Monday, May 26, 2008

Color Correction with the Datacolor Spyder 3 Studio

In getting ready for another year of art shows, I'm addressing some issues I faced last year. One of those issues is a lack of color consistency throughout my workflow process. Last year I learned all about color profiling and got OK results by using generic paper profiles from various sources. This year, I've decided to address the issue further. After researching various options, I decied to try out the Spyder 3 Studio from Datacolor. So how has it turned out? So far so good....

A brief history of my color workflow

For years, color correction was a non-existent topic for me. What I got out of my Canon i960 was mostly close enough for my satisfaction. If not, I'd tweak the photo in Photoshop to give me what I wanted. That "worked" well enough (only because my definition of "close enough" was pretty lax) until it came time to upgrade printers.

When I got my Canon IPF5000, I decided if I was going to spend over $1500 on a printer, I was going to do it right and learn how to use color profiles. I used generic color profiles from a variety of sources. That worked a lot better, but I found myself results still lacking. For one, none of the prints matched my monitor exactly. Second, I still found every profile gave me (at best) slightly different results. Some of the profiles that I got straight from the paper manufacturers were outright terrible.

Making the move to custom profiles

Over this last winter, I decided that from here on out, I needed to do things right and get custom profiles made. I had debated about paying to have them made, but I decided against it for a few reasons. The thought of paying out money for something like that just didn't strike me right. I'd rather pay more up front and then have the flexibility to create profiles whenever I wanted. I tend to like to experiment with different papers, and this gives me the option to create as many profiles as I want on a whim without wasting money. Any new paper I want to evaluate can be evaluated using a properly made profile, so theres no need to waste money, or to decide on a paper based on non-representative results.

I looked around at various options. The i1 Color from X-rite is well regarded as the leader in color management solutions, but in the $1000+ range, it was a bit out of my budget. I was looking for about half that price. I found a few solutions, including the x-rite ColorMunki and the Datacolor Spyder 3 Studio. From reviews, both seemed to do a good job, but the ColorMunki seemed to be a bit too basic for my tastes.

I decided to go with the Spyder 3 Studio. I found it through for only $545 shipped. It included both the Spyder 3 Elite (for profiling your monitor) and the Spyder 3 Print (for profiling your printer/paper).

Profiling my monitor

The first step of the process is getting the colors on your monitor corrected, so that your workflow begins with a known configuration. Otherwise, getting the print to match your monitor is both next-to-impossible and pointless. I installed the software for the Spyder 3 Elite and gave it a go. It was relatively painless. However, when I was done, I ended up with what I thought was a terrible monitor profile. The screen had a definite magenta cast to it. I repeated the process and got identical results.

At this point I was feeling like I had purchased junk and was about to say screw it and just make my monitor look good again. However, I remembered that it sometimes takes your eyes a while to adjust to different lighting, and thought maybe I should give the profile a chance to sink in. After about 10 minutes the color started to look really natural. Moving to a different light source and then coming back, I could see the magenta cast each time, but my eyes quickly readjusted. After a few days of working with it, somehow my mind has gotten adjusted to the new color settings, and now, even after coming in from a different light source, I can't really notice the color unless I'm consciously looking for it. Almost instantly, it looks like pure white to me.

I don't believe I saw any tips like this in the program. It would probably be useful, to set new user's mind at ease.

Profiling my printer

Now, with the monitor calibrated, it was time to move onto the printer. I started with my Canon HW Satin paper. I printed a basic 225 patch target and scanned it in. The scanning is a bit time consuming and tedious, but the included guide makes it easier and less error prone. After scanning in all the patches for the target, it generated a profile and I went to print a test page. When I hit the soft proof button, it showed very little change, so I figured that was a good sign. When I then printed it out, what I got was quite surprising. For the first time ever, my print looked virtually identical to what I saw on the screen. This was really a breakthrough moment for me.

Next I moved onto the "paper" that previously gave me the most problems: the Breathing Color Chromata White Matte Canvas. Breathing Color doesn't have a profile available for the iPF5000, but they do for the iPF8000. Being that the 2 printers are nearly identical (same ink set, same print heads) the profile should have worked fine, but it didn't. It had terrible contrast and saturation, so I had to correct for that in the printer's export module configuration. I was hoping the Spyder 3 Print could do better.

I printed out the 729 patch color target plus the 238 patch extended gray target. Scanning the nearly 1000 patches took me somewhere in the range of 30 minutes. As I got a little way into it, I was noticing that the results for each patch as displayed on the screen was quite terrible. Like my previous profile, they were all low in contrast and saturation. When I finished, I soft proofed my test print and it was horrible...much like what my uncompensated prints with the old profile had given me. I backed up and experimented a little with the advanced options, but nothing could give me a soft proof that had acceptable results.

Once again, feeling disappointed and thinking I had wasted money, I decided to see just how bad it is on actual paper. I restored the profile to all of it's default settings and generated my test print. Once again I was shocked when, despite the terrible looking proof on the monitor, the actual printout look great. It was virtually identical to the monitor, and also virtually identical to the HW Satin print.

I then threw together a sampler page of my own photos and printed it out, and the results were virtually identical for every single image. Everything from a black and white to the most saturated sunset all look wonderful and accurate.


This is far from a thorough test, but so far I am extremely pleased with the results I've received. It's really opened my eyes to exactly what is possible. In fact, I feel quite stupid for having tolerated what I did before. If you have the money to spend and aren't completely satisfied with your current output results, I wouldn't hesitate to recommend this product.


Brooks said...

This is extremely helpful, especially to know that despite what seem to be initially disappointing soft proofs, the actual prints come extremely close to what is displayed on the monitor. Considering that a lot of time and costly paper are involved in trying to "wing it" in order to get a good match, investing in the proper tools seems infinitely better.

PTKen said...

Hi. I just read your blog on calibrating your monitor and seeing a magenta cast that you eventually adapted to. I am a photographer and have been working in digital imaging for 15 years, I am a color management specialist, and I am a color scientist with an M.S. Color Science from the Munsell Color Science Laboratory.

I can tell you that if after calibrating your monitor you see a magenta cast that it takes 10 minutes to adapt to, then something is wrong with your calibration. First of all, there should not be such a strong cast. Second, you should be almost fully adapted to the new white point after only 60 seconds.

My recommendation is to try calibrating with a different measuring device and/or different software. You should be able to achieve much better results than you have described. If you still have trouble, then it is probably worth it to hire a consultant to sort it out for you.

Ron Frazier said...

Actually, it's more a matter of the way individuals (or at least myself) respond to lighting color. I know you are color scientist and maybe this sounds crazy, but hear me out as I have other proof of it.

I very recently changed the lighting in my home. The house lighting used to be all incandescent bulbs or CFL's with an equivalent color temperature (ie: around roughly 3000K). Well, I decided to experiment with daylight CFLs (5500K). I first started in the bathroom. I put in the new lights, and it looked really funny. Even if I was in there for several minutes, I could tell the lighting color was totally different. Yet, after leaving them in there, I eventually got accustomed to it, and it after a few days it appeared fine.

So, next I moved on to the kitchen, and put in the same lightbulbs. Now here is the funny thing. Even though I've got the same lighting color in both room, the bathroom looked perfectly fine while the kitchen looked very odd for many minutes after I entered the room. Again, it took me several days to get accustomed to the change in lighting before it looked natural relatively quickly after entering.

I then did the same thing in each other room of the house, and ended up with the same results. My adaptation was very slow for several days, but gradually it improved until it was nearly instant.

So here is the way it seems to work in the mind (at least for me...I don't want to speak for anyone else). My brain seem to have expectations about what each area should look, and when I change the lighting, it takes a good while for it get used to it. But as time goes on, and I've been exposed to the new lighting more, I begin to adapt to it quicker and quicker until it eventually is nearly instant.

Now, with respect to the monitor color calibration, part of the calibration process is that the software tells you to change the color temperature of the monitor to 6500K. However, my monitor was previously set significantly lower than that (but I don't recall exactly what). So now, what have I done? Well, I've changed the color temperature of the lighting to something that my brain hadn't been adapted to, it took a lot of exposure before it would adapt to it more easily. This is the exact same experience I had in changing my house lighting.

Perhaps this sounds crazy, and perhaps I'm a statistical oddity in the matter, but that's how it is for me. I might have believed your post that something was wrong if I hadn't gone through the exact same thing in changing my house lighting.