What's up with Fuji ISO values?

[Mads Bjerke]

I was out shooting two different systems today.

Setup 1: X-T3 with XF100-400mm
Setup 2: Nikon D500 with 200-500mm

Both cameras set to the same exposure values: iso2000 - 1/2000s - f5.6

The X-T3 is at least one full stop darker:

This means I can shoot the Nikon at 1/4000s or f8.0 or iso1000 and get the same exposure.
That is a big advantage for wildlife.

Anyone else found a similar comparison with other cameras?
Not really sure what to read into this.

I would be interested in your comments and experiences.

madsbjerke.com

 

[DarnGoodPhotos]

Fuji underexposes to protect highlights. Nothing to worry about.

 

[Garthom]

I'll let people with more technical knowledge than me explain it in detail, but my understanding is that Fuji uses a different ISO standard than some of the other manufacturers. A quick Google search for "Fuji ISO" results in a number of articles about this, as well as previous threads on this forum and others. Here's the first one I saw from DPR, though surely there have been others:

https://www.dpreview.com/forums/post/53391280

 

[atalwar]

Metering Variations.

I know that my D500 usually overexposes by 1/3rd to 2/3 and easily blows up speculars.

So Mine I have permanently biased its meter by -1/3 (fine tune exposure setting in menus)

Fuji, OTOH on X-T3 seem a bit too conservative, perhaps 1/3 to 2/3 Under.

If you add them up and you get about 1 - 11/3 stop of difference.

I havent tested, but spot metering same subject (a wall perhaps) with both could shed more light.

 

[Ysarex]

I'll let people with more technical knowledge than me explain it in detail, but my understanding is that Fuji uses a different ISO standard than some of the other manufacturers. A quick Google search for "Fuji ISO" results in a number of articles about this, as well as previous threads on this forum and others. Here's the first one I saw from DPR, though surely there have been others:

https://www.dpreview.com/forums/post/53391280

Yep, ISO standards for digital cameras were updated in 2006. There isn't one single standard. Fuji uses SOS (Standard Output Sensitivity) for their cameras and Nikon uses REI (Recommended Exposure Index).

 

[Morris0]

I agree with your observation. Don't worry about as you are now shooting mirror less and when you make exposure adjustments they will appear before your eyes in the EVF or on the LCD. You will soon wonder why you ever used and SLR.

Morris

 

[Mads Bjerke]

Fuji underexposes to protect highlights. Nothing to worry about.

Well, actually this isn't about the cameras metering bias.
This is about what the camera says it will do compared to what it actually does.

Both systems where set to manual exposure.
I always use manual exposure, so the meter doesn't enter into it.

What I am getting at is something like this;

• Two racing cars are doing a drag race.
• Both are supposed to be 1000bhp and evenly matched in power.
• Turns out one is 500bhp, but the manufacturers used a different measuring standard to make it look it is 1000bhp.
• The real 1000bhp car pulls away and wins.

If I can do double the shutter speed (or half the iso) in one system and still get equal exposure, that is a very real difference and something I would personally care about.

Granted this only matters where speed is important, landscape photographers don't need to worry.
For sports and wildlife this is VERY important.

We have to pay a lot of money to gain 1 stop of extra light when buying lenses.
The difference in price between a f5.6 and f4.0 lens can be thousands.

If the camera takes that stop away in effective exposure it makes the rest of the system very expensive.

madsbjerke.com

 

[Mads Bjerke]

I agree with your observation. Don't worry about as you are now shooting mirror less and when you make exposure adjustments they will appear before your eyes in the EVF or on the LCD. You will soon wonder why you ever used and SLR.

Morris

As I have both systems I can attest to the benefits of both OVF and EVF's.
I added the mirrorless systems about 18 months ago and they are great.

It is however a worry that the X-T3 underexposes and as a result I have to sacrifice either shutter speed or iso to regain the exposure (compared to my D500).

 

[Mads Bjerke]

I'll let people with more technical knowledge than me explain it in detail, but my understanding is that Fuji uses a different ISO standard than some of the other manufacturers. A quick Google search for "Fuji ISO" results in a number of articles about this, as well as previous threads on this forum and others. Here's the first one I saw from DPR, though surely there have been others:

https://www.dpreview.com/forums/post/53391280

Yep, ISO standards for digital cameras were updated in 2006. There isn't one single standard. Fuji uses SOS (Standard Output Sensitivity) for their cameras and Nikon uses REI (Recommended Exposure Index).

Sounds like Fujifilm should play "fair" and use the REI standard since that is what appears to be a common standard.

My iso160 - 12800 X-T3 behaves like an iso320 - 6400 Nikon D500 in exposure terms.

 

[darinb]

Have you tried comparing the images (shot at the same manual settings) after making whatever post processing adjustments required? Compared them with the Fuji ISO boosted one stop?

I don't know which one will look better, but that is the question.

--Darin

 

[Henrik Helmers]

I think this has been discussed several times, and DPReview also points it out from time to time in camera reviews. Last I read it was measured to be a difference of about 1/3 EV.

I knew this buying my X-H1, and I think it is fair to say that with a base ISO that is 100 or 60 higher and slightly darker exposures you do get a bit more noise. My 5DmkIII produced clean images at base ISO, but in return the dynamic range was terrible, and the color noise was especially bad.

I used Magic Lantern on my 5D to try boosting dynamic range via dual ISO , but that was not exactly a silver bullet.

 

[Mads Bjerke]

Have you tried comparing the images (shot at the same manual settings) after making whatever post processing adjustments required? Compared them with the Fuji ISO boosted one stop?

I don't know which one will look better, but that is the question.

I did a quick test when out on a shoot.
It isn't done under lab conditions, but it's real enough for me.

Here is an example - note that there was a difference in camera settings:

• D500 was set to ISO2000 - 1/2500s - f6.3
• X-T3 was set to ISO2000 - 1/2000s - f5.6

I removed all NR and sharpening in Capture One Pro - including output sharpening.
This isn't a test for sharpness, but noise is relevant.

D500



X-T3



The X-T3 still needs another 1/4 stop of exposure to match the D500 despite already having about 2/3 stop advantage in the shutter speed and aperture.

This is applicable for the centre of the frame.
The XF100-400 vignettes heavily at f5.6 @ 400mm so away from the centre it probably needs another 1/2 stop exposure to match the D500 image (shot with Nikon 200-500/f5.6).

In terms of noise I don't see much of a difference between the two.

Here is the X-T3 image with +0.2 exposure and full anti-vignette (Light Fall-off) applied in C1P:



I will let you judge for yourself if there is a noticeable difference in image quality.
There is certainly a difference in shutter speed and aperture.

madsbjerke.com

 

[Iuvenis]

You are treating ISO as part of exposure, so feel 'cheated' that, when set for the same shutter speed and aperture, the Fuji needs a higher ISO value (or to be boosted in post processing.

However, the exposure is the amount of light falling on the sensor, divided by the sensor area. It is therefore determined by the f number and the shutter speed, nothing else.

The ISO setting is not a 'noise' setting. Noise seems to increase as you increase the iso setting, because you increase the iso setting in low light, and low light causes noise.

ISO is just a way of making sure the brightness of the image matches what you want. If you wanted to, you could boost the image in post processing instead, and get very similar results.

With modern sensors, the amount of light is the main factor impacting noise. As a result, f5.6 at 1/500th on the Fuji will have roughly the same noise as f5.6 at 1/500th on the Nikon. The results may not be identical - some cameras do perform better in low light than others - but that has nothing to do with the iso setting.

In England, when they moved to the Gregorian calendar, they moved the calendar forward 11 days. There were riots by people who thought their lives had been cut short, chanting 'give us back our 11 days!' Worrying about the iso number is the same thing; if Fuji added a nought to every iso number on the dial, it wouldn't make a blind bit of difference to its performance, positively or negatively.

As for the different ISO standards, it isn't a question of iso 200 on the Fuji = iso 100 on the Nikon. Your results will differ depending on the scene and its highlights and shadow areas. If you want to play around, imaging resource in their camera reviews include a number of images with controlled lighting, so you can see how differently the cameras expose them. You will find a whole range of values, Nikon being on one extreme and Fuji being on the other. Which is better is more or less irrelevant, as it doesn't influence the output.

 

[Iuvenis]

Fuji underexposes to protect highlights. Nothing to worry about.

Well, actually this isn't about the cameras metering bias.
This is about what the camera says it will do compared to what it actually does.

Both systems where set to manual exposure.
I always use manual exposure, so the meter doesn't enter into it.

What I am getting at is something like this;

• Two racing cars are doing a drag race.
• Both are supposed to be 1000bhp and evenly matched in power.
• Turns out one is 500bhp, but the manufacturers used a different measuring standard to make it look it is 1000bhp.
• The real 1000bhp car pulls away and wins.

No, no, no. This is like a race between two cars that are practically equal in performance. One manufacturer says their car can reach 240 on the straight, the other says they can only reach 150. But the first manufacturer is measuring speed in kph, the other in mph.

If I can do double the shutter speed

Absolutely

(or half the iso)

Why would that matter? Only if iso setting was a proxy for 'amount of noise in the photo', which it isn't.

in one system and still get equal exposure, that is a very real difference and something I would personally care about.

Granted this only matters where speed is important, landscape photographers don't need to worry.
For sports and wildlife this is VERY important.

We have to pay a lot of money to gain 1 stop of extra light when buying lenses.
The difference in price between a f5.6 and f4.0 lens can be thousands.

If the camera takes that stop away in effective exposure

It isn't, because iso is not part of exposure. Exposure is about how much light you have, as you noted above. Iso setting does not influence how much light you have.

it makes the rest of the system very expensive.

madsbjerke.com

 

[Iuvenis]

Have you tried comparing the images (shot at the same manual settings) after making whatever post processing adjustments required? Compared them with the Fuji ISO boosted one stop?

I don't know which one will look better, but that is the question.

I did a quick test when out on a shoot.
It isn't done under lab conditions, but it's real enough for me.

Here is an example - note that there was a difference in camera settings:

• D500 was set to ISO2000 - 1/2500s - f6.3
• X-T3 was set to ISO2000 - 1/2000s - f5.6

I removed all NR and sharpening in Capture One Pro - including output sharpening.
This isn't a test for sharpness, but noise is relevant.

Look again at the heron's white feathers. You can see that the Fuji records those feathers as white, the Nikon as pale grey. This is the result of the extra exposure.

What you seem to be looking at is the dark twigs. However, that is the point about iso calibration. Do you base it on 18% grey, or the highlights, or the shadows? In this case, the herons are meant to be the subject, so the Fuji result is better, but in other scenes it could be the opposite way round.

D500

X-T3

The X-T3 still needs another 1/4 stop of exposure to match the D500 despite already having about 2/3 stop advantage in the shutter speed and aperture.

This is applicable for the centre of the frame.
The XF100-400 vignettes heavily at f5.6 @ 400mm so away from the centre it probably needs another 1/2 stop exposure to match the D500 image (shot with Nikon 200-500/f5.6).

In terms of noise I don't see much of a difference between the two.

It's hard to tell from an image like this as it is not sharp. Sharp images show noise more prominently. If you are interested, DPreview have a studio scene you can use for comparisons. Use RAW mode as that takes iso calibration out the equation. Use comp mode, as the pixel pitch is also different. Honestly though, there will be very little difference between modern cameras of the same sensor size.

Here is the X-T3 image with +0.2 exposure and full anti-vignette (Light Fall-off) applied in C1P:

I will let you judge for yourself if there is a noticeable difference in image quality.
There is certainly a difference in shutter speed and aperture.

madsbjerke.com

 

[Mads Bjerke]

You are treating ISO as part of exposure, so feel 'cheated' that, when set for the same shutter speed and aperture, the Fuji needs a higher ISO value (or to be boosted in post processing.

However, the exposure is the amount of light falling on the sensor, divided by the sensor area. It is therefore determined by the f number and the shutter speed, nothing else.

The ISO setting is not a 'noise' setting. Noise seems to increase as you increase the iso setting, because you increase the iso setting in low light, and low light causes noise.

ISO is just a way of making sure the brightness of the image matches what you want. If you wanted to, you could boost the image in post processing instead, and get very similar results.

With modern sensors, the amount of light is the main factor impacting noise. As a result, f5.6 at 1/500th on the Fuji will have roughly the same noise as f5.6 at 1/500th on the Nikon. The results may not be identical - some cameras do perform better in low light than others - but that has nothing to do with the iso setting.

In England, when they moved to the Gregorian calendar, they moved the calendar forward 11 days. There were riots by people who thought their lives had been cut short, chanting 'give us back our 11 days!' Worrying about the iso number is the same thing; if Fuji added a nought to every iso number on the dial, it wouldn't make a blind bit of difference to its performance, positively or negatively.

As for the different ISO standards, it isn't a question of iso 200 on the Fuji = iso 100 on the Nikon. Your results will differ depending on the scene and its highlights and shadow areas. If you want to play around, imaging resource in their camera reviews include a number of images with controlled lighting, so you can see how differently the cameras expose them. You will find a whole range of values, Nikon being on one extreme and Fuji being on the other. Which is better is more or less irrelevant, as it doesn't influence the output.

I agree in your statements above.

The question is where it leaves you when light is marginal and you need to keep the shutter speed up.

I am ok with ignoring the iso value in itself, but when you are pushing the sensor towards its limits at the high iso end you will inevitably end with a poorer quality image.

Fast action requires a high shutter speed.
There is no substitute for that.

 

[darinb]

I'm sorry, I wasn't clear. Shoot the scene with the Fuji with whatever your requirements are for shutter speed an aperture. Don't go smaller or slower than you need to--the goal si to judge whether the Fuji can meet your requirements (and then compare it to the result from the Nikon). So let the ISO vary to compensate for whatever exposure differences there are. Then process and compare the result.

If the Fuji can't get to the aperture and shutter speed you require--if you don't have sufficient depth of field or sufficient action-stopping--then nothing else really matters.

However, I suspect that at the same shutter speed and aperture that after post-processing the differences in the two cameras will be minor. But maybe not! I haven't compared them myself.

--Darin

 

[Iuvenis]

You are treating ISO as part of exposure, so feel 'cheated' that, when set for the same shutter speed and aperture, the Fuji needs a higher ISO value (or to be boosted in post processing.

However, the exposure is the amount of light falling on the sensor, divided by the sensor area. It is therefore determined by the f number and the shutter speed, nothing else.

The ISO setting is not a 'noise' setting. Noise seems to increase as you increase the iso setting, because you increase the iso setting in low light, and low light causes noise.

ISO is just a way of making sure the brightness of the image matches what you want. If you wanted to, you could boost the image in post processing instead, and get very similar results.

With modern sensors, the amount of light is the main factor impacting noise. As a result, f5.6 at 1/500th on the Fuji will have roughly the same noise as f5.6 at 1/500th on the Nikon. The results may not be identical - some cameras do perform better in low light than others - but that has nothing to do with the iso setting.

In England, when they moved to the Gregorian calendar, they moved the calendar forward 11 days. There were riots by people who thought their lives had been cut short, chanting 'give us back our 11 days!' Worrying about the iso number is the same thing; if Fuji added a nought to every iso number on the dial, it wouldn't make a blind bit of difference to its performance, positively or negatively.

As for the different ISO standards, it isn't a question of iso 200 on the Fuji = iso 100 on the Nikon. Your results will differ depending on the scene and its highlights and shadow areas. If you want to play around, imaging resource in their camera reviews include a number of images with controlled lighting, so you can see how differently the cameras expose them. You will find a whole range of values, Nikon being on one extreme and Fuji being on the other. Which is better is more or less irrelevant, as it doesn't influence the output.

I agree in your statements above.

The question is where it leaves you when light is marginal and you need to keep the shutter speed up.

I am ok with ignoring the iso value in itself, but when you are pushing the sensor towards its limits at the high iso end you will inevitably end with a poorer quality image.

Fast action requires a high shutter speed.
There is no substitute for that.

You set the aperture and shutter speed your shot requires, or your camera allows.

That leaves you with a higher iso value on the Fuji, or with having to push the Fuji file in post processing. The question is whether that is a disadvantage.

The answer to that is no. The iso is abitrary, and doesn't amount to a measure of the amount of noise in a photograph. Take this example, from the DP comparison tool, first with JPEGs:


JPEG test

Look at the words on the colour wheel - do you see how blurry the Nikon image is? That is the JPEG noise reduction kicking in at iso 6400. However, the Fuji is at iso 12800, and suffers less. It is more on par with the full frame cameras at the bottom at iso 12800.

However, that is misleading, as the Fuji image had more exposure. Have a look at the situation in RAW:


RAW test

With an equal amount of exposure, and with the image adjusted to the same brightness in post processing, the Fuji is now more or less equal with the Nikon (it has less colour noise due to the X-trans array, but doesn't have less overall). The full frame cameras now pull ahead, as you would expect.

So if you are out with the Fuji, just let the iso climb higher than you normally would. You will not suffer from more noise overall.

 

[Mads Bjerke]

You are treating ISO as part of exposure, so feel 'cheated' that, when set for the same shutter speed and aperture, the Fuji needs a higher ISO value (or to be boosted in post processing.

However, the exposure is the amount of light falling on the sensor, divided by the sensor area. It is therefore determined by the f number and the shutter speed, nothing else.

The ISO setting is not a 'noise' setting. Noise seems to increase as you increase the iso setting, because you increase the iso setting in low light, and low light causes noise.

ISO is just a way of making sure the brightness of the image matches what you want. If you wanted to, you could boost the image in post processing instead, and get very similar results.

With modern sensors, the amount of light is the main factor impacting noise. As a result, f5.6 at 1/500th on the Fuji will have roughly the same noise as f5.6 at 1/500th on the Nikon. The results may not be identical - some cameras do perform better in low light than others - but that has nothing to do with the iso setting.

In England, when they moved to the Gregorian calendar, they moved the calendar forward 11 days. There were riots by people who thought their lives had been cut short, chanting 'give us back our 11 days!' Worrying about the iso number is the same thing; if Fuji added a nought to every iso number on the dial, it wouldn't make a blind bit of difference to its performance, positively or negatively.

As for the different ISO standards, it isn't a question of iso 200 on the Fuji = iso 100 on the Nikon. Your results will differ depending on the scene and its highlights and shadow areas. If you want to play around, imaging resource in their camera reviews include a number of images with controlled lighting, so you can see how differently the cameras expose them. You will find a whole range of values, Nikon being on one extreme and Fuji being on the other. Which is better is more or less irrelevant, as it doesn't influence the output.

I agree in your statements above.

The question is where it leaves you when light is marginal and you need to keep the shutter speed up.

I am ok with ignoring the iso value in itself, but when you are pushing the sensor towards its limits at the high iso end you will inevitably end with a poorer quality image.

Fast action requires a high shutter speed.
There is no substitute for that.

You set the aperture and shutter speed your shot requires, or your camera allows.

That leaves you with a higher iso value on the Fuji, or with having to push the Fuji file in post processing. The question is whether that is a disadvantage.
The answer to that is no. The iso is abitrary, and doesn't amount to a measure of the amount of noise in a photograph. Take this example, from the DP comparison tool, first with JPEGs:


JPEG test

Look at the words on the colour wheel - do you see how blurry the Nikon image is? That is the JPEG noise reduction kicking in at iso 6400. However, the Fuji is at iso 12800, and suffers less. It is more on par with the full frame cameras at the bottom at iso 12800.

However, that is misleading, as the Fuji image had more exposure. Have a look at the situation in RAW:


RAW test

With an equal amount of exposure, and with the image adjusted to the same brightness in post processing, the Fuji is now more or less equal with the Nikon (it has less colour noise due to the X-trans array, but doesn't have less overall). The full frame cameras now pull ahead, as you would expect.

So if you are out with the Fuji, just let the iso climb higher than you normally would. You will not suffer from more noise overall.

Thank you all for your contributions.

With regards to the DPreview studio lit test shots above, are they taken at the same shutter speed and aperture for all cameras?
Or do they match the exposure visually?

Just curious.

 

[Rmcp20]

Thank you all for your contributions.

With regards to the DPreview studio lit test shots above, are they taken at the same shutter speed and aperture for all cameras?
Or do they match the exposure visually?

Just curious.

It doesn't matter the shutter speed or aperture. It shows that the noise on the Fuji at 12800 is similar to the noise on the Nikon at 6400.

So if you set both to 1/1000 at f5.6, the Nikon will give you ISO 200 and Fuji 320 for example. In the end the image quality is similar (plus Fuji tends to have less color noise).

What he is showing you is that you judged your manual exposure incorectly. You reduced shutter speed and/or aperture to match your Nikon, when you should have raised the ISO by 2/3 stop (which by the same standards is the same as you would have on the Nikon).