Downside of using below native ISO ?

Joined
Jul 28, 2016
Messages
46
Location
California
Lower noise due to (effectively) longer exposure, giving higher signal to noise ratio.
You cannot have higher SNR but same DR. DR is defined as the range from max highlights to shadows with a specified SNR.
Extended low ISO is using native ISO (200), and overexposing. Sometimes that improves noise in the shadows without clipping the highlights. If you shoot RAW, you are better off taking control of exposure correction yourself, e.g., using ETTR/expose-for-the-highlights.
 
Joined
Apr 20, 2020
Messages
1,220
Location
Beaumaris, Melbourne, Australia
You cannot have higher SNR but same DR. DR is defined as the range from max highlights to shadows with a specified SNR.
Extended low ISO is using native ISO (200), and overexposing. Sometimes that improves noise in the shadows without clipping the highlights. If you shoot RAW, you are better off taking control of exposure correction yourself, e.g., using ETTR/expose-for-the-highlights.
.

Tell it to Bill Claff ...

I am telling you what I have observed, not some theoretically predicted outcome. However, it is also what Bill has found from his testing.

I also always shoot RAW + LSF JPEG. I use the resulting two files for different purposes.

You appear to be ignoring the specific effects of the sensor and signal processing chain. What you have said holds good for my E-30 and E-M1 MkI. It does not hold good for my E-M1 MkII.
 

doady

Mu-43 Regular
Joined
May 18, 2020
Messages
106
Location
Canada
DR and SNR of E-M1 mk2 is not maximized at ISO 200, because native ISO is actually ISO 250. Highest DR and the least noise is at ISO 250, not ISO 200. Olympus shifts the true ISO 1/3 stop or something, I don't understand the technical details, but yeah I don't go below 250 to maximize SNR. I always ETTR and shoot RAW anyways so I don't see the point of using fake low ISO values. I'd rather just proper techniques and forget about shortcuts, especially those that can come at the cost of DR, which I try to maximize as much as SNR.
 
Joined
Jul 28, 2016
Messages
46
Location
California
.

Tell it to Bill Claff ...

I am telling you what I have observed, not some theoretically predicted outcome. However, it is also what Bill has found from his testing.

I also always shoot RAW + LSF JPEG. I use the resulting two files for different purposes.

You appear to be ignoring the specific effects of the sensor and signal processing chain. What you have said holds good for my E-30 and E-M1 MkI. It does not hold good for my E-M1 MkII.
Not sure what you mean. I agree with Bill Claff's methodology and learned from him quite a bit.
The best DR you get is using native iSO and saturating sensors. This has been demonstrated empirically and theoretically for any sensor and camera.
I assume that we have a misunderstanding.
 
Joined
Apr 20, 2020
Messages
1,220
Location
Beaumaris, Melbourne, Australia
Not sure what you mean. I agree with Bill Claff's methodology and learned from him quite a bit.
The best DR you get is using native iSO and saturating sensors. This has been demonstrated empirically and theoretically for any sensor and camera.
I assume that we have a misunderstanding.
See here:

https://www.photonstophotos.net/Charts/PDR.htm#Olympus E-30,Olympus OM-D E-M1,Olympus OM-D E-M1 Mark II

I try things. If they work, they work. If they don't, I find out why. Then I change what I do so that it works better.
 
Joined
Jul 28, 2016
Messages
46
Location
California
DR and SNR of E-M1 mk2 is not maximized at ISO 200, because native ISO is actually ISO 250. Highest DR and the least noise is at ISO 250, not ISO 200. Olympus shifts the true ISO 1/3 stop or something, I don't understand the technical details, but yeah I don't go below 250 to maximize SNR. I always ETTR and shoot RAW anyways so I don't see the point of using fake low ISO values. I'd rather just proper techniques and forget about shortcuts, especially those that can come at the cost of DR, which I try to maximize as much as SNR.
Agreed. One thing that many do not realize is that, simplified, if you expose one stop below sensor saturation, you loose one stop of DR.
I think most cameras have enough DR so that is not a big issue, but I also do not complain about m43's lack of DR :).
 
Joined
Apr 20, 2020
Messages
1,220
Location
Beaumaris, Melbourne, Australia
I know these charts and have provided Bill with data for several of his charts. Could you please elaborate how they disagree with what I have said or show that you have less noise with pulled ISO? Thanks!
With such knowledge and experience, I would suggest you share it. The chart I linked to shows pretty much what I previously wrote.

I do not upload full size images to my web site in the public areas, or for public use.

If I wanted to be the victim of an Inquisition, I would have kept posting at DPR ...
 
Joined
Jul 28, 2016
Messages
46
Location
California
With such knowledge and experience, I would suggest you share it. The chart I linked to shows pretty much what I previously wrote.

I do not upload full size images to my web site in the public areas, or for public use.

If I wanted to be the victim of an Inquisition, I would have kept posting at DPR ...
I apologize, I did not mean to offend or to sound like inquisition. My point in posting in this forum is either to learn from others, or to share what I know.

Bill's chart show same or less DR for pulled ISO. This means that, you do not get higher SNR with pulled ISO. Also, the SNR is not increased by longer exposure alone. The higher the signal the better the SNR, but there is a max value that the signal can reach. The max signal is reached when sensor is saturated at native ISO. That is why DR can never be better with pulled ISO, unless some NR is added.
I hope this clarifies what I meant.
 
Joined
Apr 20, 2020
Messages
1,220
Location
Beaumaris, Melbourne, Australia
I apologize, I did not mean to offend or to sound like inquisition. My point in posting in this forum is either to learn from others, or to share what I know.

Bill's chart show same or less DR for pulled ISO. This means that, you do not get higher SNR with pulled ISO. Also, the SNR is not increased by longer exposure alone. The higher the signal the better the SNR, but there is a max value that the signal can reach. The max signal is reached when sensor is saturated at native ISO. That is why DR can never be better with pulled ISO, unless some NR is added.
I hope this clarifies what I meant.
No worries. I suffer from chronic pain, and both it, and the painkillers I take, can make me a bit short tempered.

I am merely saying what I have observed with my cameras.

While most shots are not particularly affected by small reductions in DR, the reduction in noise by effectively doubling the exposure can be significant.

This is not really an ETTR method, as the camera is compensating for this effective doubling of exposure by changes to the amplification by the ADC. This is different from 'fooling' the camera by using an ETTR approach.

The E-30 is quite a noise machine even at ISO 200, and the noise is quite noticeably reduced by shooting at ISO 100 when possible. It also has some other peculiarities, in that the intermediate digital ISOs are noisier than going to the next full stop ISO setting. The full stop settings appear to be done using the ADC.
 

pdk42

One of the "Eh?" team
Joined
Jan 11, 2013
Messages
7,243
Location
Leamington Spa, UK
High DR helps in post when you need to recover details in the shadows so that they fit into the DR of the printer or the screen.
I understand your point here. It's not difficult to find scenes in nature with very high DR. Our eyes, by darting around, and working with our brain can make it appear as if the whole scene is visible - we can see detail in both shadows and highlights. Our cameras can't do that and our output media can't either. So we need to compress the DR of the real scene. That's one of the things that landscape photographers in particular do all the time. We are trying to somehow get a lifelike rendering of a scene but crushing maybe 16 stops of DR in the wild into 8 stops on a monitor or print.

No amount of metering will completely help here btw - whatever exposure we make can't possibly be right for all parts of the image. Best we can do on a single exposure is expose for the highlights (saturate but not over saturate them) and hope we can pull enough detail later from the shadows before noise becomes an issue. A multi exposure stack is often a better approach.
 
Last edited:

BushmanOrig

Mu-43 Top Veteran
Joined
Mar 31, 2015
Messages
535
Location
Switzerland
Real Name
Siegfried
A bit more of a technical explanation:

- A sensor works by turning incoming photos into electric charge. As photons arrive at the sensor sites (sensels), they liberate electrons which are then accumulated in a charge well - one for each sensel. At the end of the exposure, the sensels will be filled with a certain charge and this is then turned into a digital value by various electronic elements - the final one being a Digital to Analogue Converter (DAC).

- A certain number of photons will completely fill the well and this will be turned into the highest numerical value for that sensel. If we have more than the necessary photons to fill the well, then they will be ignored since once the well is full, it can't be filled any more. If this happens, we get "blown highlights".

- At low light levels, the charge accumulated will be small and various noise elements will be visible as well as the signal from the light. These noise elements will include those from random photon effects (shot noise) through to noise introduced by the electronics (e.g. read noise).

- The DR of the sensor will be in effect the range of values between the point at which noise is not problematic and when we get blown highlights. The definition of "problematic" varies a little so DR is somewhat of a subjective measure according to who is measuring it.

- The efficiency of turning photons into charge is a feature of the sensor alone and cannot be changed by controls on the camera (not even the ISO control). Camera manufacturers will express this sensitivity in terms of the camera's base ISO. The base ISO represents an equivalence to the old photographic "speed" rating we all know and love.

- In principle, a scene exposed using an ISO-calibrated light meter at the same ISO as the sensor will generate an image where the final output has the "correct" brightness. If the contrast is not too high, this means that the sensels won't be overfilled and we won't have any "blown highlights". Likewise, there will be enough light in the dark areas that preserve some detail and don't lose it to the noise.

- If we underexpose the sensor by not giving it enough light then the wells will be underfilled and the photo will look dark unless we do something. We do that by increasing the ISO dial on the camera. But that's not going to change the sensor sensitivity - all it does is amplify the signal. The amplification could be done before the DAC (so an analogue amplification), or after (so a digital amplification). Amplification will brighten the image, but it'll also brighten the noise so we'll see more noise as the ISO increases.

- But what if we decrease the ISO dial below the base ISO? What's happening there? Well, what's happening is that the camera is exposing the sensor as if it's less sensitive - i.e. it's boosting the exposure. It records this fact in the information with the image and then when the image is rendered for viewing, it pulls the brightness back to compensate for this over exposure. This pulling back will happen in-camera for JPEGs, but in your favourite raw editor if you use the raw. The camera is telling the raw processor what the mid-grey numerical value is on the image and the processor will do as it's told. In effect, it's negative amplification.

- You can do your own negative amplification at base ISO by doing your own over-exposing and your own pulling back - the effect will be the same. It's called ETTR - Expose To The RIght - from the effect on the histogram of moving it to the right.

- This negative amplification trick works fine, and it will reduce the noise in the image - which is great. But, the over exposure that's at its base risks over-filling the wells and giving blown highlights. So, it's a technique that only works on some scenes. In fact, you can see this yourself very easily by looking at the histogram or the over-exposure indicators when you choose a low ISO setting since you will see it showing over exposure more readily than at base ISO.

- So, overall it's just a trick - not at all the same as the camera having a lower base ISO.
Great explanation - you welcome to use this image next time you explain ISO and dynamic range...

Paul-Bucket-Theory.jpg
Subscribe to see EXIF info for this image (if available)
 
Last edited:
Links on this page may be to our affiliates. Sales through affiliate links may benefit this site.
Mu-43 is a fan site and not associated with Olympus, Panasonic, or other manufacturers mentioned on this site.
Forum post reactions by Twemoji: https://github.com/twitter/twemoji
Copyright © 2009-2019 Amin Forums, LLC
Top Bottom