Fast lenses, and High ISO

Started 3 months ago | Discussions thread
Great Bustard
Forum ProPosts: 24,076
Like?
Re: Exposure and brightness.
In reply to bobn2, 3 months ago

bobn2 wrote:

Great Bustard wrote:

bobn2 wrote:

Great Bustard wrote:

OK, this is a little more work. Tell me how you like this:

A certain number of photons fall on a pixel releasing a certain number of electrons which generates a charge. A gain may, or may not, be applied to this charge as a function of the ISO setting on the camera. The charge is then converted into a digital number by the ADC (Analog to Digital Converter). The RAW converter (or in-camera JPG engine) processes groups of digital numbers into RGB values for the image file where the ISO setting on the camera maps these values so that they have the brightness that the corresponds to the exposure and ISO setting.

You're making it too complex, getting bogged down on the detail of the technology, rather than what is the process being undertaken.

Is that not necessary, however? I mean, if people are to understand what an ISOless sensor is all about, for example, then the ADC needs to be discussed, as well as understanding why, for example, a photo at f/2.8 1/100 ISO 1600 is less noisy than a photo of the same scene at f/2.8 1/100 ISO 100 pushed four stops for a sensor that is not ISOless.

They are important when they are important, but not before they are important.

The problem we have is that people think that ISO is the implementation details. That misunderstanding is not their fault, it is because a great many sources, including this site have told them that it is so. So, the first thing people need to get is what ISO actually is. Once they have that, they might want to know about how the ISO control works on their camera, because it may have side effects, which are not ISO but are worth knowing.

OK -- I can absolutely agree with that.

I would say that those side effects are not relevant in terms of this discussion. The difference really is between there being no ill-effect of 'turning up the ISO' and it being beneficial. The mindset equivalence deniers tend to have is that turning up the ISO as of itself causes 'ISO noise'.

Sure. But after that, I do think it's important to discuss the role the camera's ISO setting plays in read noise via the ADC.

The camera counts the photons, in three different bags (red, green or blue). Then it translates those photon counts into grey scale values. The exact translation done depends on the ISO setting, or to reverse that the translation done determines the ISO. The rest is implementation details.

Sure, but at least some of the details are important, methinks.

Depends on the discussion whether or not they are important. In terms of this discussion, I'd think they add unnecessary detail, not needed unless someone raises the old saw about 'amplifier noise'. As I said, first get across what ISO actually is, then look at implementations details on a case by case basis. Lumping them together just makes people think that ISO is 'amplification', which is where we came in.

OK, sure.

If we could engineer a sensor that directly output a count from each pixel, the basic process would be the same, and no-one would get bogged down in issues of 'ADC' or 'gain' or any of that stuff. That is just the way we engineer a photon counter at present. Film was also a photon counter. A pair of photons would reduce a reactive spot in a silver grain which catalysed its chemical reduction to silver. Each grain represented two photons collected.

Ideally, of course, we would have a sensor that recorded every photon that landed on it along with the wavelength (and momentum) of each photo that added no additional electronic noise.

However, in my opinion, I think the processing chain, from the capture to the image file, is necessary to describe as it does apply to how current cameras are used. Like I said, there is a reason, other than operational convenience, to use ISO 1600 rather than ISO 100 and push, on many cameras, and that shouldn't be glossed over.

You have to be very sure not to mix the what with the how to the extent that people think that 'ISO' is amplification, because that is where a lot of these misunderstandings come from. In my experience as an educator, I would say, try to teach one concept at a time. First let's get an understanding of the black box 'ISO Engine', which takes a photon count and translates it to a grey scale value. That is something most people simply don't have. They think ISO magnifies the light. When people are clear what the ISO engine does, they can learn about the innards as they are today. It's like telling a learner driver what the engine does and thinking that then you have to go straight into pistons and camshafts and stuff.

I hear what your saying, and it appears that I was using the term "amplification" in a misleading manner. I simply meant that higher ISOs would result in higher values in the image file, not that there was a linear mapping or that higher ISO settings increased the light.

Do you think it's correct to say that the density of light falling on the sensor (photons / mm²) is mapped into the exposure via the luminosity function which is then mapped into the image file by an ISO function?

Reply   Reply with quote   Complain
Post (hide subjects)Posted by
(unknown member)
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark post MMy threads
Color scheme? Blue / Yellow