First profile with i1Profiler

[tony22]

So I had to post this as a new thread only because of how pleased I am with the result (there, I spoiled the surprise ).

I've been using printer profile creation products for years, and have always been disappointed. From the early versions of ProfilerPlus (Datacolor), to SpyderPrint, to Argyll CMS, to i1Match, and a couple of others I either got poor results, or (in the case of Argyll) was not really inclined to go in the direction of a non-wizard based product (I did try it). And in the case of Argyll I blame myself, because there are certainly those who've used it to create excellent results. For the others, though, I went around and around with tweaking, reading from other users, tweaking, adjusting, and still could not get a result I was satified with. This is partly my problem I think because I seem to be super picky (although many on the dpreview Printer forum over the years have reported results similar to mine with the same tools).

I bit the bullet and got the i1Pro and i1Match, and still was not happy. Before I go any further let me be sure to state that I have a properly calibrated monitor (Sony GDM-FW900), a good work area with indirect neutral lighting, all the things you need to do for proper printing. I always got results which were enough off target in neutral grays to be noticeable to me. Print darkness (or lightness) was not an issue.

I bit the other bullet and spent the money on i1Profiler. Took sheets from my trusty stack of Epson Photo Paper (I use this for everyday casual printing), printed out a 1727 patch target (reading something here about picking a number that was not a multiple of 3), and let it sit for 48 hours before measuring with i1Profiler.

Long story short, while I haven't measured every color yet in a test target, I am now getting a dE2K of only .7 in pure grays, and visually at least the test images look spot on under D50 lighting conditions. dE2K of .7 may not seem all that great to you professionals here, but to me it's heaven sent. I never got a profile this good in all my years of hacking away with these other tools. Of course I don't think I would have been able to use these tools (even the ones from the past) effectvely without the help of the smart people here and on dpreview.

I only hope that I have the same kind of success with the "real" papers that I'll be profiling next (fingers are definitely crossed).

[smilem]

It is nice you have found your program, you might like to read and look at my test results at:
http://www.luminous-landscape.com/forum/index.php?topic=58277.msg472435#msg472435

It seems that i1Profiler optimizes grays like expected But I can tell you other color management package from heidelberg called color toolbox 2010 works even better by producing gamut for xrga file is 638.191 vs. i1profiler 625.57.
Color toolbox even optimizes the RGB ICC curves they are smoother, and you get loads of manual generation adjustments.

I'm trying to get it to read the targets to no avail, it seems to support some strange CGATS format because the RGB values are like CMYK, the largest value is 100. And when I import my own targets it will not open them. I can however make profiles with it but I read that color toolbox 2010 supports re-measuring algorithm that means it re-measures (if you have i1iO automatically) the patches that are problematic.

i1profiler optimizes peceptual and relative colorimetric intents grayscale where color toolbox only perceptual.

[tony22]

It is nice you have found your program, you might like to read and look at my test results at:
http://www.luminous-landscape.com/forum/index.php?topic=58277.msg472435#msg472435

It seems that i1Profiler optimizes grays like expected But I can tell you other color management package from heidelberg called color toolbox 2010 works even better by producing gamut for xrga file is 638.191 vs. i1profiler 625.57.

Yes, I was reading this thread. I was waiting to see how it would turn out. If this works with little or no twaeking I may consider it. I looked for info on Prinect Color Toolbox but could not find a price anywhere. It looks expensive.

The grays are the only thing I measured so far. I go to that first because my eye seems to be drawn to color offsets in grays. But as I said above, visually the whole range of colors looks much better than anything I've gotten before. I use the same images as test prints, and the ones I've made with other profiling tools never came out looking this close to dead on by just measuring and pressing the "create" button. I haven't evaluated the gradient smoothness and deatil in the dark areas yet.

Of course now the color gods will make sure my efforts on my higher end papers will fail miserbaly.

[VitOne]

A bigger gamut doesn’t’ necessarily mean a better profile. A bigger gamut could mean less saturated colors, for example. And, generally speaking, a bigger gamut does produce less saturated colors. You can verify this with a simple print test. The gamut of a device should be right, not bigger and not smaller.

I did several tests too and I found that i1Profiler is, at the moment, the software that gives me the best results (better than ArgyllCMS for many applications, and I was happy with ArgyllCMS).

I will never stop repeating that, when looking DeltaEs, you should understand what errors is worth to care about. For photography I have found that in many cases an higher DeltaE is better than a profile without smooth gradients, for example.

I strongly suggest you to do your evaluations giving the numbers the right importance, do not overestimate them.

[smilem]

A bigger gamut doesn’t’ necessarily mean a better profile.

Yes I agree, bigger gamut most of the time means less accurate profile, and on the printer I was testing on one of the papers I noticed that while the grayscale was better with i1profiler by a big leap, the photographic reproduction was not.

This was seen on PDI test chart where you see a blue vase with flowers on the top right corner, I think the glass transparency effect is a very nice profile performance test method. If I only had more tests with glass, waiting for suggestions

I strongly suggest you to do your evaluations giving the numbers the right importance, do not overestimate them.

The test I did was because I wanted to know what the best tool numerically was, and that doesn't always provide best photographic results like you said. But building profiles from test charts that were measured the best I can is important.

[octagon]

So I had to post this as a new thread only because of how pleased I am with the result (there, I spoiled the surprise ).

Glad to here it. This is pretty much my experience as well; ie, pleased.

[smilem]

@tony22

What printer, ink , paper did you make profile for? What setting did you choose is i1profiler, all sliders were at zero etc.

[tony22]

Canon Pro9000 Mk.II, OEM inks, Epson Photo Paper Glossy. I used a 1727 patch target, and the profile was built D50 with all other settings at their default. For this paper I use the Photo Paper Pro selection in the Canon print driver because in my past testing this seemed to produce the best ink coverage (smoothness, good base color accuracy etc) with this Epson paper.

When the profiles were done I tested Percpetual without BPC and Relative Colorimetric without BPC. The RC looked and tested with my i1Pro marginally better than Perceptual.


** edited to correct my dyslexia **

[smilem]

When the profiles were done I tested Percpetual with BPC and Relative Colorimetric without BPC. The RC looked and tested with my i1Pro marginally better than Perceptual.

As far as I know perceptual does not need the BPC checked as it is allready mapped wpt and bpt, the Relative Colorimetric needs BPC checked. As your printer does not have gray cartridge the gray will never be ideal, if you compare your best print to say R2880 with even lower quality profile, you will see that R2880 gray is better.

[tony22]

As far as I know perceptual does not need the BPC checked as it is allready mapped wpt and bpt, the Relative Colorimetric needs BPC checked.

Duh, sorry. Dyslexia on my part. In fact it is my RC print that has BPC turned on.