Perry R writes:
About calibration, white balance, getting accurate prints from a lab.
It looks to me from your NEC PA302W color gamut diagram that you have your white point calibration set to 5500 (D55)? [diglloyd: no, that's just a comparison of calibrated to other; I calibrate at 6500°K / gamma 2.2].
I have been under the impression that the recommended and commonly used "industry standards" for monitor calibration have been D65, gamma 2.2, for photographers, but that the prepress graphics people use D50, 2.2. Which of the 3 is correct for photographers and why?
I'm also trying to sort out if and how this relates to printed images from a commercial photo lab. I shoot perfectly exposed, custom white balanced studio images in camera (i.e. black/white/gray tri-color target histogram's black and white spikes are equidistant from ends of the histogram displayed on the camera LCD, then custom white balance shots x 2 of the gray area only). I save both JPEG and RAW versions, the JPEG for fast volume job work, the RAW in case something else needs to be generated for other reasons).
The resultant JPEGs look fine on my NEC PA322UHD calibrated as above.
I've used several different commercial photo labs for volume printing/packaging. Some of them offer "color correction" (at additional cost) and some do not. My understanding of the "color correction" they do (in my case) is really just adjustments to make the printed images look right on the paper, meaning they have to adjust for the peculiarities of the white point of the paper, inks/type of printer, etc. which would vary from lab to lab. For customers who submit non-perfectly exposed /non-custom white balanced images they would also have to first adjust for exposure errors and improper white balance. Usually this mostly involves increasing the brightness of the image to compensate for transmitted vs reflected light luminosity differences of viewing the image on a monitor vs paper, and maybe some selective color boosts/shifts to compensate for the ink set used.
I've found that the labs that don't offer color correction expect the photographer to do one of two things:
1. Submit test images, get the prints back, and fudge/adjust their monitor settings, until the on screen and printed versions are similar - this can require several rounds of testing, and results in a non-calibrated monitor; or
2. Stick with D65, 2.2 calibration, but formulate a Photoshop/LR finishing correction to adjust for the printer, and apply it to all submitted images as a last step.
A few labs provide 1-3 ICC profiles for different papers/presses that can be used to "soft proof" in Photoshop for the latter, but not many.
My feeling is that if they are going to be in the printing business, it is their responsibility to adjust properly exposed and white balanced images for their own printing process, since this is operator/equipment/substrate dependent, and there should be no charge. If the photographer is submitting improperly prepared images, there should be a charge for adjustment. If I make a 40x50 inkjet canvas on my Epson 9900, I don't charge the client for the profile correction work separately, it is built in the price of the print.
So I'd be curious to hear your thoughts on all this, and back to the original issue, if I'm calibrating to D65 vs. D55, which is the correct setting that results in better output for these commercial labs?
DIGLLOYD: I calibrate as shown in Calibrating the NEC PA302W and other NEC Displays. Both D50 and D65 are correct, but I cannot stand working with D50; it looks dull and ugly to me. So I have used D65/2.2 for years.
I can shoot a scene, print it, and hold the print next to the original subject and get a perfect match, and nothing fancy involved—Epson printer or Canon printer. That’s what I do to verify output, but I do not own a large format printer and my preferred print medium is skillfully done coated canvas in large format. See Printing On Canvas. The lab I use is Picture Element (Mike Chambers, no relation). Superb work done there.
My core advice: if a lab cannot achieve consistent results, vote with your wallet and use another lab. Send various labs a perfectly balanced test photo and have it printed on at least two distinctly different papers (matte/semi-matte/gloss for example). To cut off the workflow-challenged places, submit the file in 16-bit ProPhotoRGB, and see which labs screw it up; if a lab cannot deal with a good original and properly color manage, then they’re cut from the candidate list. I suggest this to ferret out claims vs reality on color management. There may be labs which insist on a fixed format (e.g., TIF in AdobeRGB); these may be acceptable too with suitable conversion of your own prior to submission. I do not like this because a good lab should be working with the full quality original, and the gamut of AdobeRGB is often a problem for me. So I want a lab that accepts 16-bit PSD or TIF in ProPhotoRGB; they can make their minor tweaks using the full quality original.
Shown below is a *scan of a print* made by Pixel Element. See Tulips #10 @ f/13 Finished Print (Nikon D800E + Zeiss 135/2 APO-Sonnar), as is the DP2 Merrill tulips image. View more: .Tulips