My profiles just suck.

[W. Fisher]

I don't know what's going on with my i1 PhotoPro 2 profiling using x-rite's i1 Profiler software, but my profiles bite the dust against the manufacturers.

Mine seem to be much smaller in overall gamut coverage. 3rd party ink (Cone)? Initial ink loading or density before profiling? Something's askew somewhere.

The attached image shows the manufacture's (wireframe) of theirs against mine which is the small solid one inside of theirs. Theirs has a better (denser) black, brighter white, and much more red and blue density. Fwiw, it is a "Pictorico White Gloss Film" which should show a large gamut.

I'll try raising the Epson Ink Density from 0% and see what happens. I don't know why x-rite or anyone else (basICColor) makes no mention of the importance of getting the density right prior to making a profile, but they all seem to omit it.

I would expect to see a somewhat better result than a manufacture's generic "One size fits all profile" - but this is maddening!

W. F.

 

[Ink stained Fingers]

How do you know that the manufacturer's profiles are correct ?
That was already discussed in this thread
http://www.printerknowledge.com/thr...o3880-and-p800-kinda-quirky.10713/#post-90623

Since you are running decent equipment I rather would trust your own results and just not use the manufacturer's profiles .

Not all printers allow changes in the ink density settings, and using that parameter would require an iterative process by the user which is not the philosophy of the profiling packages - print - measure - and let the computer calculate the profile.

 

[W. Fisher]

hmmm...

Turns out you're right! The manufacturer's was worse than the one I made. Go figger!?!

Still seems odd the ColorThink Pro 3 shows a bigger gamut with the Factory profile above, but doesn't work out that way on paper with the home-brew one. Puzzling.

Anyway, here's the image of "Bill's Balls" I found on LuLa ( http://forum.luminous-landscape.com/index.php?topic=79799.0 ). Seems the lower one (Factory) has a odd magenta ring around the darkest blue ball (Bottom right. Maybe too much blue per the Colorthink Pro wireframe above?), and the greens aren't all that green either. Cyan is sort of faint too. Few other color banding issues showing up in it too.

So much for "The Factory Profile."

Back to printing...

W. F.

 

[Ink stained Fingers]

Such images with synthetically generated color graduations are not that easy to interpret, you would have to look and identify those areas which are out of gamut, you may try to look to those in proof mode with PS, some colors may be out of the colorspace of the monitor and/or the printer, and with such images you may start thinking about the color space used for display. And you may try rel. colorimetric to compare with for the printout, banding may be gone, or at another color area.
Quite some detail about printing, profiling etc can be found at Noman Koren's web pages, not all up to date but still valid since that's general explanations independent of actual hardware or software
http://www.normankoren.com/makingfineprints.html
http://www.normankoren.com/printer_calibration.html
http://www.normankoren.com/makingfineprints1A.html

And comments about banding here , in the middle of the article

http://www.outbackphoto.com/printinginsights/pi049/essay.html

 

[RogerB]

hmmm...

Turns out you're right! The manufacturer's was worse than the one I made. Go figger!?!

Still seems odd the ColorThink Pro 3 shows a bigger gamut with the Factory profile above, but doesn't work out that way on paper with the home-brew one. Puzzling.

Anyway, here's the image of "Bill's Balls" I found on LuLa ( http://forum.luminous-landscape.com/index.php?topic=79799.0 ). Seems the lower one (Factory) has a odd magenta ring around the darkest blue ball (Bottom right. Maybe too much blue per the Colorthink Pro wireframe above?), and the greens aren't all that green either. Cyan is sort of faint too. Few other color banding issues showing up in it too.

So much for "The Factory Profile."

Back to printing...

W. F.

Maybe I'm misunderstanding, but aren't you using Cone inks? If so, you wouldn't expect OEM profiles to give a good result, especially on such a demanding image as Bill's Balls.

 

[The Hat]

I reckon using the Bill’s Balls test for your new profiles is way over the top, when you have sorted out your profiles and your very happy and satisfied with your results, only then should you consider this all most impossible test print..

 

[W. Fisher]

Maybe I'm misunderstanding, but aren't you using Cone inks? If so, you wouldn't expect OEM profiles to give a good result, especially on such a demanding image as Bill's Balls.

Yes, this was with Cone's inks. Both of Bill's Ball were same ink, paper, printer, output via Qimage Ultimate (Not a fan of printing from Adobe PS/LR since QU shows me the profiles and sets the Epson driver up correctly based from prior profile's memory.), etc. Just the profiles are Mine verses Factory.

Fwiw, I just read on Cone's site they are coming out with a blacker black, maybe two L* points blacker, but they want to figure out how that will play in standard profiles before marketing it.

Back to Bill's Balls, what seems odd is although the factory profile shows such a huge 3D gamut 'color bubble' in Colorthink Pro 3 in the wireframe above - both in color and density - then why wouldn't that also show up in the Bill's Ball Factory image above which seems opposite? With my profile being much smaller in both color and density within the Factory wireframe, the opposite happens on the print as it is both denser and more colorful?!!? I would think the Factory should be darker in the cyans and greens, but not lighter. Something else is going on in the ICC profiles someplace that I'm missing (i.e. "Bigger gamut is less on print??").

Fwiw, the image on the monitor is correct and I do not see that magenta ring around the Factory dark blue ball on the print either, nor the lightness of the Factory greens and cyans in the print. So monitor is okay, and the output of Mine of Bill's Balls are almost in visual sync, but Factory print is "huh?"

I don't know if Bill's Balls is the cruelest test out there for printing, but it was an eye opener. I found an old dpreview.com thread on the Bill's Ball Test here that is entertaining where someone argued that the blue should be black: http://www.dpreview.com/forums/post/56448594


W. F.

 

[Ink stained Fingers]

Oh, lots of things are going on, just to give you an example what problems a profile can create and how that can be analyzed. You are using Colorthink, I don't know whether this analysis done with Gamutvision is possible as well with Colorthink
http://www.gamutvision.com/docs/smudged_pines.html
Applying this approach around the colors in vicinity of the banding should show what the problem is.
It could be a read error of a particular patch, some software like basiccolor has a smoothing option for the read data from your photometer, X-Rite with Profile Maker gave an option to average - to smooth effectively - data by reading one patch sheet several times , not a big problem with an automatic patch sheet reader/scanner, you could as well average data from more than one patch sheet to overcome potential print errors when printing the target sheet.

I'm ignoring OEM canned profiles altoghether, I'm not using their papers.

Doesn't Colorthink have a display option showing how colors are mapped by a profile - vectors from the
original color to the profiled output color ? That should show why one profile yields more saturated colors than another one.

 

[W. Fisher]

Oh, lots of things are going on, just to give you an example what problems a profile can create and how that can be analyzed. You are using Colorthink, I don't know whether this analysis done with Gamutvision is possible as well with Colorthink
http://www.gamutvision.com/docs/smudged_pines.html
.....
Doesn't Colorthink have a display option showing how colors are mapped by a profile - vectors from the
original color to the profiled output color ? That should show why one profile yields more saturated colors than another one.

Interesting reading in the link about the trees. I don't know if Colorthink has GamutVision's ability to read a profile as it does 'spatially,' but haven't found it yet. I'm still fairly new to it.

Fwiw, I am making some profiles now with basICColor's dropRGB and their Display software as well. I may have some issues with the Catch part of it that is needed for the scanning that goes into dropRGB that makes the profile and it adds the "optimize profile" (??) as an option once making it, but their German tech support will not answer ANYTHING at all in my experience with them. It shows promise as software that you can build a system on (i.e. Very pricey plug-ins to do certain things.), but they do need far, far better customer support, imho. I do not see a remarkable difference between it and i1 Profiler, but it's very small differences. X-rites software and bundled hardware seems a bargain against them - and x-rite does provide support.

The two companies test charts/targets are a lot different in design. basICColor has some black divider bars, and white between the all patches where x-rite does not with theirs. Plus the basICColor patches are larger. So I'm thinking basICColor must have seen some residual color lag in scanning or something that required the black bars near some faint colors along with the white dividers against x-rite's patches? The basICColor Display part is far more adjustable and validates the results in Advanced mode, not that it could be a good thing at times with x-rite being simpler in operation there. Also, manual scanning is different: x-rite is top left to right and then down, and basICColor is bottom right to left and then up. Weird.

W. F.

 

[RogerB]

I think you are chasing shadows here. You are unhappy that your profiles show a smaller colour gamut than the manufacturer's profiles, but (for the same printing conditions) the colour gamut is determined by the ink, not by the profile. The fact is that the Cone inks give a smaller colour gamut than the Epson inks on the same media. Here's how Gamutvision compares Cone's published profile for Epson PPPG with Epson's profile. The difference in gamut volume is very obvious.

If you want to compare your profiles to someone elses you should really be comparing them with an equivalent Cone profile.

Unfortunately your prints of Bill's Balls don't help much either. They just show that the Cone inks are very different from Epson inks and that printing using Cone inks and the Epson profile gives pretty poor results. To really evaluate your profile you can start with a software tool like ColorThink or Gamutvision. If you are very serious about evaluating accuracy then the real test is to print a range of known colours, to measure the printed colours and compare them with the file values. Or you could just print a standard test image and see if it looks nice!