Color Calibration (7/25/08) In the previous note, I parenthetically carry on about this Samsung 243T monitor. It's about three years old; it originally sold for 6-8x what I paid for it. It worked very well right out of the box; I fixed the base (although it worked OK with the broken base, as my eBay seller promised it would) with just $6 worth of JB Weld epoxy; and I just now calibrated it.

First I downloaded Samsung's original driver for the monitor and replaced the driver for the 2943HM which I was still using. Images that had looked pretty good now looked like crap. This is progress: at least I knew something had changed, and I knew I could always revert to the driver for the other monitor if needed.

Then I used Pantone Colorvision SpyderPro, a calibratrion package that uses a USB-connected colorimeter and software to mess with the video card's transfer function (the soft- firm- and hardware that says that this value in maps to that value out). As instructed, I set the monitor to factory default values (brightness 100, contrast 50), and then spent five or six minutes holding the spyder to the monitor face. I named the resulting profile; made it the default for the monitor; hit "apply" and...

Voila! I was blind, but now I see!

My old Trinitron CRT calibrated nicely using this suite, but the VP930B never quite did (I always had to fudge its settings "manually" to get really close). The Samsung 243T works like a charm with the resulting profile.

I built a set of five calibration targets years ago. Using them, I can see that the image on my screen is very, very, very like the resulting prints. And by using Photoshop to measure what I can and cannot see, I can tell you that the difference between Black (0,0,0) and (4,4,4) is right there. Prior to calibrating I could just distinguish (0,0,0) from (17,17,17) or (12,12,12) with fudging.

In the detail at left from the image above, the black bar is actually a gradiant running from (0,0,0) at left to something brighter at right; the "saw teeth" are all (0,0,0). The first cuts into the gradiant in the vicinity of (2,2,2) and ends at (4,4,4). In Photoshop, I can see that! No more inverting images to check the shadow detail. And the highlights in the main image hold up well, too. And the colors seem dead on. It passes the gamma test in the previous slowblog entry (Lagom LCD Monitor Test) with flying [non]colors: the gamma setting is spot on and constant from edge to edge and top to bottom. Ah... Sometimes you win one.

:: back to the slow blog ::