Does the camera ISO change exposure?

Joined
Mar 31, 2015
Messages
436
Location
Switzerland
Real Name
Siegfried
  • Thread Starter Thread Starter
  • #61
Those interested, this video is one of the best I saw to date explaining what exposure is and the role of ISO. This is a potential photography life changer...

It also clears the well-known Exposure Triangle - this video explain a better way to understand exposure...

My original question was - does ISO change the exposure?

 
Last edited:

John King

Mu-43 Top Veteran
Joined
Apr 20, 2020
Messages
553
Location
Beaumaris, Melbourne, Australia
Sorry, Siegfried. I do not have the time or patience (or tolerance ...) for 26 minutes of another hyperactive American.

While not strictly correct, the exposure triangle has utility.

The hard-line alternative has little utility in assisting ordinary photographers in taking better photographs.

Take the classic exposure triangle, modify it by stating that one should almost always keep the ISO as low as is reasonable, and you have a good, valid, working rule of thumb. One that has a lot of utility. It is also extremely easy to understand.

The alternative view takes reams of explanation, even more reams of contention, and has little working utility.

I know which I prefer ...
 

swifty

Mu-43 Regular
Joined
Jan 30, 2013
Messages
158
Location
Singapore
Real Name
Dave
Sorry going slightly OT but since it was raised. I know many won’t agree with me but I’ll put this out there for people to research themselves and make up their own minds.

With regards to large/small pixels, it is largely an issue of fill factor. Fill factor is the proportion of the sensor surface area available for light capturing. Not 100% of the surface is available because things like supporting electronics and sensor walls take up room.
But with gapless microlenses and the move to BSI, the fill factor is approaching 100% for newer sensors. If they maintained the fill factor going to a higher pixel density sensor (smaller pixels), it won’t perform any worse. The sensor doesn’t get noisier if you increase pixel count but you potentially reduce the signal capturing surface area if you don’t maintain the fill factor.
Conversely if you reduce the pixel count (bigger pixels), it doesn’t make the sensor any better if the fill factor was already close to 100% before. You’d might be that little bit closer to 100% but it was already high before.

Fat/large pixels doesn’t reduce noise. It may only make it easier to achieve a higher fill factor so it was largely an issue with older sensors when fill factor was low.

Now what does increase noise is making it read faster. Global shutters aren’t a new thing, they’re just too noisy for use in photography compared to what we currently enjoy. So increasing total pixel count does have a bandwidth problem offloading more data.
 

exakta

Mu-43 Top Veteran
Joined
Jun 2, 2015
Messages
687
I know those explaining would already know this but using the terminology small sensors are noisier leads to the perception that small sensors are inherently noisy which I believe contributes to the negative perception of m43 by some sectors.
I used to own a Fuji X10 which has a 12MP sensor that is 1/2 the area of a 1" sensor, 1/4 the area of m43 and 1/16 the area of FF. It was still good enough for 16x20 prints.

It had a strange mode called "EXR" in which the camera reduced resolution to 6MP by using pixels in pairs, thus acting like a larger pixel that captured twice as much light. There were a few options for EXR mode, one was higher dynamic range and one was lower noise. In both modes, the improvement was roughly 2 f-stops.

Compared to my 16MP E-M10, shooting with the X10 at 6MP at the same ISO, noise and DR were as good or sometimes better despite the much smaller sensor. The lowered resolution only mattered with large prints.
 

swifty

Mu-43 Regular
Joined
Jan 30, 2013
Messages
158
Location
Singapore
Real Name
Dave
I used to own a Fuji X10 which has a 12MP sensor that is 1/2 the area of a 1" sensor, 1/4 the area of m43 and 1/16 the area of FF. It was still good enough for 16x20 prints.

It had a strange mode called "EXR" in which the camera reduced resolution to 6MP by using pixels in pairs, thus acting like a larger pixel that captured twice as much light. There were a few options for EXR mode, one was higher dynamic range and one was lower noise. In both modes, the improvement was roughly 2 f-stops.

Compared to my 16MP E-M10, shooting with the X10 at 6MP at the same ISO, noise and DR were as good or sometimes better despite the much smaller sensor. The lowered resolution only mattered with large prints.
Hi Exakta, I’m not familiar with that model but a quick google says it’s an 8+ year old camera so it’s hard to relate to current discussions. I used to own a Fuji S5 Pro that used a unique S-R dual pixel configuration where the R pixels were used to capture additional highlight information and combined to expand DR. It’s the only configuration where DR was expanded in the highlight region as far as I know whereas today we mostly talk about engineering DR where the extra DR is in the shadows and largely influenced by the read noise. So Fuji are known to do a few funky things with their sensors.

In terms of making comparisons, printing is a great neutraliser and a good reminder that we often over obsess too much about noise. But in terms of digital comparisons, it’s important to also compare things on an image or same output level, rather than only on the pixel level because two unequal resolution images will be magnified to a different extent when compared on a pixel level on-screen. Remember also that a higher resolution image can also be subjected to more aggressive NR compared to lower resolution image.

But just a small addition to my previous post.
So as fill factor approaches 100%, the entire sensor area becomes the best predictor of sensor performance. The assumption being fill factor is maintained at a high level but if you take things to extremes eg. a billion pixels on a m43 sensor, the model breaks down.
Much of today’s sensor improvements are now driven to improve sensor bandwidth and read speeds whilst maintaining the same, already excellent read noise levels.
Look at the development in 1” sensor as a precursor to what might occur in larger sensors.
 

exakta

Mu-43 Top Veteran
Joined
Jun 2, 2015
Messages
687
It's a great article, but he doesn't really say why the triangle works for film but not digital. The answer is that different films do have different sensitivities to light. Same camera, different film, different results.

This is by far the most important statement in the article:

set exposure according to your photographic goals, raising it to the limit (constraints: DoF, aberrations / resolution / diffraction, blur, clipping of highlights that are important to the composition of the shot)
In other words, assuming the sensor is truly linear, one could shoot in extreme darkness even with tiny apertures, say f22, and simply keep the shutter open as long as necessary. With Olympus cameras, the Live Composite mode works something like that. Just ETTR at base ISO and you have made the optimal exposure.

Why this isn't practical, of course, is two constraints mentioned in that sentence...the need to deal with blur (use faster shutter) or diffraction (use wider aperture). By abberations, I assume the reference is to an optimum aperture for each lens where it is the sharpest and has the least vignetting. DOF control of course is obvious.

That leads back to the dance between shutter speed vs. aperture for a given effect. Sometimes the shot does call for blurred motion, sometimes the shot does call for hyperfocal DOF and sometimes the combination of shutter and aperture leads to blown highlights even at base ISO...at that point, it's time for ND filters :roflmao:

So it could be reduced to:

1. Always ETTR at base ISO if possible. Use filters if shutter is too fast or aperture too small.

2. Only increase ISO if a faster shutter speed is required because of the chosen aperture.
 
Last edited:

PakkyT

Mu-43 Hall of Famer
Joined
Jun 20, 2015
Messages
3,329
Location
New England
Personal insults are not professional,
Wait, you are getting paid to be here? Man I missed out on that.

My comments about your behavior are based on your actual behavior here and you own posts of your battles over at the other site. Your dismissal of others' opinions as just repeating information without understanding it or your arrogance towards people here implying they are not smart enough to decipher your ramblings is easily just as insulting toward them as you are perceiving posts discussion your behavior as personal insults. Basically if you don't want people to say you are being a schmuck, then stop acting like a schmuck. Talking down to people doesn't make you look any smarter.


I gave very clear arguments for why I say Sony proofed the popular "large is better" narrative is not 100% relevant.
No, you did not. You alluded to it or so I thought but wasn't entirely sure. So I specifically asked you and you went off on one of your tangents again still not directly answering the question. It was only this last time asking your again to directly answer the question that you finally specified points you imagined you had made previously instead of the mind-reading games you wanted us to play with you.

Two full-frame sensors (same sensor size, photocell diameter, MegaPixels) and the one performs way better than the other.
I refer again to the A7 II and the A7 III example. This example is really important because it explains so much. Two full-frame sensors, both 24MP, same photocell size... But they perform hugely different... Why?
The huge difference is because of the very thing several of us have pointed out to you and you continue to ignore in order to support your narrow point of view. And that is bigger photocells produce better results than smaller photocells (now pay attention here) on sensors of the same type and generation. I know that bold part is very difficult for you to accept since it ruins your arguments. In your poorly picked example that does not support your argument, the two models were released four years apart so the newer model has years of sensor development over the old one and more importantly it is not even the same sensor type, the former being a CMOS sensor and the latter being a much better technology BSI-CMOS sensor. One of the fundamental changes the newer Back Side Illumination sensor does is allows more of the incoming light to actually get to the photocells. So even without making the photocells any larger the newer sensor collects more light.

It should be very clear that you are trying to compare two very different sensor types (apples and oranges) and there is a technical/scientific reason (not just marketing) why "they perform hugely different".
 
Joined
Mar 31, 2015
Messages
436
Location
Switzerland
Real Name
Siegfried
  • Thread Starter Thread Starter
  • #68
It's a great article, but he doesn't really say why the triangle works for film but not digital. The answer is that different films do have different sensitivities to light. Same camera, different film, different results.

This is by far the most important statement in the article:



In other words, assuming the sensor is truly linear, one could shoot in extreme darkness even with tiny apertures, say f22, and simply keep the shutter open as long as necessary. With Olympus cameras, the Live Composite mode works something like that. Just ETTR at base ISO and you have made the optimal exposure.

Why this isn't practical, of course, is two constraints mentioned in that sentence...the need to deal with blur (use faster shutter) or diffraction (use wider aperture). By abberations, I assume the reference is to an optimum aperture for each lens where it is the sharpest and has the least vignetting. DOF control of course is obvious.

That leads back to the dance between shutter speed vs. aperture for a given effect. Sometimes the shot does call for blurred motion, sometimes the shot does call for hyperfocal DOF and sometimes the combination of shutter and aperture leads to blown highlights even at base ISO...at that point, it's time for ND filters :roflmao:

So it could be reduced to:

1. Always ETTR at base ISO if possible. Use filters if shutter is too fast or aperture too small.

2. Only increase ISO if a faster shutter speed is required because of the chosen aperture.
Great points, same to Swifty, thank you.

It's amazing to think in both threads, ETTR and the role of ISO, optimizing exposure is so basic, interesting, and significant in terms of improving IQ. This basic know-how can help photographers improve image quality.

To summarize:

1. Exposure:
- Shutter Speed and Aperture determines how much light reach the sensor
- One can optimize the external environment with a flash or reflector plus plus
- Image sensors perform optimally when it receives the right amount of exposure
- The more accurate one expose the sensor, the better it performs - this is key
- Most of the time Olympus cameras will select safe & conservative exposures
- ETTR is a simple technique to help photographers find the "best" exposure
- Call it ETTR or "best" exposure @ any ISO, sensors are better at "best" exposure

2. The ISO setting
- The ISO setting adjusts image brightness (it's not important how it's done)
- The ISO does not change the amount of light reaching the sensor
- In most cases, image noise increase as the ISO increase
- A basic method to keep noise to a minimum is to select the exposure accurately,
and to keep the ISO as low as possible (IBIS is a great help to lower ISO)

One can spend hours talking more about the relevant inputs by Swifty and Exacta and yes it is very interesting, especially for the technical minded.

In most cases, I rely on my camera metering, in challenging conditions I prefer to apply these simple techniques to get higher-quality results...

To optimize your camera exposure, apply this simple method:-
  1. Fix the ISO to the value you require (Go to the SCP and select an ISO)
  2. By selecting an ISO you ensuring its fixed, the next steps only adjust exposure
  3. Set your aperture or shutter speed (A or S mode) and press the shutter halfway
  4. Check the histogram (the exposure meter should read 0EV)
  5. The camera "auto" mode will generally select a lower or conservative exposure
  6. If there is free space on the right of the histogram, increase the exposure
  7. Follow the histogram as it moves to the right. (Do not move it too far right)
  8. The exposure meter will show you overexposing by a 1/3, 2/3 or 1 stop
  9. When you are satisfied with the histogram, your exposure is optimized...
  10. You can now safely take the image
Your image will be a little overexposed. This is easy to correct in the camera with curves or in Photoshop Raw. The image will now have less noise and more image data to work with in post-processing.
 
Last edited:

barry13

Mu-43.com Editor
Joined
Mar 7, 2014
Messages
10,028
Location
Southern California
Real Name
Barry
Hi,
There's a lot of arguing here; some would say it's unfriendly.

In order to keep the site friendly, we have rules against repetitive arguing. Please keep that in mind.

Also, please note you can use the Ignore function to stop seeing posts from particular users.

Thanks
 

cjoliprsf

Mu-43 Regular
Joined
May 28, 2019
Messages
100
Location
Quebec, Canada
  1. Check the histogram (the exposure meter should read 0EV)
  2. The camera "auto" mode will generally select a lower or conservative exposure
  3. If there is free space on the right of the histogram, increase the exposure
  4. Follow the histogram as it moves to the right. (Do not move it too far right)
  5. The exposure meter will show you overexposing by a 1/3, 2/3 or 1 stop
  6. When you are satisfied with the histogram, your exposure is optimized...
Question: would I get same result by using the zebra function, and increase exposure until I see the zebra in the brighest portions of the image, and then decrease exposure just to make the zebra go away?
 

twigboy

Mu-43 Top Veteran
Joined
Sep 10, 2016
Messages
909
Location
Virginia
It's a great article, but he doesn't really say why the triangle works for film but not digital. The answer is that different films do have different sensitivities to light. Same camera, different film, different results.

This is by far the most important statement in the article:



In other words, assuming the sensor is truly linear, one could shoot in extreme darkness even with tiny apertures, say f22, and simply keep the shutter open as long as necessary. With Olympus cameras, the Live Composite mode works something like that. Just ETTR at base ISO and you have made the optimal exposure.

Why this isn't practical, of course, is two constraints mentioned in that sentence...the need to deal with blur (use faster shutter) or diffraction (use wider aperture). By abberations, I assume the reference is to an optimum aperture for each lens where it is the sharpest and has the least vignetting. DOF control of course is obvious.

That leads back to the dance between shutter speed vs. aperture for a given effect. Sometimes the shot does call for blurred motion, sometimes the shot does call for hyperfocal DOF and sometimes the combination of shutter and aperture leads to blown highlights even at base ISO...at that point, it's time for ND filters :roflmao:

So it could be reduced to:

1. Always ETTR at base ISO if possible. Use filters if shutter is too fast or aperture too small.

2. Only increase ISO if a faster shutter speed is required because of the chosen aperture.
What is base ISO and how do I know what my camera's sensor's base ISO is? Is it the lowest ISO the camera does without resorting to fiddly-diddly electronic magic to make a lower ISO? (E-P5 and E-PL2).
 

cjoliprsf

Mu-43 Regular
Joined
May 28, 2019
Messages
100
Location
Quebec, Canada
I remember "a long time ago" when boosting a film to increase its sensitivity. Sensitivity was then written as ASA (which I gather is pretty equivalent to the ISO we have now). There was a formula that permitted to take a 400 ASA film and shoot it setting the camera up to 6400, and then by leaving the film longer in the developer solution it was possible to compensate the fact the film was in fact grossly underexposed. The operation had as side effect a great increase of the grain size, pretty much as today we get an increase of noise.

These things were quite a lot more work in those days!
 

John King

Mu-43 Top Veteran
Joined
Apr 20, 2020
Messages
553
Location
Beaumaris, Melbourne, Australia
What is base ISO and how do I know what my camera's sensor's base ISO is? Is it the lowest ISO the camera does without resorting to fiddly-diddly electronic magic to make a lower ISO? (E-P5 and E-PL2).
"Base ISO" is generally the ISO at which the camera has its best dynamic range ("DR").

e.g. Base ISO for my E-30 was ISO 200, but it had noticeably less noise at ISO 100, and only slightly less DR.

Base ISO for most FTs and mFTs cameras is ISO 200.
 

John King

Mu-43 Top Veteran
Joined
Apr 20, 2020
Messages
553
Location
Beaumaris, Melbourne, Australia
I remember "a long time ago" when boosting a film to increase its sensitivity. Sensitivity was then written as ASA (which I gather is pretty equivalent to the ISO we have now). There was a formula that permitted to take a 400 ASA film and shoot it setting the camera up to 6400, and then by leaving the film longer in the developer solution it was possible to compensate the fact the film was in fact grossly underexposed. The operation had as side effect a great increase of the grain size, pretty much as today we get an increase of noise.

These things were quite a lot more work in those days!
ASA = American Standards Association, IIRC. ISO = International Standards Organisation.

The same principles apply to both digital and analog - More exposure (Shutter Speed x Aperture), the less noise. The ISO control on the camera changes how the camera electronics react to less exposure, but less exposure always equals more noise. Same for digital as for film, even though the processes are quite different in most ways.
 

Vivalo

Olympus loser
Joined
Nov 16, 2010
Messages
1,391
Has there been any research if the exposure time is affected by the day of the week? To my knowledge on weekends time passes by much faster, so does that make the exposure darker on weekends?
 
Joined
Mar 31, 2015
Messages
436
Location
Switzerland
Real Name
Siegfried
  • Thread Starter Thread Starter
  • #78
For those coming late to the thread and feeling as if they're going mad, I refer you to an earlier post I made:
There is no need to complicate a simple and friendly exposure technique that will only help fellow M43 photographers. See the basic and simple summary below... I cannot imagine that anybody would like to prevent photographers to learn new basic techniques...

It's amazing to think, optimizing exposure is so basic, interesting, and significant in terms of improving IQ. This basic know-how can help photographers improve image quality.

To summarize:

1. Exposure:
- Shutter Speed and Aperture determines how much light reach the sensor
- One can optimize the external environment with a flash or reflector plus plus
- Image sensors perform optimally when it receives the right amount of exposure
- The more accurate one expose the sensor, the better it performs - this is key
- Most of the time Olympus cameras will select safe & conservative exposures
- ETTR is a simple technique to help photographers find the "best" exposure
- Call it ETTR or "best" exposure @ any ISO, sensors are better at "best" exposure

2. The ISO setting
- The ISO setting adjusts image brightness (it's not important how it's done)
- The ISO does not change the amount of light reaching the sensor
- In most cases, image noise increase as the ISO increase
- A basic method to keep noise to a minimum is to select the exposure accurately,
and to keep the ISO as low as possible (IBIS is a great help to lower ISO)

One can spend hours talking more about the relevant inputs by Swifty and Exacta and yes it is very interesting, especially for the technical minded.

In most cases, I rely on my camera metering, in challenging conditions I prefer to apply these simple techniques to get higher-quality results...

To optimize your camera exposure, apply this simple method:-
  1. Fix the ISO to the value you require (Go to the SCP and select an ISO)
  2. By selecting an ISO you ensuring its fixed, the next steps only adjust exposure
  3. Set your aperture or shutter speed (A or S mode) and press the shutter halfway
  4. Check the histogram (the exposure meter should read 0EV)
  5. The camera "auto" mode will generally select a lower or conservative exposure
  6. If there is free space on the right of the histogram, increase the exposure
  7. Follow the histogram as it moves to the right. (Do not move it too far right)
  8. The exposure meter will show you overexposing by a 1/3, 2/3 or 1 stop
  9. When you are satisfied with the histogram, your exposure is optimized...
  10. You can now safely take the image
Your image will be a little overexposed. This is easy to correct in the camera with curves or in Photoshop Raw. The image will now have less noise and more image data to work with in post-processing.
 
Joined
Mar 31, 2015
Messages
436
Location
Switzerland
Real Name
Siegfried
  • Thread Starter Thread Starter
  • #79
Question: would I get same result by using the zebra function, and increase exposure until I see the zebra in the brighest portions of the image, and then decrease exposure just to make the zebra go away?
Hi, interesting question, thank you.

When setting the Zebra to 100% or 105%, the Zebra's will indicate areas already clipped, those areas that will go white.

Setting the Zebra to 95%, any indication will show pre-clipping areas.

I set my camera to 100%. The reason is I then know I am at the edge of the best exposure (histogram to the right), plus it helps me to quickly decide if I am OK with some parts clipping to white... For example, if there is a house with white walls in the image, I know the walls will clip first and I am OK if the walls are at the point of clipping, especially if other details are important to me. But this is a personal decision...

What I do not like about the Pansonic histogram, it's too small. The Olympus histogram is wider and its easier to read...

Tip:- In those cases, that you need to check the exposure more accurately, my advice is to fix the ISO first. For example in a city scene at night, select the ISO and then select the exposure using Zebras and the histogram. The moment you fix the ISO you able to follow changes to exposure (shutter & aperture) more easily or accurately...

Hope this helps

Siegfried
 

3dpan

Mu-43 Top Veteran
Joined
Mar 11, 2017
Messages
615
Location
Far North, New Zealand
Real Name
Alec
ISO in digital affects brightness or darkness of a JPG. The RAW file gets the same light regardless of the ISO setting, for a given f stop and shutter speed setting. F stop and shutter speed control how much light is hitting the sensor. Exposing using f stop and/ or shutter speed, at the sensors base ISO, with ETTR but not clipping highlights, is theoretically getting the most out of a sensor.
A larger sensor collects a greater total amount of light by area for the same f stop and shutter speed of a smaller sensor. A larger sensor thus has less noise because less amplification is required... The SNR is better. Think... 4 solar panels gather more than 1 solar panel although in the same light for the same time.
Sometimes, when we use Auto ISO with exposure compensation dial, and there is enough light to keep the shutter speed well above our defined slowest limit, true over or under exposure is achieved by letting in a different amount of light. If we hit say 1/60 as our slowest limit and the aperture is already wide open, and then use exposure compensation, a plus 1 exp comp is just making the ISO rise a stop to lighten the image... no different to boosting the RAW in post a stop.
Are you saying that increasing ISO has no effect on the brightness of a RAW file ?
 
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