Thoughts on dynamic range in cameras

Here is an out of camera JPG taken of a lamp in a room, an experiment. The first thing that stands out is how dark the right hand side of the image is.

A lamp in a room – original out of camera JPG

Sitting in the room and looking at the curtain what I see is more like this:

Processed single RAW file. Handheld, E-M1 in HDR2 mode.

This image is based on five exposures. The camera produced a single RAW file which is interesting. I expected five which I’d have to combine. Still, the five exposures at 2ev spacing provided more dynamic range. The out of camera JPG in “HDR2” mode is similar.

We know that printing photos increases contrast and reduces dynamic range. We had to dodge and burn in the black and white days to give an image with enough local contrast to look interesting while preserving information in highlight and shadow detail. But at least we captured a lot of information on film. Does the camera sensor manage this?

This is the result of processing the RAW file, with Lightroom set to squash the dynamic range almost as much as it could. The camera has captured a lot of detail and maybe some work with Lightroom’s gradients and brush tool could have helped. I’ve not preserved the orange of the lampshade as well as the camera, which is pretty well spot on.

Scene recreated from a single exposure RAW

DXO report that the Olympus E-M1 has a dynamic range of 12.7 stops at ISO 200. That’s pretty impressive. I remember when I started out in the world of digital reading that we’d have reached some kind of holy grail once we reached 11. We’re past that now, and cameras such as the Sony A7S take it even further.

It’s clear that the E-M1’s JPEG engine sacrifices range. This could be a mode setting, though I have it set to Gradation Auto which I believe is its most range preserving setting. DPReview show JPEG ranges for the cameras they review and they all give a similar small range. I guess the JPEG engines are trying to make pleasing photos that pop.

Some articles on the human eye have claimed a range of 20 to 30 stops, though claims vary. It could be that the eye adapts as it looks around the scene, so our instantaneous range is relatively low. This is an unusual scene. Maybe it needed a grad filter in front of the lens to capture.

Similarly with those old black and white films – I used to process following the instructions. Sometimes I’d push process. Apparently film is more non-linear with a gentle tail at the top and bottom of its ranges, perhaps leading to easier recovery.

Leave a Reply

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

This site uses Akismet to reduce spam. Learn how your comment data is processed.