Utility of dynamic range beyond a certain point
p.5 #1Utility of dynamic range beyond a certain point
EB-1 wrote:
Most of the time I have enough or I need more than the sensor will provide. It's rarely that a single stop or so will save the day. At ISO 100 we want the best IQ and it looks like the a7rVI delivers. It's not much to complain about. I have not tried bracketing with the a7rVI yet.
EBH
Make sure you are using the mechanical shutter, take advantage of the whole range...
Most of the time I have enough or I need more than the sensor will provide. It's rarely that a single stop or so will save the day. At ISO 100 we want the best IQ and it looks like the a7rVI delivers. It's not much to complain about. I have not tried bracketing with the a7rVI yet.
EBH
Make sure you are using the mechanical shutter, take advantage of the whole range...
Jun 13, 2026 at 04:49 PM
p.5 #2Utility of dynamic range beyond a certain point
EB-1 wrote:
Most of the time I have enough or I need more than the sensor will provide. It's rarely that a single stop or so will save the day.
That’s been my point for quite a while… but you made it much more concisely. :-)
I’m all for manufacturers continuing to improve DR along with all other aspects of camera performance, but the differences in DR are rarely decision points in the way that other factors can be.
Most of the time I have enough or I need more than the sensor will provide. It's rarely that a single stop or so will save the day.
That’s been my point for quite a while… but you made it much more concisely. :-)
I’m all for manufacturers continuing to improve DR along with all other aspects of camera performance, but the differences in DR are rarely decision points in the way that other factors can be.
Jun 14, 2026 at 02:27 AM
p.5 #3Utility of dynamic range beyond a certain point
gdanmitchell wrote:
^^^
Concerning the above, you are right that the DR of jpg files isn’t really directly relevant here, especially in light of the fact that (as mentioned earlier in this thread) the real value of large DR in cameras is to provide working space in post so that final images in DR limited formats (which all are) can include values form darkest and lightest areas.
There are two ways, however, in this mentioning jpg DR in a thread about camera DR kind of makes sense:
1. Today’s cameras do a fair amount of (sometimes a lot of) internal processing of the initial raw...Show more →
phones, in-camera bracketing jpegs, the x100iv is also a good example of what great incamera jpegs are like, also slog 3 video from sony cameras is better than any raw recording from other brands.
^^^
Concerning the above, you are right that the DR of jpg files isn’t really directly relevant here, especially in light of the fact that (as mentioned earlier in this thread) the real value of large DR in cameras is to provide working space in post so that final images in DR limited formats (which all are) can include values form darkest and lightest areas.
There are two ways, however, in this mentioning jpg DR in a thread about camera DR kind of makes sense:
1. Today’s cameras do a fair amount of (sometimes a lot of) internal processing of the initial raw...Show more →
phones, in-camera bracketing jpegs, the x100iv is also a good example of what great incamera jpegs are like, also slog 3 video from sony cameras is better than any raw recording from other brands.
Jun 15, 2026 at 03:30 AM
p.5 #4Utility of dynamic range beyond a certain point
chiron wrote:
Naturally any photography medium can be done poorly but HDR images on HDR displays can look amazing as well. If you're not planning to view or distribute HDR images then your SDR thesis is reasonable.Can you link to some examples of HDR images that you think have been well done?
For most landscape type images the HDR variant looks more true to life by far.
Not much processing required, just load the raw and export a jxl file.
This way we see what the camera sees.
The problem with HDR is that in 2026 different screens display HDR at completely different fidelity levels. It may look good on my screen but blown out on yours if the monitor cant display the highlight detail and it gets crushed by the display's EOTF curve
Edited on Jun 16, 2026 at 02:23 AM · View previous versions
snapsy wrote
old-gregg wrote
Sure, there are 10-bit screens and true HDR HEIC, but that shit looks garish to me.For most landscape type images the HDR variant looks more true to life by far.
Not much processing required, just load the raw and export a jxl file.
This way we see what the camera sees.
The problem with HDR is that in 2026 different screens display HDR at completely different fidelity levels. It may look good on my screen but blown out on yours if the monitor cant display the highlight detail and it gets crushed by the display's EOTF curve
Edited on Jun 16, 2026 at 02:23 AM · View previous versions
Jun 16, 2026 at 02:07 AM
p.5 #5Utility of dynamic range beyond a certain point
old-gregg wrote:
Everyone is excited about the increased dynamic range of the Sony A7RVI at nominal ISO. Meanwhile, I am asking myself whether I ever bump into the DR limitations of my existing cameras in practice?
When I look at my favorite landscape shots, for example, I see that - while there's plenty of unused room to lift shadows - lifting them even further results in unrealistic "HDR" look. I think you all are familiar with that look. In other words, at least in my work, there's a natural limit of how much the luminance range of a scene can be "compressed" into...Show more →
greg, have you never blown out the sky, lace details on a white dress or feather detail on a white bird before?
I have had to do sky replacement "first aid" on images before
Everyone is excited about the increased dynamic range of the Sony A7RVI at nominal ISO. Meanwhile, I am asking myself whether I ever bump into the DR limitations of my existing cameras in practice?
When I look at my favorite landscape shots, for example, I see that - while there's plenty of unused room to lift shadows - lifting them even further results in unrealistic "HDR" look. I think you all are familiar with that look. In other words, at least in my work, there's a natural limit of how much the luminance range of a scene can be "compressed" into...Show more →
greg, have you never blown out the sky, lace details on a white dress or feather detail on a white bird before?
I have had to do sky replacement "first aid" on images before
Jun 16, 2026 at 02:17 AM
p.5 #6Utility of dynamic range beyond a certain point
aCuria wrote:
When I look at my favorite landscape shots, for example, I see that - while there's plenty of unused room to lift shadows - lifting them even further results in unrealistic "HDR" look. I think you all are familiar with that look. In other words, at least in my work, there's a natural limit of how much the luminance range of a scene can be "compressed" into an 8-bit JPEG (or into a 200:1 print) by lifting the shadows. Sure, there are 10-bit screens and true HDR HEIC, but that shit looks garish to me.
In other words, the Nikon Z8 and the Sony A7RVI offer the same practical dynamic rangne for my landscapes, and I can't see how I can use the extra 2EV of the Sony. Which brings my question: what makes you excited about the additional stop of DR at ISO 100, and where are you **practically** bumping into DR limits of your current camera?
P.S. I forgot to mention the highlights. On that front my experience is the opposite: comparing A7RV to cameras with 2EV less DR, I do not see much "unused room" there, and there aren't any meaningful differences in highlight recovery across models. I suspect that the metering systems in all cameras have gotten pretty reliable at nailing exposure within ~1EV of highlight clipping, which means you get 1EV of highlight recovery regardless of a camera.greg, have you never blown out the sky, lace details on a white dress or feather detail on a white bird before?
I have had to do sky replacement "first aid" on images before
I agree,
i shoot studio dance photoshoots in my studio and i can easily shoot a black lace tutu with a white lace top at the same time with all the detail and DR one could ever want without that HDR look.
old-gregg wrote
Everyone is excited about the increased dynamic range of the Sony A7RVI at nominal ISO. Meanwhile, I am asking myself whether I ever bump into the DR limitations of my existing cameras in practice?When I look at my favorite landscape shots, for example, I see that - while there's plenty of unused room to lift shadows - lifting them even further results in unrealistic "HDR" look. I think you all are familiar with that look. In other words, at least in my work, there's a natural limit of how much the luminance range of a scene can be "compressed" into an 8-bit JPEG (or into a 200:1 print) by lifting the shadows. Sure, there are 10-bit screens and true HDR HEIC, but that shit looks garish to me.
In other words, the Nikon Z8 and the Sony A7RVI offer the same practical dynamic rangne for my landscapes, and I can't see how I can use the extra 2EV of the Sony. Which brings my question: what makes you excited about the additional stop of DR at ISO 100, and where are you **practically** bumping into DR limits of your current camera?
P.S. I forgot to mention the highlights. On that front my experience is the opposite: comparing A7RV to cameras with 2EV less DR, I do not see much "unused room" there, and there aren't any meaningful differences in highlight recovery across models. I suspect that the metering systems in all cameras have gotten pretty reliable at nailing exposure within ~1EV of highlight clipping, which means you get 1EV of highlight recovery regardless of a camera.
I have had to do sky replacement "first aid" on images before
I agree,
i shoot studio dance photoshoots in my studio and i can easily shoot a black lace tutu with a white lace top at the same time with all the detail and DR one could ever want without that HDR look.
Jun 16, 2026 at 02:30 AM
p.5 #7Utility of dynamic range beyond a certain point
aCuria wrote:
When I look at my favorite landscape shots, for example, I see that - while there's plenty of unused room to lift shadows - lifting them even further results in unrealistic "HDR" look. I think you all are familiar with that look. In other words, at least in my work, there's a natural limit of how much the luminance range of a scene can be "compressed" into an 8-bit JPEG (or into a 200:1 print) by lifting the shadows. Sure, there are 10-bit screens and true HDR HEIC, but that shit looks garish to me.
In other words, the Nikon Z8 and the Sony A7RVI offer the same practical dynamic rangne for my landscapes, and I can't see how I can use the extra 2EV of the Sony. Which brings my question: what makes you excited about the additional stop of DR at ISO 100, and where are you **practically** bumping into DR limits of your current camera?
P.S. I forgot to mention the highlights. On that front my experience is the opposite: comparing A7RV to cameras with 2EV less DR, I do not see much "unused room" there, and there aren't any meaningful differences in highlight recovery across models. I suspect that the metering systems in all cameras have gotten pretty reliable at nailing exposure within ~1EV of highlight clipping, which means you get 1EV of highlight recovery regardless of a camera.greg, have you never blown out the sky, lace details on a white dress or feather detail on a white bird before?
I have had to do sky replacement "first aid" on images before
That would be an exposure issue and the reduced shadow noise (in the A7RVI compared, e.g., with the A1II), which is what the extra DR is all about, is not likely to help much with that. You can choose to give the subject less exposure and by doing this you can avoid the highlight clipping but also whenever you give yourself more room in the highlights for brighter spots to be correctly imaged, you increase noise in the darker parts of the image due to increased photon shot noise. The differences in DR are likely to be visible in the darkest shadows at low ISO and it's not clear they would make a difference in most practical use cases. The differences in DR (between A7RVI and A1II) are present in the ISO 100-500 range while photonstophotos.net shows practically identical PDR at ISO 640 and higher; how often would bird photographs be made at max ISO 500? It's certainly possible but I think most bird photography happens at medium to high ISO these days.
I think it's brilliant that such developments still occur in camera sensors and think people who shoot at low ISO will find uses for the reduced shadow noise. However, if you actually take a landscape and bracket the exposures (assuming there is little movement in the scene) and merge the exposures carefully, you will get a much larger reduction of noise and improvement of the final image, but of course, this requires a lot of skill and a good eye when doing the editing. For shots where there is movement in the scene, single-exposure DR can matter, but because of the ever-present photon shot noise, the differences between camera generations might not be as important as they once were in this area of technical image quality improvement. At high ISO, different technical implementations might be more effective in optimizing the DR. Certainly if shooting flying birds at ISO 25600 before the sunrise, any image quality improvement that can be had is welcome, as the starting point is not very good, but may still show interesting behavior.
Examples of where base ISO DR might matter (and when bracketing is still a useful tool) is landscapes that include the setting or rising sun in the image. People love sunsets and sunrises and depending on the foreground, there can be a huge lighting contrast between some rocks and the brightest part of the sun. Whether this type of photography is any particular photographer's cup of tea is another matter, of course.
old-gregg wrote
Everyone is excited about the increased dynamic range of the Sony A7RVI at nominal ISO. Meanwhile, I am asking myself whether I ever bump into the DR limitations of my existing cameras in practice?When I look at my favorite landscape shots, for example, I see that - while there's plenty of unused room to lift shadows - lifting them even further results in unrealistic "HDR" look. I think you all are familiar with that look. In other words, at least in my work, there's a natural limit of how much the luminance range of a scene can be "compressed" into an 8-bit JPEG (or into a 200:1 print) by lifting the shadows. Sure, there are 10-bit screens and true HDR HEIC, but that shit looks garish to me.
In other words, the Nikon Z8 and the Sony A7RVI offer the same practical dynamic rangne for my landscapes, and I can't see how I can use the extra 2EV of the Sony. Which brings my question: what makes you excited about the additional stop of DR at ISO 100, and where are you **practically** bumping into DR limits of your current camera?
P.S. I forgot to mention the highlights. On that front my experience is the opposite: comparing A7RV to cameras with 2EV less DR, I do not see much "unused room" there, and there aren't any meaningful differences in highlight recovery across models. I suspect that the metering systems in all cameras have gotten pretty reliable at nailing exposure within ~1EV of highlight clipping, which means you get 1EV of highlight recovery regardless of a camera.
I have had to do sky replacement "first aid" on images before
That would be an exposure issue and the reduced shadow noise (in the A7RVI compared, e.g., with the A1II), which is what the extra DR is all about, is not likely to help much with that. You can choose to give the subject less exposure and by doing this you can avoid the highlight clipping but also whenever you give yourself more room in the highlights for brighter spots to be correctly imaged, you increase noise in the darker parts of the image due to increased photon shot noise. The differences in DR are likely to be visible in the darkest shadows at low ISO and it's not clear they would make a difference in most practical use cases. The differences in DR (between A7RVI and A1II) are present in the ISO 100-500 range while photonstophotos.net shows practically identical PDR at ISO 640 and higher; how often would bird photographs be made at max ISO 500? It's certainly possible but I think most bird photography happens at medium to high ISO these days.
I think it's brilliant that such developments still occur in camera sensors and think people who shoot at low ISO will find uses for the reduced shadow noise. However, if you actually take a landscape and bracket the exposures (assuming there is little movement in the scene) and merge the exposures carefully, you will get a much larger reduction of noise and improvement of the final image, but of course, this requires a lot of skill and a good eye when doing the editing. For shots where there is movement in the scene, single-exposure DR can matter, but because of the ever-present photon shot noise, the differences between camera generations might not be as important as they once were in this area of technical image quality improvement. At high ISO, different technical implementations might be more effective in optimizing the DR. Certainly if shooting flying birds at ISO 25600 before the sunrise, any image quality improvement that can be had is welcome, as the starting point is not very good, but may still show interesting behavior.
Examples of where base ISO DR might matter (and when bracketing is still a useful tool) is landscapes that include the setting or rising sun in the image. People love sunsets and sunrises and depending on the foreground, there can be a huge lighting contrast between some rocks and the brightest part of the sun. Whether this type of photography is any particular photographer's cup of tea is another matter, of course.
Jun 16, 2026 at 04:00 AM
p.5 #8Utility of dynamic range beyond a certain point
gdanmitchell wrote:
On the other hand, if there is, as you suggest, a specific and objectively measurable difference in the color from the systems you should be able to simply create a preset (or some kind of profile) that could be applied in bulk with hardly any effort or time required.
In other words, making the color from camera A look like the color from camera B would not likely require a lot of image-specific processing at all, FWIw.
That's not really true as the full continuous spectrum of light striking the sensor is filtered by the color filter array and the spectra of different filters can be different in different implementations and this reduces the full spectrum of light into just three values. The three values will be different between the different CFAs, depending on the spectrum of the incoming light and the CFA spectra, and there is no way to un-filter the light after those CFAs to make a mechanistic algorithm that would reproduce the colors from one camera based on images from a different one. The information is simply not there in the recorded file to make this happen.
ilkka_nissila wrote
Yes, you can edit the colors but for high-volume photography it may not be that practical to spend a lot of time doing that, depending on the priorities of the photographer.In other words, making the color from camera A look like the color from camera B would not likely require a lot of image-specific processing at all, FWIw.
That's not really true as the full continuous spectrum of light striking the sensor is filtered by the color filter array and the spectra of different filters can be different in different implementations and this reduces the full spectrum of light into just three values. The three values will be different between the different CFAs, depending on the spectrum of the incoming light and the CFA spectra, and there is no way to un-filter the light after those CFAs to make a mechanistic algorithm that would reproduce the colors from one camera based on images from a different one. The information is simply not there in the recorded file to make this happen.
Jun 16, 2026 at 04:05 AM
p.5 #9Utility of dynamic range beyond a certain point
aCuria wrote:
greg, have you never blown out the sky, lace details on a white dress or feather detail on a white bird before?
I have had to do sky replacement "first aid" on images before
I added a clarification at the bottom, sharing my observation that all modern cameras, regardless of what photonstophotos shows, have about the same headroom for highlights. Probably the result of metering+sensor pair tuning. I guess what you're saying is these newer DGO cameras allow for more negative exposure compensation with no shadow noise penalty to protect the highlights even further. That's a great answer!
greg, have you never blown out the sky, lace details on a white dress or feather detail on a white bird before?
I have had to do sky replacement "first aid" on images before
I added a clarification at the bottom, sharing my observation that all modern cameras, regardless of what photonstophotos shows, have about the same headroom for highlights. Probably the result of metering+sensor pair tuning. I guess what you're saying is these newer DGO cameras allow for more negative exposure compensation with no shadow noise penalty to protect the highlights even further. That's a great answer!
Jun 16, 2026 at 09:46 AM
p.5 #10Utility of dynamic range beyond a certain point
ilkka_nissila wrote:
On the other hand, if there is, as you suggest, a specific and objectively measurable difference in the color from the systems you should be able to simply create a preset (or some kind of profile) that could be applied in bulk with hardly any effort or time required.
In other words, making the color from camera A look like the color from camera B would not likely require a lot of image-specific processing at all, FWIw.That's not really true as the full continuous spectrum of light striking the sensor is filtered by the color filter array and the spectra of different filters can be different in different implementations and this reduces the full spectrum of light into just three values. The three values will be different between the different CFAs, depending on the spectrum of the incoming light and the CFA spectra, and there is no way to un-filter the light after those CFAs to make a mechanistic algorithm that would reproduce the colors from one camera based on images from a different one. The...Show more →
In reality, no sensor provides the Perfect Holy Grail Of Color, and when we include variations in lighting color (directly, reflections, and light coming through things like leaves) color is highly variable. I suppose that if you are working under color temperature controlled lighting all of the time you can minimize these issues, but otherwise…
Given all of those variables, in my experience you can get pretty much any sensor to look the way you want it too if you are proficient at post.
YMMV.
Edited on Jun 17, 2026 at 01:08 AM · View previous versions
gdanmitchell wrote
ilkka_nissila wrote
Yes, you can edit the colors but for high-volume photography it may not be that practical to spend a lot of time doing that, depending on the priorities of the photographer.In other words, making the color from camera A look like the color from camera B would not likely require a lot of image-specific processing at all, FWIw.
In reality, no sensor provides the Perfect Holy Grail Of Color, and when we include variations in lighting color (directly, reflections, and light coming through things like leaves) color is highly variable. I suppose that if you are working under color temperature controlled lighting all of the time you can minimize these issues, but otherwise…
Given all of those variables, in my experience you can get pretty much any sensor to look the way you want it too if you are proficient at post.
YMMV.
Edited on Jun 17, 2026 at 01:08 AM · View previous versions
Jun 16, 2026 at 02:15 PM
|
p.5 #11Utility of dynamic range beyond a certain point
I agree. Most of the last 2 or 3 generations of cameras have more than enough DR and raw color data that with proper raw processing you can pretty much make any file look extremely close to another, even considering manufacturer lens color cast differences...
Jun 16, 2026 at 02:27 PM
p.5 #12Utility of dynamic range beyond a certain point
ilkka_nissila wrote:
That's not really true as the full continuous spectrum of light striking the sensor is filtered by the color filter array and the spectra of different filters can be different in different implementations and this reduces the full spectrum of light into just three values. The three values will be different between the different CFAs, depending on the spectrum of the incoming light and the CFA spectra, and there is no way to un-filter the light after those CFAs to make a mechanistic algorithm that would reproduce the colors from one camera based on images from a different one. The...Show more →
While what you say is technically true, I do not think it is practically relevant. If you use appropriate color profiles, the colors (assuming typical continuous spectrum colors) recorded by different Bayer sensors can be mapped with an accuracy of 1∂e or better. Which means human eyes can't distinguish them. Where you do get different colors between sensors is when you do not use appropriate color profiles, if you happen to record monochromatic (or otherwise discontinuous) light or when the color is near (or beyond) the gamut limit.
That's not really true as the full continuous spectrum of light striking the sensor is filtered by the color filter array and the spectra of different filters can be different in different implementations and this reduces the full spectrum of light into just three values. The three values will be different between the different CFAs, depending on the spectrum of the incoming light and the CFA spectra, and there is no way to un-filter the light after those CFAs to make a mechanistic algorithm that would reproduce the colors from one camera based on images from a different one. The...Show more →
While what you say is technically true, I do not think it is practically relevant. If you use appropriate color profiles, the colors (assuming typical continuous spectrum colors) recorded by different Bayer sensors can be mapped with an accuracy of 1∂e or better. Which means human eyes can't distinguish them. Where you do get different colors between sensors is when you do not use appropriate color profiles, if you happen to record monochromatic (or otherwise discontinuous) light or when the color is near (or beyond) the gamut limit.
Jun 16, 2026 at 04:26 PM
p.5 #13Utility of dynamic range beyond a certain point
Do all camera have the same SMI, and if they don’t, can it be worked around solely with profiles?
Daran wrote:
ilkka_nissila wrote
That's not really true as the full continuous spectrum of light striking the sensor is filtered by the color filter array and the spectra of different filters can be different in different implementations and this reduces the full spectrum of light into just three values. The three values will be different between the different CFAs, depending on the spectrum of the incoming light and the CFA spectra, and there is no way to un-filter the light after those CFAs to make a mechanistic algorithm that would reproduce the colors from one camera based on images from a different one. The information is simply not there in the recorded file to make this happen.Jun 16, 2026 at 06:51 PM
p.5 #14Utility of dynamic range beyond a certain point
Daran wrote:
While what you say is technically true, I do not think it is practically relevant. If you use appropriate color profiles, the colors (assuming typical continuous spectrum colors) recorded by different Bayer sensors can be mapped with an accuracy of 1∂e or better. Which means human eyes can't distinguish them. Where you do get different colors between sensors is when you do not use appropriate color profiles, if you happen to record monochromatic (or otherwise discontinuous) light or when the color is near (or beyond) the gamut limit.
Daran, while I believe you (that the differences are not practically relevant), I am confused by the Cobalt Image "Camera Sensor Color Accuracy Benchmarks" where Nikon D3 is the top scorer (93 in "Overall Hardware" and 89.70 in "Sensor Separation"). The topmost-ranked Sony is A1II (75 in "Overall Hardware" and 72.5 in "Sensor Separation").
Are these differences negated by "use appropriate color profiles" as you suggested? That is, the "the sensor’s potential for accurate profiling" (using the wording from Cobalt) is of no practical concern?
The following is a quote from Cobalt on the meaning of their measurements:
"These measurements describe the colour behaviour of the sensor at its most fundamental level. Sensor Separation under D50 and StdA shows how effectively the camera distinguishes chromatic signals under daylight and tungsten illumination, highlighting its response across different spectral conditions. The Luther–Ives metric indicates how closely the sensor can approximate the human visual observer, while the Vora Index measures the similarity between the sensor’s native colour subspace and the CIE reference model. Together, these values provide an objective summary of colour discrimination, spectral coherence, and the sensor’s potential for accurate profiling."
ilkka_nissila wrote
That's not really true as the full continuous spectrum of light striking the sensor is filtered by the color filter array and the spectra of different filters can be different in different implementations and this reduces the full spectrum of light into just three values. The three values will be different between the different CFAs, depending on the spectrum of the incoming light and the CFA spectra, and there is no way to un-filter the light after those CFAs to make a mechanistic algorithm that would reproduce the colors from one camera based on images from a different one. The information is simply not there in the recorded file to make this happen.Daran, while I believe you (that the differences are not practically relevant), I am confused by the Cobalt Image "Camera Sensor Color Accuracy Benchmarks" where Nikon D3 is the top scorer (93 in "Overall Hardware" and 89.70 in "Sensor Separation"). The topmost-ranked Sony is A1II (75 in "Overall Hardware" and 72.5 in "Sensor Separation").
Are these differences negated by "use appropriate color profiles" as you suggested? That is, the "the sensor’s potential for accurate profiling" (using the wording from Cobalt) is of no practical concern?
The following is a quote from Cobalt on the meaning of their measurements:
"These measurements describe the colour behaviour of the sensor at its most fundamental level. Sensor Separation under D50 and StdA shows how effectively the camera distinguishes chromatic signals under daylight and tungsten illumination, highlighting its response across different spectral conditions. The Luther–Ives metric indicates how closely the sensor can approximate the human visual observer, while the Vora Index measures the similarity between the sensor’s native colour subspace and the CIE reference model. Together, these values provide an objective summary of colour discrimination, spectral coherence, and the sensor’s potential for accurate profiling."
Jun 16, 2026 at 08:17 PM
p.5 #15Utility of dynamic range beyond a certain point
ilkka_nissila wrote:
When I look at my favorite landscape shots, for example, I see that - while there's plenty of unused room to lift shadows - lifting them even further results in unrealistic "HDR" look. I think you all are familiar with that look. In other words, at least in my work, there's a natural limit of how much the luminance range of a scene can be "compressed" into an 8-bit JPEG (or into a 200:1 print) by lifting the shadows. Sure, there are 10-bit screens and true HDR HEIC, but that shit looks garish to me.
In other words, the Nikon Z8 and the Sony A7RVI offer the same practical dynamic rangne for my landscapes, and I can't see how I can use the extra 2EV of the Sony. Which brings my question: what makes you excited about the additional stop of DR at ISO 100, and where are you **practically** bumping into DR limits of your current camera?
P.S. I forgot to mention the highlights. On that front my experience is the opposite: comparing A7RV to cameras with 2EV less DR, I do not see much "unused room" there, and there aren't any meaningful differences in highlight recovery across models. I suspect that the metering systems in all cameras have gotten pretty reliable at nailing exposure within ~1EV of highlight clipping, which means you get 1EV of highlight recovery regardless of a camera.greg, have you never blown out the sky, lace details on a white dress or feather detail on a white bird before?
I have had to do sky replacement "first aid" on images beforeThat would be an exposure issue and the reduced shadow noise (in the A7RVI compared, e.g., with the A1II), which is what the extra DR is all about, is not likely to help much with that. You can choose to give the subject less exposure and by doing this you can avoid the highlight clipping but also whenever you give yourself more room in the highlights for brighter spots to be correctly imaged, you increase noise in the darker parts of the image due to increased photon shot noise. The differences in DR are likely to be visible in the...Show more →
There seems to be something contradictory in "You can choose to give the subject less exposure and by doing this ... you increase noise in the darker parts of the image due to increased photon shot noise."
Statistically, the shot noise increases as the square root of the signal. E.g., for a signal of 100 electrons, the shot noise is 10, and for a signal of 10000, the shot noise is 100.
The absolute value of the shot noise is not important (except in the highlights), because the (engineering or photographic) signal-to-noise ratio increases with more light despite the concomitant increase in the shot noise. E.g., quadrupling the amount of light captured by a sensor gives a two-fold increase in the signal-to-noise ratio.
I agree with "The differences in DR are likely to be visible in the darkest shadows at low ISO and it's not clear they would make a difference in most practical use cases" with reference to the A7RVI. One practical question about the A7RVI is in what tonal region the noise (that must be the shot & read noise in the signal prior to the analogue-to-digital conversion) is reduced?
If we divide the region below midtones into (a) Light shadows, (b) Deep shadows, (c) Extreme shadows, I wonder which one(s) should have have reduced noise due to the dual gain? You use the expression "the darkest shadows" that probably corresponds to my "Extreme shadows", however, I don't have the knowledge of how dual gain is implemented in the A7RVI to make any concrete statement.
aCuria wrote
old-gregg wrote
Everyone is excited about the increased dynamic range of the Sony A7RVI at nominal ISO. Meanwhile, I am asking myself whether I ever bump into the DR limitations of my existing cameras in practice?When I look at my favorite landscape shots, for example, I see that - while there's plenty of unused room to lift shadows - lifting them even further results in unrealistic "HDR" look. I think you all are familiar with that look. In other words, at least in my work, there's a natural limit of how much the luminance range of a scene can be "compressed" into an 8-bit JPEG (or into a 200:1 print) by lifting the shadows. Sure, there are 10-bit screens and true HDR HEIC, but that shit looks garish to me.
In other words, the Nikon Z8 and the Sony A7RVI offer the same practical dynamic rangne for my landscapes, and I can't see how I can use the extra 2EV of the Sony. Which brings my question: what makes you excited about the additional stop of DR at ISO 100, and where are you **practically** bumping into DR limits of your current camera?
P.S. I forgot to mention the highlights. On that front my experience is the opposite: comparing A7RV to cameras with 2EV less DR, I do not see much "unused room" there, and there aren't any meaningful differences in highlight recovery across models. I suspect that the metering systems in all cameras have gotten pretty reliable at nailing exposure within ~1EV of highlight clipping, which means you get 1EV of highlight recovery regardless of a camera.
I have had to do sky replacement "first aid" on images before
There seems to be something contradictory in "You can choose to give the subject less exposure and by doing this ... you increase noise in the darker parts of the image due to increased photon shot noise."
Statistically, the shot noise increases as the square root of the signal. E.g., for a signal of 100 electrons, the shot noise is 10, and for a signal of 10000, the shot noise is 100.
The absolute value of the shot noise is not important (except in the highlights), because the (engineering or photographic) signal-to-noise ratio increases with more light despite the concomitant increase in the shot noise. E.g., quadrupling the amount of light captured by a sensor gives a two-fold increase in the signal-to-noise ratio.
I agree with "The differences in DR are likely to be visible in the darkest shadows at low ISO and it's not clear they would make a difference in most practical use cases" with reference to the A7RVI. One practical question about the A7RVI is in what tonal region the noise (that must be the shot & read noise in the signal prior to the analogue-to-digital conversion) is reduced?
If we divide the region below midtones into (a) Light shadows, (b) Deep shadows, (c) Extreme shadows, I wonder which one(s) should have have reduced noise due to the dual gain? You use the expression "the darkest shadows" that probably corresponds to my "Extreme shadows", however, I don't have the knowledge of how dual gain is implemented in the A7RVI to make any concrete statement.
Jun 16, 2026 at 09:06 PM
p.5 #16Utility of dynamic range beyond a certain point
old-gregg wrote:
I have had to do sky replacement "first aid" on images beforeI added a clarification at the bottom, sharing my observation that all modern cameras, regardless of what photonstophotos shows, have about the same headroom for highlights. Probably the result of metering+sensor pair tuning. I guess what you're saying is these newer DGO cameras allow for more negative exposure compensation with no shadow noise penalty to protect the highlights even further. That's a great answer!
This is qualitatively true that "all modern cameras... have about the same headroom for highlights," although your use of the "headroom" is somewhat ambiguous (this can be the headroom above the middle grey - an 18% grey exposure - and sensor saturation, or the highlight headroom as determined by the manufacturer's ISO calibration, to avoid clipping).
Nevertheless, the "headroom" and the dynamic range are different quantities and should not be conflated. Indeed, two cameras can have ca. 3.3 stops of headroom above the middle grey but one of the two can have a better dynamic range and, thus, cleaner shadows.
This whole subject is possibly confusing because the dynamic range is calculated from the signal-to-noise ratio and, therefore, can be manipulated on both sides, by changing either the signal or the noise, or both. Thus, a "better dynamic range" can mean two completely different things: an increased well capacity or a reduced noise. Both are useful in their different ways. However, those who argue that the dynamic range of modern cameras is fully adequate and no further increase should be of a meaningful practical significance (to them) might want to be more specific as to whether they are happy about the existing low level of noise in their photos, or if they already have no problem capturing brightly-lit scenes with strong highlights.
You gave a general title to this thread: "Utility of dynamic range beyond a certain point", rather than a more specific one, for example "Utility of the extended dynamic range of the A7RVI". Therefore, my posts in this thread have been about the utility of the dynamic range in the broader sense. I do value the ability to capture more light and achieve cleaner shadows with the help of the increased, this way, signal-to-noise ratio, while shooting at or close to the base ISO. Naturally, this is a personal preference. I well understand that many photographers, e.g. the birders relentlessly pursuing birds in flight, simply cannot take advantage of a high dynamic range of a camera when they shoot at ISO values way above the base. Thus, depending on the ISO used by a photographer, the "Utility of dynamic range beyond a certain point" can be absolutely NONE.
aCuria wrote
greg, have you never blown out the sky, lace details on a white dress or feather detail on a white bird before?I have had to do sky replacement "first aid" on images before
This is qualitatively true that "all modern cameras... have about the same headroom for highlights," although your use of the "headroom" is somewhat ambiguous (this can be the headroom above the middle grey - an 18% grey exposure - and sensor saturation, or the highlight headroom as determined by the manufacturer's ISO calibration, to avoid clipping).
Nevertheless, the "headroom" and the dynamic range are different quantities and should not be conflated. Indeed, two cameras can have ca. 3.3 stops of headroom above the middle grey but one of the two can have a better dynamic range and, thus, cleaner shadows.
This whole subject is possibly confusing because the dynamic range is calculated from the signal-to-noise ratio and, therefore, can be manipulated on both sides, by changing either the signal or the noise, or both. Thus, a "better dynamic range" can mean two completely different things: an increased well capacity or a reduced noise. Both are useful in their different ways. However, those who argue that the dynamic range of modern cameras is fully adequate and no further increase should be of a meaningful practical significance (to them) might want to be more specific as to whether they are happy about the existing low level of noise in their photos, or if they already have no problem capturing brightly-lit scenes with strong highlights.
You gave a general title to this thread: "Utility of dynamic range beyond a certain point", rather than a more specific one, for example "Utility of the extended dynamic range of the A7RVI". Therefore, my posts in this thread have been about the utility of the dynamic range in the broader sense. I do value the ability to capture more light and achieve cleaner shadows with the help of the increased, this way, signal-to-noise ratio, while shooting at or close to the base ISO. Naturally, this is a personal preference. I well understand that many photographers, e.g. the birders relentlessly pursuing birds in flight, simply cannot take advantage of a high dynamic range of a camera when they shoot at ISO values way above the base. Thus, depending on the ISO used by a photographer, the "Utility of dynamic range beyond a certain point" can be absolutely NONE.
Jun 16, 2026 at 09:42 PM
p.5 #17Utility of dynamic range beyond a certain point
old-gregg wrote:
I have had to do sky replacement "first aid" on images beforeI added a clarification at the bottom, sharing my observation that all modern cameras, regardless of what photonstophotos shows, have about the same headroom for highlights. Probably the result of metering+sensor pair tuning. I guess what you're saying is these newer DGO cameras allow for more negative exposure compensation with no shadow noise penalty to protect the highlights even further. That's a great answer!
CRW_9943-SDR by acurian, on Flickr
Greg, you can try taking this kind of shot and see what results you get.
What I would want is for the sky to be blue and the inside of the building to be retain detail.
CRW_9948-SDR by acurian, on Flickr
This is what the sky should have looked like, was taken a few frames after.
DSC02718-SDR-noedit by acurian, on Flickr
Backlit shots like this is also a problem. The above image has no editing, you can see that the subject was already under exposed in an attempt to preserve highlights, however the highlights are still blown out.
DSC02718-SDR by acurian, on Flickr
This is the same shot with some simple editing, highlights are -100 in lightroom
DSC02716-SDR by acurian, on Flickr
This is what the sky should have looked like in the squirrel shot, it was taken immediately before the squirrel
To me, having more dynamic range is generally helpful. We are still a long way off from what black and white film was capable of, which is about 15 stops
aCuria wrote
greg, have you never blown out the sky, lace details on a white dress or feather detail on a white bird before?I have had to do sky replacement "first aid" on images before
CRW_9943-SDR by acurian, on FlickrGreg, you can try taking this kind of shot and see what results you get.
What I would want is for the sky to be blue and the inside of the building to be retain detail.
CRW_9948-SDR by acurian, on FlickrThis is what the sky should have looked like, was taken a few frames after.
DSC02718-SDR-noedit by acurian, on FlickrBacklit shots like this is also a problem. The above image has no editing, you can see that the subject was already under exposed in an attempt to preserve highlights, however the highlights are still blown out.
DSC02718-SDR by acurian, on FlickrThis is the same shot with some simple editing, highlights are -100 in lightroom
DSC02716-SDR by acurian, on FlickrThis is what the sky should have looked like in the squirrel shot, it was taken immediately before the squirrel
To me, having more dynamic range is generally helpful. We are still a long way off from what black and white film was capable of, which is about 15 stops
Jun 16, 2026 at 11:45 PM
p.5 #18Utility of dynamic range beyond a certain point
aCuria wrote:
I have had to do sky replacement "first aid" on images beforeI added a clarification at the bottom, sharing my observation that all modern cameras, regardless of what photonstophotos shows, have about the same headroom for highlights. Probably the result of metering+sensor pair tuning. I guess what you're saying is these newer DGO cameras allow for more negative exposure compensation with no shadow noise penalty to protect the highlights even further. That's a great answer!https://live.staticflickr.com/65535/55339165627_6cd2763a34_b.jpgCRW_9943-SDR by acurian, on Flickr
Greg, you can try taking this kind of shot and see what results you get.
What I would want is for the sky to be blue and the inside of the building to be retain detail.
https://live.staticflickr.com/65535/55339189807_04104e2b07_b.jpgCRW_9948-SDR by acurian, on Flickr
This is what the sky should have looked like, was taken a few frames after.
https://live.staticflickr.com/65535/55340516420_782e85f863_b.jpgDSC02718-SDR-noedit by acurian, on Flickr
Backlit shots like this is also a problem. The above image has no editing, you can see that the subject was already under exposed in an attempt to preserve highlights, however the highlights are still blown out.
https://live.staticflickr.com/65535/55340248108_70cf601d60_b.jpgDSC02718-SDR by acurian,...Show more →
Haveing more experience is better than throwing money for extra DR, you could have shot a better image with a 20 yo DSLR with 8 stops of DR and popped a little fill flash on the squirrel. 🙄
old-gregg wrote
aCuria wrote
greg, have you never blown out the sky, lace details on a white dress or feather detail on a white bird before?I have had to do sky replacement "first aid" on images before
Greg, you can try taking this kind of shot and see what results you get.
What I would want is for the sky to be blue and the inside of the building to be retain detail.
https://live.staticflickr.com/65535/55339189807_04104e2b07_b.jpgCRW_9948-SDR by acurian, on Flickr
This is what the sky should have looked like, was taken a few frames after.
https://live.staticflickr.com/65535/55340516420_782e85f863_b.jpgDSC02718-SDR-noedit by acurian, on Flickr
Backlit shots like this is also a problem. The above image has no editing, you can see that the subject was already under exposed in an attempt to preserve highlights, however the highlights are still blown out.
https://live.staticflickr.com/65535/55340248108_70cf601d60_b.jpgDSC02718-SDR by acurian,...Show more →
Haveing more experience is better than throwing money for extra DR, you could have shot a better image with a 20 yo DSLR with 8 stops of DR and popped a little fill flash on the squirrel. 🙄
Jun 17, 2026 at 01:21 AM
p.5 #19Utility of dynamic range beyond a certain point
A74me wrote:
I have had to do sky replacement "first aid" on images beforeI added a clarification at the bottom, sharing my observation that all modern cameras, regardless of what photonstophotos shows, have about the same headroom for highlights. Probably the result of metering+sensor pair tuning. I guess what you're saying is these newer DGO cameras allow for more negative exposure compensation with no shadow noise penalty to protect the highlights even further. That's a great answer!https://live.staticflickr.com/65535/55339165627_6cd2763a34_b.jpgCRW_9943-SDR by acurian, on Flickr
Greg, you can try taking this kind of shot and see what results you get.
What I would want is for the sky to be blue and the inside of the building to be retain detail.
https://live.staticflickr.com/65535/55339189807_04104e2b07_b.jpgCRW_9948-SDR by acurian, on Flickr
This is what the sky should have looked like, was taken a few frames after.
https://live.staticflickr.com/65535/55340516420_782e85f863_b.jpgDSC02718-SDR-noedit by acurian, on Flickr
Backlit shots like this is also a problem. The above image has no editing, you can see that the subject was already under exposed in an attempt to preserve highlights, however the highlights are still blown out.
https://live.staticflickr.com/65535/55340248108_70cf601d60_b.jpgDSC02718-SDR by acurian, on Flickr
This is the same shot with some simple editing, highlights are -100 in lightroom
https://live.staticflickr.com/65535/55339172402_f4a8a4b88e_b.jpgDSC02716-SDR by acurian, on Flickr
This is what the sky should have looked like in the squirrel shot, it was taken immediately before the squirrel
To me, having more dynamic range is generally helpful. We are still a long way off from what black and white film was capable of, which is about 15 stopsHaveing more experience is better than throwing money for extra DR, you could have shot a better image with a 20 yo DSLR with 8 stops of DR and popped a little fill flash on the squirrel. 🙄
To be fair the first picture was also shot like 20 years ago. 8 stops of DR on that APSC body is equivalent to a Z6iii at iso 640, and iso 640 is well within the scope of this discussion
Did not have the fill flash with me, i did ask what gear to bring and nobody suggested flash units for DC haha.
Flash is also banned at some wildlife conservatories, Its not a solution that works for everything.
aCuria wrote
old-gregg wrote
aCuria wrote
greg, have you never blown out the sky, lace details on a white dress or feather detail on a white bird before?I have had to do sky replacement "first aid" on images before
Greg, you can try taking this kind of shot and see what results you get.
What I would want is for the sky to be blue and the inside of the building to be retain detail.
https://live.staticflickr.com/65535/55339189807_04104e2b07_b.jpgCRW_9948-SDR by acurian, on Flickr
This is what the sky should have looked like, was taken a few frames after.
https://live.staticflickr.com/65535/55340516420_782e85f863_b.jpgDSC02718-SDR-noedit by acurian, on Flickr
Backlit shots like this is also a problem. The above image has no editing, you can see that the subject was already under exposed in an attempt to preserve highlights, however the highlights are still blown out.
https://live.staticflickr.com/65535/55340248108_70cf601d60_b.jpgDSC02718-SDR by acurian, on Flickr
This is the same shot with some simple editing, highlights are -100 in lightroom
https://live.staticflickr.com/65535/55339172402_f4a8a4b88e_b.jpgDSC02716-SDR by acurian, on Flickr
This is what the sky should have looked like in the squirrel shot, it was taken immediately before the squirrel
To me, having more dynamic range is generally helpful. We are still a long way off from what black and white film was capable of, which is about 15 stops
To be fair the first picture was also shot like 20 years ago. 8 stops of DR on that APSC body is equivalent to a Z6iii at iso 640, and iso 640 is well within the scope of this discussion
Did not have the fill flash with me, i did ask what gear to bring and nobody suggested flash units for DC haha.
Flash is also banned at some wildlife conservatories, Its not a solution that works for everything.
Jun 17, 2026 at 05:00 AM
p.5 #20Utility of dynamic range beyond a certain point
tsdevine wrote:
Do all camera have the same SMI, and if they don’t, can it be worked around solely with profiles?
While what you say is technically true, I do not think it is practically relevant. If you use appropriate color profiles, the colors (assuming typical continuous spectrum colors) recorded by different Bayer sensors can be mapped with an accuracy of 1∂e or better. Which means human eyes can't distinguish them. Where you do get different colors between sensors is when you do not use appropriate color profiles, if you happen to record monochromatic (or otherwise discontinuous) light or when the color is near (or beyond) the gamut limit.
No, they don't. But it doesn't matter as much as one might think. To put this in perspective: while Bayer sensors and the human eye both use three color sensitivities in the same frequency range, they are quite different from each other, complicating the mapping. Much more different than the differences among modern Bayer sensors. And even this big difference can obviously be compensated very well, as long as the light spectrum is continuous (hence the CRI quality signifier of lights made for photography).
The reason sensors are so different from human sensitivities is in order to more evenly space them. This is more similar to how these sensitivities are spread in birds (which is perfectly even, they can do four colors and for *much* longer). In terms of biological development the human eye having three colors is a comparatively recent development and still vastly inferior (too much overlap) compared to birds. If mapping Bayer colors would be a significant issue, sensors could be made that mimic the human sensitivity more closely. But this is obviously not necessary, Bayer colors are good enough as they are.
Do all camera have the same SMI, and if they don’t, can it be worked around solely with profiles?
Daran wrote
ilkka_nissila wrote
That's not really true as the full continuous spectrum of light striking the sensor is filtered by the color filter array and the spectra of different filters can be different in different implementations and this reduces the full spectrum of light into just three values. The three values will be different between the different CFAs, depending on the spectrum of the incoming light and the CFA spectra, and there is no way to un-filter the light after those CFAs to make a mechanistic algorithm that would reproduce the colors from one camera based on images from a different one. The information is simply not there in the recorded file to make this happen.No, they don't. But it doesn't matter as much as one might think. To put this in perspective: while Bayer sensors and the human eye both use three color sensitivities in the same frequency range, they are quite different from each other, complicating the mapping. Much more different than the differences among modern Bayer sensors. And even this big difference can obviously be compensated very well, as long as the light spectrum is continuous (hence the CRI quality signifier of lights made for photography).
The reason sensors are so different from human sensitivities is in order to more evenly space them. This is more similar to how these sensitivities are spread in birds (which is perfectly even, they can do four colors and for *much* longer). In terms of biological development the human eye having three colors is a comparatively recent development and still vastly inferior (too much overlap) compared to birds. If mapping Bayer colors would be a significant issue, sensors could be made that mimic the human sensitivity more closely. But this is obviously not necessary, Bayer colors are good enough as they are.
Jun 17, 2026 at 05:23 AM