ISO invariance - do you use it?

IPTAK

Forum Enthusiast
Messages
302
Reaction score
68
I recently was made aware of ISO invariance (yeash, a bit behind the times I suppose). If you take advantage of this, and are using it in your workflow, I would be curious to hear about how and when how and when you use (or choose not to use it).
 
I use ISO invariance at higher ISOs. After setting the max exposure, I set ISO one or two stop below metered in order to be on the safe side of highlight clipping.
 
For the vast majority of what I shoot, my exposures are based on Keeper Rates.

Which for me means maintaining sufficient shutter speed to stop motion.

I aim for high keeper rates, so my shutter speeds are generally on the high side, which means that I'm underexposing the vast majority of the time (often by quite a bit).

(Keep in mind that Exposure is determined by only your shutter speed and aperture)

ISO determines how bright you want your output to be (It's an amplification of your exposure). When shooting RAW, ISO merely serves as a marker for your software.

So in-camera, I set my ISO to protect my highlights (using the Blinkies as a guide).

Keeping in mind that software does a bad job at bringing down those clipped highlight levels while still maintaining natural tonality (if they've been over-brightened by the ISO setting).

This is kind of a quasi "Expose To The Right" (ETTR) process ("quasi" since my ISO is almost never at base). The effects on the end result are basically the same though.

I'm often at max aperture (or stopped down when I need more DOF).

That's my process. It does incorporate ISO invariance, but it's really only incidental. ;-)

R2

ps. Noise reduction becomes an integral part of the process for me, but in the end it mainly determines how much I can underexpose and still get acceptable results.

--
Good judgment comes from experience.
Experience comes from bad judgment.
http://www.pbase.com/jekyll_and_hyde/galleries
 
Last edited:
I recently was made aware of ISO invariance (yeash, a bit behind the times I suppose). If you take advantage of this, and are using it in your workflow, I would be curious to hear about how and when how and when you use (or choose not to use it).
I do when shooting RAW in manual mode.

If shooting JPG, you can't.

Better to brighten photos in post where you have more control over the tone curve etc.

Digital amplification in-camera is no different from digital amplification by brightening the photo in Lightroom, except that the former decreases your dynamic range and makes it harder to retain detail in the highlights.

 
Last edited:
I do not like that term. If there were true invariance - why not glue your ISO lever, figuratively speaking, to ISO 100 or ISO 64000?

About choosing ISO below what the camera suggests - I use it when I shoot concerts. Without it, the extreme contrast could blow the highlights I wanted to keep. I was doing this even when l was shooting with lower DR cameras but I was more careful then.
 
I recently was made aware of ISO invariance (yeash, a bit behind the times I suppose). If you take advantage of this, and are using it in your workflow, I would be curious to hear about how and when how and when you use (or choose not to use it).
I do when shooting RAW in manual mode.

If shooting JPG, you can't.

Better to brighten photos in post where you have more control over the tone curve etc.

Digital amplification in-camera is no different from digital amplification by brightening the photo in Lightroom, except that the former decreases your dynamic range and makes it harder to retain detail in the highlights.

https://photographylife.com/iso-invariance-explained
You'd better skip the technical explanations in that article as they seem to be a mixture of information taken from here https://dpreview.com/articles/0388507676/sources-of-noise-part-two-electronic-noise and author's own misunderstandings.

In practice you should be aware that your camera may have an ISO-invariant range (I think most of current top end Canon cameras do) and you can check it here


Also be aware that noises introduced by camera may be not fully random so it's best to try and test workflow with ISO invariance before using it in the field.
 
I do not like that term. If there were true invariance - why not glue your ISO lever, figuratively speaking, to ISO 100 or ISO 64000?
It's not ideal, but I like it more than 'ISOlessness'. It's not ISO that's invariant, it's the visible noise that's invariant against ISO setting in-camera after pushing the 'exposure' slider in Lightroom.

We change the ISO value, the noise stays the same - that's invariance.
About choosing ISO below what the camera suggests - I use it when I shoot concerts. Without it, the extreme contrast could blow the highlights I wanted to keep. I was doing this even when l was shooting with lower DR cameras but I was more careful then.
 
I do not like that term. If there were true invariance - why not glue your ISO lever, figuratively speaking, to ISO 100 or ISO 64000?
It's not ideal, but I like it more than 'ISOlessness'. It's not ISO that's invariant, it's the visible noise that's invariant against ISO setting in-camera after pushing the 'exposure' slider in Lightroom.

We change the ISO value, the noise stays the same - that's invariance.
I understand that. What I do not like about this term is that the highlight clipping is ISO dependent, and this remains hidden.
 
I recently was made aware of ISO invariance (yeash, a bit behind the times I suppose). If you take advantage of this, and are using it in your workflow, I would be curious to hear about how and when how and when you use (or choose not to use it).
I don’t believe in ISO “invariance”. ISO is a function that controls the lightness of jpgs. I mostly ignore it.
 
I do not like that term. If there were true invariance - why not glue your ISO lever, figuratively speaking, to ISO 100 or ISO 64000?
It's not ideal, but I like it more than 'ISOlessness'. It's not ISO that's invariant, it's the visible noise that's invariant against ISO setting in-camera after pushing the 'exposure' slider in Lightroom.

We change the ISO value, the noise stays the same - that's invariance.
I understand that. What I do not like about this term is that the highlight clipping is ISO dependent, and this remains hidden.
Exactly. Canon cameras typically have two levels of signal amplification that affects noise in shadows. For the R5 the second one is at ISO 400 if I'm remembering correctly. ISO settings higher than that just push the data to the right.

For settings like a jazz club where the lighting is poor, I typically use ISO 3200 (sometimes as low as 800) with manual shutter speed and aperture set to give the visual effects that I want. Usually that means the photos are underexposed. I then adjust in post. If I boosted the ISO to 6400 or higher, highlights would get blown. Lighting is usually be colored flood lights, and when highlights blow, the color changes.
 
I recently was made aware of ISO invariance (yeash, a bit behind the times I suppose). If you take advantage of this, and are using it in your workflow, I would be curious to hear about how and when how and when you use (or choose not to use it).
I don’t believe in ISO “invariance”. ISO is a function that controls the lightness of jpgs. I mostly ignore it.
If the camera is not ISO invariant then increasing ISO reduces noise as well.
 
I recently was made aware of ISO invariance (yeash, a bit behind the times I suppose). If you take advantage of this, and are using it in your workflow, I would be curious to hear about how and when how and when you use (or choose not to use it).
I don’t believe in ISO “invariance”. ISO is a function that controls the lightness of jpgs. I mostly ignore it.
If the camera is not ISO invariant then increasing ISO reduces noise as well.
Turning the ISO dial has no impact on shot noise, only changing the amount of light collected while the shutter is open impacts shot noise.
 
I recently was made aware of ISO invariance (yeash, a bit behind the times I suppose). If you take advantage of this, and are using it in your workflow, I would be curious to hear about how and when how and when you use (or choose not to use it).
I don’t believe in ISO “invariance”. ISO is a function that controls the lightness of jpgs.
The ISO standard does that, but the ISO setting also controls internal camera parameters. In particular, it affects the input-referred read noise. The effect is specific to the design of the camera.

It's not irrelevant to consider the ISO setting when trying to optimize image quality, although doing so may be overrated.
 
I recently was made aware of ISO invariance (yeash, a bit behind the times I suppose). If you take advantage of this, and are using it in your workflow, I would be curious to hear about how and when how and when you use (or choose not to use it).
I don’t believe in ISO “invariance”. ISO is a function that controls the lightness of jpgs. I mostly ignore it.
If the camera is not ISO invariant then increasing ISO reduces noise as well.
Turning the ISO dial has no impact on shot noise, only changing the amount of light collected while the shutter is open impacts shot noise.
What matters is the SNR, not shot noise alone. Shot and read noise matter. The larger the exposure, the larger the shot noise (but the higher the signal). The lower the exposure, the more read noise matters.

Increasing ISO can reduce in-camera noise. On the one hand, by switching gain (dual conversion gain). On the other hand, by decreasing the noise from conversion to digital. The Canon EOS 5D is an example where increasing ISO improves noise in shadows:

 
I recently was made aware of ISO invariance (yeash, a bit behind the times I suppose). If you take advantage of this, and are using it in your workflow, I would be curious to hear about how and when how and when you use (or choose not to use it).
I don’t believe in ISO “invariance”. ISO is a function that controls the lightness of jpgs.
The ISO standard does that, but the ISO setting also controls internal camera parameters. In particular, it affects the input-referred read noise. The effect is specific to the design of the camera.

It's not irrelevant to consider the ISO setting when trying to optimize image quality, although doing so may be overrated.
Dual conversion gain, controlled by ISO settings, significantly influences SNR.
 
I recently was made aware of ISO invariance (yeash, a bit behind the times I suppose). If you take advantage of this, and are using it in your workflow, I would be curious to hear about how and when how and when you use (or choose not to use it).
I don’t believe in ISO “invariance”. ISO is a function that controls the lightness of jpgs.
The ISO standard does that, but the ISO setting also controls internal camera parameters. In particular, it affects the input-referred read noise. The effect is specific to the design of the camera.

It's not irrelevant to consider the ISO setting when trying to optimize image quality, although doing so may be overrated.
Dual conversion gain, controlled by ISO settings, significantly influences SNR.
Yes it does, but I suspect that sometimes people may pay too much attention to it.
 
I recently was made aware of ISO invariance (yeash, a bit behind the times I suppose). If you take advantage of this, and are using it in your workflow, I would be curious to hear about how and when how and when you use (or choose not to use it).
I don’t believe in ISO “invariance”. ISO is a function that controls the lightness of jpgs.
The ISO standard does that, but the ISO setting also controls internal camera parameters. In particular, it affects the input-referred read noise. The effect is specific to the design of the camera.
You raise an excellent point. There is the ISO standard, which relates to the lightness of jpgs, and there is whatever other occult, undocumented, and idiosyncratic stuff the manufacturers also do with that camera control from time to time. Still no “invariance”, though. I get it, there is empirical evidence that suggests that there is …something… they do maybe once but not at every turn of that dial, and that they don’t document. Whatever it is that they do, they don’t think I need to know about it, and neither do I.
It's not irrelevant to consider the ISO setting when trying to optimize image quality, although doing so may be overrated.
Agree that it is, at least, overrated.
 
I recently was made aware of ISO invariance (yeash, a bit behind the times I suppose). If you take advantage of this, and are using it in your workflow, I would be curious to hear about how and when how and when you use (or choose not to use it).
I don’t believe in ISO “invariance”. ISO is a function that controls the lightness of jpgs. I mostly ignore it.
If the camera is not ISO invariant then increasing ISO reduces noise as well.
Turning the ISO dial has no impact on shot noise, only changing the amount of light collected while the shutter is open impacts shot noise.
What matters is the SNR, not shot noise alone. Shot and read noise matter. The larger the exposure, the larger the shot noise (but the higher the signal). The lower the exposure, the more read noise matters.

Increasing ISO can reduce in-camera noise. On the one hand, by switching gain (dual conversion gain). On the other hand, by decreasing the noise from conversion to digital. The Canon EOS 5D is an example where increasing ISO improves noise in shadows:

https://www.photonstophotos.net/Charts/PDR_Shadow.htm#Canon EOS 5D
Perhaps, but what you are talking about is not part of what the relevant ISO standard specifies. Anyone who wants less noise should seek more light and forego all speculation as to occult processes.
 
I recently was made aware of ISO invariance (yeash, a bit behind the times I suppose). If you take advantage of this, and are using it in your workflow, I would be curious to hear about how and when how and when you use (or choose not to use it).
I don’t believe in ISO “invariance”. ISO is a function that controls the lightness of jpgs.
The ISO standard does that, but the ISO setting also controls internal camera parameters. In particular, it affects the input-referred read noise. The effect is specific to the design of the camera.

It's not irrelevant to consider the ISO setting when trying to optimize image quality, although doing so may be overrated.
Dual conversion gain, controlled by ISO settings, significantly influences SNR.
Yes it does, but I suspect that sometimes people may pay too much attention to it.
Indeed, way too much. None of the occult stuff will do you better than just getting more light.
 
Whatever it is that they do, they don’t think I need to know about it, and neither do I
So in certain (pretty common) conditions you'll be getting highlights blown, and I won't be getting highlights blown under the same exposure settings but lower ISO setting.

I think it's an issue with the camera manufacturers that they don't provide a documented way to use ISO invariance. It's not good when users have to use hacks and workarounds. Same as with the lack of a raw histogram and UniWB.
It's not irrelevant to consider the ISO setting when trying to optimize image quality, although doing so may be overrated.
Agree that it is, at least, overrated.
It's overrated for those who don't need it, in the technical sense, in their typical shooting conditions.
 

Keyboard shortcuts

Back
Top