ISO in digital cameras

[NMGPhotos]

This might have been asked before, but I am curious how digital cameras handle ISO. Obviously, in manual cameras, the ISO is a property of the film that is used. But in a digital camera, there is no such film. Is there a specific formula that the camera uses to emulate(?) different film speeds (ISO values)? If so, is it proprietary or well known?

[paddler4]

The difference in ISO between base ISO and any other simply measures the amount that the camera's electronics have amplified the signal captured by sensor, as well as any noise in that capture. It's scaled so that a doubling of ISO is a doubling of luminosity, that is, so that the impact on brightness is the same as a doubling of exposure.

Just google "iso in digital cameras" without the quotes and you'll find many further explanations.

[NMGPhotos]

I understand how ISO works. I was wondering more about the technical / programming stuff behind it.

[Bill C]

On 11/29/2022 at 4:35 PM, NMGPhotos said:

I understand how ISO works. I was wondering more about the technical / programming stuff behind it.

I've not read the ISO method for digital cameras, but have formerly read some technical papers related to the topic.

To get an overall view, of a very technical nature, I would suggest doing a search for "Jack Holm" and "ISO speed." I think that he may have been on the ISO committee. If you have the background to follow it, here's an older IS&T paper: https://www.google.com/url?sa=t&source=web&rct=j&url=https://www.imaging.org/site/PDFS/Papers/1999/PICS-0-42/987.pdf&ved=2ahUKEwjZyor8sdj7AhUCj4kEHZttDV0QFnoECBMQAQ&usg=AOvVaw0Cnw8dCOTA3UsO9hiQbEcA

My rough understanding is that there are two general methods. The first, called base ISO, is roughly based on how much light energy to "saturate" an image sensor. (Image sensors can "fill up" with a charge, and then can not detect any further increase in light; they are then said to be saturated.) A particular Kodak sensor that I was familiar with was based on being able to reach 170% of a perfect white (diffuse) reflector. So, for example, an 18% gray reflector, when photographed, should put the sensor at about 18/170 of the saturation level. (One would need to know the actual amount of light energy being delivered to the sensor at that point.) This is a simplification. 

The second method, for higher ISO speeds, is based on some "signal/noise" ratio. I don't know any specific details, but basically the camera electronics can amplify the sensor output until it eventually becomes "too noisy" to be useful. This is loosely like what paddler4 is saying. 

Ps, I should mention that the nominal correct exposure should remain roughly the same between a film camera and a digital camera at the same ISO speed setting. One big difference is that a "daylight-balanced" color film is ideally used at the correct color temperature of light. Whereas the digital camera can be adapted for different color temperatures, within certain "noise" limits.

Edited December 1, 2022 by Bill C

[paddler4]

 

Quote

The second method, for higher ISO speeds, is based on some "signal/noise" ratio. I don't know any specific details, but basically the camera electronics can amplify the sensor output until it eventually becomes "too noisy" to be useful.

Exactly. When you boost the captured data, either by increasing ISO in camera or by boosting "exposure" in postprocessing, you are amplifying both signal and noise. That's why high-ISO images appear noisier: the amount of noise is roughly the same as at base ISO, but the signal is much weaker.

Unfortunately, some explanations of this, including some from very reputable organizations, refer to an increase in ISO as boosting the sensitivity of the capture. That's sloppy language and is incorrect.

[NMGPhotos]

I think I heard also that digital sensors expose differently to where you would want to "overexpose" a manual camera vs a digital one for the same scene. Is that true? And, if so, what are the specifics around it?

[Dustin McAmera]

I think you may be talking about reciprocity failure , which is a thing for film. Under normal conditions of light, if you close down the aperture by one stop (from f/8 to f/11, say, or from f/2 to f/2.8), doubling the exposure time gives the same overall exposure. However, if the light is very low, and you're using a long exposure time, this relationship may break down. The film doesn't record light as effectively if it arrives at a very low intensity. Typically, you don't need to think about this for exposure times shorter than about a second. Where it begins, and how much you need to increase the exposure by, varies from film to film. If you need to know this sort of thing, the film makers usually have a data sheet that you can download; or it may even be printed somewhere in the information on the inside of the film box.

[NMGPhotos]

20 minutes ago, Dustin McAmera said:

I think you may be talking about reciprocity failure , which is a thing for film. Under normal conditions of light, if you close down the aperture by one stop (from f/8 to f/11, say, or from f/2 to f/2.8), doubling the exposure time gives the same overall exposure. However, if the light is very low, and you're using a long exposure time, this relationship may break down. The film doesn't record light as effectively if it arrives at a very low intensity. Typically, you don't need to think about this for exposure times shorter than about a second. Where it begins, and how much you need to increase the exposure by, varies from film to film. If you need to know this sort of thing, the film makers usually have a data sheet that you can download; or it may even be printed somewhere in the information on the inside of the film box.

Thanks 🙂 I guess, the question which I should have asked at the beginning: is there any way I could use a manual camera app on my phone (which lets me change the ISO and shutter speed), and then do some calcuations as to what settings I should use on my manual camera? Or is there too much of a difference? I know what the aperture the lens is (as well as the model of sensor).

[Dustin McAmera]

In normal daylight conditions, exposure for digital and film will be the same. You could certainly use the readings from your phone or any digital camera like a light meter. I have installed an app called Light Meter, by WBPhoto  (my phone is Android - I suppose the same app is available for IPhone). It uses our phone's camera as a meter, and presents the results in a convenient way that doesn't need you to do calculations (convenient-ish - I preferred another app that I can't get any more). It has a few other features, mostly that I haven't used.

 

Edited December 3, 2022 by Dustin McAmera
Adding missing letters - dying keyboard...

[NMGPhotos]

26 minutes ago, Dustin McAmera said:

In normal daylight conditions, exposure for digital and film will be the same. You could certainly use the readings from your phone or any digital camera like a light meter. I have installed an app called Light Meter, by WBPhoto  (my phone is Android - I suppose the same app is available for IPhone). It uses our phone's camera as a meter, and presents the results in a convenient way that doesn't need you to do calculations (convenient-ish - I preferred another app that I can't get any more). It has a few other features, mostly that I haven't used.

 

My issue is, my camera only goes down to 1/200 for shutter speed. According to the Sunny 16 rule, I feel like that stops me from using ISO 400 film. Maybe if I go up to f/22 (the far end of what my camera can do) it would be better?

My phone camera, on the other hand, is a static f/1.7, a dramatic difference. I got an exposure calculator app that lets you plug in an initial set of ISO, F-stop, and shutter speed, and then a desired new settings for two out of three, and it will give you the value for the third. Is this a good tool?

To note, aperture and shutter speed on my camera both have to be set manually.

Edited December 3, 2022 by NMGPhotos

[Dustin McAmera]

In the bright sun that Sunny 16 assumes, you could just do it: 1/400 at f/16 is equivalent to 1/200 at f/22; but you wouldn't want to use ISO 400 in bright sun.

I wouldn't try to use Sunny 16 as your only guide. We don't only photograph in bright sunshine (it's not the most interesting light). Here in the UK, for most of the year we get a rather oblique sunlight, distinctly weaker than what people in (say) California call sunshine. So if we try to use Sunny 16 in our sunshine, we often underexpose.

There are printed exposure guides - used to be on the inside of many film boxes.

Give your calculator tool a go! It sounds a bit laborious to me, but hey - it's maths..

[NMGPhotos]

I just tried the app you mentioned and I see what you were saying about it making calculations easier 🙂 Might end up using that one.

I would assume it does take in my lenses aperture into account with its calculations.

Edited December 4, 2022 by NMGPhotos

[JDMvW]

I will add that if you search for "sweet-16" you will not find out about sunny-16, although you might learn something about early Apple ][.

[glen_h]

On 11/28/2022 at 7:06 PM, NMGPhotos said:

This might have been asked before, but I am curious how digital cameras handle ISO. Obviously, in manual cameras, the ISO is a property of the film that is used. But in a digital camera, there is no such film. Is there a specific formula that the camera uses to emulate(?) different film speeds (ISO values)? If so, is it proprietary or well known?

The ISO methods for film are partly based on the non-linearities of the medium.

It is designed, as well as possible, to generate a straight line for some range

on the log-log B-H curve.   Among others, film handles overexposure better than

underexposure.  (Except for reversal films ...)   But even for reversal films, the

effect on the highlights isn't completely sharp.

 

Typical scenes have, overall, an 18% reflectance, and film exposure

is partly based on that.

 

Electronic sensors, and the attached analog to digital converter

use a linear representation for intensity.  There is a hard saturation

at the maximum. 

 

ISO likes to charge a lot of money for their standards. 

The idea, though, is that if you use the same metering system

used with film cameras, based on the ISO value, that you get

reasonable results.

 

At some point, that depends on the properties of average

scenes. 

[SCL]

You should easily be able to use your phone app to determine proper exposure, with a simple caveat. That being you need to determine carefully exactly what is being measured...highlights, shadows, a combination of both. In a film situation, negative film best results typically occur when one exposes for shadows, while in positive film one typically exposes for the highlights. In digital there seems to be differences of opinion, and I suggest you try it out yourself with your own equipment bracketing exposures of subjects you typically shoot to determine what works best in your lighting with your equipment.

[Ricochetrider]

I use the app, My Light Meter and it works wonderfully well. It fails when, as mentioned above, I’m photographing something bright or brightly colored, in bright light. I was told by some one here at P.net a while back that rather than meter the whole scene, I should meter the shadows in this instance. 

 

[Ricochetrider]

Posting an example of app failure er, that is I mean-  operator error! Metered the overall scene here, with the double indemnity of using Fuji Provia reversal film. Provia's a little bit forgiving- but not very much so! 

Live & learn, with any luck

Edited February 7, 2023 by Ricochetrider