reference data and measure there as well (the same color patches). As long as the targets are identical in terms of layout, you can use either.

[shewhorn]

What I can tell you is I’ve built targets in i1Profiler and can read them fine there. I can then use MeasureTool to load the reference data and measure there as well (the same color patches). As long as the targets are identical in terms of layout, you can use either.

You may introduce some error there if you're using the i1Profiler generated CGATS file. The CGATS reference file that i1Profiler generates has 16 bit values. i1Profiler does not use 16 bit files though, it generates 8 bit files for the targets and the mxf file which it uses for reference data is also 8 bits so if the Measure Tool can actually read 16 bit reference files, you'll introduce some error when doing the comparison. If you look at the mxf file, THAT is in 8 bits and is what i1Profiler uses for the reference data. Before loading the CGATS reference file in to the Measure Tool, you first need to bring it into a spreadsheet, truncate the floating point precision (no rounding, just lob off the precision so 2.9 = 2, 2.3 = 2, 3.1=3, 3.8=3, etc.), and save it, otherwise you can have reference data that says 127.99 but with a patch that's actually 127.

During beta I asked why the mxf file uses 8 bit data, and the CGATS file uses 16 bit data but I never got a response.

Cheers, Joe

[digitaldog]

You may introduce some error there if you're using the i1Profiler generated CGATS file.

You might but I suspect it would be lower than device variation.

[smilem]

And here are my results. What do you think?
Perhaps there is a way to load i1Profiler data to colorthink for deltaE generation I am not aware of?

[Wayne Fox]

Hello, I'm looking for best way to read targets with i1iO.

Target generation:

i1Profiler uses all patches in a target to generate a profile, uses thin white and normal black separators
...
i1Profiler has so far unreliable target reading, frequent pauses, re-reads some strips that are fine.

Had the same problem with my i1i0 until i increased the size of the patches.  It then started working perfectly (until the entire table died ... does not see the connected i1 anymore)

[smilem]

Had the same problem with my i1i0 until i increased the size of the patches.

Well, what patch size did you had the problems with ? And what size was fine ?

It then started working perfectly (until the entire table died ... does not see the connected i1 anymore)

Very sad how long did it last then?

After all the testing I may have exaggerated the reading problems too much, but it still reads better (number wize, and how the thing reads) with profilemaker5. Test results confirm this. The i1Profiler and colorport re-reads perhaps because of the new algorithm that should make better measurements, so far I do not see that measurements were better. The profile however seems to have bettet gamut. Hmm.... Colortoolbox gamut for xrga file is 638.191 ! Perceptual grayscale looks only 0.5 worse (at lightest shade), and curves seem better.

Very nice photos (captured moments in time) on your website gallery I would seriously consider upgrading gallery to something like "http://www.fotoplayer.com/demo.html" That way you could sell, digital or prints without any problems

[Wayne Fox]

Well, what patch size did you had the problems with ? And what size was fine ?

Very sad how long did it last then?

After all the testing I may have exaggerated the reading problems too much, but it still reads better (number wize, and how the thing reads) with profilemaker5. Test results confirm this. The i1Profiler and colorport re-reads perhaps because of the new algorithm that should make better measurements, so far I do not see that measurements were better. The profile however seems to have bettet gamut. Hmm.... Colortoolbox gamut for xrga file is 638.191 ! Perceptual grayscale looks only 0.5 worse (at lightest shade), and curves seem better.

Very nice photos (captured moments in time) on your website gallery I would seriously consider upgrading gallery to something like "http://www.fotoplayer.com/demo.html" That way you could sell, digital or prints without any problems

Wasn't around the machine so I didn't post the size ... sry.  Just looked it up, I changed the patches to 9mm wide and 8mm high.  Rarely had to read a row the second time after that.  The i0 table I've had for quite some time, but it was pretty flakey with PM5 so I didn't use it much.  really shouldn't have died.  I haven't decided whether to send it in or not yet.  I already had made the switch to an iSis.

Are you referring to my gallery?

[smilem]

Wasn't around the machine so I didn't post the size ... sry.  Just looked it up, I changed the patches to 9mm wide and 8mm high.  

Well, the ISO 5-4:2009 recommended patch size for i1iO would be 1.08cm x 0.93cm, your size is very close, so that explains why you do not have any problems now.

I contacted about this Xrite - as always they did not reply why recommemded default minumum is 0.64 x 0.64 and why standart IT9.18 chart uses these small 0.64 x 0.64 patches when they are obviously too small to read on i1iO.

Rarely had to read a row the second time after that.  The i0 table I've had for quite some time, but it was pretty flakey with PM5 so I didn't use it much.  really shouldn't have died.  I haven't decided whether to send it in or not yet.  

Hmm, isis is nice, if you buy it in US, in europe the price is 2x more. It can't build profiles for very thin film like, or thick like CD's DVD's, plastic media. I would have gotten it too, but at the price difference and reduced flexibility I could not justify it.

I already had made the switch to an iSis.

So how is it compared to i1IO, I mean do you notice that your profiles have improved, any info that would confirm it ?

Are you referring to my gallery?

Yes, you have very nice photos.

[smilem]

You may introduce some error there if you're using the i1Profiler generated CGATS file. The CGATS reference file that i1Profiler generates has 16 bit values. i1Profiler does not use 16 bit files though, it generates 8 bit files for the targets and the mxf file which it uses for reference data is also 8 bits so if the Measure Tool can actually read 16 bit reference files, you'll introduce some error when doing the comparison. If you look at the mxf file, THAT is in 8 bits and is what i1Profiler uses for the reference data. Before loading the CGATS reference file in to the Measure Tool, you first need to bring it into a spreadsheet, truncate the floating point precision (no rounding, just lob off the precision so 2.9 = 2, 2.3 = 2, 3.1=3, 3.8=3, etc.), and save it, otherwise you can have reference data that says 127.99 but with a patch that's actually 127.

During beta I asked why the mxf file uses 8 bit data, and the CGATS file uses 16 bit data but I never got a response.

Cheers, Joe

My workflow was:

Generate target in i1Profiler, save to known folder, this saves mxf and txt files.
Import the txt file in Colorport -> generate target - > save new reference and target
Import the txt file in profilemaker as reference -> generate target - > save new reference and target

Measure tartget in colorport - > save as CGATS Profilemaker -> Average in measuretool - > save
Measure tartget in colorport - > save as CGATS Profilemaker XRGA -> Average in measuretool - > save

Load the saved data that is a reference and measurement information in one to i1Profiler and profilemaker5 for ICC generation.

Do you see any flaws here? I re-generated the reference so I did not introduce any errors no?