Profile Prism and my Canons

[JV]

Grandad,

www.colorvision.com does list iP8500 as supported on PrintFIX. There is not much description of the PrintFIX scanner and the Canon LiDE 80 scanner appears fairly good from what I read so far, so maybee it should be Profile Prism unless I could mail you a target profiling chart printed on my printer for your spectrophotometer printer profiling.

JV

 

[Grandad35]

JV,
Normally, the only reason to buy a profiling system is if you plan on generating a number of profiles over a few years time. If you only need one or two profiles, the most cost effective approach is to have custom profiles run at a place like Dry Creek or Cathy's.

Your i8500 is just a narrower version of my i9900, so a profile should work about the same on either printer. As such, there are already 5 profiles for your printer and Kirkland paper posted on this forum:
Canon OEM carts
Old InkGrabber carts
New InkGrabber (G&G) carts
Atlanticinkjet (IS?) bulk ink
Alotofthings (Formulabs) bulk ink

I would first try these 5 profiles to see how they work on your BulkInkJetCarts - maybe one of them uses the same ink set as BIJC.

I would be willing to run a few profiles on widely used printer/ink/paper combinations for the group, but I don't want to get into a situation where it becomes a job - remember that my goal in all of this is to be able to print lots of pictures of my grandchildren at a reasonable cost.

BJIC certainly have attractive prices for their carts, but their web site doesn't mention bulk ink for us refillers and they don't say where they buy their ink. Do these carts have any markings that are similar to other carts (e.g. G&G)? What is your experience with ink feeding and "clogging" problems? Are BIJC carts commonly used by others in the forum?

 

[JV]

Grandad,

Thank you. I will try your 5 profiles as soon as I get my monitor profiled. If that does not work maybee you could do a profile of BIJC carts from Chinese manufactured #2 (www.inklabcartridge.com) with Kirkland paper.These carts look like a close copy of OEM carts. I have printed 44 photos +50 plain 8.5 x 11 without problems and soon will dissasemble an empty cart to see how close they are to OME carts. So far two persons from the forum have ordered BIJC carts, but there could be others if the results are good. Some people may want to start with BIJC carts before refilling.

JV

 

[CollegeDad71]

Grandad,

I have the PrintFix software. After profiling my Canon i960 the ColorVision test image following a cfriends.com refill kit looks the same as that printed with OEM ink. However my daughters printed macro image of a panzy has way too much magenta in it. The point of all this is that ColorVision tech support made the statement to me that PrintFix was not meant to function with 3rd party ink. I was at a loss for words for a few seconds. If not for 3rd party inks and paper, what is the point?? As there are other other posts in this thread on i960's having a magenta bias, is it advisable to approach this by makeing a color adjustment to the profile even though their test image prints as expected?

CollegeDad

 

[Grandad35]

CollegeDad71,

The first word that popped into my mind in response to their statement has to do with what I watch out for when walking in a field used by mature male cows. I agree with you that one of the main reasons why you buy a profiling package is the ability to use 3rd party inks.

There IS one scenario that I can think of where a 3rd party ink could cause a problem that a profile couldn't correct, and I ran into this problem with my first bulk inks for my i9900. The PM/magenta and PC/Cyan combinations must match each other and have color density ratios approximately the same as the OEM inks. The printer only gets R/G/B data, and the printer's internal software decides when and in what ratios to use the light and dark magenta and cyan inks. The light (photo) inks are used for the lighter magentas and cyans, the regular inks are used for the darker magentas and cyans, and both inks are used in varying ratios for the intermediate colors. The software that controls this transition is designed specifically for Canon's inks, so if the ratio of the color densities for your bulk ink is not reasonably close to Canon's ratio, strange things will happen in the middle colors.

In my case the PM was way too dark, so much so that the magenta printed at 30% was darker than the magenta printed at 60%. In my case, I could get the same magenta color by printing at (approximately) 20,40 and 70%. This will obviously drive the profiling software crazy, especially when you remember that the inks interact with each other such that 40 and 70% magenta (which look the same when printed by themselves) give two very different colors when printed along with 30% cyan and 30% yellow.

To make sure that your inks don't suffer from this problem, download the test chart below and print it with no color management (just like you do to print the test charts for your profiling software). After it dries, scan it in. If you have Photoshop, convert it to CMYK and split the channels (giving you 4 separate grayscale images for the C/M/Y/K channels. The cyan and magenta gradients on the cyan and magenta channels should have a gradually increasing density from left to right. If you don't have PS, post your scanned image on the forum and I'll separate the channels for you.

You didn't say, but I assume that you have already calibrated your monitor.

Another test is to print the test image using your new profile. The bottom gradients should be gray, and the middle grays are normally the most sensitive to showing a color cast. Don't expect perfection - even Canon inks and paper don't give a perfect gray.

Another thought - is it possible that your profile is working properly, but that your daughter's macro image doesn't have an embedded profile? To check this, convert her image to sRGB and see if that makes any difference when you print it. It is also very possible that the image of the flower has colors that are out of gamut for your printer, and that the shift in color has to do with how these colors are handled. What software are you using to print, and what "Rendering Intent" are you using - they can both affect how extreme colors are printed? If you have PS or Qimage, "Soft Proof" the image with your profile - does it look like the print? If so, your profile may not have a problem at all.

Just like the adage about getting old, "Color Management isn't for sissies".

 

[CollegeDad71]

Grandad,

A few days of vacation later, let see where were we... Yes the monitor was calibrated with Spyder2 Pro. My print of the above Color Test file with no color management show some obvious problems

I found the forum instructions for uploading images but as this is the first time, I'm not entirely sure how to get the split channel images side by side.

If this workes the CMYK Cyan chanel is on the left and Magenta chanel is on the right. I am going to include the calibration charts as well. No profiling on the left, profiled on the right

There are some obvious differences from the original below

The differences from original to uncorrected print are not surprizing since this is the input to the profiling software. The profiled output of that same file is a little depressing, one expect a match. The top bar of your color test image are obviously over saturated (RGB image) but the grayscale progression on the split chanel images seems OK to me.

Lastly, I am including the Test Image from Colorvision. This is with profiling turned on and it actually looks good compared to the same file with OEM ink (sorry I don't have that image with me at present).

If this has been addressed previously please feel free to send a link to that discussion. I really am not intending that you answere this to every individual who wanders into this learning curve.


I do have PS-CS. The imbedded profile for the camera image is sRGB but the output uses Adobe RGB(1998)

On the issue of Soft Proofs. PS looks for .PFS files but PrintFix generated profiles are .ICC The help system in PS doesn't seem to address this although the discussion expects that the output device profile can be used. How does one address this?

Thanks,

CollegeDad

 

[CollegeDad71]

Grandad,

Sorry, I didn't consider that the file names would not show in print, but since they show up in the URL link you will figure out that the top grayscale is the cyan chanel and lower is magenta. The images were scanned with an Epson 4870 that was profiled using MonacoEZcolor against a Kodak IT8 transparency.

CollegeDad

 

[Grandad35]

CollegeDad71,

In re-reading your post #14, I note that you refilled the carts just before running the profile. If you used virgin carts or carts that were completely purged of all ink, this isn't a problem. However, if the carts were previously filled with some other ink before they were refilled, you will not get consistent color until after they are refilled several times. Consider that there are several CCs of ink in the sponge, so refilling with a new ink will not even begin to change the color for a number of prints. Even then, there will be a gradual color change as the old ink in the sponge is replaced with the new ink.

Before printing a set of targets, it is important to run a nozzle test and an extended print test to make sure that everything is printing properly. I also print a "dummy" target immediately before printing the real targets to make sure that the printer is "warmed up" and printing consistently.

I split the image that you posted into C/M/Y/K, and the PM/Magenta and PC/Cyan transitions both look good (I got the same images as you posted for the cyan and magenta channels, but I also wanted to look at the yellow and black channels). The cyan channel looked close to the cyan channel in the original image that you printed, but the yellow and magenta were both about 10-20% weak. This can also be seen in the black gradient at the bottom of your color image, where there is a blue color cast (a typical set of R/G/B values in this area is 155/162/174). This is not a problem, as the profile is designed to correct such color imbalances.

BTW, When I want the images to appear side-by-side in a post I combine them into a single image and scale it down to 600 pixels for the total image width.

I assume that the last target image that you posted was the original target image that you printed. This is a blow-up of a section of the image that shows severe jpeg artifacts. Was this the image that you printed, or a compressed version that has a reduced file size for posting purposes? If this was the image that you printed, the colors on many of the swatches would be corrupted.

PS uses the same .icc profiles for soft proofing as for printing (I haven't heard of .pfs files, and I couldn't find any on my system). Use "View/Proof Setup/Custom", and the pull down menu under "Device to Simulate" should list all of the profiles installed on your system. Where did you install your profile? On XP they go in "CWINDOWS\SYSTEM32\SPOOL\DRIVERS\COLOR".

References:
http://www.nifty-stuff.com/forum/viewtopic.php?pid=625#p625
http://www.normankoren.com/color_management.html
http://www.drycreekphoto.com/

 

[Grandad35]

CollegeDad71,

I forgot to mention one other important point. Only one of the images that you posted had an embedded profile. Without an embedded profile, there is no information telling how the color numbers are to be interpreted. This is similar to the crash of the space probe caused by one team working in meters while the other team worked in feet. An embedded profile defines what the color numbers mean compared to real world colors - without a profile, the color values are just numbers. This may be involved in the problem that you mentioned where your daughter's prints were oversaturated in magenta.

As a demonstration of what can happen, I loaded your last image as "sRGB" and also as "aRGB". I cropped them down to just the picture of the girl on the bottom left corner and included the color histograms for reference. These two images are included below. It is clear that the flesh tones on the two profiles are different. You will have to carefully compare the histograms to see the differences, but the red is clearly different in the bulge about 70% toward the right, and the green and blue are different at the left side. These differences are most easily seen on the tri-color histogram at the top. The fact that the differences occur in different places for different colors also demonstrates that the sRGB and aRGB profiles are not simple linear expansions/contractions of each other that can be corrected by an "intensity" adjustment on the printer driver.

Color management isn't an easy subject to master, but it is necessary if you want to get accurate and repeatable colors.

 

[CollegeDad71]

Grandad,

This is going to be an adventure, let me start with the easy issues. The last PrintFix calibration target that I included in post #16 is the unprinted original from ColorVision (but as you pointed out, it is a compressed jpg version rather than the supplied .tif file), it seemed useful to include a "standard" for post-profile comparison of the printed output.

The soft-proof worked as expected when at home where the printer lives. If from Proof Setup>Custom you try to "load" a profile (because the appropriate .ICC wasn't on my laptop) the default file type is .PSF but that is now a non-issue.

The images that were scanned in did not have an embedded profile because I hadn't told the SilverFast AI software that drives the scanner to embed one. So thank you, I need to add the profile used for printing to the available list for the scanner.

The camera embedds sRGB.

CFriends made such a point of saying their inks matched OEM colors that I did take the easy route and refilled the Canon carts. I thought if they were close and I had the ability to adjust the profile through the transition that it would all work out....I really liked the blue plugs. That brings me to what I thought was the really striking difference between the Color Test image that you posted in #15 and my printed output. Although the transitions are OK the actual Cyan top bar of the original (RGB 1/255/253) is not even close to the same region of the printed (RGB 24/103/182) and Magenta original (RGB 255/0/255) vs 196/61/102. It does point out that I need to re-profile the scanner with the printed IT8 target from Monaco because the saturated blue is not black as it appears, but the Cyan and Magenta are really no where close to your original.

So...how can the profile be so far off? How does one begin to address the problem?

In case your wondering, I used the PrintFix scanner for the profiling and calibrated it before use.