Is there a benefit of a camera with superior dynamic range?

[Shadowman]

Nicely summarized and I feel it answers the original question and if the camera functions admirably it opens up opportunity to capture energetic shots as well as somewhat static shots.

[Manfred M]

While discussing dynamic range we need to also remember two time-honoured weapons photographers have used over the years when it was too bright out for their camera to capture an image; the ND filter and on the other end of the scale, when the shadow detail was too dark to be captured, we would "add light" via flash.

[george013]

While discussing dynamic range we need to also remember two time-honoured weapons photographers have used over the years when it was too bright out for their camera to capture an image; the ND filter and on the other end of the scale, when the shadow detail was too dark to be captured, we would "add light" via flash.

They don't change the dynamic range. It are tricks so you can capture a certain scene within the dynamic range of the sensor.
One should distinguish between the definition and how to deal with it.

George

[Manfred M]

They don't change the dynamic range. It are tricks so you can capture a certain scene within the dynamic range of the sensor.
One should distinguish between the definition and how to deal with it.

George

You are 100% correct, George. I was merely trying to point out that HDRI is not the only solution when parts of the scene lie outside the range of what the camera can record. As Dave pointed out, the camera's dynamic range is determined by the component's the camera's builder uses and is at its widest at base ISO.

[Dave Humphries]

True that adding an ND filter won't change the DR of the scene as presented to the sensor, it'll just 'move it down a bit'.

However, if you think about it, adding light to the shadows with flash (or even a reflector) could reduce the DR of the scene as seen by the sensor (lots of caveats of course).

i.e. if the 'maximum' light (sunlight perhaps) at the top of the scene DR range is capturable by the camera and light is added to all the shadow areas (at the lower end of the DR), it may then fit within the sensor's DR capability.

Fairly unlikely to happen in real life, but theoretically possible.

[Manfred M]

True that adding an ND filter won't change the DR of the scene as presented to the sensor, it'll just 'move it down a bit'.

With a straight neutral density (ND) filter, that is true, but if you use a graduated neutral density (GND) filter you have a closer approximation to a flash (with drop-off) as it doesn't affect the whole scene. While I don't use a reverse GND, I believe it is designed primarily for sunsets. The bottom half of the reverse GND is clear, and then you get a hard line of density that decrease from the middle of the filter to the top. That way the bright spike of light from the setting sun is cut drastically, so that the camera's sensor can record it without clipping.

[Mike Buckley]

True that adding an ND filter won't change the DR of the scene as presented to the sensor,

Actually, the ND filter doesn't change the dynamic range of the scene itself but it definitely does change the DR of the scene as presented to the sensor.

However, if you think about it, adding light to the shadows with flash (or even a reflector) could reduce the DR of the scene as seen by the sensor

More important, it could reduce the dynamic range of the scene itself. Unlike when using a filter, a flash and/or reflector changes the scene.

[Tronhard]

I thought that CiC's own tutorial on the subject shed some dynamic light on the whole discussion. Perhaps you might review and reflect on it! (and no, I'm not just playing it for laughs, the tutorial and the one on camera sensors it refers to is quite pertinent.

https://www.cambridgeincolour.com/tutorials/dynamic-range.htm

To me the dynamic range of the camera defines the potential range of shadows and highlights within the captured image, What happens in PP makes use of, but cannot exceed that potential.

[Dave Humphries]

Actually, the ND filter doesn't change the dynamic range of the scene itself but it definitely does change the DR of the scene as presented to the sensor.

Well, a GND or RGND will/can, as Manfred has now clarified, but an "ND" filter won't (and that's what was being discussed).
If a simple ND filter; the EV range at the sensor will still be what it was, just the 'top' will be x stops lower, as will the 'bottom' end.

More important, it could reduce the dynamic range of the scene itself. Unlike when using a filter, a flash and/or reflector changes the scene.

Err, I though that's what I said
(but I was probably adding more text while you were replying)

[xpatUSA]

If you shoot at the camera's 'base ISO', that will give an output file (if shooting RAW and using the largest bit range) with the widest dynamic range the camera's sensor is capable of capturing <>

Please pardon a little pedantry, Dave. "base ISO" is one term we see a lot but does not exist in ISO 12232, well not in my 2006 edition anyway. But, assuming as many do that it refers to a value of ISO defined by their sensor saturation-based method (section 6), there is a built-in headroom of a 1/2-stop (41%). That means for a scene like ISO's 18% test card (i.e. the Kodak) there is a loss of exposure which could for example be re-gained by exposing to the right and thereby gaining a bit more DR.

I was moved to comment because today, by coincidence, I was trying to determine the 'base ISO' of one of my cameras and am momentarily expert on the subject. One thing I found out was that the actual value is by no means cut and dried! Another method (Kodak's ?) allows a mere 6% headroom for example. I've also read that folks can create their own 'base ISO' by deciding on a personal image acceptability. So, these days I take the phrase 'base ISO' with a pinch of salt!

My camera's 'base ISO' turned out to be either 107 or 84 depending on whose constant I use in the equation S = K.f-number^2 divided by (t.L). It'll be different again under different light, no doubt. I'm still playing . . . beats shooting Real World stuff anytime . .

[Tronhard]

To extend Ted's observation, DxO Mark, the camera and lens testing site, plot (among other things) the effective ISO profiles of the bodies they test, some of which vary considerably from the specs supplied by the manufacturer...

[Dave Humphries]

Hi Ted,

If I may also be a little pedantic, what I think you're largely talking about is defining a numerical value for it and how that might be compared to other cameras. That's not directly relevant to this discussion - although I agree that if absolute and comparable figures were known to buyers, it would better inform the actual choice of a camera.

My use of the term does not relate to this, nor to ISO standards, it is about how the analogue and digital circuitry works inside the camera.

What I meant by 'base ISO' is that electronic camera control setting where the analogue signal is passed to the A-D converter with the lowest amplification and all bits of the width of the data bus are used. It is my understanding that often manufacturers will fudge their lower minimum ISO figure(s) by simply/effectively ignoring the least significant bit (LSB) from the A-D converter and shifting the other bits along one, thus apparently making the sensor less sensitive by one stop.

You'll recall that say 5-10 years ago, cameras ISO ranges were often labelled something like "L1", 200, 400, 800, 1600, 3200, "H1", "H2", etc. - the "L" and "H" settings were achieved by bit shifting.

It may be that my information is outdated in these days of 'ISO-less cameras', I accept that I haven't looked in to their internal circuitry.

Disclaimer: It's late and my brain is not entirely convinced I have explained the concept 'the right way round'

Cheers, Dave

[Mike Buckley]

If a simple ND filter; the EV range at the sensor will still be what it was, just the 'top' will be x stops lower, as will the 'bottom' end.

Thanks for proving my point. When measuring the dynamic range in terms of EV, you're right of course that it hasn't changed. When measuring the dynamic range in terms of luminosity values of the light entering the lens, the dynamic range indeed has changed. At least that's my understanding of the term, dynamic range.

[Dave Humphries]

The light levels change Mike, but their range (extent from max to min) does not - and since we're discussing a 'range', I say it doesn't change, but it's late and I am arguing semantics with you, so - time for bed!

G'night all,
Dave

[xpatUSA]

Hi Ted,

If I may also be a little pedantic, what I think you're largely talking about is defining a numerical value for it and how that might be compared to other cameras. That's not directly relevant to this discussion - although I agree that if absolute and comparable figures were known to buyers, it would better inform the actual choice of a camera.

Yes, my most useful reference to it comes from a Technical Note:

"There are two basic types of ISO: saturation-based and noise-based. The saturation-based ISO is also referred to as the ‘base ISO’."

http://kronometric.org/phot/temp/Bas...0Reference.PDF

I put it there for you to download if you're interested. Point being that it mentions 'base ISO' and goes on to say what that means (to them).

My use of the term does not relate to this, nor to ISO standards, it is about how the analogue and digital circuitry works inside the camera.

What I meant by 'base ISO' is that electronic camera control setting where the analogue signal is passed to the A-D converter with a nominal 'unity' gain and all bits of the width of the data bus are used. It is my understanding that often manufacturers will fudge their lower minimum ISO figure(s) by simply/effectively ignoring the least significant bit (LSB) from the A-D converter and shifting the other bits along one, thus apparently making the sensor less sensitive by one stop.

You'll recall that say 5-10 years ago, cameras ISO ranges were often labelled something like "L1", 200, 400, 800, 1600, 3200, "H1", "H2", etc. - the "L" and "H" settings were achieved by bit shifting.

Sorry, I'm a bit of a newcomer to serious photography; my last 'normal' camera was a Nikon D50 and I often wondered why it's lowest ISO was 200.

It may be that my information is outdated in these days of 'ISO-less cameras', I accept that I haven't looked in to their internal circuitry.

All my Sigmas are truly ISO-less - the camera ISO setting is passed as meta-data to their conversion conversion software which does the gain thing digitally. So the gain between the Foveon sensor chip and the 3 ADCs is fixed. Nice and simple. However, it does mean that I can't use your personal definition of 'base ISO' which I did understand fully, BTW.

[Manfred M]

"base ISO" is one term we see a lot but does not exist in ISO 12232

Base ISO is the lowest numeric ISO setting. While some cameras will can be set to values below the "base" setting, they are not assigned a numeric ISO value and instead are given values like Lo -1, etc. As others have suggested the actual ISO values may differ from the ones stated by camera. My source for this data is the DxO Mark website.


https://www.dxomark.com/


Sigma appears to be an exception here, but as you have written many times before, it is the exception rather than the rule in some of its features as it uses a Foveon sensor rather than the more common Bayer array type sensor.

[Shadowman]

You are 100% correct, George. I was merely trying to point out that HDRI is not the only solution when parts of the scene lie outside the range of what the camera can record. As Dave pointed out, the camera's dynamic range is determined by the component's the camera's builder uses and is at its widest at base ISO.

Nikon's D-Lighting function also achieves in-camera the same dynamic range modification that ND filters and flash attempts to do, however when used the image can lose some of its snap, but on most camera models you can limit the amount of modification.

[Manfred M]

Nikon's D-Lighting function also achieves in-camera the same dynamic range modification that ND filters and flash attempts to do, however when used the image can lose some of its snap, but on most camera models you can limit the amount of modification.

Active D-Lighting is something that raw photographers need to approach with extreme caution, as it is primarily meant to enhance jpeg highlight and shadow details, much like the Adobe Camera Raw and Lightroom Shadows and Highlights sliders do for raw files. In order to optimize the results, the actual exposure is affected, which means the raw data, even though it does not get any of the Active-D processing, is exposed to optimize this in-camera functionality. This suggests that the actual raw data will be suboptimal.

If you are primarily a jpeg shooter, Active-D can give you better straight of of camera results by opening up shadow details and suppressing highlights. If you are a raw or jpeg + raw shooter, I suggest you deactivate this Nikon camera setting and expose properly to get high quality raw. You will get better base raw data to work with that way.

[Stagecoach]

Active D-Lighting is something that raw photographers need to approach with extreme caution, as it is primarily meant to enhance jpeg highlight and shadow details, much like the Adobe Camera Raw and Lightroom Shadows and Highlights sliders do for raw files. In order to optimize the results, the actual exposure is affected, which means the raw data, even though it does not get any of the Active-D processing, is exposed to optimize this in-camera functionality. This suggests that the actual raw data will be suboptimal.

Manfred, for info tests undertaken on a D300 showed that the exposure was reduced between 1/3 to 2/3 stop that affected the RAW file when Active D Lighting was used on High setting and metering in Matrix mode.

I found the variation was dependent upon scene content.

[Dave Humphries]

I did my own tests (years ago) and determined much as Manfred states, for RAW (only) processing, there were only disadvantages to enabling Active D-Lighting and hence I leave mine OFF permanently.

Except - on the very odd occasions I shoot jpg (which can be years between), if the DR is large, I guess I could turn it on - if I remember!