Discussion in 'AP Magazine Feedback & Suggestions' started by PC60, Jan 13, 2018.
That's a fourth!
I'll have to read it again - I skimmed AP after getting in after a week away and the cumulative impression of all the articles wasn't that great - but I was tired last night.
I agree I am far from likely to go sourcing any new iso standard.
The acknowledgement of noise associated with the sensors and supporting circuitary is a step on from the original article in jan 2018.
The collective articles do much better justice to the whole business but demonstrate different personal nuances which can make combining the whole a bit of a task. I empathise with the need for re-reads (extra value from AP).
It's been a while since I had access to an electron-microscope to look at molecular and crystal structures. Never looked at a ccd sensor but film emulsions were interesting long ago.
Nigel Atherton did say they planned a followup last year, it just shows that a lot of effort went in to the subject this time. So thank you.
https://dslrbodies.com/cameras/camera-articles/image-sensors/is-iso-fake.html is another article that might be useful although the style of writing will not be to everyone's taste.
So: (and I know I'm going to regret this) what do electron microscopes have to do with taking real world pictures?
Not by my count:
Subject brightness, shutter speed & f-stop.
ISO will change the cameras response to the total light reaching the sensor but not the amount of light itself.
Originally you had incident light as third then you added the subject reflectivity. But if by amount of light present you meant reflected light then that is three.
I see your thinking, in the more general case quite correct, but @Chester AP did say 'for any particular photo' which effectively fixes subject reflectivity. This leaves incident light & subject brightness as directly related.
I suppose the possible addition of an ND filter is technically a fourth if working with incident light, but we can consider this to be changing the subject brightness and keep it at 3. No point making things too complicated, threads on ISO definitions will generally manage that without any help!
I've discovered an error in my original post (which I can no longer edit):
ISO 2240 was for colour slides, and there's ISO 6 for Black & white film - So 3 standards for film & 5 procedures for digital. No wonder people struggle to grasp it!
I've just checked the preview for ISO12236:2019 , it might be that one of the procedures has been dropped & it appears higher values of ISO have been added.
Less relevant perhaps we also have:
ISO 2720 for exposure meters,
ISO 2721 for camera auto metering,
ISO 5763 for flash,
ISO 7187 for direct positive colour film,
ISO 10157 for flash meters
It should be clear by now that iso is technically frought and does not apply to raw anyway.
However we have to set something when we expose raw. The histogram is not as helpful as it should be. It is a histogram of a tone mapped jpeg.
Apparently vidiographers have an alternative method that does not yet appear on still cameras.
For now, I don't get worked up about it. I just set auto iso with a conservative top limit of 6400 on my D500. If that isn't enough then I manually take it a couple of stops higher. The theory is very nice to talk about in ignorance, but results are really all that matters.
Overall, aftet the half of nothing bit done in January 2018 the magazine issue 16th Mar 2019 collective is more thorough and was a better read.
ISO ratings of film also had a link to grain. pushing processing had limitations not only due to the 'exposure' but down to the emulsions and chemistry. A comparison of film(s) at molecular level with sensors would be an interesting thing if you have a mind for it. However sensors are affected by the circuitary and algoritha (quality) for example banding issues seen on some sensors like the D7100.
I wonder what old Muhammad_ibn_Musa_al-Khwarizmi would think of that plural. Most people use “algorithms”.
Another from you.... I had you blocked, but clearly the forum blocking isn't working. Shame.
Another thought about this... what about working with RAW files of images taken at high ISO settings?
If I understand correctly, the ISO setting for a digital image is stored in the RAW data and only applied (by 'amplification' of RAW image data) when the RAW file is converted to a JPG for display on a monitor (or by the camera's software to create a JPG version of the image), because when we work with a RAW file what we see on the screen is a JPG file derived from it. If I've got this wrong, more information would be welcome.
Alternatively, we can all carry on calling ISO adjustment 'sensitivity' adjustment, and take pictures.
On the processing side the ISO setting, stored in the raw file, essentially determines where the white point is set when converting raw data to image format*. There is little "amplification" involved as the articles in AP were discussing. This is adjustable in post-processing via the exposure slider.
On the exposure side: increasing ISO decreases the exposure i.e. the quantity of light getting to the sensor. You should use the lowest value consistent with getting a workable F number and exposure time in order to maximise dynamic range and reduce noise. This also is what the articles were discussing.
*raw image processing programs may use an internal format and then convert to jpg when you export the file. So what you see on the computer screen depends on the program. Raw files also have a jpg embedded. Some viewers will display this.
I never block anyone. Most people here have something of value to state even when I disagree very strongly.
Separate names with a comma.