Have you found it useful to know your camera's ISO invariance point?

The Grumpy Snapper said:

What an amusing thread this has become. A bunch of photo forum "experts" each trying to out expertise each other.

Click to expand...

Grumpy, you must be new to DPR. This is normal for all ISO threads on all boards. In fact, this one is very mild in comparison to most.

ISO is the most misunderstood thing in all of digital photography. The more you know about it, the more likely you are to make an incorrect statement about it when you start getting really technical.

I've been reading about it for years and am very interested in it. But I'm not a scientist and I stay away from getting into highly technical explanations.

I'm a photographer and I know how to use ISO in combination with the aperture and speed I want for that shot, and I know the cost of going above base and when it hurts you the most or when it won't hurt as much.

I'm a GFX photographer, so I therefore am obsessed with image fidelity and I know when DOF, diffraction and ISO starts affecting my expensive image fidelity.

The subject that causes the most argument on the Medium Format Board is DOF. GFX DOF is limited compared to APSC and FF. That means on a lot of shots you get out of focus pixels that you don't think you are going to get. To get more DOF, you shoot smaller apertures. That causes diffraction (but I shoot a lot of F11 w GFX). So we bump up ISO to get a smaller aperture at hand held speeds or in order to get a little more speed for shots where stuff is moving a bit. That has a cost.

Just know this. Moving above base ISO has a cost. Know when to spend that money.

It's all part of the EV decision. A lot of old school guys say that ISO is not a part of the EV decision.

They are wrong.

I think about it on every shot and it affects my aperture / speed decision on every shot.

No, it hasn't been useful. All my cameras are Sonys and ISO invariant but to be honest it makes no difference to me. Maybe if someone is very sensitive to noise they might care but I never think about it.

--
Tom

Hello...

Chris R-UK said:

I am giving a talk to my camera club next month and I will be covering, amongst other things, ISO invariance and how to find out whether your camera is ISO invariant. I don't have a particularly ISO invariant camera myself (Olympus E-M1.2), but I would like to include some examples of the situations where it might be useful to know the point at which your camera effectively become ISO invariant.

So, can anybody with an ISO invariant camera give me examples of situations that I could include as examples in my talk?

Click to expand...

IDK but with your Olympus E-M 1.2 and iPhone/whatever - you should be fine. IMHO

Greg7579 said:

Just know this. Moving above base ISO has a cost.

Click to expand...

A negative one. The whole point of high ISO is to improve the noise.

Greg7579 said:

Bill, I have read your posts for years and you are usually right on these crazy ISO threads.

But if you say climbing the ISO ladder is free of charge, then you and I live on different planets.

I need to carefully reread every word again because I must be missing something and we are talking past each other. Either that, or you have discovered an alternative universe.

Click to expand...

This.

tony field said:

Greg7579 said:

I never allow 2 stops of highlight headroom. What am I? A rookie?

I know my camera and its meter. I know what the jpeg histo in the EVF does to my raw perception because I use raw digger to learn in post.

But you know what you are doing and your lessons are good. No worries.

I just don't want new raw shooters casually jacking up the ISO when they don't need to. I want them to think about their EV.

Click to expand...

If that is your concern then why not take the easy and obvious way out and tell them to use their desired Auto expose mode of choice or manual exposure and use auto ISO with a maximum ISO set something realistic for the camera, possibly at the point of the cameras ISO invariance.

Click to expand...

I don't like auto ISO. But that is just me. I have had this argument with some of the best GFX photographers in the world who use auto ISO. Not me. No way. When I go above base, I want to control it.

But I am not shooting fashion in a fluid situation and I don't shoot sports or BIF. My shooting is more static (and certainly not as good as those guys). I think long and hard about DOF on every shot.

Yesterday, I jumped out of the car with the converted GFX 50r and did not use a monopod on that shot like I always do. I shot at ISO 400 because I needed another two stops of hand-held speed, and I needed F11.

So know this. I can name 5 pro GFX photographers who use auto ISO in the way you describe. I can name 20 who don't.

But your point is well taken.

I was speaking at a more basic level. This is not the Medium Format Board. I would not be lecturing guys there to think about their ISO decisions and to try to stay at base if they can.

tbcass said:

No, it hasn't been useful. All my cameras are Sonys and ISO invariant but to be honest it makes no difference to me. Maybe if someone is very sensitive to noise they might care but I never think about it.

Click to expand...

You don't need to. Just shoot as low as you can get in the situation you are in.

FingerPainter said:

When I am shooting RAW and using Auto-ISO, I set the maximum ISO to the level at which the camera goes invariant. I'll get no read noise benefit from using a higher ISO setting, and there is always the risk that I'll blow a highlight with a higher ISO setting.

Click to expand...

It could be somewhere in-between, though, but manufacturers don't give us the choice.

Typically we get something like this:

ISO 100 reference analog gain, reference headroom

ISO 200 2.0 analog gain, reference headroom

ISO 400 4.0 analog gain, reference headroom

ISO 800 8.0 analog gain, reference headroom

or:

ISO 100 reference analog gain, reference headroom

ISO 200 reference analog gain, +1.0 stop headroom

ISO 400 reference analog gain, +2.0 stops headroom

ISO 800 reference analog gain, +3.0 stops headroom

but why not:

ISO 100 reference analog gain, reference headroom

ISO 200 1.4 analog gain, +0.5 stops headroom

ISO 400 2.0 analog gain, +1.0 stop headroom

ISO 800 2.8 analog gain, +1.5 stops headroom

This would result in increasing headroom at higher ISOs, but minimize the post-gain read noise at the higher ISOs, too, compared to just suddenly all-arithmetic at a fairly low ISO. In fact, it might be even better just to do this partially-proportional gain up to a point where headroom is sufficient, and start increasing gain proportional to ISO above that point:

ISO 1600 4.0 analog gain, +2.0 stops headroom

ISO 3200 8.0 analog gain, +2.0 stops headroom

ISO 6400 16.0 analog gain, +2.0 stops headroom

ISO 12800 32.0 analog gain, +2.0 stops headroom

Greg7579 said:

tbcass said:

No, it hasn't been useful. All my cameras are Sonys and ISO invariant but to be honest it makes no difference to me. Maybe if someone is very sensitive to noise they might care but I never think about it.

Click to expand...

You don't need to. Just shoot as low high as you can get in the situation you are in.

Click to expand...

Fixed it for you.

J A C S said:

J A C S said:

J A C S said:

J A C S said:

It pushes both ISO 100 and ISO 400 (+6 and +4ev respectively). That's not exactly the same as pushing ISO 100 + 2ev in post and comparing it to ISO 400.

Click to expand...

Why should I (and they) push 2 stops only? What you are doing is circular reasoning.

Click to expand...

Because ISO 100 pushed + 6 stops and ISO 400 pushed +4ev are well beyond the claimed ISO-invariant range of 100-500. They're within the second claimed range of 640+.

Click to expand...

I do not know how to say it differently. What they claimed was based on exactly that comparison. True or not? You are trying to redefine their claim.

Click to expand...

They should have taken shots at ISO 500 and 640 in order to make claim on ISO-invariant ranges 100-500 and 640-51200.

And there's a more detailed explanation right under 'ISO invariance' chart. https://www.dpreview.com/reviews/sony-a7-iii-review/6

Apparently they used Bill Claff's data, and comparison tool in the review just doesn't cover the ISO values in question (500, 640).

J A C S said:

J A C S said:

It's very simple. ISO-invariance is rarely perfect so ISO 100 pushed +2 stops may look very close to ISO 400,

Click to expand...

True for just about every camera made in the last 15 years. Are they all ISO invariant in that range?

Click to expand...

I don't know. It's not really relevant for this talk as we're looking into the a7III review above.

J A C S said:

J A C S said:

but if you push both by 4 more stops, any difference in noise will be literally magnified by 16.

Click to expand...

So? Then they should not make that claim. "By 16" is 4 stops and we all know that this is what you are supposed to do with your shots, right?

Click to expand...

From that review:

"Our friend Bill Claff has actually measured any noise benefit to - for the same exposure - using a higher ISO in-camera vs. brightening a lower ISO Raw in post. He's found that there's essentially no benefit to using ISO 500 vs. boosting an ISO 100 Raw 2.3 EV, and no benefit to using ISO 51,200 vs. boosting an ISO 640 Raw 6.3EV in post. There is, however, nearly a stop improvement in shadow noise going from ISO 500 to ISO 640, because of the gain switch at ISO 640."

J A C S said:

J A C S said:

Maybe A7II is not ISO-invariant in the 100-500 range. It's just that pushing ISO 100 + 6ev doesn't prove that.

Click to expand...

Why are they doing it then and what meaning they assign to ISO-invariant?

Click to expand...

That's in the quote above, and I think it matches exactly what I was saying in the previous messages.

J A C S said:

J A C S said:

You need an example that follows the concept of ISO-invariance: shoot at ISO 100, push by 2ev in post and compare to non-pushed ISO 400.

Click to expand...

Great, my 5D2 was ISO invariant then.

What you are presuming is that you must stay with standard processing (after 2EV). This goes against the whole notion of DR. Standard processing does not reveal DR differences.

Click to expand...

What do you mean by standard processing?

J A C S said:

J A C S said:

Pushing both images to some crazy limits doesn't help answer the practical question: can I shoot at ISO 100 and then push +2ev (within the claimed invariant range) and get the same result as with ISO 400?

Click to expand...

They are not so crazy. I would push by a few stops some high ISO images occasionally in some part of the tonal range. Next, that craziness is what we call DR in the first place.

Click to expand...

I use ISO invariance when shooting action in changing light. Exactly because lower ISOs have greater DR. It's a protection against accidentally blown out highlights.

Quarkcharmed said:

Click to expand...

I give up on that part, I said enough.

Quarkcharmed said:

From that review:

"Our friend Bill Claff has actually measured any noise benefit to - for the same exposure - using a higher ISO in-camera vs. brightening a lower ISO Raw in post. He's found that there's essentially no benefit to using ISO 500 vs. boosting an ISO 100 Raw 2.3 EV, and no benefit to using ISO 51,200 vs. boosting an ISO 640 Raw 6.3EV in post. There is, however, nearly a stop improvement in shadow noise going from ISO 500 to ISO 640, because of the gain switch at ISO 640."

Click to expand...

This says something about Bill's so-called PDR, doesn't it? His chart shows no benefits at ISO 400 vs. ISO 100 either - and their own comparison disproves it.

[No message]

J A C S said:

I give up on that part, I said enough.

J A C S said:

From that review:

"Our friend Bill Claff has actually measured any noise benefit to - for the same exposure - using a higher ISO in-camera vs. brightening a lower ISO Raw in post. He's found that there's essentially no benefit to using ISO 500 vs. boosting an ISO 100 Raw 2.3 EV, and no benefit to using ISO 51,200 vs. boosting an ISO 640 Raw 6.3EV in post. There is, however, nearly a stop improvement in shadow noise going from ISO 500 to ISO 640, because of the gain switch at ISO 640."

Click to expand...

This says something about Bill's so-called PDR, doesn't it?

Click to expand...

It says something about A7III's performance at different ISO settings.

J A C S said:

His chart shows no benefits at ISO 400 vs. ISO 100 either - and their own comparison disproves it.

Click to expand...

Why does it have to show benefits? This chart https://photonstophotos.net/Charts/PDR_Shadow.htm#Sony ILCE-7M3 shows (almost) flat line in the ISO ranges 100-500 and 640-51200.

The only use case from having an ISO-invariant range in your camera is to shoot at a lower ISO and push in postprocessing. It doesn't affect the final image quality/noise level and it preserves highlights in case of inaccurate metering or suddenly changed light.

Outside of ISO-invariant range, it's best to push ISO setting in camera for lower noise, instead of pushing it in postprocessing.

Quarkcharmed said:

J A C S said:

I give up on that part, I said enough.

Quarkcharmed said:

From that review:

"Our friend Bill Claff has actually measured any noise benefit to - for the same exposure - using a higher ISO in-camera vs. brightening a lower ISO Raw in post. He's found that there's essentially no benefit to using ISO 500 vs. boosting an ISO 100 Raw 2.3 EV, and no benefit to using ISO 51,200 vs. boosting an ISO 640 Raw 6.3EV in post. There is, however, nearly a stop improvement in shadow noise going from ISO 500 to ISO 640, because of the gain switch at ISO 640."

Click to expand...

This says something about Bill's so-called PDR, doesn't it?

Click to expand...

It says something about A7III's performance at different ISO settings.

Click to expand...

That something is not what dpreview wanted it to be.

Quarkcharmed said:

Quarkcharmed said:

His chart shows no benefits at ISO 400 vs. ISO 100 either - and their own comparison disproves it.

Click to expand...

Why does it have to show benefits? This chart https://photonstophotos.net/Charts/PDR_Shadow.htm#Sony ILCE-7M3 shows (almost) flat line in the ISO ranges 100-500 and 640-51200.

Click to expand...

This is what I meant. Raising ISO usually shows benefits, that is the whole point. When it does not, it is "ISO-invariant".

J A C S said:

Quarkcharmed said:

J A C S said:

I give up on that part, I said enough.

J A C S said:

From that review:

"Our friend Bill Claff has actually measured any noise benefit to - for the same exposure - using a higher ISO in-camera vs. brightening a lower ISO Raw in post. He's found that there's essentially no benefit to using ISO 500 vs. boosting an ISO 100 Raw 2.3 EV, and no benefit to using ISO 51,200 vs. boosting an ISO 640 Raw 6.3EV in post. There is, however, nearly a stop improvement in shadow noise going from ISO 500 to ISO 640, because of the gain switch at ISO 640."

Click to expand...

This says something about Bill's so-called PDR, doesn't it?

Click to expand...

It says something about A7III's performance at different ISO settings.

Click to expand...

That something is not what dpreview wanted it to be.

J A C S said:

J A C S said:

His chart shows no benefits at ISO 400 vs. ISO 100 either - and their own comparison disproves it.

Click to expand...

Why does it have to show benefits? This chart https://photonstophotos.net/Charts/PDR_Shadow.htm#Sony ILCE-7M3 shows (almost) flat line in the ISO ranges 100-500 and 640-51200.

Click to expand...

This is what I meant. Raising ISO usually shows benefits, that is the whole point. When it does not, it is "ISO-invariant".

Click to expand...

Raising ISO is beneficial for IQ only in that it allows you to set higher speeds and smaller apertures to get the dshot, which you could do anyway by underexposing and pushing it in post. But few work that way or want to see that mess in the EVF. But climbing the ISO ladder costs Image fidelity and DR in post. Nothing is free. This ISO invariance is the most misstated and misunderstood subject in all of photography. Everyone but the shooting scientists talk past each other on the subject. I can name you three guys on the MF Board who have a deep and thorough understanding of this and two of them are scientists. The pure shooters know it intuitively from experience.

When you peep every shot in a variety of situations at full res with GFX files on a 32 inch pro IPS calibrated high color gamut monitor you see the truth and you learn fast.

Greg7579 said:

J A C S said:

Quarkcharmed said:

J A C S said:

I give up on that part, I said enough.

Greg7579 said:

From that review:

"Our friend Bill Claff has actually measured any noise benefit to - for the same exposure - using a higher ISO in-camera vs. brightening a lower ISO Raw in post. He's found that there's essentially no benefit to using ISO 500 vs. boosting an ISO 100 Raw 2.3 EV, and no benefit to using ISO 51,200 vs. boosting an ISO 640 Raw 6.3EV in post. There is, however, nearly a stop improvement in shadow noise going from ISO 500 to ISO 640, because of the gain switch at ISO 640."

Click to expand...

This says something about Bill's so-called PDR, doesn't it?

Click to expand...

It says something about A7III's performance at different ISO settings.

Click to expand...

That something is not what dpreview wanted it to be.

Greg7579 said:

Greg7579 said:

His chart shows no benefits at ISO 400 vs. ISO 100 either - and their own comparison disproves it.

Click to expand...

Why does it have to show benefits? This chart https://photonstophotos.net/Charts/PDR_Shadow.htm#Sony ILCE-7M3 shows (almost) flat line in the ISO ranges 100-500 and 640-51200.

Click to expand...

This is what I meant. Raising ISO usually shows benefits, that is the whole point. When it does not, it is "ISO-invariant".

Click to expand...

Raising ISO is beneficial for IQ only in that it allows you to set higher speeds and smaller apertures to get the dshot, which you could do anyway by underexposing and pushing it in post. But few work that way or want to see that mess in the EVF. But climbing the ISO ladder costs Image fidelity and DR in post. Nothing is free. This ISO invariance is the most misstated and misunderstood subject in all of photography. Everyone but the shooting scientists talk past each other on the subject. I can name you three guys on the MF Board who have a deep and thorough understanding of this and two of them are scientists. The pure shooters know it intuitively from experience.

When you peep every shot in a variety of situations at full res with GFX files on a 32 inch pro IPS calibrated high color gamut monitor you see the truth and you learn fast.

Click to expand...

I said nothing about underexposing and pushing in post in that sentence. What I meant that for a given exposure (choice of f-stop and speed), the highest ISO which does not clip the highlights you want to keep is the best choice (with little benefit or not).

You see ISO as primary, and SS and f-stop as afterthoughts. I see the latter two as primary and ISO as secondary.

Greg7579 said:

tbcass said:

No, it hasn't been useful. All my cameras are Sonys and ISO invariant but to be honest it makes no difference to me. Maybe if someone is very sensitive to noise they might care but I never think about it.

Click to expand...

You don't need to. Just shoot as low as you can get in the situation you are in.

Click to expand...

Which is exactly what I do.

J A C S said:

Greg7579 said:

J A C S said:

Quarkcharmed said:

J A C S said:

I give up on that part, I said enough.

J A C S said:

From that review:

"Our friend Bill Claff has actually measured any noise benefit to - for the same exposure - using a higher ISO in-camera vs. brightening a lower ISO Raw in post. He's found that there's essentially no benefit to using ISO 500 vs. boosting an ISO 100 Raw 2.3 EV, and no benefit to using ISO 51,200 vs. boosting an ISO 640 Raw 6.3EV in post. There is, however, nearly a stop improvement in shadow noise going from ISO 500 to ISO 640, because of the gain switch at ISO 640."

Click to expand...

This says something about Bill's so-called PDR, doesn't it?

Click to expand...

It says something about A7III's performance at different ISO settings.

Click to expand...

That something is not what dpreview wanted it to be.

J A C S said:

J A C S said:

His chart shows no benefits at ISO 400 vs. ISO 100 either - and their own comparison disproves it.

Click to expand...

Why does it have to show benefits? This chart https://photonstophotos.net/Charts/PDR_Shadow.htm#Sony ILCE-7M3 shows (almost) flat line in the ISO ranges 100-500 and 640-51200.

Click to expand...

This is what I meant. Raising ISO usually shows benefits, that is the whole point. When it does not, it is "ISO-invariant".

Click to expand...

Raising ISO is beneficial for IQ only in that it allows you to set higher speeds and smaller apertures to get the dshot, which you could do anyway by underexposing and pushing it in post. But few work that way or want to see that mess in the EVF. But climbing the ISO ladder costs Image fidelity and DR in post. Nothing is free. This ISO invariance is the most misstated and misunderstood subject in all of photography. Everyone but the shooting scientists talk past each other on the subject. I can name you three guys on the MF Board who have a deep and thorough understanding of this and two of them are scientists. The pure shooters know it intuitively from experience.

When you peep every shot in a variety of situations at full res with GFX files on a 32 inch pro IPS calibrated high color gamut monitor you see the truth and you learn fast.

Click to expand...

I said nothing about underexposing and pushing in post in that sentence. What I meant that for a given exposure (choice of f-stop and speed), the highest ISO which does not clip the highlights you want to keep is the best choice (with little benefit or not).

You see ISO as primary, and SS and f-stop as afterthoughts. I see the latter two as primary and ISO as secondary.

Click to expand...

I agree with you. A lot of people see ISO as something that allows you to use higher SS, but that's why M mode exists, you can set SS/f to whatever you want, ISO is a slave to these settings.

Chris R-UK said:

I am giving a talk to my camera club next month and I will be covering, amongst other things, ISO invariance and how to find out whether your camera is ISO invariant. I don't have a particularly ISO invariant camera myself (Olympus E-M1.2), but I would like to include some examples of the situations where it might be useful to know the point at which your camera effectively become ISO invariant.

So, can anybody with an ISO invariant camera give me examples of situations that I could include as examples in my talk?

Click to expand...

Understanding your camera’s ISO invariant capability is especially important if you are shooting RAW in low light/high dynamic range situations. If you’ve minimized your SS (while still preventing motion blur) and maximized you aperture (while maintaining the necessary DOF and/or sharpness), then you’ve already maximized your sensor exposure/minimized shot noise and are left with ISO manipulation for ensuring an optimal RAW file to work with in post.

I (used to) shoot a lot of theatre which would often be poorly and unpredictably lit. With wildly varying light levels and important spotlit highlights to capture, an invariant sensor allows you to significantly “underISO” to greatly improve the effective dynamic range/highlight headroom while suffering no significant noise penalty vs. shooting at a higher ISO in-camera. What would normally be an ISO 6400 scene, but shot at ISO 800 (the “mostly” invariant starting point of my Fujis) and selectively brightened in post will still have about the same noise had you shot at ISO 6400, but will retain ISO 800 dynamic range with significantly more highlight headroom to work with which can yield a far better result.
Astrophotography is another situation where this is a big help. Instead of shooting at ISO 6400 which might work for the night sky, but will blow out any terrestrial light sources, shoot at ISO 800 to preserve the lights on the ground without clipping and just push the sky in post - again with no significant noise penalty relative to shooting at ISO 6400 (or whatever).

Thanks Erik, your examples will be useful for my talk.