Colors. Who knew they could be so tricky.
One night, a week or so ago, I was making a print of this photo on my photo printer, a Canon iP6600D:
Simple enough, you’d think. I even was using Canon’s specific ICC profiles, based on these instructions for Photoshop, for my specific printer and paper combination.
The colors weren’t even close. The white was white, but the green background was very yellow-ish brown…to the point that it pretty much killed the picture. I held the print up to the screen (I was using a Dell 2001FP display), and the print was much more yellow/brown than the nice green-ish color on the screen.
Hmm. Maybe the printer profiles were off, I thought. So I uploaded a copy to ritzpix.com, and the next day I picked up the print at a Wolf Camera store near my office, after asking them to turn off all of their auto-levels and color correction. I took the picture home that night, and compared it to the screen and the other print…it, too, was very yellow-ish brown, nothing like the screen. It was actually pretty close in color to the print I made on my own printer.
Well, it must be that my monitor needs to be calibrated, I thought. No problem. The next day after work, I stopped and picked up a ColorVision ColorPlus colorimeter for monitor calibration. Got home that night, ran through the setup and calibration, looked the before/after on the screen (there is a button you can press at the end of the calibration to do this), and saw that there was indeed a difference…although it was subtle.
So then I made another print of the photo above. Why, exactly, I wasn’t sure…but maybe there was some kind of interaction between the monitor and printer ICC profiles, right? Well, I held the new print up to the screen, and it was pretty much the same old yellowish-brown, compared to the nice green on the screen. Crap.
Two more calibrations, and 5 more prints, and still no love. I did discover the “rendering intent” feature in Photoshop, which is a very nice tool to know (had to do some research, but I think I understand what it does now), and that affected the output colors quite a bit…but not enough to get it close to what I saw on the screen. Sigh.
About now, it’s about midnight, and I’m ready to punt on the whole thing. The 2001FP has independent RGB gain controls, and I’m thinking I can eyeball this thing better than the colorimeter can do. So I download and print Smugmug’s calibration print. Using that, and my own print of the photo above, I adjusted the gains on the monitor until it just about matched. Woo-hoo! Or so I thought. The one thing I noticed after doing this is all of my windows were kind of reddish where they used to be white…kind of annoying for reading your email. But at 1:00 in the morning, I no longer cared too much, and went to bed.
The next morning, I was quickly checking my email (yes, I know), and the reddish color everywhere was really pronounced – and really annoying. Annoying enough that I reset the monitor back to its default settings, which were the same settings that I used when running the calibration the night before…so we were now back to the calibrated settings. Ahh, the email window looked white again. Whew!
Just for giggles, I opened Photoshop and pulled up the photo again. Not sure why…just seemed like the thing to do at the time. Ahh, nice and green. I reached over and grabbed one of the prints I had made the night before, and held it up to the screen. Man, this looks…
The colors matched almost perfectly. WTF?
Turns out it was the light in the room. Under daylight conditions, during the day, the colors on the print look almost exactly like the screen. But at night, with my “regular” incandescent lighting, the prints look different. WAY different.
Sheesh…there went 4 hours I’m never going to get back…but in the glass-is-half-full department, I won’t soon forget the effects of different-colored light. And hey, turns out my monitor is calibrated pretty darn well now. :-)
Full-spectrum lighting, here I come! Either that, or leave work earlier when it’s still daylight… ;-)
Make sure you get Daylight Balance Full spectrum Flourescent tubes if you plan to work at night. Just changing out your incandescent bulbs with the “full” spectrum bulbs won’t do – you’ll have an even tougher time getting color to match. I converted my entire work space to use Daylight Balance Flourescent Tubes (using High Performance shop lights) and I can now work at any time of day. This is great when you get a creative itch you need to scratch. Good luck!
“Full-spectrum lighting, here I come! Either that, or leave work earlier when it’s still daylight…”
hahahahaahhahahahahahahah and ha! If I stay you stay!
this is a good piece. do you think this problem could have been solved if the colorimeters for monitor calibration also had a sensor on their back which measured outside environment and adjusted the monitor according to that?
Turgan – it’s possible…but I’m not sure you’d _want_ to calibrate your monitor for changes in ambient light. This would change throughout the day…and plus, you’d be seeing the image as it would look printed in the current lighting, rather than what it would look like under daylight.
So is changing the lights the only option? Even if the light changes through the day, the device can help you update your computer’s settings…
although a sensor that measures color temperature sounds rather expensive: