Since exposure happens at the sensor, before ISO gain is applicable, it seems that shutter speed, aperture size, and available light are the components that impact exposure. Isn't changing ISO just a way to help compensate for over or under exposure that happens at the sensor?
You're getting close. One recommendation I'll make, is to remove subjective evaluations of
incorrect exposure,
overexposure or
underexposure. More on this, later.
Yes, scene brightness, f-stop and shutter speed determine exposure: the brightness of the scene per unit area focused upon the sensor. Strictly speaking, ISO is a standard that relates exposure to lightness in the rendered image. Here's a relevant excerpt from a recent post by Bobn2:
First, ISO, the standard, doesn't apply to raw files (which has long been correctly interpreted from 12232:2006 by Iliah Borg and others, and has now been made explicit in 12232:2019). By extension, it also doesn't apply to sensors, sensor electronics or the read chain in general. Thus talking about 'Fake ISOs' as Jim does, and 'base ISO' as a characteristic of read chain gain are both strictly speaking wrong. 'ISO' relates exposure at the image plane to lightness in the final processed image, and doesn't put any requirements whatsoever on how that mapping is achieved.
Full post:
https://www.dpreview.com/forums/poshttps://www.dpreview.com/forums/post/62390194t/62390194
The last sentence in the above excerpt is key. It's the reveal that ISO does not
directly impact either
exposure (amount of light focused upon the sensor) or
gain (amplification of signal). In a nutshell, I think of ISO as assigning an exposure value to a scene. The camera uses this value to determine how light to render each pixel. The rendering is effected by a combination of a manufacturer's algorithm relating the exposure value of a scene to the on-sensor exposure and the manufacturer's in-camera electronics which effect the application of the derived lightness.
So, while ISO can be accurately thought of as one element among many
resulting in a post-exposure lightness adjustment, ISO is
not the same thing as that lightness adjustment. Nor is ISO the same thing as the means used by the camera to effect the lightness adjustment. This is the best way this non-engineer can think of to describe ISO in a way that encompasses both its nature and the outcome it has a role in producing.
From what I have learned about ISO, through reading and use, it seems to function similar to an audio amplifier while available light, shutter speed, and aperture work together to produce the music track (image) being made audible by the amplifier.
Am I misunderstanding something about the function of ISO gain?
In a public photography forum where folks use conversational language to discuss everything from gear, to technique and even - on rare occasions - actual photographs, I'm a big believer that we need create room for discussions about technical subjects to be broadly accurate without insisting upon the rigor of a peer-reviewed journal paper.
For example, there is value in acknowledging the distinction between exposure settings (f-stop and shutter speed) and ISO. As a general rule, the more light delivered to the sensor without clipping highlights, the better the image quality tends to be. If a person understands that increasing or decreasing ISO does not, in and of itself, result in a change to the amount of light delivered to the sensor, that's useful information which can be applied when using a camera.
Similarly, I think it's of some value to disconnect ISO from the concept of gain. Gain, is one link in a chain of processes applied in the rendering of an image. However, it does not - strictly speaking - directly determine the lightness of a pixel. For this reason, I've taken to describing ISO as a tool
used to manage post-exposure lightness adjustment.
To be fair, it's a phrase that conflates a number of processes that are distinct from each other. If I were being pedantic, I'd maintain a strict usage of the term as a standard while avoiding language characterizing ISO as an outcome. However in a public forum, we need ways to talk about technical subjects that might bend reality just a bit...but not so much as to break it. While ISO is not the same as the tools or processes used to adjust image lightness, it is the one control available to us that most directly influences those tools and processes.
I don't describe ISO as gain and certainly not as a sensitivity to light adjustment. The latter implies an ISO setting can, in and of itself, affect how much light from the scene is captured during an exposure, which is incorrect. I use the term, lightness, to distinguish ISO from exposure, which is determined and defined by brightness...an amount of light. I use the term, adjustment, to imply change without assigning a value. It is neither complimentary nor a pejorative. It's a neutral term.
For example, when photographing birds in flight, I might choose an f-stop of f/8 to ensure a depth of field in which the majority of the bird will appear in focus. I might choose a shutter speed of 1/2000 to ensure freezing the movement of the wings. In addition to controlling depth of field and the degree to which motion is frozen, these settings also determine how much light from the scene will be captured during the exposure.
Even on a bright, sunny morning, these settings can result in a dark image at base ISO. If I change the exposure settings to capture more light, the depth of field may be too shallow or the motion blur of the wings too prominent for my liking. While the image at base ISO may be darker than I want, it's not due to an
incorrect or even an
under exposure. The exposure settings are correct for the image I want to capture. Describing the image as underexposed suggests, in my opinion, an error of some sort was made in choosing the exposure settings. However, that wasn't the case. The settings were chosen to achieve a desired look.
ISO can be used to manage the lightness of the image. We don't go into the entire chain of in-camera processes from ISO establishing an exposure value for the scene through front-end analog amplification to back-end digital amplification and so on. This is a public forum, the online equivalent to folks having a conversation over a meal, a couple of beers or the beverage of one's choice. There's a lot of background information that's just assumed. How else are we're going to solve all the world's problems in an hour?
IMO, talking about ISO as managing the post-exposure lightness adjustment adequately summarizes the
outcome. While not a strictly accurate description, it does focus the conversation on the outcome...the thing we're actually interested in. It avoids confusing ISO with exposure. It also avoids certain weighted terminology (e.g. incorrect, over, under) and leaves the door open to compare the effects of in-camera ISO to the "exposure compensation" tool in your favorite photo editing app.
That's a subject for another post.
