If you are of a similar generation to me, you will remember when ISO was a vital element in deciding which film to put in our cameras. Too low and we would be stuck with slow shutter speeds or tripods. Too high and the image quality suffers.
Much of the mentality of the film era has carried over into digital photography. We still consider ISO when shooting, it still plays on our mind that image quality may suffer as we increase our ISO.
But what if we are worrying too much? What if ISO is somewhat irrelevant in a modern digital workflow? What if we have been doing it wrong all the time?
Recently I have written about the incredible advancement of AI denoise tools such as those found in Lightroom and DXO Photolab. However, what if we could shoot without actually changing the ISO from a low figure. Today we are going to take a simplified look at ISO Invariance.
An ISO Primer
The issue many of us older, and some newer photographers have is that we still think of ISO in the wrong way. ISO or ASA as it was known in the days of films was a way to measure the sensitivity of film. Wait I hear you cry, that’s the same for digital sensors. Yes it is, however the way that sensitivity works is very different.
Film is composed of silver halide crystals that are sensitive to light. The bigger those crystals are, the more sensitive to light they become. However, this is at the cost of a more pronounced grain. That grain manifests itself in the film when developed and shows up in the final image whether printed or projected. There were limitations as to how high an ISO film could go with the fastest black and white films topping out at 3200 and most colour films 800 ASA. At those limits we would get quite considerable noise in the image.
On digital sensors, we technically only have one or two ISO levels, the base ISO and a higher base ISO (Known as dual ISO). The way the sensitivity is increased is by amplifying the output of each pixel before it is converted into the raw data to create the image. The higher the amplification, the more interference we get between pixels and the more noise we get in our images. There are other factors in play however. The size of each pixel and also the distance between each pixel is an important element in how the noise plays out. The smaller and closer pixels are, the more noise we will have for a given sensitivity. The process of increasing the sensitivity of a sensor is called analogue gain as the signal analogue as it is amplified and then converted to digital for output.
So What Is ISO Invariance?
The way to think of ISO invariance is as a digital gain. Instead of allowing the electronics of the camera to amplify the signal, we take our images at the base ISO of the camera regardless of the light levels and then amplify them in post production. This is generally done by pushing the exposure slider to the right to raise the overall image, then using the other tools such as highlight and shadows as you would normally.
There are major caveats to this though. You need to be using a camera that has an ISO invariant sensor and you must shoot in RAW. JPEG files bake in the noise reduction regardless of ISO and as such will not give you ISO invariance. Many newer models have ISO invariant sensors but a lot of older cameras may not have the ability. Those that do have it will have varying numbers of stops that can be recovered in editing.
If your camera is not ISO invariant, when you push the exposure levels in Lightroom, the image will brighten, possibly clip but most importantly introduce a lot of noise, particularly in the darker, shadow areas of the shot.
However with an invariant sensor, you can set the base ISO and then shoot an image that is technically multiple stops under exposed. In Lightroom or other editing apps, you then light the exposure slider to recover the image and it will maintain the highlights and contrast but most importantly not introduce noise in the shadows.
There is another caveat to this. Because ISO invariance is not perfect across the whole range of ISO levels, many cameras now feature dual ISO sensitivities. The Sony A7iv for example has base ISOs of 100 and 400. Using 400 natively may give better results than pushing the exposure at ISO 320. My Sony a7Rv has a second base ISO of 320 for photography but when shooting video in S-Log it has base ISOs of 800 and 3200. The feature is called dual native ISO and you need to check online whether your camera features it.
Why and Where Would You Use ISO Invariance?
So far so interesting, but why might we use ISO Invariance over just raising the ISO? Well firstly, raising the ISO above a certain level, depending on the camera, decreases the dynamic range. So if we are shooting a high contrast scene, perhaps where we need a high shutter speed or small aperture, using ISO invariance will allow us to hold the highlights better whilst maintaining detail in the shadows.
Concerts and events are where we might have low light and also high contrast. By keeping the ISO at one of the base levels we can shoot freely and have leeway in post production to isolate the parts of the image where we want to nail the exposure on the subject.
In astrophotography the higher dynamic range at base ISOs allows for stars to punch out better from the dark skies. There is a caveat that ISO invariance is not recommended for stacked astro shots.
Sometimes it’s just better to keep the ISO low, maximise the dynamic range and make the corrections in post production.
Unlike the days of film, ISO is not as vital to the way we shoot our images nowadays. With ISO invariance we can maintain dynamic range by shooting at base ISO and cranking exposure in post. But also, as mentioned at the top, we can do incredible things to very high ISO images using AI denoise to bring the quality close to a low ISO image.
Is ISO invariance the way forward? No, it’s one of several ways forward, including actually using high ISO to add atmosphere and realism to our photos. It’s certainly an amazing time to be a photographer.
