Unfiltered Story #262951

While taking a class on image encoding methods, the professor decided to take advantage of the captive student audience in order to get more research data. Thus, everyone in the class had to take a test to see at what point different image compression algorithms break down and become noticeable to the human eye. This test consisted of looking at two copies of the same image side by side, then answering whether they are the same or whether one of them has noticeable artifacts from whatever compression system was used. Simple, straightforward, and the only real way to see if the inevitable defects caused by a lossy compression algorithm were actually noticeable by people or were tiny enough to pass without notice. As an example, think of a picture straight from your phone vs. that same picture converted into a good quality .jpg image to post on the internet. Then compare the original image to a .jpg so blotchy that you can barely tell whether the orange splotch is supposed to be your orange cat or a pumpkin.

We’re sent to a tiny computer lab with half a dozen computers and set up to look through the test images. I start working through the images, and end up marking a lot of them as significantly different due to weird color shifts. This seemed a bit odd since I hadn’t thought that color matching was often wrong in image compression, but I figured they might be testing out some odd new compression algorithm that had color issues. Then more of the images have the same issue, and I realize that I’m halfway through the test and all of the images that aren’t primarily greenish seem to have color matching issues.

Confused, I look at the computer to my left and immediately see the problem. On that computer, the background border around the two images is a blank white color. On my computer, there is a mottled green splotch that I thought was just background texturing and someone trying to pretty up the interface (and failing). As it turns out, my computer monitor had a large blotch of the worst color shifting I’ve ever seen on any piece of electronics. Since this blotch only covered one of the images in the test, there was a significant difference between them purely due to the monitor that had nothing to do with the actual image files and compression systems being measured.

I immediately went to talk to the student overseeing the lab, but they didn’t seem to think it was that big of a deal. I then asked if I could start over, because obviously my data was horrible. Nope. They didn’t have any way to do that. They also didn’t have any way to discard or invalidate my data, or the data from anyone else who had used that same monitor. I ended up finishing up the test on that computer, trying to ignore the color differences, completely baffled at their lack of concern.

I’m still not sure whether this was some other sort of psychological experiment from a psych professor, because I still find it hard to believe that an image processing computer lab would have a monitor that was so bad. Usually those types of labs spend a lot of money on the fancier monitors with better color rendering. Since there were so few computers, someone had to have noticed this before me, right? And someone in that lab must’ve cared, right? . . . right?