2nd A7R6 or an A1ii?
p.2 #12nd A7R6 or an A1ii?
old-gregg wrote:
Don't be dismissive because you only shoot RAW. Your camera's metering intentionally slightly under-exposes because the dumb JPEG engine simply ignores the upper bits of highlights, even if JPEGs are turned off. This means that most cameras can benefit from a permanent exposure compensation of +0.5 to +1EV. This is definitely the case with Sonys. With some scenes you need +1.5EV.
Sorry, but I have to ask - are you intentionally trying to make people overexpose their RAWs so that their DR is as limited as jpegs?
Don't be dismissive because you only shoot RAW. Your camera's metering intentionally slightly under-exposes because the dumb JPEG engine simply ignores the upper bits of highlights, even if JPEGs are turned off. This means that most cameras can benefit from a permanent exposure compensation of +0.5 to +1EV. This is definitely the case with Sonys. With some scenes you need +1.5EV.
Sorry, but I have to ask - are you intentionally trying to make people overexpose their RAWs so that their DR is as limited as jpegs?
Aug 31, 2026 at 04:36 PM
p.2 #22nd A7R6 or an A1ii?
old-gregg wrote:
Dan, you're poking at the core here. 100% relevant. Remember the film days? Anyone familiar with the zone system intuitively understood the point of compression and expansion: the contrast ratio (what we call DR now) across media weren't the same, and was always (!) far narrower than the real world. And photographers carefully managed the zones from a scene (14EV+) to a negative (10EV), and finally to a print (6EV). What I am saying is the JPEG engines in modern cameras are horrible, horrible printers! They behave as if they're "printing" from a transparency onto a fixed-grade paper.
Meanwhile, the smartphone...Show more →
This gets at yet another set of issues.
All of those measures you write about from the film/transparency era that ensured that final images fit within the DR of the reproduction system (from print to publication to...?) required some intelligent input from the photographer or a film post-processing (e.g. darkroom) specialist.
The world certainly contains much wider DR than jpg can reproduce, just as the data collected by a 14 stop DR sensor does. But this issue isn't really new. A painting also compresses the DR of the real world into a much smaller space.
Part of that is due to the inherent limitations of reproduction systems — paint, prints, printing presses, etc. But an even bigger part has to do with our our human vision system deals with the very wide DR found in the physical world. Our vision is highly adaptive. We don't actually see the entire scene in from t of us as an immediate whole. Instead, we create a kind of mental analog of the scene by scanning it with our eyes. And as we do this we adapt to local luminosity levels without even thinking about it. (Though I'm thinking about it a LOT today, as I returned a few hours ago from an ophthalmology appointment at which my eyes were dilated!)
The reason that we can see detail in shadows and the brilliant sunshine shining on the top of a mountain peak is not because our eyes can "see" all of that in one go. Instead, we adapt. When we look at the bright peak our pupils close up (we stop down our lens) to accommodate that brightness, and when we shift our gaze into the deep shadows, assuming that we are interested enough in that part o the scene to try to see it, our pupils dilate and we see more detail.
Photographers cannot do that, and viewers don't really respond quite that way to dynamic contrasts in photographs. So we find ways to compress the real world's dynamic range into a much smaller space.
Part of the "problem" with the jpg format is that its dynamic space, while large enough to hold all of the range of a print (and more, I think), is not as large as that of the real world. And because it doesn't benefit from the intelligence that a good photographer brings to the process of compressing dynamic range (control thise highlights, bring up those shadows) we are stuck with what the automatic process in the camera does, which is a sort of best guess.
JPG is a fine display format still, just as a print or a painting is.
gdanmitchell wrote
I’m not certain this is responsive to your point, but one issue is that while sensors have improved a great deal and can record higher dynamic ranges… the jpg format hasn’t really changed. It has always had a much more limited DR, and today it falls even further behind what sensors are capable of.Meanwhile, the smartphone...Show more →
This gets at yet another set of issues.
All of those measures you write about from the film/transparency era that ensured that final images fit within the DR of the reproduction system (from print to publication to...?) required some intelligent input from the photographer or a film post-processing (e.g. darkroom) specialist.
The world certainly contains much wider DR than jpg can reproduce, just as the data collected by a 14 stop DR sensor does. But this issue isn't really new. A painting also compresses the DR of the real world into a much smaller space.
Part of that is due to the inherent limitations of reproduction systems — paint, prints, printing presses, etc. But an even bigger part has to do with our our human vision system deals with the very wide DR found in the physical world. Our vision is highly adaptive. We don't actually see the entire scene in from t of us as an immediate whole. Instead, we create a kind of mental analog of the scene by scanning it with our eyes. And as we do this we adapt to local luminosity levels without even thinking about it. (Though I'm thinking about it a LOT today, as I returned a few hours ago from an ophthalmology appointment at which my eyes were dilated!)
The reason that we can see detail in shadows and the brilliant sunshine shining on the top of a mountain peak is not because our eyes can "see" all of that in one go. Instead, we adapt. When we look at the bright peak our pupils close up (we stop down our lens) to accommodate that brightness, and when we shift our gaze into the deep shadows, assuming that we are interested enough in that part o the scene to try to see it, our pupils dilate and we see more detail.
Photographers cannot do that, and viewers don't really respond quite that way to dynamic contrasts in photographs. So we find ways to compress the real world's dynamic range into a much smaller space.
Part of the "problem" with the jpg format is that its dynamic space, while large enough to hold all of the range of a print (and more, I think), is not as large as that of the real world. And because it doesn't benefit from the intelligence that a good photographer brings to the process of compressing dynamic range (control thise highlights, bring up those shadows) we are stuck with what the automatic process in the camera does, which is a sort of best guess.
JPG is a fine display format still, just as a print or a painting is.
Aug 31, 2026 at 05:59 PM
p.2 #32nd A7R6 or an A1ii?
j4nu wrote:
Sorry, but I have to ask - are you intentionally trying to make people overexpose their RAWs so that their DR is as limited as jpegs?
What people? Sorry, but I am not following you at all here. I am not "trying", I am simply telling you that your in-camera JPEG engine clips highlights for no reason other than laziness/incmpetence of its programmers. To compensate for this, your in-camera meter intentionally under-exposes most of the time.
@gdanmitchell, JPEG as a final destination medium is fine. In fact, it is a luxury: JPEG can hold more information (DR and gamut) than all papers and most monitors are capable of. The problem isn't the medium, the problem is that the JPEG engine in the A7RVI is the same tech as in the Canon 1D, as if 20 years of progress between them didn't happen.
Sorry, but I have to ask - are you intentionally trying to make people overexpose their RAWs so that their DR is as limited as jpegs?
What people? Sorry, but I am not following you at all here. I am not "trying", I am simply telling you that your in-camera JPEG engine clips highlights for no reason other than laziness/incmpetence of its programmers. To compensate for this, your in-camera meter intentionally under-exposes most of the time.
@gdanmitchell, JPEG as a final destination medium is fine. In fact, it is a luxury: JPEG can hold more information (DR and gamut) than all papers and most monitors are capable of. The problem isn't the medium, the problem is that the JPEG engine in the A7RVI is the same tech as in the Canon 1D, as if 20 years of progress between them didn't happen.
Aug 31, 2026 at 06:16 PM
p.2 #42nd A7R6 or an A1ii?
old-gregg wrote:
What people? Sorry, but I am not following you at all here. I am not "trying", I am simply telling you that your in-camera JPEG engine clips highlights for no reason other than laziness/incmpetence of its programmers. To compensate for this, your in-camera meter intentionally under-exposes most of the time.
Are you saying that the camera underexposes by default, when looking at zebras? And about which metering are we talking about?
Your advise only makes sense when using highlight-metering, and even then 1-1.5EV might not be enough.
In general the camera should underexpose a bit, because that's what gives you more available DR to work with in post in most cases...
j4nu wrote
Sorry, but I have to ask - are you intentionally trying to make people overexpose their RAWs so that their DR is as limited as jpegs?Are you saying that the camera underexposes by default, when looking at zebras? And about which metering are we talking about?
Your advise only makes sense when using highlight-metering, and even then 1-1.5EV might not be enough.
In general the camera should underexpose a bit, because that's what gives you more available DR to work with in post in most cases...
Aug 31, 2026 at 06:24 PM
p.2 #52nd A7R6 or an A1ii?
j4nu wrote:
In general the camera should underexpose a bit, because that's what gives you more available DR to work with in post in most cases...
This is either incomplete or incorrect, but since I don't know your assumptions, I can only guess. You are right if you're saying that a camera should avoid highlight clipping in RAW data at all costs. If you are calling this "underexpose a bit", fine.
But this "at all costs" approach is dumb by 2026 standards, because, surprisingly, it reduces the useful dynamic range. Apologies if you're already familiar with Bill Claff's concept of PDA (photographic dynamic range). But if not, the basic idea is that not all shadows are created equal, and above certain noise level, those low values aren't useful. Modern digital cameras are overly aggressive at highlight protection, leaving a lot of sensor PDA on the table.
Putting PDA aside for a second, consider the following thought experiment. We all know that dynamic range is the ratio of the brightest and the darkest pixels in an image. Now imagine a photo with 1 million pixels. One of them, just one, has the highest brightness of 15 (units don't matter). All others are under 5, and about 20% of them have a value of 1. Assume that your sensor has a DR of 14. You need to select the exposure, knowing that you can't capture the entire brightness range with DR of 14. What would be better: optimize for that one lone brightest pixel and clip 20% of the image, or capture everything and clip just that one pixel?
The modern cameras lean towards the former. Back in the day they lacked processing power to be intelligent about it, but it is no longer an acceptable excuse in 2026. Now they have all the computing power they need, plus AI, to literally count pixels and make more intelligent decisions when metering. Just like an iPhone.
Lloyd Chambers goes into this in more detail here:
https://diglloyd.com/prem/prot/ML/NikonZ/NikonZ8-metering-ETTR.html
It is paywalled, but his conclusion is similar, quoting: "As of 2023, no vendor has offered a metering method appropriate for RAW capture with digital—strange as that sounds. They’re all using improved versions of metering for film circa 1979."
To see for yourself, download RawDigger and examine various RAW files, look at the histogram and how much unused room there's usually on the right.
In general the camera should underexpose a bit, because that's what gives you more available DR to work with in post in most cases...
This is either incomplete or incorrect, but since I don't know your assumptions, I can only guess. You are right if you're saying that a camera should avoid highlight clipping in RAW data at all costs. If you are calling this "underexpose a bit", fine.
But this "at all costs" approach is dumb by 2026 standards, because, surprisingly, it reduces the useful dynamic range. Apologies if you're already familiar with Bill Claff's concept of PDA (photographic dynamic range). But if not, the basic idea is that not all shadows are created equal, and above certain noise level, those low values aren't useful. Modern digital cameras are overly aggressive at highlight protection, leaving a lot of sensor PDA on the table.
Putting PDA aside for a second, consider the following thought experiment. We all know that dynamic range is the ratio of the brightest and the darkest pixels in an image. Now imagine a photo with 1 million pixels. One of them, just one, has the highest brightness of 15 (units don't matter). All others are under 5, and about 20% of them have a value of 1. Assume that your sensor has a DR of 14. You need to select the exposure, knowing that you can't capture the entire brightness range with DR of 14. What would be better: optimize for that one lone brightest pixel and clip 20% of the image, or capture everything and clip just that one pixel?
The modern cameras lean towards the former. Back in the day they lacked processing power to be intelligent about it, but it is no longer an acceptable excuse in 2026. Now they have all the computing power they need, plus AI, to literally count pixels and make more intelligent decisions when metering. Just like an iPhone.
Lloyd Chambers goes into this in more detail here:
https://diglloyd.com/prem/prot/ML/NikonZ/NikonZ8-metering-ETTR.html
It is paywalled, but his conclusion is similar, quoting: "As of 2023, no vendor has offered a metering method appropriate for RAW capture with digital—strange as that sounds. They’re all using improved versions of metering for film circa 1979."
To see for yourself, download RawDigger and examine various RAW files, look at the histogram and how much unused room there's usually on the right.
Aug 31, 2026 at 07:33 PM
p.2 #62nd A7R6 or an A1ii?
old-gregg wrote:
@gdanmitchell@, JPEG as a final destination medium is fine. In fact, it is a luxury: JPEG can hold more information (DR and gamut) than all papers and most monitors are capable of. The problem isn't the medium, the problem is that the JPEG engine in the A7RVI is the same tech as in the Canon 1D, as if 20 years of progress between them didn't happen.
To some extent I'm not as concerned about this as you are since I prefer to shoot raw and get the full benefit of the camera's capabilities.
On the other hand, I suspect that before long we'll start to see mainline camera companies develop (or buy!) software like that used on smartphones, which does all kinds of "intelligent" things with the raw capture before feeding us the final result.
Regarding exposure, the idea usually (with room for some exceptions) is to protect the highlights even if that means lowering the original exposure and then bringing back dark tones later. In digital capture, once lost, highlight tones over the maximum value that can be recorded cannot be reclaimed. They are gone. ON the other hand, you can recover shadows quite effectively — you may have to apply NR but that's not nearly the problem it used to be.
Let me give you a recent example of this. Last week I was photographing in the Sierra Nevada in the Yosemite high country. I found a forested hill with very strong backlight, and I wanted to photograph the trunks of the trees with some of that bright light coming between them.
This is the photograph that I had in mind, more or less, as I made the exposure:
Evening Forest and Hill
But the dynamic range in the actual scene was huge. About 1/3 of the way down from the top is almost direct light from the sun, only slightly muted by some branches. In the foreground are very dark shadows, not only because they are on the dark side of the trees but additionally because they are inside the forest.
So I knowing that I had to protect those highlights and that I could recover the shadows in post, I used an exposure that put the brightest tones right around the level at which they would blow out. That left the dark tones looking essentially black — which was intentional. Here's a copy of the image with no adjustment applied:
UnprocessedOriginal
By the way, while you can never trust these things absolutely — actual scenes and images don't always fit neatly within boundaries — the camera's histogram display is your friend.
Edited on Aug 31, 2026 at 08:39 PM · View previous versions
@gdanmitchell@, JPEG as a final destination medium is fine. In fact, it is a luxury: JPEG can hold more information (DR and gamut) than all papers and most monitors are capable of. The problem isn't the medium, the problem is that the JPEG engine in the A7RVI is the same tech as in the Canon 1D, as if 20 years of progress between them didn't happen.
To some extent I'm not as concerned about this as you are since I prefer to shoot raw and get the full benefit of the camera's capabilities.
On the other hand, I suspect that before long we'll start to see mainline camera companies develop (or buy!) software like that used on smartphones, which does all kinds of "intelligent" things with the raw capture before feeding us the final result.
Regarding exposure, the idea usually (with room for some exceptions) is to protect the highlights even if that means lowering the original exposure and then bringing back dark tones later. In digital capture, once lost, highlight tones over the maximum value that can be recorded cannot be reclaimed. They are gone. ON the other hand, you can recover shadows quite effectively — you may have to apply NR but that's not nearly the problem it used to be.
Let me give you a recent example of this. Last week I was photographing in the Sierra Nevada in the Yosemite high country. I found a forested hill with very strong backlight, and I wanted to photograph the trunks of the trees with some of that bright light coming between them.
This is the photograph that I had in mind, more or less, as I made the exposure:
Evening Forest and Hill
But the dynamic range in the actual scene was huge. About 1/3 of the way down from the top is almost direct light from the sun, only slightly muted by some branches. In the foreground are very dark shadows, not only because they are on the dark side of the trees but additionally because they are inside the forest.
So I knowing that I had to protect those highlights and that I could recover the shadows in post, I used an exposure that put the brightest tones right around the level at which they would blow out. That left the dark tones looking essentially black — which was intentional. Here's a copy of the image with no adjustment applied:
UnprocessedOriginal
By the way, while you can never trust these things absolutely — actual scenes and images don't always fit neatly within boundaries — the camera's histogram display is your friend.
Edited on Aug 31, 2026 at 08:39 PM · View previous versions
Aug 31, 2026 at 08:18 PM
p.2 #72nd A7R6 or an A1ii?
Wow, this discussion sure took a turn down a bright alley. It never occurred to me that the meter was responsible for providing the correct exposure. Now that you point out that with the computational power of the modern camera that the typical 7-14 stops of real world dynamic range.. assuming that you aren’t shooting into the sun… the camera should be able to capture the full dynamic range of a typical daylight scene. Being old, outdated and the like, I have viewed my internal exposure meter to be much like the evaluative meters at the end of the film era, where it is up to me to decide what to emphasize and knowing my meter I can productively ride the exposure compensation dial. Sure, I can change the meter bias by changing the metering mode. I might even pull out my external incident, or spot meter. It seems to me that the histogram probably remains one of the best exposure tools. I have to admit, the zebra patterns just irritates me. Also, I’ve never used JPEG, but the HDR output of modern cameras might just change my opinion of lossy formats. The point taken above is that outmoded computational metering propagated by lazy programmers just might make RAW lossy with respect to the real world.
Aug 31, 2026 at 08:21 PM
p.2 #82nd A7R6 or an A1ii?
old-gregg wrote:
Don't be dismissive because you only shoot RAW. Your camera's metering intentionally slightly under-exposes because the dumb JPEG engine simply ignores the upper bits of highlights, even if JPEGs are turned off. This means that most cameras can benefit from a permanent exposure compensation of +0.5 to +1EV. This is definitely the case with Sonys. With some scenes you need +1.5EV.
I've never analyzed this in the way you are doing here, but I have always found it to be true that adding 0.5 to 1.0 stops of exposure compensation often improves the RAW file and the final image.
Don't be dismissive because you only shoot RAW. Your camera's metering intentionally slightly under-exposes because the dumb JPEG engine simply ignores the upper bits of highlights, even if JPEGs are turned off. This means that most cameras can benefit from a permanent exposure compensation of +0.5 to +1EV. This is definitely the case with Sonys. With some scenes you need +1.5EV.
I've never analyzed this in the way you are doing here, but I have always found it to be true that adding 0.5 to 1.0 stops of exposure compensation often improves the RAW file and the final image.
Aug 31, 2026 at 08:30 PM
p.2 #92nd A7R6 or an A1ii?
The jpeg preview in camera is ALWAYS tricking me. The extremes from the jpeg preview to unprocessed raw is too extreme imo. I feel like I have to work too hard in lightroom to get close to what I was seeing in the back of the camera... That's with all the Sony cameras I've used, and to a very slightly lesser extent, some of my Canons. Like I get the purpose of a raw, but wouldn't it make more sense to be able to see what your raw will look like in camera, rather than some cooked up jpeg? I'd like to have an option to toggle the jpeg look or the raw look. That would help me a lot. :shrug:
Aug 31, 2026 at 09:03 PM
p.2 #102nd A7R6 or an A1ii?
BigBabyMoses06 wrote:
The jpeg preview in camera is ALWAYS tricking me. The extremes from the jpeg preview to unprocessed raw is too extreme imo. I feel like I have to work too hard in lightroom to get close to what I was seeing in the back of the camera... That's with all the Sony cameras I've used, and to a very slightly lesser extent, some of my Canons. Like I get the purpose of a raw, but wouldn't it make more sense to be able to see what your raw will look like in camera, rather than some cooked up jpeg? I'd...Show more →
I’ve never used the EVF/LCD for anything other than framing/composition. Perhaps that comes from shooting rangefinder cameras for decades. If I wanted to judge exposure, then HDR mode and the X2D2’s LCD screen are the way to go. Unfortunately there are some ISO limitations imposed.
Aug 31, 2026 at 09:19 PM
|
p.2 #112nd A7R6 or an A1ii?
bwcolor wrote:
I’ve never used the EVF/LCD for anything other than framing/composition. Perhaps that comes from shooting rangefinder cameras for decades. If I wanted to judge exposure, then HDR mode and the X2D2’s LCD screen are the way to go. Unfortunately there are some ISO limitations imposed.
Quit tempting me
.
BigBabyMoses06 wrote
The jpeg preview in camera is ALWAYS tricking me. The extremes from the jpeg preview to unprocessed raw is too extreme imo. I feel like I have to work too hard in lightroom to get close to what I was seeing in the back of the camera... That's with all the Sony cameras I've used, and to a very slightly lesser extent, some of my Canons. Like I get the purpose of a raw, but wouldn't it make more sense to be able to see what your raw will look like in camera, rather than some cooked up jpeg? I'd like to have an option to toggle the jpeg look or the raw look. That would help me a lot. :shrug:Quit tempting me
. Aug 31, 2026 at 09:23 PM
p.2 #122nd A7R6 or an A1ii?
gdanmitchell wrote:
To some extent I'm not as concerned about this as you are since I prefer to shoot raw and get the full benefit of the camera's capabilities.
I am starting to suspect you haven't read my posts above fully. Again: it doesn't matter whether you shoot RAW or not. If your exposure is not optimal, you are not getting the benefit of the camera capabilities. It's about metering, not choosing the output format.
One more time: unless you are using exposure compensation or spot metering, your camera always under-exposes, which means you are literally losing data.
gdanmitchell wrote:
the camera's histogram display is your friend.
It is not your friend, it is your enemy. Your camera histogram is based on the output of the lobotomized JPEG engine, not the RAW sensor data. Why do you think its shape shrinks/stretches as you adjust exposure, instead of just shifting left/right? It should never be used to evaluate exposure if you shoot RAW.
RawDigger is your friend. You will be "pleased" to discover how many of your RAW files only used the bottom 10bits of the 14bit ADC output. I am convinced now that you are not getting the full benefit of your camera capabilities.
P.S. And by the way, all of this is just an engineer rambling about users under-utilizing their equipment because of poor quality software. In the real world, even the truncated output you're getting is apparently enough to produce stunning images, as your samples above demonstrated. I'm simply saying hey, there's even more potential!
To some extent I'm not as concerned about this as you are since I prefer to shoot raw and get the full benefit of the camera's capabilities.
I am starting to suspect you haven't read my posts above fully. Again: it doesn't matter whether you shoot RAW or not. If your exposure is not optimal, you are not getting the benefit of the camera capabilities. It's about metering, not choosing the output format.
One more time: unless you are using exposure compensation or spot metering, your camera always under-exposes, which means you are literally losing data.
gdanmitchell wrote:
the camera's histogram display is your friend.
It is not your friend, it is your enemy. Your camera histogram is based on the output of the lobotomized JPEG engine, not the RAW sensor data. Why do you think its shape shrinks/stretches as you adjust exposure, instead of just shifting left/right? It should never be used to evaluate exposure if you shoot RAW.
RawDigger is your friend. You will be "pleased" to discover how many of your RAW files only used the bottom 10bits of the 14bit ADC output. I am convinced now that you are not getting the full benefit of your camera capabilities.
P.S. And by the way, all of this is just an engineer rambling about users under-utilizing their equipment because of poor quality software. In the real world, even the truncated output you're getting is apparently enough to produce stunning images, as your samples above demonstrated. I'm simply saying hey, there's even more potential!
Aug 31, 2026 at 10:04 PM
p.2 #132nd A7R6 or an A1ii?
BigBabyMoses06 wrote:
The jpeg preview in camera is ALWAYS tricking me. The extremes from the jpeg preview to unprocessed raw is too extreme imo. I feel like I have to work too hard in lightroom to get close to what I was seeing in the back of the camera... That's with all the Sony cameras I've used, and to a very slightly lesser extent, some of my Canons. Like I get the purpose of a raw, but wouldn't it make more sense to be able to see what your raw will look like in camera, rather than some cooked up jpeg? I'd...Show more →
"What it is going to look like" is not as simple as it sounds. In particular if you have a very wide dynamic range scene the raw is like to look pretty awful and not much at all like what you actually saw with your eyes.
If you are shooting raw it is far better to rely on the histogram.
Or if you don't want to bother, just shoot jpg.
- - -
old-gregg wrote:
I am starting to suspect you haven't read my posts above fully. Again: it doesn't matter whether you shoot RAW or not. If your exposure is not optimal, you are not getting the benefit of the camera capabilities. It's about metering, not choosing the output format.
One more time: unless you are using exposure compensation or spot metering, your camera always under-exposes, which means you are literally losing data.
It is not your friend, it is your enemy. Your camera histogram is based on the output of the lobotomized JPEG engine, not the RAW sensor data. Why do you think its shape...Show more →
Regarding "always underexposes," it is true that you can push your exposure a bit beyond what the camera probably suggests. My approach (since I'm not carrying a computer into the field to download and check each exposure with rawdigger) is to understand that no rear panel histogram display is perfect. However, from using the camera and learning to understand the relationship between what the histogram says, what the image displayed on the screen tells me, and what my experience with different subject colors and so on tells me... the histogram provides the most accurate information source in the field. And when it doubt, I'll bracket... or at leas err on the side of slightly underexposing.
From my perspective, there is no perfectly engineered solution available to me... just as was the case with film. Quite often we must rely on experience, instinct, and interpretation... and in the end that works pretty will in conjunction with the remarkable tools that we have in modern digital cameras. Believe me, this was all much more difficult and much less accurate in the old days.
I understand the desire, especially on the part of those with the engineering mindset, to make the system perfect. And I'm sure that those like you who take that approach will continue to make our systems better. But in the meantime, at least for me, it isn't worth getting upset about the issues you bring up. They just aren't interfering with my photography.
Edited on Sep 01, 2026 at 07:49 AM · View previous versions
The jpeg preview in camera is ALWAYS tricking me. The extremes from the jpeg preview to unprocessed raw is too extreme imo. I feel like I have to work too hard in lightroom to get close to what I was seeing in the back of the camera... That's with all the Sony cameras I've used, and to a very slightly lesser extent, some of my Canons. Like I get the purpose of a raw, but wouldn't it make more sense to be able to see what your raw will look like in camera, rather than some cooked up jpeg? I'd...Show more →
"What it is going to look like" is not as simple as it sounds. In particular if you have a very wide dynamic range scene the raw is like to look pretty awful and not much at all like what you actually saw with your eyes.
If you are shooting raw it is far better to rely on the histogram.
Or if you don't want to bother, just shoot jpg.
- - -
old-gregg wrote:
gdanmitchell wrote
To some extent I'm not as concerned about this as you are since I prefer to shoot raw and get the full benefit of the camera's capabilities.One more time: unless you are using exposure compensation or spot metering, your camera always under-exposes, which means you are literally losing data.
gdanmitchell wrote
the camera's histogram display is your friend.Regarding "always underexposes," it is true that you can push your exposure a bit beyond what the camera probably suggests. My approach (since I'm not carrying a computer into the field to download and check each exposure with rawdigger) is to understand that no rear panel histogram display is perfect. However, from using the camera and learning to understand the relationship between what the histogram says, what the image displayed on the screen tells me, and what my experience with different subject colors and so on tells me... the histogram provides the most accurate information source in the field. And when it doubt, I'll bracket... or at leas err on the side of slightly underexposing.
From my perspective, there is no perfectly engineered solution available to me... just as was the case with film. Quite often we must rely on experience, instinct, and interpretation... and in the end that works pretty will in conjunction with the remarkable tools that we have in modern digital cameras. Believe me, this was all much more difficult and much less accurate in the old days.
I understand the desire, especially on the part of those with the engineering mindset, to make the system perfect. And I'm sure that those like you who take that approach will continue to make our systems better. But in the meantime, at least for me, it isn't worth getting upset about the issues you bring up. They just aren't interfering with my photography.
Edited on Sep 01, 2026 at 07:49 AM · View previous versions
Aug 31, 2026 at 10:19 PM
p.2 #142nd A7R6 or an A1ii?
So now people are debating about ETR like it was 2002?
It's one option, but not always the best option.
EBH
It's one option, but not always the best option.
EBH
Aug 31, 2026 at 10:48 PM
p.2 #152nd A7R6 or an A1ii?
old-gregg wrote:
This is either incomplete or incorrect, but since I don't know your assumptions, I can only guess. You are right if you're saying that a camera should avoid highlight clipping in RAW data at all costs. If you are calling this "underexpose a bit", fine.
But this "at all costs" approach is dumb by 2026 standards, because, surprisingly, it reduces the useful dynamic range. Apologies if you're already familiar with Bill Claff's concept of PDA (photographic dynamic range). But if not, the basic idea is that not all shadows are created equal, and above certain noise level, those low values aren't...Show more →
I'm a afraid I'm still a bit lost in your train of thought.
The issue with that one singel pixel is a non-issue if you choose the right metering mode. The case you describe would only affect highlight or spot metering, and even then I don't think 1 pixel would be enough, so let's say it's 1%.
As for the other issue you're describing, if you have a lot of "unused room on the right", that means you didn't expose correctly. That's why zebras are the only way to get actually close to correct exposure, as with the right settings they show the actual RAW DR latitude...
j4nu wrote
In general the camera should underexpose a bit, because that's what gives you more available DR to work with in post in most cases...But this "at all costs" approach is dumb by 2026 standards, because, surprisingly, it reduces the useful dynamic range. Apologies if you're already familiar with Bill Claff's concept of PDA (photographic dynamic range). But if not, the basic idea is that not all shadows are created equal, and above certain noise level, those low values aren't...Show more →
I'm a afraid I'm still a bit lost in your train of thought.
The issue with that one singel pixel is a non-issue if you choose the right metering mode. The case you describe would only affect highlight or spot metering, and even then I don't think 1 pixel would be enough, so let's say it's 1%.
As for the other issue you're describing, if you have a lot of "unused room on the right", that means you didn't expose correctly. That's why zebras are the only way to get actually close to correct exposure, as with the right settings they show the actual RAW DR latitude...
Sep 01, 2026 at 01:25 AM
p.2 #162nd A7R6 or an A1ii?
I would challenge both absolutes in this exchange. “Always underexpose slightly” leaves image quality unused in many situations, while saying that “zebras show the actual raw latitude” overstates what the camera is measuring. The camera’s exposure aids are valuable, but they remain processed indicators that must be interpreted—or experimentally calibrated against the raw file.
When making an image in the field without a large calibrated display and the ability to inspect the raw file, the photographer has to decide which highlights must be preserved and which ones can be allowed to clip. We all understand that.
The problem, in my opinion, is that an in-camera meter and its exposure aids come with the manufacturer’s assumptions and processing choices built in. A separate reflected-light meter also provides a reading based on its calibration standard; it still cannot know which parts of the scene you consider important.
If you had the time—and were working from a tripod with a calibrated MacBook and RawDigger—you could examine the raw data and determine precisely which areas and color channels had clipped. Under normal field conditions, however, most of us rely on experience, including lessons learned from past mistakes, to estimate the exposure. With familiar subjects and lighting, that estimate is usually quite good.
Modern cameras can handle many backlit situations remarkably well, but they cannot eliminate the dynamic-range problem. With a scene such as the backlit woodland examples in this discussion, you may need to make several exposures and blend them to taste in post-processing. In that sense, we have come full circle to the film era, when photographers bracketed exposures, combined negatives, accepted a compromise—or simply turned the camera 90 to 180 degrees and found something else to photograph.
I digress, but for sunrise and sunset scenes, a 0.9 or 1.2 reverse graduated neutral-density filter remains one practical solution. It is not perfect, but it is often good enough. In suitable scenes with limited subject movement, an in-camera HDR or exposure-blending feature can also do much of the heavy lifting.
Ultimately, neither the meter, histogram, nor zebras can decide which light matters to the photograph. They provide information; the photographer still has to interpret it.
When making an image in the field without a large calibrated display and the ability to inspect the raw file, the photographer has to decide which highlights must be preserved and which ones can be allowed to clip. We all understand that.
The problem, in my opinion, is that an in-camera meter and its exposure aids come with the manufacturer’s assumptions and processing choices built in. A separate reflected-light meter also provides a reading based on its calibration standard; it still cannot know which parts of the scene you consider important.
If you had the time—and were working from a tripod with a calibrated MacBook and RawDigger—you could examine the raw data and determine precisely which areas and color channels had clipped. Under normal field conditions, however, most of us rely on experience, including lessons learned from past mistakes, to estimate the exposure. With familiar subjects and lighting, that estimate is usually quite good.
Modern cameras can handle many backlit situations remarkably well, but they cannot eliminate the dynamic-range problem. With a scene such as the backlit woodland examples in this discussion, you may need to make several exposures and blend them to taste in post-processing. In that sense, we have come full circle to the film era, when photographers bracketed exposures, combined negatives, accepted a compromise—or simply turned the camera 90 to 180 degrees and found something else to photograph.
I digress, but for sunrise and sunset scenes, a 0.9 or 1.2 reverse graduated neutral-density filter remains one practical solution. It is not perfect, but it is often good enough. In suitable scenes with limited subject movement, an in-camera HDR or exposure-blending feature can also do much of the heavy lifting.
Ultimately, neither the meter, histogram, nor zebras can decide which light matters to the photograph. They provide information; the photographer still has to interpret it.
Sep 01, 2026 at 06:14 AM
p.2 #172nd A7R6 or an A1ii?
“Zebras … show the actual RAW DR latitude.”
Usually they do not. Zebras are generally calculated from the camera’s processed preview or video signal, just like the histogram. They can become a good proxy for raw clipping after testing and calibration, but they are not inherently a direct measurement of raw photosite values. Their relationship to raw clipping can change with picture profile, white balance, ISO, gamma, and zebra threshold.
Metering determines the camera’s exposure recommendation.
The JPEG histogram and zebras estimate what the processed image is doing.
Raw clipping is what determines whether recoverable sensor data has actually been lost.
Usually they do not. Zebras are generally calculated from the camera’s processed preview or video signal, just like the histogram. They can become a good proxy for raw clipping after testing and calibration, but they are not inherently a direct measurement of raw photosite values. Their relationship to raw clipping can change with picture profile, white balance, ISO, gamma, and zebra threshold.
Metering determines the camera’s exposure recommendation.
The JPEG histogram and zebras estimate what the processed image is doing.
Raw clipping is what determines whether recoverable sensor data has actually been lost.
Sep 01, 2026 at 06:28 AM
p.2 #182nd A7R6 or an A1ii?
For me and my style of shooting, the past 4 or so generations of cameras with multi-metering have been consistently good enough. My empirical confirmation of this is that in nearly every image I am lifting shadows, pulling down highlights and adding black and white; or trimming both ends of the levels slider to meet the histo’s tails.
With Sony however, it does appear they have pre-baked something like Fuji’s and Nikon’s DR400, so the left end of the histo is almost always very near 0 while the right tail has more headroom. I have not experimented with Sony’s exposure calibration menu item, but I am close to digging into it and figuring out how to bake in a +⅓ EV.*
* Edit: I just went ahead and did this. For those that have not looked into it, it appears pretty slick. First, you adjust by each individual metering pattern. IOW you can dial in a correction for multi metering and it doesn't change spot or average, etc. Next, you can dial in by 1/6th stop EV's, a nice fine-tune touch. Finally, it is not a system wide adjustment, so only affects the chosen custom shooting bank.
Anyway, I have dialed in +2/6 EV to "multi" in all my shooting banks. I will use it today this way and see how it works out.
With Sony however, it does appear they have pre-baked something like Fuji’s and Nikon’s DR400, so the left end of the histo is almost always very near 0 while the right tail has more headroom. I have not experimented with Sony’s exposure calibration menu item, but I am close to digging into it and figuring out how to bake in a +⅓ EV.*
* Edit: I just went ahead and did this. For those that have not looked into it, it appears pretty slick. First, you adjust by each individual metering pattern. IOW you can dial in a correction for multi metering and it doesn't change spot or average, etc. Next, you can dial in by 1/6th stop EV's, a nice fine-tune touch. Finally, it is not a system wide adjustment, so only affects the chosen custom shooting bank.
Anyway, I have dialed in +2/6 EV to "multi" in all my shooting banks. I will use it today this way and see how it works out.
Sep 01, 2026 at 07:17 AM
p.2 #192nd A7R6 or an A1ii?
EB-1 wrote:
So now people are debating about ETR like it was 2002?
It's one option, but not always the best option.
EBH
No option that I know of is always the best option.
For example, I would not ETR for a low dynamic range subject, something I used to sort of do years back. “For reasons” it is generally better, I think, to keep the luminosity values away from the dark and bright extremes of the histogram when possible. (Note the word “generally” in that sentence.)
Technically, ETR was/is not what I’m talking about in my example. The ETR folks would have you align the brighter tones with the right end of the histogram in almost all cases, even those in which the DR was not that large. That’s not my approach at all. (That theory wasn’t just about retaining highlight details. It also was based on the observation that more bits were allocated to a given range of dynamic contrast at the bright ends of the luminosity spectrum than at the dark end, suggesting that it might help keep tonal transitions smoother.)
But generally — unless you are in a situation where you want to lose details in the brightest areas — if you are dealing with a very large dynamic range subject where you know that you’ll work in post to get the tonal balance you are after you’ll want to pay a lot of attention to the exposure of the highlights.
The issue, of course, is that once you feed a photo site more light than it can record, all of the values beyond that limit end up the same and you lose highlight details. At the same time, what happens at the dark end of the luminosity scale (within limits) is different — mostly the noise levels increase there. While you cannot get back anything from those blown highlights in post you can get quite a bit back from the “underexposed” shadows, especially with today’s improved cameras and improved post-processing software.
(This is kind of the reverse of the approach with old-school film negatives, where things “rolled off” at the bright end of the subject’s dynamic range but might be blocked in the very dark areas. In a way, I suppose we might say that the highlights are “blocked” with an overexposed digital capture.)
Make sense?
Edited on Sep 01, 2026 at 10:23 AM · View previous versions
So now people are debating about ETR like it was 2002?
It's one option, but not always the best option.
EBH
No option that I know of is always the best option.
For example, I would not ETR for a low dynamic range subject, something I used to sort of do years back. “For reasons” it is generally better, I think, to keep the luminosity values away from the dark and bright extremes of the histogram when possible. (Note the word “generally” in that sentence.)
Technically, ETR was/is not what I’m talking about in my example. The ETR folks would have you align the brighter tones with the right end of the histogram in almost all cases, even those in which the DR was not that large. That’s not my approach at all. (That theory wasn’t just about retaining highlight details. It also was based on the observation that more bits were allocated to a given range of dynamic contrast at the bright ends of the luminosity spectrum than at the dark end, suggesting that it might help keep tonal transitions smoother.)
But generally — unless you are in a situation where you want to lose details in the brightest areas — if you are dealing with a very large dynamic range subject where you know that you’ll work in post to get the tonal balance you are after you’ll want to pay a lot of attention to the exposure of the highlights.
The issue, of course, is that once you feed a photo site more light than it can record, all of the values beyond that limit end up the same and you lose highlight details. At the same time, what happens at the dark end of the luminosity scale (within limits) is different — mostly the noise levels increase there. While you cannot get back anything from those blown highlights in post you can get quite a bit back from the “underexposed” shadows, especially with today’s improved cameras and improved post-processing software.
(This is kind of the reverse of the approach with old-school film negatives, where things “rolled off” at the bright end of the subject’s dynamic range but might be blocked in the very dark areas. In a way, I suppose we might say that the highlights are “blocked” with an overexposed digital capture.)
Make sense?
Edited on Sep 01, 2026 at 10:23 AM · View previous versions
Sep 01, 2026 at 07:59 AM
p.2 #202nd A7R6 or an A1ii?
2613pch wrote:
“Zebras … show the actual RAW DR latitude.”
Usually they do not. Zebras are generally calculated from the camera’s processed preview or video signal, just like the histogram. They can become a good proxy for raw clipping after testing and calibration, but they are not inherently a direct measurement of raw photosite values. Their relationship to raw clipping can change with picture profile, white balance, ISO, gamma, and zebra threshold.
Metering determines the camera’s exposure recommendation.
The JPEG histogram and zebras estimate what the processed image is doing.
Raw clipping is what determines whether recoverable sensor data has actually been lost.
Not sure what you intended to convey here, but my point was that zebras are good enough to protect highlights from clipping (if zebra level lower limit is properly set), which in turn maximizes captured DR.
“Zebras … show the actual RAW DR latitude.”
Usually they do not. Zebras are generally calculated from the camera’s processed preview or video signal, just like the histogram. They can become a good proxy for raw clipping after testing and calibration, but they are not inherently a direct measurement of raw photosite values. Their relationship to raw clipping can change with picture profile, white balance, ISO, gamma, and zebra threshold.
Metering determines the camera’s exposure recommendation.
The JPEG histogram and zebras estimate what the processed image is doing.
Raw clipping is what determines whether recoverable sensor data has actually been lost.
Not sure what you intended to convey here, but my point was that zebras are good enough to protect highlights from clipping (if zebra level lower limit is properly set), which in turn maximizes captured DR.
Sep 01, 2026 at 09:18 AM