OK, my hands are high in the air. I’m as guilty of this as everyone else. I test cameras, present results here and elsewhere and I report on my observations. Typically providing frame grabs of what I have shot, maybe some nice scenic shots, maybe people shots, but not often shots of test charts. I use them all the time in my workshops and when setting up cameras for myself so why don’t I shoot test charts for my reviews? Well because they are boring and often don’t reflect the real world that we shoot.
But here’s the problem. In my previous post I commented about how some footage of Panasonic’s V-Log from the GH4 looked. The shooter, like me, had chosen to shoot some random shots and then grade them. But here’s the problem. I don’t know what those scenes actually looked like when they were shot and I don’t know what look the reviewer was trying to achieve. Plus there is always the issue that a look that one person finds pleasing may be the next persons least attractive look.
This is where charts should be used. If you shoot a known test chart you eliminate a lot of variables. One of my all time favourite charts is the DSC labs Cambelles.
It’s not a chart in the traditional sense, it’s just a picture of some attractive young ladies on a beach. But it’s a very clever image as it contains some very useful markers. It has very deep shadows and bright highlights (with a 5 stop dynamic range). It has a very broad range of skin tones and the skin tones have the same shade as the industry standard Chroma Du Monde charts. It has all the colors that you are going to encounter in normal shooting, blues, green. reds, they are all there. But here’s one of the key things – repeatability. Every one of these charts that you buy from DSC is exactly the same. So Fred in Australia could get a new camera, shoot his copy of the chart and straight away I know exactly what it should look like here on the other side of the world. Also because it’s a picture of people on a beach, we all know what that looks like and when it isn’t right we notice it. How many of you right now without looking it up can tell me exactly what color the 3rd bar from the left on an SMPTE color bars chart should be?
The use of a decent chart also becomes especially important when looking at cameras where the final image isn’t derived in the camera but in post production. By shooting a known chart and providing a frame grab of that unprocessed image those reading the review can make their own assessment of how easily the footage will grade and how it will respond to different lookup tables. Also when providing ready graded examples of footage, by including a chart in a corner of the shot the reader can see exactly what you have done to the footage. Maybe the reviewer likes strong reds, if there is a chart or other known color reference in the shot then the person reading the review will be able to see this from the strong reds on the chart in the graded footage.
Of course test charts do only show a part of the story about how a camera will behave in the real world. Real shots of real scenes are still incredibly important. But well designed charts brings a known reference and helps both the reviewer and those reading the review see through a lot of the variables that creep into creative shooting of real world scenes.
So: Note to self: Try to include charts in your test shots more often.
One thought on “Testing without charts leads to questionable results.”
I couldn’t agree more with the Cam Belles chart. For me, it’s real-world and not so “scientific” as it’s things my eyes are used to looking at. Specifically, I look at the skin tone on the second girl from the left. If she goes too pink, I have issues. Same with the green in her dress and the red on the beach ball. If those areas look correct, then i’m good to go. It’s a great chart and I use it a LOT.