Noise at low ISO

yousaf

Senior Member
Messages
1,206
Solutions
3
Reaction score
86
Location
lahore, PK
why is there Noise even at low ISO ? It is especially visible under the wings.

cb28c06cb5c34103b31b74180156c5b2.jpg




b9711d452cf54f169fca1cf96c0aa857.jpg




b46fb5235d8f4810bae22d27b24b21b7.jpg
 
It’s difficult to tell anything at the tiny size pictures you posted. But it’s the shadows, where you’d expect noise.



The ISO isn’t the deciding factor in noise, the amount of light gathered is. ISO is a rough indicator of that. In the shadows, there’s less light gathered, so noise appears. The shadows are effectively exposed at a higher ISO.
 
Simple answer is that they're not exposed properly (under exposed). You'll notice the properly expose image doesn't have the same problem with noise.
 
Simple answer is that they're not exposed properly (under exposed). You'll notice the properly expose image doesn't have the same problem with noise.
Which just demonstrates that the source of noise is not high ISO but insufficient exposure.
 
It’s difficult to tell anything at the tiny size pictures you posted. But it’s the shadows, where you’d expect noise.
2 of the 3 pictures posted are 20 megapixels.
The ISO isn’t the deciding factor in noise, the amount of light gathered is. ISO is a rough indicator of that. In the shadows, there’s less light gathered, so noise appears.
Agreed with all of the above in most real-world scenarios on modern cameras.
The shadows are effectively exposed at a higher ISO.
You lost me here. Using the phrase "exposed at" implies at the time of capture. At the time of capture, the shadows were exposed at the same ISO as the rest of the image was-- ISO 200. Also, there's no indication the OP boosted shadows in post. So whether we're speaking literally or figuratively, the shadows were exposed at ISO 200 and no higher.
 
Noise is inherent in light itself. Light is made up of photons which arrive completely randomly and these random variations in the number of photons captured makes up most of the noise seen in photographs.

The brighter the light and the greater the exposure, the greater the number of photons captured by the camera. If a large number of photons are captured per pixel, then the random fluctuations are less significant and the noise is less noticeable.
 
The shadows are effectively exposed at a higher ISO.
You lost me here. Using the phrase "exposed at" implies at the time of capture. At the time of capture, the shadows were exposed at the same ISO as the rest of the image was-- ISO 200. Also, there's no indication the OP boosted shadows in post. So whether we're speaking literally or figuratively, the shadows were exposed at ISO 200 and no higher.
If there's ISO invariance at play, setting a higher ISO just multiplies the recorded pixel values to make the image brighter.

If the shadows are exposed at -5EV, you could declare the ISO was 6400, and the shadows would be mid grey.

The noise in the shadows is the same as if the whole picture was expose "correctly" for the shadows at 6400.

That's just a fact of there being less light in the shadows. Basically shadows are noisier, and you can get an idea of how noisy, by estimating what ISO would be needed for "correct" exposure.
 
Last edited:
When we talk about things looking noisy, we're describing tones with a low signal-to-noise ratio.

There are two main sources of noise: photon shot noise (the randomness of the light), and (electronic) read noise.

Darker tones are made up from less light (less signal) so, as you look at darker parts of your image, the impact made by photon shot noise and read noise starts to have more and more impact on the signal to noise ratio.

Or, to put it another way: darker parts of your image will always be noisier. The only solution is to capture more light to describe them.

Richard - dpreview.com
 
The shadows are effectively exposed at a higher ISO.
You lost me here. Using the phrase "exposed at" implies at the time of capture. At the time of capture, the shadows were exposed at the same ISO as the rest of the image was-- ISO 200. Also, there's no indication the OP boosted shadows in post. So whether we're speaking literally or figuratively, the shadows were exposed at ISO 200 and no higher.
If there's ISO invariance at play, setting a higher ISO just multiplies the recorded pixel values to make the image brighter.
First, the OP's 70D is not ISO invariant:

http://www.photonstophotos.net/Charts/PDR_Shadow.htm#Canon EOS 70D

So I really don't know why you're bringing ISO invariance into the discussion.

Second, I see what you're trying to get at above, but as written you're confusing what ISO-invariance is like for an end user versus how ISO invariance is achieved by sensor engineers. You're taking a figurative concept ("it's just like multiplying the pixel values from a lower ISO image!") and declaring it to be literally how these sensors work. I'd be careful with that.
If the shadows are exposed at -5EV, you could declare the ISO was 6400, and the shadows would be mid grey.

The noise in the shadows is the same as if the whole picture was expose "correctly" for the shadows at 6400.

That's just a fact of there being less light in the shadows. Basically shadows are noisier, and you can get an idea of how noisy, by estimating what ISO would be needed for "correct" exposure.
Labeling shadow noise with an ISO value different from the ISO at which the image was shot is convoluted, obfuscatory, unnecessary, and will perpetuate the mistaken belief that ISO causes noise. Please don't do that.
 
Last edited:
Simple answer is that they're not exposed properly (under exposed). You'll notice the properly expose image doesn't have the same problem with noise.
Which just demonstrates that the source of noise is not high ISO but insufficient exposure.
These are really two ways of looking at the same thing: signal-to-noise ratio is higher at higher Exposure Index, which means less light gathered. Underexposing by 2 stops at ISO 200 (and then maybe pushing up to the desired level in the final output) is EI=800, the same as exposing correctly at ISO 800: same light gathered, just different routes then taken to getting the desired output level placement. (And even if the underexposed image is not pushed up, its S/N ratio is about as bad as in the ISO 800 image, at least as far as photon shot noise is concerned.)

Also noise is most visible in very uniform parts of a scene, like those blue skies.

--
Smaller lenses, better in low light, more telephoto reach:
you can have any _two_ at one time.
 
Last edited:
The shadows are effectively exposed at a higher ISO.
You lost me here. Using the phrase "exposed at" implies at the time of capture. At the time of capture, the shadows were exposed at the same ISO as the rest of the image was-- ISO 200.
Perhaps it helps to use the correct jargon for amount of exposure, which is Exposure Index (as in the setting on a light meter). Roughly, the EI is the ISO setting that would render a particular subject at a suitable mid-tone level in a default JPEG conversion (around 18% of maximum). So one way to describe it is that when part of the scene is two stops darker than the mid-tones, and comes out two stop darker in default JPEGs (3.4%?) it gets two stops less exposure and that part is "exposed" at four times the exposure Index — and has the worse S/N that goes with that higher EI/low amount of exposure.

("ISO" is used for at least four related but different measures; usually the ambiguity is irrelevant, but sometimes it helps to distinguish exposure index from various measures of sensitivity and gain.)
 
ISO doesn't cause, create or add noise to a digital photo.

The predominant source of noise in a digital image is in the light captured at the sensor. Known as shot noise, it is proportional to the square root of the volume of light falling upon the sensor. If nine photons hit the sensor, there will be three bits of noise and a resulting signal-to-noise ratio of 3:1. If 81 photons strike the sensor, 9 bits of noise will also be captured. Strictly speaking, more noise is captured in the second image. However, the signal-to-noise ratio is higher at 9:1 and, as a result, noise is less prominent.

Exposure is strictly a function of scene brightness, f-stop and shutter speed. Two of the three photos you posted are of inherently dark subjects. The sides of the jets facing you were in shadow with very little light falling on them. At the exposure settings (f-stop and shutter speed) used, very little light was captured and this produced photos having a low SNR. Thus, the noise in the photos is prominently visible.
 
Simple answer is that they're not exposed properly (under exposed). You'll notice the properly expose image doesn't have the same problem with noise.
For this, I choose Av mode. Set aperture at F8 to keep all of the aircraft in focus. ISO was set at 200 (to keep the noise at a minimum, Poor soul :-()

Shutter speed was set by the camera itself, and still, it is underexposed. should I have used Ev compensation?
 
Simple answer is that they're not exposed properly (under exposed). You'll notice the properly expose image doesn't have the same problem with noise.
For this, I choose Av mode. Set aperture at F8 to keep all of the aircraft in focus. ISO was set at 200 (to keep the noise at a minimum, Poor soul :-()

Shutter speed was set by the camera itself, and still, it is underexposed. should I have used Ev compensation?
Yes. That's how it's usually done.

With propeller aircraft, it is often recommended to use exposure times long enough to blur the propeller blades.
 
why is there Noise even at low ISO ? It is especially visible under the wings.
Your camera does not have noise-free shadow areas at any ISO, does it?

d28f41ca04f746b2bd66fa9e5abe9f13.jpg
I don't think so. Which modern Mirrorless has noise-free shadows at ISO 800 or above?
Forget about ISO 800 or above. I'm showing you that you'll have noisy shadows not only at ISO 200 but also at ISO 100. If you never want to see noisy shadows you'll just have to apply some good noise reduction.
 
Last edited:

Keyboard shortcuts

Back
Top