Eureka! (ehh.... I think.....)

After reading what you guys have said about the quantity of ink vs type of paper on Epson printers, I have tried yesterday printing a target using a different paper setting than the one recommended by the manufacturer.

i.e. Red River recommends using the "luster" paper with their "Artic Satin Luster" paper. I have printed a target using the "Semi-Gloss" paper in the printer dialog, on the same RR paper.

And you know what? It ain't bad at all....!!!! The Lab test print came out about just right, yellow in the grays, just a little bit oversaturated in the yellows.

Now, I'm wondering two things:

1. Should I "edit" the profile and change the yellows? When I select, say, a yellow flower, how wide the range of yellow gets selected? Precisely just this flower's yellow? Should I try playing with the initial settings before creating the profile and add some blue? What's the compromise/advantages of each? I know that the initial settings will affect the whole profile, but what I'm not sure is how narrow or wide is the editor's selection

2. Is this going to work when I try other papers, or is it going to be a question of fiddling again..... 

Sorry guys, didn't have the time yesterday to do screen print-outs, I was on a production deadline.

For the Scanner, I used a Canon scanner, an HP Scanner and the best results (in terms of numbers) I'm getting from the Epson scanner.

I will post histograms and also try to see if I can find saturation and color settings in the initial screens of PP that could correct scanner lighting issues. However, by having tried this before, you sort of have to do this for every paper you profile.... 

Did it again with the scan I printed this morning (with color management off). Worst results than this a.m.

Just to give you an idea, I have attached a scanned version of the printout I get using the profile, vs the print I get using the RedRiver Paper profile which is pretty much right on what I have on screen. They're only 100dpi, but is should give you an idea.

Is there a significant difference when printing with Perceptual vs Relative colorimetric? I don't see much difference, should I???

Hy Ray,

Thanks for your input. And I just want to say thay maybe my answer will look defensive, it is not.... I just want to say what I've been doing in the hope I will be found wrong somewhere...... 

First off, I should say I'm getting good results with Epson paper with Epson ink using Epson's preset. I have purchased PP to depart from manufacturer's paper, which I though (from everything I've read around) was possible. And I calibrate my monitor before every attempt at defining a new profile... I've ben caught before.....  So since I'm getting a "match" print with this, I assume my workflow is ok, or should I??? 

I'm sticking to glossy and semi-gloss paper. I've heard that the R1900 is not all that great with matte paper and, in any case, my photo clients all want glossy or semi-gloss. And I have to say that I really like the quality (and the price...) of the Redriver paper. They have great profiles too, so what I'm doing with you guys is to try to at least come close to those, as an exercise to learn what I'm doing wrong while using Profile Prism. Once I master Profile Prism, I'll be more adventurous with other papers.

Thanks for the "High Speed" vs other. Will remember this and make sure it consistent.

I'm using the PP supplied test image, as well as Bill Atkinson's Lab Test image to check colors.

Could you explain a bit more about prophoto? I'm using Adobe RGB, but I know that Lightroom works with Prophoto (although I'm testing everything from Photoshop, I'd like to use the profiles with Lightroom as well, eventually... when it works.....)

The images I print with the profile generated from Profile Prism have the same issues in PS or in Qimage

Thanks for your reply guys.

Terry/Fred, you are correct: there is a "No color management" option in PS that I had not noticed. I just re-printed a target with it. We'll see after it dries. However, at first sight, it does not appear significantly different from the others I've printed. But we'll see.

Also, I've tried the demo version of Qimage, same problems.

Would you guys know why the full black and full white patches of the IT8 target always come up with an X??? Is this normal?

I've been trying to use profile print for over four months her without success!!! Bought a new printer (Epson R1900, don't tell me this one is not a wide-gammut enough...!), a photo scanner (Epson Perfection 1660, I agree it's used, but does very good scan!) Bought a copy of VueScan and piles an piles of paper and inkjet cartridges later, I still get awfull results!!! If all of you are getting good results, I figure the problem ought to be between the chair and the computer, but can't figure it out.....

Here is, step by step, what I do:

Start from a calibrated screen (Spyder2). I'm running Vista Premium, 64 bits. In this case, let's say I'm trying to calibrate some "Red River Paper" Artic Polar Luster. I know profiles exist on the RR website, but I should be able to come close with Profile Prism, shouldn't I???

Print a 5x7 target with photoshop: select "Printer manages color", and in the printer I choose (as recommanded by RedRiver) Ultra Premium Photo Luster as paper, make sure I'm not printing High Speed (unckecked) I'm using the Gloss Optimizer (checked), selected ICM and checked the Off button (No Color Adjustments). I print the target. Let it dry overnight. I've been using the Bright target, incidentally.


Next day, I open VueScan, put the IT8 target on top, my printed output below and a piece of black cardboard over the whole thing. Within VueScan, I scan at 48 bits, 400DPI, Output is a Raw file, 48 bits, Raw compression off, and Raw size reduction =1. No filters are selected, and I understand that what is selected in the Color tab is irrelevant since we're using raw. I do three or four scans, and then use the last one.

Now, within Profile Prism, Device to profile  is Printer, Check that Reference Target is the same as my IT8 target, that Printer-Target-Bright is selected, and that in Initial Adjustments, everyting is set to Normal, no biases.  I open the .TIF scaned file, select the corners of the IT8 target and get "Image properly exposed", 4% While Balance accuracy, 3% lighting variance, only 2 patches at minimum/max brightness (the extreme white with a red X and extreme black with a black X on the gray scale at the bottom).

After selecting the corners of the target, I get Exposure 224, and 5 patches at minimum/maximum brightness. Both histograms look similar to those in the instructions. I click "Create Profile". I then get Paper Dynamic Range = 84.3, Lab space coverage = 16.3%, 1 pass smoothing, and all profile usable ranges 0-255.

Yet!

When I print using the generated profile, I get a wayyyyy over saturated print, with reds and greens in particular, and a neutral gray that is yellow-ish.

I know I can adjust those and I have been playing with the "Initial Adjustment " settings. I have come close, but nothing really close. I've tried to edit the profile, but what I'm getting is too way off to adjust one color at the time......

So!! What am I doing wrong? Any clues???

Mike, I think we're confused here.

In the scaned image above, the picture with the color shift (magenta) and banding, the one on the left, is what I am getting with Profile Prism. The picture on the right, more accurate but less saturated, is what I'm getting from the Canon profile. It is not Canon that's taking true red and shifting it to Magenta, it is Profile Prism that does that and that's what I'm puzzled with...

I have not been able to get some Canon Pro paper around here. I've ordered some. In the mean time, I have purchased some Ilford paper. Already better, but now, I'm getting a greenish hue once I print with Profile Prism generated profile.....  

Here is a link to the 177Mb Tiff file scan I'm using to generate the profile:
http://francoisgagnonphoto.com/ip4700_ilford_galerie_classic_gloss.tif

Also, I'm not sure we're talking about the same printer model. The ip4700 is about 100$, and uses the Canon Chromalife 100+ inks, the same as their Pro 9000.

Guys, no matter how good/bad color accuracy might be as compared to the monitor, I'm looking for a balanced output. There is no reason for the PP profile (left) to give me consistantly reddish pictures (not only slightly, see attached), when the PP profile (right) gives me consistantly a correctly balanced profile, however too contrasty and undersaturated. Look at the color band (third row from the printout): with Profile Prism (left), cyan is almost absent. However, the printer is capable of producing cyan, as seen with the Canon Profile (right)


There's gotta be a way with PP to get at least what the Canon profile gives me. Capability of the printer is at the very least that. If I cannot get better than the Canon profile, then so be it. But I would have hoped that PP be at least capable of giving me the equivalent. Capability of the printer has nothing to do. I'm not trying to get the equivalent of a $5000 printer out of that $50 printer (which by the way is $100, not $50). I'm trying to get what I know it is capable to do (because I've seen it produce it!) then tweek it to my taste.

Since the last post, I have also used Ilford paper, with the Ilford profile. The color space that paper is capable of is much better, but still colors are off. 12.5%, according to PP. However, this one gives me a greenish bias (as opposed to a reddish one with Canon paper)!!!

I'm getting only 5-7 clipped patches, excellent exposure and excellent contrast (according to PP after reading my scan).

Why, oh why can't I get a reliable profile?

I have tried to use Canon paper with one of the profiles that came with the printer (Photo Paper Plus Glossy). The prints are OK looking, slightly undersaturated, slightly over-contrasty. Which is why I wanted to have my own profile, adjusted to my taste.

I will play tonight and try to adjust the Canon's original profiles (making a copy before, of course...). Maybe that will be an easier start to get to a result.

Also, Mike, if the printer gives me a consistant color shift, whatever it's quality, there's gotta be a way to correct it, no?

Ok, I played with the metamerism compensation. It did help a bit. I did several print, including one witht the slider all the way to the right.

Still, on the test photo included with PP, the skin tones of the hand and of the lady have a pink/magenta cast, and the D of the words "The Digital Dog" is dark green instead of the fluo green it is on screen. The background of the hand holding jewels has a shade of green to it.

I wish I could scan it and show you, but that's exactly the point, colors are not accurate!.... lol!

Don't judge the timestamp! I am reporting over two weeks of trial and errors/unsuccess..... :-\

Thanks for help and your replies guys,

First, for the monitor, I am aware that this does nothing for printer calibration. What I wanted to absolutely confirm is that, if the color on the screen didn't match the printer color, it was not because the monitor was poorly calibrated. Having eliminated that possibility, that's now the printer's (or profile) fault. Otherwise, the first question you would have asked me would have been: Is your monitor calibrated? 

I'm using original Canon paper (Glossy II). Canon inks. Setting CS4 the exact same way: photoshop manages color, select the PP-generated profile, within the printer driver, select "none" on color management, and select the proper type of paper and high quality..

I tried Qimage. Same result.....

I will read indeed on metamerism! Hopefully that will help.

Hi!

I have received my copy of PP abit over two weeks ago and am no further ahead.....  All the profile generated give me something way oversaturated, contrasty, with an overall reddish/magenta slight color cast...

I am using Vista 64, with a screen calibrated three times! with a Spyder, Photoshop CS-4, Lightroom 3.4. I have a IP4700 printer, and two scanners: one that is integrated with my Canon MP750 multi-function, and one that is an old Scanjet 5300C, from HP. I am using VueScan. I have tried PP with output from both scanners: same result.

First thing, I tried to calibrate my scanner from within Vuescan. Used VueScan and tried the IT8 target that came with PP. Results were unsatisfactory: the colors as they appeared from the IT8 target were not as vivid on screen. Tried with PP: scanning the IT8 target from VueScan, with the recommended settings, as per the the PP help recommandation. When I read the scan from the MP750, I kept getting a high amount of "patches with an X" and clipping (like over 60). I was puzzled, this is a decent scanner, I've done photocopies of photos with it and it was pretty good! Decided to do it with the HP scanner. Got a good scan (less than 6 items with an X), all other settings were good. Generated the profile and tried to scan with that profile using VueScan: very washed-out colors.....   

The hell with it! I know PP looks at the differences between expected results (from the IT8 target) and the printed output.  I generated a printer output using the recommended settings from the PP Help using Photoshop to print (you know, Printer manage color, and within the printer driver, set color management to "none"). Let the print dry overnight.

The next day, used BOTH scanners (without any color management, let it to Factory within VueScan) to try to generate printer profiles. Got good scans, according to PP (good white balance, exposure, only 5-6 squares marked with an X). Generated the profile both times. then printed the PP test image from within Photoshop: ouch! Pink/reds come out wayyyy to bright, contrast is too strong. Tried this procedure several times with the same kind of results.

Tried to blame the monitor, re-calibrated, tried PP again, re-calibrated screen again, tried PP again..... same thing!

Anything I'm doing wrong here? Is this as good as can be expected with Profile Prism?

Thanks Terry and Ray. I just purchased a licence. Now, I need to wait for it in the mail.

I'll give it a try and follow-up in this tread!