Get less ISO noise at higher ISO sensitivity

dolina

dolina

Messages
631
Edit My Images
Yes
This has been verified by me with EF bodies from 2003 EOS 10D up to 2015 EOS 5Ds R.

RF bodies like the 2020 EOS R5 & 2022 EOS R7 may have this. Owners may want to verify.

To reduce whatever ISO noise Canon has a weird ISO increments of every ISO 160.

So ISO

- 160
- 320
- 640
- 1250
- 2500
- 5000
- 10,000
- so on and so forth

These ISOs are cleaner than the next 1/3rd bump or bump down
 
dolina said:
This has been verified by me with EF bodies from 2003 EOS 10D up to 2015 EOS 5Ds R.

RF bodies like the 2020 EOS R5 & 2022 EOS R7 may have this. Owners may want to verify.

To reduce whatever ISO noise Canon has a weird ISO increments of every ISO 160.

So ISO

- 160
- 320
- 640
- 1250
- 2500
- 5000
- 10,000
- so on and so forth

These ISOs are cleaner than the next 1/3rd bump or bump down
Click to expand...
This is something that has been known about for years. ISO125 is ISO100 amplified so that will increase the noise, but ISO160 is ISO 200 de-amplified so that decreases the noise. It is something that has been discussed several times on the other forum that you have posted this on.
 
L320Rio said:
This is something that has been known about for years. ISO125 is ISO100 amplified so that will increase the noise, but ISO160 is ISO 200 de-amplified so that decreases the noise. It is something that has been discussed several times on the other forum that you have posted this on.
Click to expand...
So we should use the base iso and its multiples for least relative gain amplification?
 
L320Rio said:
This is something that has been known about for years. ISO125 is ISO100 amplified so that will increase the noise, but ISO160 is ISO 200 de-amplified so that decreases the noise. It is something that has been discussed several times on the other forum that you have posted this on.
Click to expand...
Not everyone is aware of this so be kind and polite. ;)
 
dolina said:
Not everyone is aware of this so be kind and polite. ;)
Click to expand...
I certainly had not heard about it before :)

I am guessing that it does not apply to all makes, but it would be interesting to know if there was a similar thing on other brands, and what the figures are.
 
As it happens, Fuji gives ISO 160 as it's base, so us Fuji users are quite used to this. I didn't realise that other manufacturers had been doing the same though.
 
Anyone please correct me if I am wrong, amplification affects read noise yes?

Read Noise in DNs versus ISO Setting

www.photonstophotos.net

Select Canon 60D to see the noise jump. The R models affected however I dont know enough and could be the invariance of the sensor and now has 2 steps instead of 1 like the Fuji
 
Last edited by a moderator:
Thanks for sharing yes it’s has been something that’s been known for a while but still worth sharing , I used to use my 40D at ISO 320
I actually recently looked at some pictures that I took on my 40D at ISO 320 and they looked very clean
with current Canon bodies though noise isn’t an issue with my R5 I just use ISO 400, 800 or 1600 :)
 
Last edited:
I read about this many years ago when I used canon , it was tried out by quiet a few well known pro toggers and found to be total tosh ..far to many variables in actual use
 
acrobatic_Citron said:
Anyone please correct me if I am wrong, amplification affects read noise yes?

Read Noise in DNs versus ISO Setting

www.photonstophotos.net

Select Canon 60D to see the noise jump. The R models affected however I dont know enough and could be the invariance of the sensor and now has 2 steps instead of 1 like the Fuji
Click to expand...
I don't think those figures are anything but numbers on paper.

If you choose two cameras on there, then look at the test shots on DPReview, the test shots don't echo what you might expect from those graphs.

And there is a warning that they are unsuitable for comparing cameras
 
ISO with a sensor is not the same as ISO with film. With film it was related to the sensitivity but with a sensor it is the amplification. But...

High ISO does not cause noise. Low light causes noise. If you have low light then you have high ISO (for jpeg).

If you shoot RAW then ISO almost has no effect, it's just a jpeg thing. On Canon camera it's not quite true because of the order in which amplification works. It means that with Canon in low light you should use a higher iso for less noise.

If you want to know more search ISO Invariance on google.
 
acrobatic_Citron said:
Anyone please correct me if I am wrong, amplification affects read noise yes?

Read Noise in DNs versus ISO Setting

www.photonstophotos.net

Select Canon 60D to see the noise jump. The R models affected however I dont know enough and could be the invariance of the sensor and now has 2 steps instead of 1 like the Fuji
Click to expand...
Yes amplification affects read noise but you only need to amplify because your signal is too weak. Increase your aperture or exposure time to let more light in then you do not need to amplify so lower noise.
 
miscat said:
ISO with a sensor is not the same as ISO with film. With film it was related to the sensitivity but with a sensor it is the amplification. But...

High ISO does not cause noise. Low light causes noise. If you have low light then you have high ISO (for jpeg).

If you shoot RAW then ISO almost has no effect, it's just a jpeg thing. On Canon camera it's not quite true because of the order in which amplification works. It means that with Canon in low light you should use a higher iso for less noise.

If you want to know more search ISO Invariance on google.
Click to expand...
ISO invariance only does not apply to all cameras, and with some where it does apply, there are also 'steps' where (similar to the OP's suggestion) the graph changes.
 
Faldrax said:
ISO invariance only does not apply to all cameras, and with some where it does apply, there are also 'steps' where (similar to the OP's suggestion) the graph changes.
Click to expand...
Indeed Canon is bit different but on the general rule is invariance. It's a bit chicken and egg but we only use high iso to compensate for not enough light and it is the not enough light that causes noise. It's a mathematical thing.

Yes there are little steps in practice.
 
miscat said:
Indeed Canon is bit different but on the general rule is invariance. It's a bit chicken and egg but we only use high iso to compensate for not enough light and it is the not enough light that causes noise. It's a mathematical thing.

Yes there are little steps in practice.
Click to expand...
It's very much a sensor design feature that is present in more modern cameras, rather than older cameras.
Sony, for example, did not have ISO invariant sensors prior to the A7ii.
 
Faldrax said:
It's very much a sensor design feature that is present in more modern cameras, rather than older cameras.
Sony, for example, did not have ISO invariant sensors prior to the A7ii.
Click to expand...
I think it's what order the amplification happens. Canon do their's in a weird order that mans in low light higher ISO is leww noisy, up to a point.

I know about Canon because that's what I have and I do some astro photography and when doing that I use quite a high ISO even with RAW because the noise is slightly less that if I'd used ISO 100 or therebouts. So high ISO better in for me in this situation where the light is low and I can't do anything about it. It's not a good feature, I'd rather it was invariant.

Here's an interesting link that's worth a read...

www.lonelyspeck.com

How to Find the Best ISO for Astrophotography: Dynamic Range and Noise

ISO that is probably most misunderstood camera setting. In this post, Ian Norman talks about the craziness surrounding ISO settings, how ISO actually affects exposure and how to find the optimal IS…
www.lonelyspeck.com www.lonelyspeck.com
 
Faldrax said:
It's very much a sensor design feature that is present in more modern cameras, rather than older cameras.
Click to expand...
Actually, many old cameras were also ISO invariant. This is the chart that shows ISO invariance, this is the first Nikon D1 and the latest Nikon Z9... their flat lines show that both are essentially completely ISO invariant (except that the Z9 is dual gain). But ISO invariance is not standard for modern cameras by any means... note the line for the Nikon D6 (D4/D5 are similar). What makes ISO invariance useful in modern cameras is a reduction in read noise AND a simultaneous reduction in conversion errors/write noise.

Screen-Shot-2022-09-15-at-9.17.49-AM.jpg


miscat said:
High ISO does not cause noise. Low light causes noise.
Click to expand...
This is true... low light is a weaker signal with more photon shot noise (poisson noise). ISO just amplifies that recorded signal along with the read errors (noise), which then can cause more conversion/write errors.
In fact, you can use a higher ISO in bright light, and record less noise than you will using the same ISO in low light; by about 2 stops generally... you just don't normally need to use a high ISO in bright light. It's also why many night images have variable noise levels in them (w/ darker areas being noisier).
 
Last edited:
Sangoma said:
I don't think those figures are anything but numbers on paper.

If you choose two cameras on there, then look at the test shots on DPReview, the test shots don't echo what you might expect from those graphs.

And there is a warning that they are unsuitable for comparing cameras
Click to expand...
Well, they are just numbers; and they have to be put into the right context to have any real meaning.
Similarly, you have to put DPR's studio shot comparisons into the right perspective... they change the ISO by changing the shutter speed, not by reducing the light available; which is not quite the same thing. It's the difference between choosing a higher ISO for a higher SS to freeze motion in good light, versus the ISO climbing higher because you are running out of light at sunset.
 
I wasn’t aware of this and whilst it’s interesting I’m not sure it’d make any difference to my shooting. Aperture and shutter speed are usually far more important to me than ISO so ISO just has to be whatever it needs to be to get the exposure I need. I don’t worry about noise until over 6400 and even then I can just use topaz these days (y)
 
snerkler said:
Aperture and shutter speed are usually far more important to me than ISO
Click to expand...
Always; and those small differences less that .5 stop aren't worth much effort trying to manage... that's about the min limit of being discernable visibly.

But if you can go up 1/3 stop and end up with less noise than a lower ISO, especially if it that lower ISO has to be pushed in post, then it's probably worth doing. So, if you are a studio/landscape/fine art photographer it might be worthwhile.
 
sk66 said:
Always; and those small differences less that .5 stop aren't worth much effort trying to manage... that's about the min limit of being discernable visibly.

But if you can go up 1/3 stop and end up with less noise than a lower ISO, especially if it that lower ISO has to be pushed in post, then it's probably worth doing. So, if you are a studio/landscape/fine art photographer it might be worthwhile.
Click to expand...
Very true, but I was only talking for myself. Landscapes I shoot at base ISO, and I don't do studio or fine art photography (y)
 
You must log in or register to reply here.
Back
Top