Canon Pro9000 ll colour shift noticeable in greys

[Emulator]

A quick question, which is the best/most accurate way to convert an A5(or other) print into a good image on the forum.

 

[Emulator]

I see no one has answered my question!

The image above is a photo (RAW) of an A5 print of the Evalation tif image, converted to sRGB and tif, reduced and converted to JPG.

This is the result of my latest effort in sorting out the grey shift problem.

I'll submit this and have a look at the image before writing any more.

 

[Emulator]

A bit on the dark side.

 

[ThrillaMozilla]

How to submit images: http://www.nifty-stuff.com/forum/help.php#img . Your link points to a nonexistent page.

 

[Emulator]

Hi TM

Yes I had read all that, it's the info about reduction from 50M files to a few K jpg, without too much degradation, that I was curious about.

In this version I used "Bicubic Sharper (best for reduction)" in PS.

The revised version above is a bit better.

 

[Emulator]

PeterBJ

Thanks for the helpful advice:

If you try to interchange the cyan cartridges, you will get an error message "One or more cartridges are in a wrong position" or something like that. That's because the chips are color coded. According to Canon that's to prevent users from making such mistakes. Not a word about making refilling more dificult

Useful to know when you buy used empty carts.

I found I had an OEM Cyan unused and a refilled Lt Cyan, so I tried those. They produced similar results, so I was inclined to think it was not the carts.

Subsequent experiments resulting in the print above tend to support this view.

 

[PeterBJ]

Thanks for the feedback.

I was convinced it was an ink flow problem, and the light cyan cartridge was my prime suspect. A test result is still valuable information, even if it is not what you expected. So most likely something else must be causing the problem. I remember having seen threads about other printing problems like banding or smearing at the beginning or end of the paper. These problems were caused by curled photo paper, so is your photo paper perfectly flat ?

I know very little about profiling and Photoshop, so if the problem is software related I have no idea of what is causing it.

When uploading images to the forum, only 80 kB is available for each image. To minimize artefacts and loss of detail due to compression, it is a good idea to crop the image to show only the area of interest. That's why I cropped my cyan starvation test image to show only the color transition.

 

[Emulator]

Hi Peter

Yes, the paper is the flatest I have ever come across!! This is one of its more attractive features, it is not too stiff for 280g and does not curl with time.

I said earlier I would summarise the changes I had made to produce the improvement in the greys colour balance. So here are some thoughts for those who are interested. They may not be appropriate to other equipment setups.

The suggested process for creating a profile is, initially to test print all the available OEM printer paper type profiles on the intended paper and to select the best one to use throughout, including subsequent printing.

This can be attempted, by printing a colour spectrum file. I used MediaOptimiser_V2.tif, which can be found on the Web. This file has all the colours and shows the effects of the different Canon media profiles. It showed emphasis in different parts of the spectrum, including some areas of excess ink, on the ink/paper combination I was using. I originally selected Photo Paper Pro ll as my media setting, rather than the suggested Glossy paper setting.

In my revised set up, I have changed this to Plain Paper.

Bearing in mind that I am using a heavy weight resin coated, micro porous paper with, I believe, a built in whitener, this may seem strange. It did to me. But I remembered a Web review, which referred to media profile selection and stated that the best option found was plain paper.

It is difficult to fathom the range of affects the choice of OEM media setting has on the overall process. Does it affect ink volume laid on the paper? Highly likely. Does it affect colour profiling and conflict with the profile you are creating? Possibly.

One of the issues with the process is that you are confronted with so many setting variables, some of which interact, that it is almost impossible to try all the combinations or keep control of the resulting comparison data. So luck plays a part.

The next issue that became apparent, in the profiling process, (having printed four sheets of colour patches), some measured colour patches, as viewed on the calibrated, comparison display, looked very different from their intended colours. It is expected that there will be differences, that is why you are profiling, but some colour patches stood out. The usual reaction is reading error, but careful checking ruled this out.

This problem resulted in the consumption of a great deal of paper and time. Initially, I thought it was all part of the profiling process, but the end results proved unsatisfactory.

I tried many ideas to solve the problem without success. More or less out of the blue (pun intended) and with a little previous experience, I thought of trying a clear film sheet between the spectrocolorimeter head and the patch printouts.

The spectrocolorimeter results were quite surprising. Certain patches became remarkably different in measured colour and very much closer, visually, to their correct shade, while the majority were little, if any changed. The resulting profiles were now looking good. The greys were now looking grey.

The crystal clear film I used is about 0.002 inch thick and as yet, of unknown plastic, it appears similar to that used by florists. (I have subsequently bought a metre square from a florist and it performs similarly). To the eye, it has no visual effect on the colours. My assumption is that it is preventing the spectro reacting to the UV end of the blue generated by the whitening agent added to the resin coated paper. Whether the software has a switch to correct for whitened paper, as speculated in my earlier post, is irrelevant if the film removes the problem.

The film is also beneficial in providing a smoother transition of the spectro across the paper when strip reading.

The remaining factor in the application of the profile was the selection of rendering intent and this seems now, to have a closer relevance to the expected influence on a printed colour patch scale and is more a process of individual choice, than it was. I am currently in favour of using of using Relative Colorimetric with Adobe RGB.

Significantly, I did not have to make any adjustments to the new colour profile (colour or tone sliders in SpyderPrint) to produce the print, unlike the earlier profiles, which tended to require extensive and often unsuccessful adjustment.

In respect of the original problem, the grey colour shift after the first half inch of printing, this appears to have been largely removed, by the change to plain paper in the media setting.

 

[Emulator]

Anyone interested in Canon EasyPhotoPrint Pro will find the Canon tutorial at:-

http://files.canon-europe.com/files/soft33490/software/EPPProV2(Metric).pdf

interesting, particularly in the later sections, it describes many of the settings in more detail, exposing aspects that are not immediately apparent.

For example Rod, did you know they recommend borders of 45mm for greyscale printing?

Regards

Ian

 

[DSchell]

I have recently had the same problem, with the same printer and same inks.

I have had the printer about 1 year, and I started using OctoInkJet inks as soon as the first OEM cartridges needed refilling.
I haven't tried any major photo printing until this month, and my first attempt was a close-up of a blue poppy, which shows a tiny bit of magenta.
The first print came out very magenta.

I posted about my problem on DPreView, and followed a suggestion to print a test photo (MatrixLarge from a UK web site that I'm not allowed to insert here) The result was similar to yours: all of the frames and the B&W images were shades of brown.

I asked tech support at OctoInkJet for suggestions, and Martin responded that he thought I might have put magenta ink into the lt magenta cartridge.
I don't think that happened.

A printout of nozzlecheck looks right.

So I bought a full set of Canon inks, and installed them.
As I changed each one, I printed the test image on 4 X 6 glossy.
No change in the result by the last install. (Probably because there was residual ink in the print head?)

I took out the cartridges and the print head and ran hot water over the print head until it ran clear.
I replaced it, and did a deep cleaning.

After that process, the test prints are no longer brown.
The problem is, I still don't know if there was wrong ink in one of the cartridges, or if the washing and cleaning fixed the problem.

I have now gone back to the OctoInkJet cartridges, waiting for the problem to reappear, so I can diagnose it.
Since then I haven't printed enough to see if the problem has returned.

Dave S.