Doing Your Own ISO Invariance Test

There’s plenty of discussion today about ISO invariance and to what degree certain cameras may be ISO invariant. This is a highly technical subject and readers who are interested in exploring this topic in-depth should research it by reading articles on more technically oriented sites. As regular readers know, I’m an experiential/experimental type of photographer, rather than being technically oriented.

In a nutshell, if a camera has perfect ISO invariance (which no camera does) there will not be any penalty in terms of noise when lightening your image in post as compared to originally capturing your image at a higher ISO. This article shares some sample images captured with Nikon 1 V2, V3 and J5 cameras, while doing a simple ISO invariance test. I’d like to thank one of our readers, William Jones, for providing the creative spark for this article.

All of the images in this article were captured tripod assisted with each camera set in Manual mode. I started shooting at ISO-6400 with each camera, setting the exposure manually to get a good exposure at that ISO. After the first image in each set was captured, I then lowered the ISO in successive steps. These progressed through ISO-3200, ISO-1600, ISO-800 and ISO-400.

I used DxO PhotoLab for initial RAW file processing, using the same custom preset for all of the images. I applied auto PRIME noise reduction to each of the images. Images were lightened using the Exposure Compensation adjustment in DxO PhotoLab. I exported DNG files into CS6 and opened them up in that program, but did not make any adjustments in CS6. The only additional adjustment I made with each image was to add Structure at a +20 level in Viveza 2. This is a setting that I use with some regularity with my Nikon 1 images.

First let’s have a look at a series of images captured with a Nikon 1 V3. This camera has an 18.4 MP 1″ sensor. All images were captured at f/5.6, 1/60 using a focal length of 100 mm (efov 270 mm).

NOTE: Click on images to enlarge.

Nikon 1 V3 ISO-6400
Nikon 1 V3 ISO-3200 lightened in post
Nikon 1 V3 ISO-1600 lightened in post
Nikon 1 V3 ISO-800 lightened in post
Nikon 1 V3 ISO-400 lightened in post

As we can see from the image series above the 18.4 MP 1″ sensor in the Nikon 1 V3 performs quite poorly when images are lightened in post.

The next series of images were captured using a Nikon 1 V2 at f/5.6, 1/50, 100 mm (efov 270mm). The V2 has a 14.2 MP 1″ sensor.

Nikon 1 V2 ISO-6400
Nikon 1 V2 ISO-3200 lightened in post
Nikon 1 V2 ISO-1600 lightened in post
Nikon 1 V2 ISO-800 lightened in post
Nikon 1 V2 ISO-400 lightened in post

The Nikon 1 V2’s 14.2 MP sensor performs better than the 18.4 MP sensor in the V3, but still has a noise penalty when images are lightened in post.

And finally, let’s take a look at a series of images captured with a Nikon 1 J5, f/5.6, 1/40, 86 mm (efov 232). The J5 has a 20.8 MP BSI 1″ sensor.

Nikon 1 J5 ISO-6400
Nikon 1 J5 ISO-3200 lightened in post
Nikon 1 J5 ISO-1600 lightened in post
Nikon 1 J5 ISO-800 lightened in post
Nikon 1 J5 ISO-400 lightened in post

As with the V2 and V3 series of images, we can see that the J5 also has a noise penalty when images are lightened in post. This is not surprising given the small pixel sizes on 1″ sensors.

Cameras that perform better in terms of ISO invariance tend to utilize larger sensors. If you are interested in seeing how the sensor in your camera performs in terms of ISO invariance under test conditions, you can use the link provided.

If you check out your camera, or other models using the link provided, keep in mind that the flatter the graph line is… the more ISO invariant the camera will be.

You may find that the graph line is flatter in certain ISO ranges. This would indicate that the camera is more ISO invariant in that portion of its ISO range.

Understanding how our cameras perform under different lighting conditions is important for all of us to know. This allows us to use different approaches when setting the exposure for our photographs.

In previous articles I’ve suggested to readers that they consider their photographic software to be part of their integrated camera systems. This can help to build your own experience base regarding how you want to handle your images in post to address the potential ISO invariance of your camera.

How we choose to use our camera gear is a matter of personal preference. For example, I know that many Nikon 1 owners darken the majority of their photographs by using  exposure compensation in order to try to hold on to highlight details. This is something that I virtually never do as I have never liked how the shadow areas in my Nikon 1 files perform when lightened. I would much rather work on bringing highlight details back, than deal with additional noise in shadow areas. But again… that’s just me. Everyone has to make their own decisions on how to best handle their photographs in post.

Technical Note:
All photographs in this article were captured tripod assisted using camera gear as noted in the article. All images were produced from RAW files using my standard process of DxO PhotoLab, CS6 and the Nik Collection.

Word of mouth is the best form of advertising. If you like our website please let your friends and associates know about our work. Linking to this site or to specific articles is allowed with proper acknowledgement. Reproducing articles, or any of the images contained in them, on another website or in any social media posting is a Copyright infringement.

My intent is to keep this photography blog advertising free. If you enjoyed this article and/or my website and would like to support my work, you can purchase an eBook, or make a modest $10 donation through PayPal. Both are most appreciated. You can use the Donate button below. Larger donations can be made to tom@tomstirr.com through PayPal.

As a reminder to our Canadian readers, you can get a special 5% discount when ordering Tamron or Rokinon lenses and other products directly from the Amplis Store.

Article and all images are Copyright 2019 Thomas Stirr. All rights reserved. No use, duplication or adaptation of any kind is allowed without written consent. If you see this article reproduced anywhere else it is an unauthorized and illegal use. Posting comments on offending websites and calling out individuals who steal intellectual property is always appreciated!

7 thoughts on “Doing Your Own ISO Invariance Test”

  1. FWIW: I ran some tests on my D750; however I ran them in the reverse fashion of this test. First I did a series of test shots from ISO 100 thru ISO 6,400, by full stops. That provided me a baseline of what an ISO X shot should look like. I then did a series of shots all forced to be at ISO 200, a series all forced to be at ISO 400, and a series all forced to be at ISO 800. Shots forced at ISO X (200, 400 or 800) could be safely pushed by two stops (200 pushed to 800; 400 pushed to 1,600; 800 pushed to 3,200) in DxO PhotoLab 2 without introducing more noise. Shots pushed more than 2 stops introduced more noise than I got with a properly exposed shot for that ISO.
    Other notes: To push an ISO 200 shot that should have been shot at 400, the Exposure compensation I needed was only 0.35, not 1.00. A two shot push only needed an Exposure compensation of 0.70, not 2.00. That amount of Exposure compensation was the same for pushed ISO 400 and 800 shots. Lastly, pushed shots also had different colors than non-pushed shots. So a properly exposed shot did not need to have the color tweaked, while a pushed shot did. Since I am partially colorblind and could still see the difference, I believe that such color differences would be even more obvious to people with proper color sight. I have no idea what would be required to fix the difference in colors (I did not try).

  2. Thanks for the info. I like the overall “look” of the higher ISO images than the lower iso images. I presume that mostly has to do with differences in post-processing.

    1. Thanks for adding to the discussion Robert!

      Since the base exposure was based on ISO-6400, the higher ISO images would have been subject to less ‘push’ in post in terms of lightening them. At the lower end, i.e. ISO-800 or ISO-400 the images would have been lightened by 3 or 4 stops, rather than just 1 or 2 stops at the higher ISOs. This demonstrates that the sensors in my Nikon 1 gear are not ISO invariant.

      Tom

Leave a Reply

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