Why Log? Part 2

The image above may look like a mistake, but it’s not. Instead, it will help me explain what “shooting log” is all about.

Last time, I talked about how differently a still photographer and a cinematographer may deal with image capture. The photographer usually tries to capture what he sees, but a cinematographer tries to capture as much image data as possible, so it can be manipulated later.

One of the tools cinematographers can use in today’s cameras is shooting using log. Log comes from “logarithm”. Without getting into too much math, think of log as non-linear, or not a straight line.

Non-linear is not solely related to image capture. Consider the volume knob on a radio or a fader on a mixing board. As you raise the level up from nothing, you might notice that it takes a lot of movement in the control to start hearing something. But as you get closer to full volume, a slight change in the control represents a large change in what you hear, even if you go up to 11! The change is not linear.

But back to image capture and log. When I said cinematographers want to capture as much data as they can, it often boils down to dynamic range—how many “stops” the camera can capture. Or, how wide a range—from the darkest part of a scene to the brightest part—the camera can capture with detail. That capture process is limited by how the image data is recorded to a file. This is where using non-linear data makes its way into the process, and also where the image above comes into play.

Take another look at the image at the top of this post. On the right-hand side of the image is a background of pure white – 100% brightness. Overlaid are three white rectangles that have white values of 99%, 98% and 97% from top to bottom.

On the left side of the image is a background of 0% brightness. It is overlaid with rectangles of 1%, 2% and 3% brightness. Can you see them? If you really increase the brightness control on your display maybe they will be more evident.

I think you will agree that the differences on the left side are not as obvious as those on the right side of the image. And even if I used fractional percentages at the bright end (ex., 99.5%) you would probably see those changes too. This is because our eyes are more sensitive to small level changes in brighter areas of an image than in dark areas. We can use this behavior to our advantage – using log.

Next time… how log takes advantage of vision limitations.