Exposure is not how dark or how light an image looks!

Getting off topic was precisely the context for my remarks. It's that context which you appear to have missed.

He was complaining that there was limited quantizing density in the midtones and shadows because of the use of a linear ADC. Never mind that that is a red herring because of noise, but that's one of his points, and a reason he advances for ETTR. ETTR is a good thing, but that's not the reason.

If you think that 64 bits is not sufficiently dense quantization for some shadow level, but want to stick with linear ADCs, then the obvious solution is to increase the precision of the ADC. That's what many people who bought Bruce's bogus quantization density argument clamored for, and one of the reasons we ended up with silly levels of precision in some Hasselblads, which ended up allowing them to more finely digitize pure noise.

The thrust of his argument about quantizing levels points towards a solution that doesn't work.

Another clear sign that he was confused was his talking about shadow posterization arising from a 12-bit ADC. The CCD sensors of the era had so much noise that that wasn't going to happen.

Bruce said this:

"Film mimics the eye’s response to light, which is highly nonlinear. Most of our human
senses display a significant compressive nonlinearity—a built-in compression that lets
us function in a wide range of situations without driving our sensory mechanisms into
overload"

It is not, in general, true.

Slide film has a expansive, not a compressive, nonlinerity.
Negative film developed to N has a compressive nonlinearity.
Negative film developed to N+x can have an expansive nonlinearity.
Negative film optically printed on paper is approximately linear when you get away from the toe and shoulder regions.

Bruce said this:

"Note that the on-camera histogram shows the histogram of the in-camera conversion to JPEG: a
raw histogram would be a rather strange-looking beast, with all the data clumped at the shadow
end..."

That's not true at all, as a few minutes spent with RawDigger will demonstrate. There's no reason why a raw histogram has to use a linear x-axis, and for many purposes, it's better to use a nonlinear one.

Nice, and now we are presented with a challenge to make it linear, because in Nature eyes are presented with a linear distribution.

I don't understand what you mean.

Are you saying your post

is on topic? 🤣

After some back and forth with chatGPT:

On first look, I am happy with that description.

It's a little strange n this context, but not incorrect, and it brings the aperture into the, ahem, picture:

en.wikipedia.org/wiki/Entrance_pupil

You know, in a general terms, it doesn't have to be camera, sensor or lens. You can expose anything on/to anything. General terms and correctness make it harder to explain and understand. There is always a reason to attack almost any word.

You caught ChatGPT on a good day. This is what chatGPT tells me (it's wrong!):
Exposure in photography refers to the amount of light that reaches the camera's sensor or film when taking a photograph. It is determined by three factors: shutter speed, aperture, and ISO sensitivity.

Well, you see - the higher you crank your ISO, the more invisible little ISO fairies pop out of your camera and start pushing and squeezing extra photons through your lens and toward the sensor. It's all perfectly sensible!

Doesn't ISO setting on the camera determines A or S in any auto modes for any given scene (lighting)?

How does one know what ISO to set?

One does not, but ISO is set to something anyway

In automatic modes, ISO guides the camera to set the proper exposure. In M mode, the human guides the camera to set the proper exposure.
You cannot say that ISO is part of exposure in A mode but not in M mode.

I'm confused ;)