P 2.0 Two Dimensions: The Contrast Sensitivity Function

This site is in the process of being updated, with extra content designed for the layperson as well as vision professionals. I’m afraid the formatting of existing pages has been affected — sorry about that. It’s still readable, but hopefully it will all be fixed up soon, better than before.

You’ll have seen this image before. It’s known as the Contrast Sensitivity Function (CSF).

A boring grey image?

I once heard an ophthalmologist refer to it as “this boring grey image“, which I thought was unjust. I feel quite fond of it. It’s one of the most important and useful aids to understanding visual function (and visual dysfunction).

It maps the relationship between a thing’s spatial frequency (how big/small it is) and its contrast, and demonstrates why we can or cannot see it. Each person viewing it has a line something like this:

Line maps the threshold between seen and unseen

The line marks the boundary between seen and not-seen. Things in the top right are low contrast and fine detail  — none of us can see them. Things down the bottom middle are high contrast and easily large enough to be seen.

I find this image helpful when considering how it relates to our clinical measurements:

The CSF unites several single-parameter clinical measurements

Everything we do with a high-contrast VA chart takes place along the very bottom of the CSF. That’s where optometrists feel at home, but it’s really important to understand that as soon as you’re dealing with someone with low vision, most of the action is taking place in the lower-contrast areas higher up on the CSF.

Up the top left is interesting. The line starts dipping down again, becoming less easy to see as the object becomes larger. This is true when you’re using a sine-wave grating as the target, which is a fancy way of saying that you’re looking at a gradation from light to dark, like this:

High contrast sine grating

If you’ve got a very wide gradation, the subtle change of shade becomes simply too hard to detect, most especially when there isn’t much difference between the peaks and troughs, like this:

Low contrast sine grating

This doesn’t apply though when you are dealing with a distinct boundary (technically a square wave function, rather than a sine wave). Print is a good example  — there’s no point where a letter gets too hard to see because it’s too big (or at least not until we start considering field of view, but that’s for later). Since much of what I’ll be talking about later relates to reading text, I’ll mostly be considering the curve as not curving down in the top left, but keep it in mind when you’re considering other visual tasks (such as being able to see changing contours on concrete pavement, for instance).

Here’s an illustration of the CSF where there is a distinct boundary, rather than a gradation:

CSF using square grating. Note there’s no dip on the left.

Note: You’ll notice that the seen/not-seen boundary isn’t drawn right, in that you’ll be able to see some detail in the unseen area. That’s mostly down to the limitations of these graphics — partly that a computer screen can’t capture the full range of contrast and brightness, but also because I’m an optometrist, not a graphic designer. But it’s also because that line will vary depending on your viewing distance, your screen brightness, your screen resolution settings, etc, and of course it varies a bit from person to person anyway (or a lot, if you have any vision impairment).

Previous Page                                                             Next Page