If you accept exposure as being the amount of light that struck the sensor per unit area while the shutter was open then there are only 3 parameters that directly control exposure.
Light coming from the scene
Aperture
Shutter speed.
Setting an iso value indirectly controls aperture and/or shutter speed because it sets the exposure that will be used to output the image lightness the camera is calibrated to.
For example, for a given scene and lighting setting a higher iso tells the camera to set a smaller exposure than setting a lower iso would have to output the image lightness the camera is calibrated to.
A. f/8, 1/400s, iso 100
And
B. f/8, 1/800s, iso 200
will both output the same image lightness but setting B will be noisier because of its lower SNR resulting from the smaller exposure.
That is exactly what was correctly stated in the video.
I have never used auto ISO, and probably never will...
I do my own calculations for any given situation and dial them in for that. Much of it is intuitive.
I am the opposite. I always use auto iso, as suggested to try in the video, and lock in base iso only when using a tripod and there is no movement in the scene.
For pointing and shooting I normally use aperture priority with a minimum shutter speed set plus auto ISO or manual plus auto ISO.
Shooting raw, the iso the camera chooses for me is irrelevant as long as highlights are not clipped because I set the image lightness in post.
Mike I have been playing about using DXO pureraw on older files and even going back many years { in digital time 😀 } the results are often extremely good
I seldom use it myself 😀 but it is a simple way to achieve very good results. Set the best aperture for the lens in use that gives you sufficient DOF for your subject. Along with a shutter speed fast enough and let the camera decide an appropriate ISO. There are certainly scenarios where the exposure can be tweaked to maximise what you want from the scene. I have used it previously in wedding shooting where lighting changed throughout the day. It would be an even more effective technique with modern NR etc
I have been using Topaz. There has been a big improvement in Topaz since the original De Noise. They keep improving it on an almost weekly basis. I'm guessing but I think the database that is used for AI within the program is enlarging enormously. Whatever, high isos no longer bother me. It's also changing my lens selection. F4 is now fast enough for my purposes. This leads to useful savings in $/weight/size.
Being old, my ISO is super-glued to 100, as is my EC to zero. Not enough light = no shot. Around here in Texas at 30 degs north, there'll be plenty more light tomorrow ...
That is why when using a tripod I lock in base ISO in low light scenes when there is no movement in them. I can slow the shutter to whatever suits my purpose. Usually it's ETTR shooting raw.
If shooting sooc jpegs I can understand setting iso because you most probably want a nice looking image sooc.
But if you shoot raw and want to maximise the quality of the raw data in terms of minimising noise what difference does the iso setting make if no highlights are clipped, especially since the final image lightness is set in post?
Very rarely, is my subject of interest comprised mostly of highlights. Most times, the highlights are in the OOF backgrounds, where I couldn't really care less if say the blue channel clips a bit. Or there's a little bit of detail lost there. I'm not going to starve my main subject of light, and create excess noise, just so the blinkies don't flash in my chosen raw processor.
Yes you're right 🙂 I wasn't totally clear. Normally I refer to important highlights but in these posts I didn't. I also normally let totally unimportant highlights clip in order to maximise the light from the important parts of the scene to hit the sensor