ArticleHow to

The science behind ETTR (expose to the right)

During my recent article about noise in digital photography I briefly mentioned the technique known as Expose To The Right, or ETTR. This technique is used to reduce the amount of noise in the photo by overexposing the shot without clipping the highlights. When you apply this method the histogram curve will be shifted to the right, hence the reference to the name.

But why is this technique helping reduce noise levels? In this article I’ll proceed to explain the science behind ETTR and show the relation to noise reduction.

Let’s start reviewing how a DSLR captures light and convert it to a digital file.

CMOS-design-01

From this process we should remember that highlights have more information than shadows. That noise is relatively constant and that signal to noise ratio is smaller on the shadows. If we increase the shadows we will also increase noise making it more visible.

Now about ETTR! This technique is based on two characteristics one of sensors and a second of human eyes. Sensors capture light in a linear way. The amount of photons captured by the photodiode will increase linearly as time (exposure) progress. We can see this in the following representation.

linear

However, the human eye does not see light in a linear way but logarithmically. We are more sensible to shadows than highlights. To correctly display an image captured by a digital sensor we have to convert the captured data to make it look as the human eye would do. To achieve this we apply a Tonal Curve to mimic the human algorithmic response. By doing this we will be stretching the information from the shadows including the noise as shown in the graphic below.

logarigmic

With ETTR we basically overexpose the photo. By doing this we will dedicate more time to capture shadows and with it we will have more data to work with. During post processing we will be shifting the exposure back to the left, reducing exposure. By doing this we shift a section with more data into the natural space of the shadows.

ettr

In this process we used larger amount of data for shadows by overexposing. More data means we bury noise and a reduced “stretching” of the shadows, and with it limiting the amount of noise gain.

Leave a Reply

Your email address will not be published. Required fields are marked *