Non OEM Inks - Why aren't they a "perfect match" to OEM? Can they be?

[mcm30114]

Here's a question to wake up this thread --

What tools do you use to help measure the L*A*B values of an ink? If I have a tool like a GretagMacbeth EyeOne Photo, can I use this to help me understand how to set print profile adjustments?

I have some OEM ink from INK4ART (BCI-6) which dont really print the way that I expect them to. The Monitor (Dell 2005FPW) is profiled and is pretty much dead-on as per the EyeOne. Printing a reference target for calibration with the printer (i8500) and the inks produced a target that EyeOne complained about.

I liked the ink cart L*A*B discussion above, and wanted to see if we could take it a little futher.

Thanks in advance

Mike

 

[Grandad35]

Mike,

The spectro used in the I1-Photo package can be used to measure LAB color value of either a light source (light bulb, sunlight, CRT, etc.) or a surface by shining its internal light source onto the surface and measuring the light that bounces back. There is a free program from G-M (Eye-One Share) that works with their spectro to allow you to make these measurements. Notice that I didn't say that it would measure the color value of an ink, only the color value of a sample (e.g. photo paper) which has been coated with the ink. This may seem like splitting hairs, but there is a BIG difference between the two.

The color that an ink produces on paper is very dependent on the paper. As an outlandish example, consider printing magenta ink onto a magenta paper - you will just get a more saturated shade of magenta. If the same magenta ink is printed onto cyan paper, you should get blue. Back in the practical world, even "white" photo papers are not true white and they absorb the ink differently, so the ink and paper must be tested as a combination. To further complicate matters, the color that you get from applying an ink to a paper depends on the amount of ink that is applied, as well as how it is applied.

Last year, a few of us tried to come up with a way to measure the colors of various ink sets by developing a method to coat the inks onto a standard paper so that the "ink" colors could be measured. The ultimate goal was to allow us to compare the color values of an unknown ink set with known color values from Canon, Formulabs, IS, etc. to see if we could identify the source of the unknown ink. We tried many different methods of applying the ink to come up with a method that gave the same (or close to the same) color values on every sample that was made with the ink, regardless of who made the sample. Unfortunately, every hand application method that we tried gave far too much sample-to-sample variance. The only thing that showed any real promise was a special test program that caused the printer to print small samples with each individual ink. Unfortunately, this program is only available (to the best of my knowledge) on the Canon i8500 and i9900 printers, so this project was abandoned.

Here are a few additional threads that you might want to read:
http://www.nifty-stuff.com/forum/viewtopic.php?id=496
http://www.nifty-stuff.com/forum/viewtopic.php?id=498
http://www.nifty-stuff.com/forum/viewtopic.php?id=499
As of this writing, some of the images on these posts don't appear, but I have e-mailed Rob to see if he can restore them.

With your I1-Photo package, you can obviously generate custom profiles for your printer/paper/ink combination (as you have already done). If you are just getting into this, read (http://www.nifty-stuff.com/forum/viewtopic.php?id=118); if you are experienced, don't bother. It isn't uncommon for the profiling software to complain about something when you generate a profile, especially with non-OEM ink. I just ignore the complaint and try the profile - usually it works reasonably in spite of the complaint. The latest I1 software also has a "profile editor" that will allow you to "tweak" a profile when it is off in certain colors - you might want to give this a try.

A little conjecture on why the profiling software sometimes complains. Consumer printers are only sent R/G/B values, and it is up to the printer's firmware RIP to convert these colors into C/M/Y/K. Obviously, the OEM optimizes this conversion for their ink set, especially how it handles the PM/M and PC/C transitions when going from light to dark magenta and cyan. If the OEM's PM is 25% of the strength of their M, and your PM is 15% of the strength of your M, obviously your transition from light to dark magenta will be very different than theirs. I remember one ink set where printing a magenta gradient without color management caused the magenta to be lighter at 60% than it was at 50%! This also meant that the printed color at 70% was about the same as it was at 50%! What can the profiling software do when it sees something like this (does it tell the printer to use 50% or 70% magenta)? It can't make a "silk purse out of a sow's ear", so it issues a warning and fudges the data as best it can. There is no way that a profile can properly correct for this situation. This type of problem is why the pros use custom RIPs (lots of $$) and expensive printers that allow each ink to be individually calibrated. (http://desktoppub.about.com/cs/rip/a/rip.htm)