Printer to Printer variance

adbsgicom

My setup is LR2.4 with Dell 2007WFP monitors, calibrated w/ Spyder3Elite on a GeForce250 (dual LUTs). I've printed to file (profile embeds properly) and printed to MPix, Costco (profiles from Drycreek), local pro camera shop, and two different Epsons, R800 and R2880. Printing on Lustre paper at each option creates a pretty wide array of color tints in the image. The starting point is the same .dng file in LR, just printed with a different profile. Is this just standard operating procedure with printers? One can easily say that my calibration is off, but even then the printers should be printing the same 'wrong' image. The Epson's came in closest to what I have, with (oddly) the Mpix sRGB w/o color correction coming in a close second, with the local photo store (not Ritz/Wolf) coming in a close third (though they seem to modify the image somewhat). The Mpix w/ 'proper' profiling was much more yellow and darker, and the costco printer came in very magenta and darker. Costco gets used by many Austin pros (based on some pros I know), though some are drifting away and using Mpix or WHCC. Bottom line, how much printer to printer variance should one expect to see when using printer profiles? Should they be within 100K of each other? The variance I'm getting is more like 250-400K.

arodney

Each printer uses different substraights, colorants (ink, toner etc) and as such have differing color gamuts. You need a good ICC profile for each printer to even think of getting close results. And process control is critical for consistency for many of the devices you mention. The Epson is going to be the most stable and consistent of the printers mentioned! You still need good profiles and their quality varies (and many of the devices you mention are used by some labs that demand you use sRGB instead of an actual profile so forget about those).

adbsgicom

arodney wrote:

Each printer uses different substraights, colorants (ink, toner etc) and as such have differing color gamuts. You need a good ICC profile for each printer to even think of getting close results. And process control is critical for consistency for many of the devices you mention. The Epson is going to be the most stable and consistent of the printers mentioned! You still need good profiles and their quality varies (and many of the devices you mention are used by some labs that demand you use sRGB instead of an actual profile so forget about those).

But for using the ICC profiles for each, the variations from red-cast to yellow-cast were pretty significant, which is what was baffling. The costco profile was very recently updated (early 2009). Is the moral of the story that the technical flow (profiling all around) only gets you so far, then one needs to adjust for one's preferred print house (assuming they are consistently shifted one way or another)?

arodney

adbsgicom wrote:

But for using the ICC profiles for each, the variations from red-cast to yellow-cast were pretty significant, which is what was baffling. The costco profile was very recently updated (early 2009). Is the moral of the story that the technical flow (profiling all around) only gets you so far, then one needs to adjust for one's preferred print house (assuming they are consistently shifted one way or another)?

There's a lot of variables. The profiles could just stink. The lab may not have used them. The lab may have used differing rendering intents. Different profile packages will provide different results with a perceptual rendering (vendor A's profile to the same device as vendor B's will appear different). The printers are not calibrated and the profile isn't defining the RGB values correctly due to poor process control.

Profiling will get you only so far, but YOU have to control the entire process before you can decide the success or failure of the profiles and the printing systems. You can have a perfect profile and the lab could have their chemistry off and there goes the color.

adbsgicom

Thanks!