Noise and ISO, One more thread

Because conducting an experiment to convince myself that I was wrong, I had a hell of a hard time in Adobe Raw making my ISO 100 image look decent. I had to lower the shadows, lower the color saturation, etc.
Some years ago, when I started shooting my daughter's hockey/skating practices, I found the value of Auto ISO in M mode ... except my camera didn't have it :) I would pick a fixed ISO and end up with shutter speeds that varied and realized that some of those could have been shot at a lower ISO. So then, one week, I tried shooting them all at the lowest ISO I expected to use and had to adjust in PP. I found this to be a ridiculous exercise in wasted time - for one thing, the darkest images gave me a useless review image. Unnecessarily having to adjust 60 or so photos at a time was annoying enough, but I found the same thing you did - a simple exposure adjustment didn't do it - I ended up needing to adjust the curve to get the image to look natural, and it the specific adjustments varied from image to image.

This was with a Sony A700 (old 12MP APS-C sensor) and whatever version of Lightroom was current at the time, so maybe things could be different now. But it wasn't long after that that I switched to Nikon to start shooting Auto ISO in M and I haven't looked back. (Not shooting hockey any more, but I use that mode most any time I'm shooting above base ISO).

- Dennis
--
Gallery at http://kingofthebeasts.smugmug.com
When I took a look at the ISO 100 image in Adobe, at first I said, "Aha, these idiots don't know what their talking about. This image SUCKS!."
Then, to be fair I started playing with the controls. And after this and that time consuming adjustment, by golly, you couldn't tell the ISO 100 from the 1600.

Fine. But as you say above, What a pain in the butt to do if you're taken a lot of images!... :-D
I wonder if you could create presets. Say your base ISO is 100. So create presets for ISO 200, 400, 800, etc. And then shoot everything at 100 and use whatever software you like with presets to "adjust" the iso in POST to whatever you need?

Still sounds more painful than doing it in-camera, though.
 
Nope. Photons are quanta, they can't be amplified.
But the A>D signal can be amplified. The photons create the charge, the digital conversion creates a number, which can be multiplied.
 
My best luck with youth sports in general, and hockey in particular, was to simply turn off all the auto-exposure features, including white balance.
How do you turn off white balance?

It has different settings but none of them is "OFF"
 
Last edited:
Ok, I am FINALLY convinced that increasing the gain of sensors by raising ISO is not the same as ISO in film. Iliah finally overcame my obtuseness. But I STILL recommend shooting in higher ISO's in dim light.
That's normally the correct strategy for noise purposes because most cameras have lower read noise at higher ISOs (up to a point). The penalty you pay for raising ISO is NOT noise, it's DR!
Why?

Because conducting an experiment to convince myself that I was wrong, I had a hell of a hard time in Adobe Raw making my ISO 100 image look decent. I had to lower the shadows, lower the color saturation, etc. Whereas the ISO 1600 image needed almost no adjustment. So, there IS an advantage in shooting at the higher ISO.
Plug your camera into this chart at PhotonstoPhotos to get good guidance about which ISO settings will yield improvements in shadow noise (i.e., mostly read noise) on your particular camera.
 
Ok, I am FINALLY convinced that increasing the gain of sensors by raising ISO is not the same as ISO in film. Iliah finally overcame my obtuseness. But I STILL recommend shooting in higher ISO's in dim light.
That's normally the correct strategy for noise purposes because most cameras have lower read noise at higher ISOs (up to a point). The penalty you pay for raising ISO is NOT noise, it's DR!
Why?

Because conducting an experiment to convince myself that I was wrong, I had a hell of a hard time in Adobe Raw making my ISO 100 image look decent. I had to lower the shadows, lower the color saturation, etc. Whereas the ISO 1600 image needed almost no adjustment. So, there IS an advantage in shooting at the higher ISO.
Plug your camera into this chart at PhotonstoPhotos to get good guidance about which ISO settings will yield improvements in shadow noise (i.e., mostly read noise) on your particular camera.
Just want to add that taking advantage of ISO-invariant ranges is only advantageous if you shoot RAW, and that the IQ advantages of doing so may not be worth the operational disadvantages.
 
My best luck with youth sports in general, and hockey in particular, was to simply turn off all the auto-exposure features, including white balance.
How do you turn off white balance?

It has different settings but none of them is "OFF"
I'm sorry, I meant turn off auto-white balance. In other words, set the white balance manually.
 
My best luck with youth sports in general, and hockey in particular, was to simply turn off all the auto-exposure features, including white balance.
How do you turn off white balance?

It has different settings but none of them is "OFF"
I'm sorry, I meant turn off auto-white balance. In other words, set the white balance manually.
Aaaah, I see. Makes sense on manual, I do all my photos in RAW , never have to worry abt WB.

Different with video, where I use Brno lens cap for WB.
 
Ok, I am FINALLY convinced that increasing the gain of sensors by raising ISO is not the same as ISO in film.
Raising ISO in a digital camera doesn't change the efficacy of the capture stage.

It's different with film. From http://www2.optics.rochester.edu/workgroups/cml/opt307/jidong/index.html

"It takes about the same number of photons to flip a large grain as a small one. Since the larger grain intercepts more light more of the larger grains will be flipped and thus less light is required to create a latent image. This later phenomenon makes coarse grained films faster (more sensitive)."
Iliah finally overcame my obtuseness. But I STILL recommend shooting in higher ISO's in dim light.

Why?

Because conducting an experiment to convince myself that I was wrong, I had a hell of a hard time in Adobe Raw making my ISO 100 image look decent. I had to lower the shadows, lower the color saturation, etc. Whereas the ISO 1600 image needed almost no adjustment. So, there IS an advantage in shooting at the higher ISO.
 
Ok, I am FINALLY convinced that increasing the gain of sensors by raising ISO is not the same as ISO in film. Iliah finally overcame my obtuseness. But I STILL recommend shooting in higher ISO's in dim light.
That's normally the correct strategy for noise purposes because most cameras have lower read noise at higher ISOs (up to a point). The penalty you pay for raising ISO is NOT noise, it's DR!
Why?

Because conducting an experiment to convince myself that I was wrong, I had a hell of a hard time in Adobe Raw making my ISO 100 image look decent. I had to lower the shadows, lower the color saturation, etc. Whereas the ISO 1600 image needed almost no adjustment. So, there IS an advantage in shooting at the higher ISO.
Plug your camera into this chart at PhotonstoPhotos to get good guidance about which ISO settings will yield improvements in shadow noise (i.e., mostly read noise) on your particular camera.
Just want to add that taking advantage of ISO-invariant ranges is only advantageous if you shoot RAW, and that the IQ advantages of doing so may not be worth the operational disadvantages.
I had to read that twice to figure out what you meant. :-) Since the OP and I were talking about the noise benefits of increasing ISO, which is due to most sensors being ISO variant (not invariant) to at least some extent, your response about "ISO-invariant ranges" is turning things on their head relative to the prior discussion. Yes, it is true that if you have an ISO invariant camera/range, there is a potential (raw) advantage to NOT increasing ISO due to the increase in DR at the lowest ISO setting in the invariant range. Whether it's worth the tradeoff in extra processing and dark in-camera previews will depend on the shooting conditions, workflow, photographer preferences, etc.
 
Perhaps each camera is different. With mine, images shot any higher than ISO 400 are going to be unacceptably noisy to me, and images shot at ISO 200 (my camera's lowest setting) have more clipped highlights than those shot at ISO 400, so my camera stays at ISO 400 all the time. In light too dim for handheld exposures at that ISO, I either mount a flash or put the camera away. Following this procedure, noise is never an issue, and I never have to think about ISO.
I'm sure they are.

But what was beaten into my head was that raising the ISO does NOT Raise the sensitivity of the sensor. What it does is increase the amplification of the photons that the sensor receives.
Photons can't be amplified.

And you don't need to amplify them. All you need to do is specify is how the number of photons captured should be translated into tones in the final image. That has nothing to do with gain.
The root of the ISO fallacy is viewing photography as 'light in-light out'. Then, the thinking goes, to make a dark image 'brighter' you need 'more light', so the light needs to be 'amplified', hence 'gain'.
Nobody in their right mind thinks that. You are attributing a misconception that does not exist to your critics and then attacking them for it, which is very disingenuous.
 
Ok, I am FINALLY convinced that increasing the gain of sensors by raising ISO is not the same as ISO in film. Iliah finally overcame my obtuseness. But I STILL recommend shooting in higher ISO's in dim light.
That's normally the correct strategy for noise purposes because most cameras have lower read noise at higher ISOs (up to a point). The penalty you pay for raising ISO is NOT noise, it's DR!
Why?

Because conducting an experiment to convince myself that I was wrong, I had a hell of a hard time in Adobe Raw making my ISO 100 image look decent. I had to lower the shadows, lower the color saturation, etc. Whereas the ISO 1600 image needed almost no adjustment. So, there IS an advantage in shooting at the higher ISO.
Plug your camera into this chart at PhotonstoPhotos to get good guidance about which ISO settings will yield improvements in shadow noise (i.e., mostly read noise) on your particular camera.
Just want to add that taking advantage of ISO-invariant ranges is only advantageous if you shoot RAW, and that the IQ advantages of doing so may not be worth the operational disadvantages.
I had to read that twice to figure out what you meant. :-)
What, did I not say some version of "advantage" enough? :-D
Since the OP and I were talking about the noise benefits of increasing ISO, which is due to most sensors being ISO variant (not invariant) to at least some extent, your response about "ISO-invariant ranges" is turning things on their head relative to the prior discussion. Yes, it is true that if you have an ISO invariant camera/range, there is a potential (raw) advantage to NOT increasing ISO due to the increase in DR at the lowest ISO setting in the invariant range. Whether it's worth the tradeoff in extra processing and dark in-camera previews will depend on the shooting conditions, workflow, photographer preferences, etc.
Yeah -- my bad on my confusing post. I'll take the hit.
 
Perhaps each camera is different. With mine, images shot any higher than ISO 400 are going to be unacceptably noisy to me, and images shot at ISO 200 (my camera's lowest setting) have more clipped highlights than those shot at ISO 400, so my camera stays at ISO 400 all the time. In light too dim for handheld exposures at that ISO, I either mount a flash or put the camera away. Following this procedure, noise is never an issue, and I never have to think about ISO.
I'm sure they are.

But what was beaten into my head was that raising the ISO does NOT Raise the sensitivity of the sensor. What it does is increase the amplification of the photons that the sensor receives.
Photons can't be amplified.

And you don't need to amplify them. All you need to do is specify is how the number of photons captured should be translated into tones in the final image. That has nothing to do with gain.
The root of the ISO fallacy is viewing photography as 'light in-light out'. Then, the thinking goes, to make a dark image 'brighter' you need 'more light', so the light needs to be 'amplified', hence 'gain'.
Nobody in their right mind thinks that.
That's the problem, there are too many wrong-minded people :-P
You are attributing a misconception that does not exist
You will need Houdini-level skills to explain the following differently from the misconception Bob describes (taken from https://www.colesclassroom.com/understanding-iso )

The job of ISO is merely to amplify the light signal that the camera receives. In the process, it can produce a similar effect as opening the aperture or using a slow shutter speed.
to your critics and then attacking them for it, which is very disingenuous.
 
Last edited:
Perhaps each camera is different. With mine, images shot any higher than ISO 400 are going to be unacceptably noisy to me, and images shot at ISO 200 (my camera's lowest setting) have more clipped highlights than those shot at ISO 400, so my camera stays at ISO 400 all the time. In light too dim for handheld exposures at that ISO, I either mount a flash or put the camera away. Following this procedure, noise is never an issue, and I never have to think about ISO.
I'm sure they are.

But what was beaten into my head was that raising the ISO does NOT Raise the sensitivity of the sensor. What it does is increase the amplification of the photons that the sensor receives.
Has that realization changed the way you take pictures? One of the primary considerations for me when choosing camera settings is the straight-from-the-camera brightness of my pictures (I do not want to fiddle with that later in an image editing program). ISO 400 needs less exposure than ISO 200 to result in the same image brightness, so even if it is amplification rather than a change in the nature of the sensor itself, the effective sensitivity of my camera has changed with the ISO setting.
Not at all, although obviously I haven't had much time to take pictures... :-)
I don't know about your camera. My Sigma DP2M is basically noiseless at ISO's from 100 to 400, and I never had any intent to change settings.

With my Nikon's I set the ISO to match the light and STILL give me a fast enough shutter speed to eliminate blur on moving objects. The greater the light, the lower the ISO. I have no reason to change that as well.
The camera I find optimal at ISO 400 is my Samsung DSLR. I also have a Sigma, but it's the earlier DP2s. With that camera, I keep the ISO at 100 for color, but I do like ISO 400 if I'm going to do black and white conversions. In that regard, it reminds me of my usual color and black and white films in years gone by, Fuji Astia for color slides and Kodak Tri-X for black and white.
 
The camera I find optimal at ISO 400 is my Samsung DSLR. I also have a Sigma, but it's the earlier DP2s. With that camera, I keep the ISO at 100 for color, but I do like ISO 400 if I'm going to do black and white conversions. In that regard, it reminds me of my usual color and black and white films in years gone by, Fuji Astia for color slides and Kodak Tri-X for black and white.
With my Nikon 7000 and 7100, ISO 6400, with good light has very little noise. I also owned a DP2, and my present DP2M isn't much different. I keep my Sigma on ISO 400.

Below are a couple of shots from the Nikon 7000 at 6400

fb09fcd8a2d6472fb8aafc34461a69ee.jpg



5810d266c160488fa2650ea314f3195d.jpg
 
You are attributing a misconception that does not exist
You will need Houdini-level skills to explain the following differently from the misconception Bob describes (taken from https://www.colesclassroom.com/understanding-iso )

The job of ISO is merely to amplify the light signal that the camera receives. In the process, it can produce a similar effect as opening the aperture or using a slow shutter speed.
Not even remotely true. Tell me, aside from lightness, what do the following photos have in common, and what is different, if photos are taken of the same scene with the same camera and lens:
  1. f/2.8 1/200 ISO 100
  2. f/5.6 1/200 ISO 400
  3. f/2.8 1/800 ISO 400
  4. f/5.6 1/800 ISO 1600
In what way is the ISO setting "amplifying the light"? In what way does it "produce a similar effect as opening the aperture or using a slow shutter speed"? Aside, of course, from resulting in photos with the same lightness.
 
Ok, I am FINALLY convinced that increasing the gain of sensors by raising ISO is not the same as ISO in film. Iliah finally overcame my obtuseness. But I STILL recommend shooting in higher ISO's in dim light.

Why?

Because conducting an experiment to convince myself that I was wrong, I had a hell of a hard time in Adobe Raw making my ISO 100 image look decent. I had to lower the shadows, lower the color saturation, etc. Whereas the ISO 1600 image needed almost no adjustment. So, there IS an advantage in shooting at the higher ISO.
Shooting long exposures increases dark signal and dark noise, so increasing ISO by a couple of stops will certainly improve the darker parts of the image, since dark signal increases with time.

There's a sweet spot on most cameras where the reduction in dark noise hits the increase in regular noise and there's no point in going any further.

My D800 looked better with a 15s exposure at ISO 800 than a 2 minute exposure at ISO 100.
 
Perhaps each camera is different. With mine, images shot any higher than ISO 400 are going to be unacceptably noisy to me, and images shot at ISO 200 (my camera's lowest setting) have more clipped highlights than those shot at ISO 400, so my camera stays at ISO 400 all the time. In light too dim for handheld exposures at that ISO, I either mount a flash or put the camera away. Following this procedure, noise is never an issue, and I never have to think about ISO.
I'm sure they are.

But what was beaten into my head was that raising the ISO does NOT Raise the sensitivity of the sensor. What it does is increase the amplification of the photons that the sensor receives.
Photons can't be amplified.

And you don't need to amplify them. All you need to do is specify is how the number of photons captured should be translated into tones in the final image. That has nothing to do with gain.
The root of the ISO fallacy is viewing photography as 'light in-light out'. Then, the thinking goes, to make a dark image 'brighter' you need 'more light', so the light needs to be 'amplified', hence 'gain'.
Nobody in their right mind thinks that.
That's the problem, there are too many wrong-minded people :-P
You are attributing a misconception that does not exist
You will need Houdini-level skills to explain the following differently from the misconception Bob describes (taken from https://www.colesclassroom.com/understanding-iso )

The job of ISO is merely to amplify the light signal that the camera receives. In the process, it can produce a similar effect as opening the aperture or using a slow shutter speed.
"Merely"? Oh my.

The job of ISO is to relate the light captured in an exposure to the tones shown in the final image. That's it. You can amplify and multiply the "signal that the camera receives" all you want. Heck, make it a million volts. But what does that mean in terms of image? Until you define that relationship between the light captured and tones in the image, you have nothing. That's what ISO does, and it does that independently of medium, whether digital, film, tattoo artist, or sunlight on skin.
to your critics and then attacking them for it, which is very disingenuous.
--
Jeff
http://www.flickr.com/photos/jck_photos/sets/
 
Last edited:

Keyboard shortcuts

Back
Top