MG6220 and ink from Precision Colors... Black is too black?

[crenedecotret]

Hi everyone,

I made a custom ICC profile for another member here. Before we started he mentioned that the black ink cartridge might be giving black levels that are way "too black" compared to the OEM ink. Nozzle checks were all good, and there are no feed issues/leaks. His printer settings were all good. Printer color management turned off, etc. He is using Canon paper so that's one less variable to worry about. The only thing that is not OEM is the ink itself.

He printed the color patches, sent them to me... I created the profile with what seems to be good results using Argyll and a Colormunki. Very very low avg DELTA-E errors, almost the best i've seen (avg. = 0.328429, RMS = 0.452294). I evaluated the profile the best I could, soft proofing in both Gimp and Photoshop. I use a calibrated monitor so the soft proofing is somewhat valid. On profiles i've done for myself, the soft proof on the screen matches the actual print. With this MG6220 profile I made, soft proofing shows shadow details that I am unable to achieve with my current setup (IP4500 and hobbicolor inks). This leads me to believe that the black and grey inks are absolutely fine.

Now when he tried the profile, he seems to still be having the same "black is too black" issue. He is seeing this on the "PDI test image" that is circulated on the web. I suggested he compare an un profiled print and a profiled print outside in daylight. He did mention the profile print looked better, with more saturated colors, etc.

Since he is not entirely happy with the results. I thought I might consult all of you here....... I suggested the prints dry for a day or two and look at them in daylight around noon to be able to accurately judge the colors and appearance. (at home, I use an ott lite). I also suggested that since his monitor is not calibrated, the colors/appearance he sees on the print might be the more accurate one. Another suggestion was to drop the PDI test image and use normal photographic material more likely to contains colors that the printer will actually be able to reproduce.

I believe this member was hoping that an ICC profile would correct what he perceives is an issue with his black ink. Some combinations of paper, printer and ink are simply hopeless even with a profile, but I don't think this should not have been the case with Canon paper and IS ink that is known to be very close to Canon's.....

Is there any chance that the black ink he was provided might not be the right one for his printer? Mikling? I'd love to hear your opinion on this, have you had any issues with MG6220?

To everyone else... Would anyone with the same printer and inkset like to try the profile? I can distribute it freely since it was done with Argyll.

 

[Smile]

This is not an advertisement, but I can make a profiles too. I can do further than that I can send custom targets, and then make RGB print certification that will show what is going on on grayscale, color etc.

Profiles will be measured on certified i1 Isis XL. I also have BARBIERI electronic LFP but that is only used for other things like transmissible measurements.

Since I do my own paper fluorescence tests, I'm not bothered with what is better UVcut- or not profiles. I just measure the paper and measure accordingly. I could also build a few different profiles to rule out profiling package influence.

Ink can't be too black, It's the high dmax we are after. Like in scanning negatives/slides. So PM me if you want to arrange this. Also the golden rule is print nozz check then print something like a purge page for all colors before printing targets, do the same when printing certification targets. Skipping this rule make all sorts of weird things happen.

You mentioned dE of avg. = 0.328429 but what target did you use. This is very related to total amount of patches. Did you make few measurements then average the data?

 

[crenedecotret]

Argyll generates it's own targets, according to the specs you give the "targen" command. In this case I specified around 1500 patches. Each argyll target is different and the patch values are recorded in a CGATS format file. I don't have the same sophisticated equipment you have (ISIS and Barbieri.. droooling here). I simply use an X-Rite Colormunki. I did one set of measurements, created the profle and then use "profcheck" to check the profile... any patches with high Delta E I re-read just to be sure it was not a measurement error.

Cool that we have a professional profile maker on this forum! I've done maybe half a dozen profiles so far for members here, free of charge. Not being a professional and doing these with entry level equipment, I would not feel comfortable charging for the service. People I've done profiles for were unsure of what settings to use to print the targets so even with guidance, there is always a chance that a profile would be unusable. When something is free, you can't complain to the person offering the service LOL! But so far it's turned pretty good for everyone

This profile was not done for my own printer, but for another member. He mailed the targets to me. Before he printed the targets, he ran a few test prints to test different media settings in the driver, he also ran a nozzle check before printing. These all looked good.

Personally, I agree with you... high DMAX is a good thing. The person I did the profile for was getting very deep blacks with the Precision Color ink simply printing out with the Canon driver. With the ICC profile, there was basically no change in that regard, but in his words "I don't think the colors could be any better". It seems the blacks were not as deep when he was using OEM ink.

He was testing out with the PDI test image that can be found here http://www.northlight-images.co.uk/downloadable_2/Test_Image_2.zip The image contains a Kodak gray scale. What he mentioned is that the darker region of that gray scale appear to all go to black. The test prints were done with Photoshop with Relative Colorimetric+BPC ON. I suggested he try it with perceptual as Ive seen shadows get "destroyed" using RC+BPC now and then especially with these test images that contain a huge range of out of gamut colors. Haven't heard back yet, I'm sure I will soon.

 

[PeterBJ]

crenedecotret wrote:

...Is there any chance that the black ink he was provided might not be the right one for his printer?...

Could your customer accidentally have refilled the CLI-226 GY cartridge using ink for CLI-226 BK? I think that would cause problems with greyscales being totally wrong.

 

[Smile]

This profile was not done for my own printer, but for another member. He mailed the targets to me. Before he printed the targets, he ran a few test prints to test different media settings in the driver, he also ran a nozzle check before printing. These all looked good.

I had some targets get damaged in shipping, they look normal but profiles were wrong. So it could be you need to re-print and make sure it's not the bad targets.

Also like I said before I always print something then print targets, certification targets etc. I don't trust the printer being accurate just sitting there. There are many variables, internal pressure inside CISS or inside cartridges, etc. When you print something "variables equalize" and what you want to capture, not a "cold started printer condition".

Certification report is very detailed something like this http://www.turboprint.info/support/viewtopic.php?f=26&t=1300
Also there is some gray scale evaluation details not listed there. One of them is shadow details analysis. I took the idea from some German site, they had done some of the similar tools online, they are not free. But evaluation version is free. Germans use different formulas is some sections of the analysis, and I've found many serious bugs in their implementation colorcheck-online.de.

I've spend countless hours with contacting different ICC, and color science professors around the world to get the spreadsheets with working color conversion, dE, etc formulas. It was very hard to make it work.

I can send you my certification target (PM me) and certify you clients prints for free too. If you take the chore of international shipping the targets to me. Or you can take online trial on German site.

Warning! They evaluate grayscale entirely different than I do, they adapt the white point to paper white (in other words the use delta S formula for grayscale evaluation in adapted white point equation) that is not very good if you judge your print near a really neutral print or struggle to get as neutral as possible evaluation. Also there are so dE calculation errors for some colors (bad formulas).

Personally, I agree with you... high DMAX is a good thing.

High DMAX ensures paper is able to show deep blacks.

It seems the blacks were not as deep when he was using OEM ink.

Did you measure the OEM ink and refill ink blacks ?

Before RGB certification I was struggling with no way to know what setting was better, perceptual, or rel colorimetric etc. Now I just waste a single A4 page print 2 targets on either end of the page let it dry for 48hrs and know what to select when printing. The certification tells me if blacks are crushed or not, grayscale is OK or not etc. It's also a good indicator for paper / ink combo quality too. If one has different printers and inks it quickly shows what prints better. EPSON usually has more neutral colors, but it's because inks are more neutral for EPSON at least in those that I had certified.

For example some InkTec ink despite being good in terms of no ink clogs, the black ink is not entirely black, it has some red/brown cast in it. Perhaps they improved lately I just seem to remember what I had seen. I now tend to use OCP ink from Germany.

 

[Smile]

He was testing out with the PDI test image. What he mentioned is that the darker region of that gray scale appear to all go to black. The test prints were done with Photoshop with Relative Colorimetric+BPC ON. I suggested he try it with perceptual as Ive seen shadows get "destroyed" using RC+BPC now and then especially with these test images that contain a huge range of out of gamut colors. Haven't heard back yet, I'm sure I will soon.

Try, Perceptual but try this image instead http://rghost.net/private/45152018/eb7b981d06390f5afbf321ad4f385759
All patches under proper lighting such as daylight should be seen.

I only use the PDI target for black evaluation, the blue vase on the top right corner shows this glass translucent effect, xrite prism engine makes it look like plastic not glass.

Best grayscale evaluation I've come across http://rghost.net/private/45152059/9be0139253cb826c192e0f9914557c17

 

[mikling]

First, I'd like to know what inks had been used. IS recommends using CLI-8 colors of the 226 carts catering to refill shops who don't want to have a plethora of various colors to refill to people who just want "color". I went a step further and I blended to the best I could what was available and trying to match the 226 OEM inks which are actually quite a bit off compared to the CLI-8. I do not ship CLI-8 colors for the 221/226 carts. I am currently working on another set for the newest 251/551 carts which are different again from the 221/226/521/526 colors again, so take note.

My 221/226 inkset is labeled as the 221 series. I send out ICCs with these inksets and have noted no issues of black not being black. The photoblack is the same used in the CLI-8 for simplicity.

If the black was insufficient, it would make itself known in softproofing and you'd have picked it up.. Of course the only way really is to create the profile and print for verification right by your side for true verification and you'd see this in the softproofed images. Maybe the printed sheets should be mailed back to you again with another set of targets to verify that nothing had changed. Also did the user flush the carts before refilling to print out the targets? Might the inks have shifted? At least three refills should have been done to stabilise the composition before profiling. If you refill once and then profile, then the ink composition would have shifted.

Could it be that the dynamic range sought cannot be matched by print, what is the user using as a reference?

 

[crenedecotret]

PeterBJ, it's the first thing I asked him before we began the profiling process and he mentioned his problem.

Smile: unfortunately I never got to mesure the OEM ink. I'm in a different province as the printer, so I have to take the person's word that there was a difference between the IS ink and the OEM ink. It's possible...... my hobbicolor ink isnt a complete match to OEM, but it does have better gamut too. I looked at the chart on zedonet/turboprint. Very very cool stuff!! I had no idea profiles could be certified this way.... If I were running a commercial service I would definitely look into it! I'm not sure the person I did this profile for will want to take this route, he was quite happy I was doing the work for free. Off topic a bit, but how do you like printfab? I'm using Turboprint on my linux installation.. Printfab is just Turboprint ported to Windows..

I'm trying a new direction now... I've been researching a bit, and there exists more than one version of the PDI image. I found one that looks exactly like the scan the person sent me.... the kodak looks plugged up starting at patch 14.. the black patch in the color checker is extremely dark compared to the gray patch next to it. The whole root of the problem could be the test image....... the issues are visible even on the screen and the printer might be printing exactly what it's supposed to. I sent him the version from Dry Creek Photo that fixes these issues and asked him to print it in perceptual mode.

I also sent him the Datacolor test image... If the print can retain good shadow detail in the darker parts of the black and white spiraling staircase, I'll be pretty convinced that his printer, ink and profile are all fine.

 

[crenedecotret]

thanks for the answer Mikling.

For the carts, he mentioned that he purged then before filling with the new ink. I had recommended he print at least 5 to 10 full color pages to not have too much of a mix in the printhead. If there is any shifting, it should be minimal. I think he purchased your ink around mid-march. Would you mind if I PMed you for your profile for his ink? If you have one for Canon Photo Glossy II, I would send it to him to have him compare the prints with the custom profile. He did pickup on the photo black being different than the OEM ink. Knowing it's CLI-8 black, at least I know none of us is crazy. Not to say the black is not sufficient.. I have a CLI-8 printer and have no issues with black on my prints.

I've purchased your inks when I still had my Epson and it was top notch, so Im not worried there. I told him I had good experience with you in the past and that you completely stand behind what you sell. I was pretty sure you would chime in on this thread with a suggestion and you did

The soft proof actually looked very very good on my screen. Like I mentioned above, Im getting a bit suspicious about the test image he's been using, considering that most of them out there show the exact same flaws on my screen as are seen in his print.........

 

[mikling]

http://www.mediafire.com/download.php?1musa0e12fjxigc

the above are not with the CLI-8 CMY