Color cast when printing B&W

Welcome to the world of black and white inkjet printing!

I don't have any specific help to offer, because I'm not familiar with Canon printers, but newbies should be warned that black and white printing with inkjet printers is extraordinarily difficult. People go out and buy $500+ RIPs, expensive color profiling products (probably what you will need) or even second printers with a dedicated set of third-party grayscale inks. Inkjets should really come with warning stickers saying "abandon all hope of black and white printing, oh ye who buy this product!"

My HP 932C does the same thing. There is no way it will ever do good B&W, because it uses color inks to produce the shades.

I do have hope for my new Epson 2200, though. The B&W sample on matte paper included with the printer has no cast to it. The Epson 2200 has photo black, matte black (photo and matte are mutually exclusive, though), and light black.

You can get an Epson 1280 for $300 new ($400 & $100 rebate) and then use an aftermarket all-black ink set in it.

I'm not familiar with Canon's printer, but if it has a separate black cartridge and the driver allows it, select Black-only ink. If you do that with Epson, a "warning" box pops up telling you the world will end. But by golly, that print will NOT have a color cast! Good luck!.. Beau

My HP 932C does the same thing. There is no way it will ever do good B&W, because it uses color inks to produce the shades.

So does your eye. In bright light you use mainly the cones in your retina but you can still see a neutral grayscale.

I do have hope for my new Epson 2200, though. The B&W sample on matte paper included with the printer has no cast to it. The Epson 2200 has photo black, matte black (photo and matte are mutually exclusive, though), and light black.

Better look again. The sample print you refer to may have no cast, but it has wicked bad metamerism! I.e., it has no cast in a specific light. Change the light source and it goes Technicolor on you. If you don't notice this get your eyes checked - I brought the Epson sample to work and everyone with a cube on the inside of the building (i.e., with fluorescent light) commented on how it looked pinkish. Those with window cubes (i.e., daylight) commented on how it looked greenish.

The 2200's black and white metamerism is a hot topic of conversation on the Yahoo Digital Black and White printing group. Epson even sells a solution for it in the form of a $200 RIP, which some people have tried and reported good results with. Other people use the Imageprint RIP to fix the 2200's black and white problems but it costs over $500. So despite what you say at the beginning, apparently it IS possible to get good black and white from a 7-ink system if you are willing to pay for it.

BTW, the Epson "light black" ink is brown, not gray.

I'm not familiar with Canon's printer, but if it has a separate black cartridge and the driver allows it, select Black-only ink. If you do that with Epson, a "warning" box pops up telling you the world will end. But by golly, that print will NOT have a color cast!

True but then it will have a fine pattern of black dots, which many people will find annoying and distracting.

The other problem with black-only printing is that ALL dithering schemes represent a tradeoff between color (or grayscale) resolution and spatial resolution. If you want to print at 300 DPI then your printer has to have high enough resolution to fit a complete tonal range into .000011". But most printers can't. The Epson 2200, for instance, claims a maximum resolution of 2880x1440, which means it can fit about 47 levels into that area using black-only. But the human eye needs 100 or so levels to see perfectly smooth tonal transitions.

So something has to give - either spatial resolution or tonal range. On the Epson it's tonal range - Black Only prints are too contrasty. I don't know what the Canon driver does but the math is simple and inexorable and must be served.

Peter, I have the B&W sample with me at work here.

I see no difference between daylight coming through my windows and the Ott-Lite Vision Saver desk lamp. I don't see any metamerism difference between the flourescent light in the computer room and in my office. The cloud layer just above the Eiffel Tower seems to have some light cyan to it, but it's like 95% grey, 5% cyan, and it's the same under the various light sources.

My sample simply doesn't have a color shift.

I wonder if the images used for the sample prints are available somewhere with printing instructions. Then we could exchange the prints as a standard of reference.

I also wonder if some of the metamerism problems could stem from printer calibration. When I calibrated my print head alignment, I had to run the #3 high-density pattern five times to get it right. I figured I had it right when the two outermost patches had the same grain, as I couldn't tell from the inner patches from each other. It would be nice if Epson had an extended pattern sequence.

It could be that the printers used at the Epson factory don't have perfectly identical calibration, so there is a difference in samples. Thus, yours has hideous metamerism, and mine doesn't.

How do you get the printer to print solid blocks of color from the individual heads? The nozzle test pattern (jagged trapezoids) appears to have the "light black" really printing as a grey.

I see no difference between daylight coming through my windows and the Ott-Lite Vision Saver desk lamp. I don't see any metamerism difference between the flourescent light in the computer room and in my office. The cloud layer just above the Eiffel Tower seems to have some light cyan to it, but it's like 95% grey, 5% cyan, and it's the same under the various light sources.

I think it's you.

On the same Epson sample print (which is apparently the same one as yours - Paris scene) that I and all my co-workers see tremendous metamerism, my wife doesn't see any at all. She's not colorbrind as far as she knows, but she just doesn't see it. Some people are like that, and perhaps you're one of them. I should also note that my company makes display systems used in medical applications so the other engineers I showed these to are probably more sensitive to color shifts and similar problems than laymen. But probably NOT more sensitive than photographers, artists, art directors, etc.

The metamerism problem on the 2200 is very well established and a great deal of discussion hes been devoted to finding workarounds for it, but some people don't seem sensitive to it. That's why on the Yahoo Digital Black and White printing group when someone says that this RIP or that workflow eliminates it the first thing we ask them is, "have you ever seen metamerism? Would you recognize it if you saw it?". If the answer is "no" then we take their claim of having found a fix with a grain of salt.

How do you get the printer to print solid blocks of color from the individual heads? The nozzle test pattern (jagged trapezoids) appears to have the "light black" really printing as a grey.

That should be a clue right there - the Epson "Light Black" is a kind of sepia brown - a "dark yellow" actually. Daniel Staver, on this forum, recently posted some images here using that ink to demonstrate this. In fact the Imageprint RIP apparently eliminates the metamerism by taking advantage of this - one of the worst metamerism offenders is the regular yellow ink. So the IP RIP doesn't use the yellow ink. Obviously, without the yellow ink that only leaves the cyan and magenta, which would create bluish prints, but the IP RIP balances out the cyan and magenta by using the "Light Black", taking advantage of its yellowness.

I would seek some second opinions on your sample print - the usual problem reported is that it looks pinkish under fluorescent light and greenish under sunlight or bright daylight.

The cloud layer just above the Eiffel Tower seems to have some light cyan to it, but it's like 95% grey, 5% cyan, and it's the same under the various light sources.

Cyan raises an interesting possibility. When did you buy this printer? Can you make black and white prints on it that have the same neutrality as the sample print?

The reason I ask is that cyan is not the usual metamerism shift reported. (pink and green are) But Epson recently released a RIP of their own that is alleged to fix the metamerism. And if it uses the same approach as the IP RIP then if the print had any color cast it would be slightly bluish, which in light areas could look cyanish. So it occured to me they may be shipping their printer with sample prints made with their new RIP - maybe not technically "false" advertising, since it was, after all, made with the 2200 - but a bit misleading since it's not made with the sw that ships with the printer. Just a thought.

I am considering the Canon i950 and through my readings from the internet, I have found this suggestion which might help you.

Site: http://www.photo-i.co.uk/

Select the Canon i950 review and go to page 7 "Black and White" in the drop down menu. The author mentioned "The slight cyan cast can be tweaked out in almost any imaging application, I think this is often overlooked by people, they expect the printer to be 100% neutral from the onset."

Please let me know if this works as I would like to know the results.