Why do camera manufacturers tout useless ISO ranges?

[Jeeter001]

Recently I was looking at adding a Panasonic GH5S, a supposed low light optimized M43 body, to my collection of M43 bodies and lenses. Despite marketing literature showing this camera can go to an "extended ISO" above 200k, looking at sample photos and the dpreview compare tool, it's MAYBE a one stop better performance than other Panasonic cameras from the same "family", and as such the highest usable ISOs are probably in the range of 6400-12800. Of course there is subjectivity and variance to this, some photographers are more tolerant than others with higher ISOs and there are some compositions that are more tolerant to higher ISOs than others.

I notice all camera manufacturers do this. What is the purpose of advertising an ISO range where MOST of that range is completely unusable? I know this might not be the case with some full frame bodies like the Sony's which are known to be able to boost ISO into the 100k range and still get an acceptable photo. Is this just somewhat dishonest advertising by APC and M43 manufacturers to make it seem like their cameras can keep up with a full frame? It just seems rather egregious to say a camera can go to 50k+ or 100k+ ISO or whatever, when really the max usable is 6,400, a tiny fraction of what is advertised.

 

[Jack Hogan]

I personally never look at them and frankly have no idea what they are for my cameras. However I can see how they could be helpful to folks just approaching digital photography.

It gives you an idea of how the manufacturers see performance of their products improving across generations. Once Iso 3200 was considered to be just acceptable, today it is run of the mill.

Manufacturers use their own IQ criteria to determine the upper limit. We don't know what those are but we trust them to be consistent in applying them.

Jack

 

[AiryDiscus]

I personally never look at them and frankly have no idea what they are for my cameras. However I can see how they could be helpful to folks just approaching digital photography.

It gives you an idea of how the manufacturers see performance of their products improving across generations. Once Iso 3200 was considered to be just acceptable, today it is run of the mill.

Manufacturers use their own IQ criteria to determine the upper limit. We don't know what those are but we trust them to be consistent in applying them.

Jack

That's pretty foolhardy. The upper limit is almost certainly a mandate from marketing to the software team.

 

[OpticsEngineer]

"Is this just somewhat dishonest advertising by APC and M43 manufacturers to make it seem like their cameras can keep up with a full frame?"

It does seem to look that way

To me it seems silly, because average consumers have no idea what ISO is so it does not affect buying decisions in the way the marketing team might imagine. Having worked in multiple large corporations over the years, my observation is marketing claims like these are usually slapped on products without informing or asking the engineers about them. I have seen situations where pains were taken to prevent the engineering team from learning what marketing claims would be from fear the engineers might somehow let upper management know claims were just wishful thinking. In large corporations, there is a mandate from above that things must be better and cheaper every year and people's bonuses get tied to checking boxes on performance reviews without regard for how things affect customers. So that is how I view high ISO claims. Meaningless, driven by internal corporate politics and metrics. If we told a small lie last year, we have to tell a little bit bigger lie next year. And it adds up.

But through all the smoke and haze of ISO specmanship, I have found that when I read the reviews here on dpreview, the editors state what in practice they find is the highest usable ISO and that matches really well with my personal experience taking travel, event and family photos.

 

[Pedagydusz]

"Is this just somewhat dishonest advertising by APC and M43 manufacturers to make it seem like their cameras can keep up with a full frame?"

It does seem to look that way

Frankly, I don't think so. It is true that I never use the upper limit ISO in my camera (Olympus OM-D cameras), but if I eventually happened to cross my path with the Pope in a dimly lit church, or with Sharon Stone running without any clothes in a hotel hallway, a 25,600 ISO picture would give a quite presentable 6x4 picture, even if it would not be acceptable in any serious photo forum.

I have chosen these too topics just as I was writing, I don't really expect any of them ever happening, but you get my drift... A poor 6x4 would be much preferable than saying "Oh, man!..."

Of course I don't buy cameras for that purpose, but knowing that I can get, if convenient, an extreme snapshot, is preferable than knowing I will never be able to. If it does not cost more money.

--
Antonio
http://ferrer.smugmug.com/

 

[Sympa]

Marketing: "Can you raise the ISO to 419430400?"

Engineer: "The software will allow it yes, well, but..."

Marketing: "Then, do it."

 

[AiryDiscus]

Marketing: "Can you raise the ISO to 419430400?"

Engineer: "The software will allow it yes, well, but..."

Marketing: "Then, do it."

You forgot the meeting before that,

"sensor performance has been stagnant for years but we need this camera to have better specs than the last one"

"what number can we bump without getting sued for false marketing?"

"ISO!"

 

[John Sheehy]

I personally never look at them and frankly have no idea what they are for my cameras. However I can see how they could be helpful to folks just approaching digital photography.

It gives you an idea of how the manufacturers see performance of their products improving across generations. Once Iso 3200 was considered to be just acceptable, today it is run of the mill.

Manufacturers use their own IQ criteria to determine the upper limit. We don't know what those are but we trust them to be consistent in applying them.

Jack

That's pretty foolhardy. The upper limit is almost certainly a mandate from marketing to the software team.

One wonders what they were thinking with the Nikon D5. The camera has very low pre-gain read noise, but significant post-gain read noise with banding, and Nikon decided to go up to ISO 3.28M on that camera with 32x multiplication of ISO 102K gain. I am all for cameras having at least an "expert mode" option to give reasonable review image and JPEG brightness for very weak exposures, but why 3.28M with just this camera? I bet there are other Nikons that are a little bit noisier than the D5 at ISO 102K, but which have much less post-gain read noise and banding that wouldn't look much (if any) worse at ISO 3.28M.

 

[John Sheehy]

"Is this just somewhat dishonest advertising by APC and M43 manufacturers to make it seem like their cameras can keep up with a full frame?"

They can, however, for focal-length-limited purposes. The person shooting wildlife with a current 20MP m43 with the largest lens they are willing to carry, is likely getting no worse noise in their subjects than someone using a top-end FF high-ISO performer, if they were to use the same focal length, f-number, and shutter speed, at the same high ISO, from the same subject distance. The FF sensors are protected from showing that they are no better (or sometimes even worse) by the circumstantial fact that people, in general, do not inspect their images on their monitors at a scale of greater than 100% pixel view, which is less demanding with larger pixels. If you could mount the telephoto lenses used to current 1" 20MP sensors, they have less noise yet, for that 1" piece of sensor area, while having a lot more resolution to boot.

 

[John Sheehy]

"Is this just somewhat dishonest advertising by APC and M43 manufacturers to make it seem like their cameras can keep up with a full frame?"

It does seem to look that way

Frankly, I don't think so. It is true that I never use the upper limit ISO in my camera (Olympus OM-D cameras), but if I eventually happened to cross my path with the Pope in a dimly lit church, or with Sharon Stone running without any clothes in a hotel hallway, a 25,600 ISO picture would give a quite presentable 6x4 picture, even if it would not be acceptable in any serious photo forum.

I have chosen these too topics just as I was writing, I don't really expect any of them ever happening, but you get my drift... A poor 6x4 would be much preferable than saying "Oh, man!..."

Of course I don't buy cameras for that purpose, but knowing that I can get, if convenient, an extreme snapshot, is preferable than knowing I will never be able to. If it does not cost more money.

I don't even really think much about ISO settings, per se, when light is limited. I have been using manual exposure with auto-ISO for many years for limited-light photography, and noise in my paradigm is down to quality of light, needed shutter speed, and the optical aperture of the lens in combination with the subject size and distance. ISOs are just incidental, and if an image is very noisy, it is because of the circumstance and my choice of shutter speed; not the ISO setting that happened. Causality vs association.

I am not saying that there are no situations in which I might choose a fixed, elevated ISO setting in low light, but that would require that the lighting was completely even and consistent, and that only the key or tone of items in the varying compositions was varying.

 

[Erik Kaffehr]

I personally never look at them and frankly have no idea what they are for my cameras. However I can see how they could be helpful to folks just approaching digital photography.

It gives you an idea of how the manufacturers see performance of their products improving across generations. Once Iso 3200 was considered to be just acceptable, today it is run of the mill.

Manufacturers use their own IQ criteria to determine the upper limit. We don't know what those are but we trust them to be consistent in applying them.

Jack

Hi Jack,

The only camera I own that has a max ISO I am aware of is my Phase One P45+ back. It has a max of ISO 800.

The way I see it, increasing ISO is essentially just reducing exposure. But increasing ISO may have some benefits:

• On older sensors, with external converters, using higher ISO could lead to increasing pre-amplification of the signal going to the ADC.
• Many modern sensors have dual gain conversion. So raising ISO causes the sensor to invoke high conversion gain mode at some ISO.
• Raising ISO may also tell camera firmware to optimize processing for low exposure/high noise. That may affect JPEG processing. In some cases that may also affect raw :-(

Now, we may have three cameras, having exactly the same CMOS sensors made by Sony.

• One may go to 6400 ISO.
• One may go to 25000 ISO.
• One may go to 250000 ISO.

Going past 25000 ISO, it is quite probable that there would be some noise reduction applied on raw data, trading resolution for SNR.

My take is that a given technology sort of sets limits on what is achievable. Apart from that we have BR (Bragging Rights). BR may be very attractive but it may also have ZE (Zero Effect) on images.

So, we may give up on some DR to achieve a bit more DR and the net results may be better ZE. Who will know?

BR

EKr

 

[RichardSwearinger]

I hope the compression doesn't ruin this photo, but you can make reasonably moody photos in dark places with ISO20,000. If I'd used a lower ISO I'm sure the colors would be a little richer but I probably wouldn't have been able to hold the camera steady long enough to keep the branches sharpish.

 

[bobn2]

I hope the compression doesn't ruin this photo, but you can make reasonably moody photos in dark places with ISO20,000. If I'd used a lower ISO I'm sure the colors would be a little richer but I probably wouldn't have been able to hold the camera steady long enough to keep the branches sharpish.

That's surprisingly similar to the photo by Pierre Sottas that started the whole 'ISOless' thing. It would have been much better had you used a lower ISO and the same exposure, because the lit windows wouldn't be clipped to white. It could probably be processed a good deal lighter with some advantage, producing some detail in the darker parts.

--
...because you know, sometimes words have two meanings.

 

[Erik Kaffehr]

Hi,

Exposure is mostly about capturing light. Sensor developments mostly help in reducing noise in the darks.

Older generations sensors used to have huge analogue gains at high ISO reducing readout noise, but it seems no longer be the case with modern CMOS.

Reducing exposure will increase noise. In older generation sensors, increasing ISO would have a significant benefit in the darks, but modern sensors are so good darks are good anyway.

With most Sony sensors it may make little sense to increase ISO beyond say 800, going past that we can either:

• Underexpose and keep DR
• Increase ISO and throw away highlight data

Many cameras will apply some kind of noise reduction at high ISOs, but, that could be better done in raw processing.

Shooting JPEG, extreme ISOs may make some sense, as they can instruct the JPEG engine to optimize noise reduction and sharpening. But shooting RAW, increasing ISO beyond 'reasonable' values may make little sense.

So, what may a reasonable max ISO be? Hard to say, but this table may give an idea:




See bottom part of Bill Claff's page on PDR .

One thing noticeable is that larger sensors allow for higher ISOs, that is because they can capture more photons at the same aperture/shutter speed combination.

Best regards

Erik

--
Erik Kaffehr
Website: http://echophoto.dnsalias.net
Magic uses to disappear in controlled experiments…
Gallery: http://echophoto.smugmug.com
Articles: http://echophoto.dnsalias.net/ekr/index.php/photoarticles

 

[bobn2]

Hi,

Exposure is mostly about capturing light. Sensor developments mostly help in reducing noise in the darks.

I would say that exposure is all about capturing light, not mostly. The question is the results which you get from an exposure, which is about sensor size (how much energy the exposure translates to), quantum efficiency (what proportion of the available photons make up the image) and read noise (how much noise intrudes into the darkest bits).

One thing noticeable is that larger sensors allow for higher ISOs, that is because they can capture more photons at the same aperture/shutter speed combination.

At the same aperture/shutter speed/angle of view combination the light captured is independent of sensor size. What you meant was f-number - the difference is important in this kind of conversation.

 

[John Sheehy]

So, what may a reasonable max ISO be? Hard to say, but this table may give an idea:

I don't think that these "highest acceptable ISO" metrics are very useful for representing the real high ISO abilities of cameras.

Yesteryear's FF sensors might score better than current m43 sensors by a good margin, and actually do better in the highlights of ISO 2000, but push both to ISO 102K, and the m43 may completely embarrass the aging FF sensor in the darker areas, because the latter has so much more read noise, which these types of metrics are not reflecting very well. These metrics fall very close to sensor area times quantum efficiency, and even to the degree to which they do weight read noise, they ignore its spatial character, which becomes extremely important in the shadows of very high ISOs.

 

[Erik Kaffehr]

Hi,

Exposure is mostly about capturing light. Sensor developments mostly help in reducing noise in the darks.

I would say that exposure is all about capturing light, not mostly. The question is the results which you get from an exposure, which is about sensor size (how much energy the exposure translates to), quantum efficiency (what proportion of the available photons make up the image) and read noise (how much noise intrudes into the darkest bits).

Thanks for correction.

One thing noticeable is that larger sensors allow for higher ISOs, that is because they can capture more photons at the same aperture/shutter speed combination.

At the same aperture/shutter speed/angle of view combination the light captured is independent of sensor size. What you meant was f-number - the difference is important in this kind of conversation.

Thanks for correction, again.

Best regards

Erik

 

[RichardSwearinger]

Yes, it would have been different with a lower ISO, but I'm not sure it would have fit the creative vision I had in mind.

Here's the thought process I went through as I shot this. It was dark and I was picking my way back to the house by feel trying not to trip over roots and fallen branches and I started to see that the house could look fairly ominous in the way it was framed by the trees. (Or what I could see of the trees in the dark).

I wanted to preserve the sinister character so I made a creative decision to use the look that comes from using a high ISO. If I'd shot at ISO 64 it would have looked too clean, more like a photo from Architectural Digest and less like "Friday the 13th."

But there was a practical consideration as well, I was kneeling down balancing the camera on a fallen log in the dark—no flashlight so I couldn't really see what I was doing. Even though the lens has one of Nikon's best VR implementations, experience has shown me that it has limits.

There was some detail in the darkest parts of the image but I crushed the blacks slightly to give the image a little better snap; Amazon's compression algorithm finished off any remaining shadow detail.

The combination of ISO 20,000, and 1/20th at f/2.8 worked out very well for the creative and technical requirements of the shot.

I hope the compression doesn't ruin this photo, but you can make reasonably moody photos in dark places with ISO20,000. If I'd used a lower ISO I'm sure the colors would be a little richer but I probably wouldn't have been able to hold the camera steady long enough to keep the branches sharpish.

That's surprisingly similar to the photo by Pierre Sottas that started the whole 'ISOless' thing. It would have been much better had you used a lower ISO and the same exposure, because the lit windows wouldn't be clipped to white. It could probably be processed a good deal lighter with some advantage, producing some detail in the darker parts.

 

[bobn2]

Yes, it would have been different with a lower ISO, but I'm not sure it would have fit the creative vision I had in mind.

So, were you working from raw, or a JPEG? If from raw, you could have captured more information about the scene and thereby fitted the creative vision, or a slightly different one. It's in the processing.

bobn2, post: 63652024, member: 1851922"]

RichardSwearinger, post: 63652024, member: 1851922"]
I hope the compression doesn't ruin this photo, but you can make reasonably moody photos in dark places with ISO20,000. If I'd used a lower ISO I'm sure the colors would be a little richer but I probably wouldn't have been able to hold the camera steady long enough to keep the branches sharpish.

That's surprisingly similar to the photo by Pierre Sottas that started the whole 'ISOless' thing. It would have been much better had you used a lower ISO and the same exposure, because the lit windows wouldn't be clipped to white. It could probably be processed a good deal lighter with some advantage, producing some detail in the darker parts.

--
...because you know, sometimes words have two meanings.
[/QUOTE]
[/QUOTE]

 

[Guy Parsons]

I have long held the belief that that top two ISO settings on any brand are always due to marketing so they can boast about bigger numbers.