-SPLIT- Discussion of Dynamic range with some tall claims.

curiousgeorge wrote in post #10396698
You can't expose for both foreground and sky when there is such a big difference. Digital sensors capture about 4-5 stops of light, if the difference is more than this than either you expose for the sky, or the foreground. Without a grad filter or using blending/ HDR techniques either the sky will be blown out or the fg will be black.

More like 10-12 stops for the latest Canon dSLRS.

Poe,

Highly unlikely, 4 to 5 is about the limit for straightforward shots without advanced manipulation after the event. Its something that Canon are supposed to be addressing in the new 1Ds mark4, but I don't expect dramatic improvements.

Poe wrote in post #10401400
More like 10-12 stops for the latest Canon dSLRS.

From a technical perspective, you are correct because the sensor data has a high enough S/N ratio that the analog sensor data can be quantized to a resolution of 14 bits when converted to a digital word. For practical purposes, the dynamic range is somewhat less.

First, when the linear response of the RAW data has a non-linear gain applied as a part of the RAW conversion process to emulate the semi-logarithmic response of human vision, the quantization levels in the very dark areas of an image are not sufficient to support discernible details -- thus the result is "blocking" (large clumps of same-value image pixels). Next, (assume that we are still working in 16 bits) when we consider the image AFTER it has been adjusted for human visual response, quantization is still a factor in discerning detail in the very dark parts of an image so a black point is usually set at some point where this "dark noise" is masked out. When the image is finally saved, most often it is converted to 8-bit resolution. Considering that the vast majority of computer monitors only support 8-bit words, but really can only display about 6 bits (not to mention their limited gamut which is typically not much more than sRGB), it is easy to see that a lot of what the camera can record is being lost before the image finally reaches our eyes.

The great majority of printers have limitations comparable to monitor limitations. Finally, it is worth mentioning that displaying an image with a very wide dynamic range on a device such as a monitor or printer with a very limited capability for displaying dynamic range only results in a very drab lifeless looking image. That is why photographers "play" with images in post processing to boost the contrast of midtones and other "tricks" such as HDR and blended images.

The bottom line is that I would prefer to have a RAW image with a wider dynamic range that gives me more "wiggle room" for post processing than a RAW image that starts out with only about six stops of dynamic range. It is always better to have data that you can throw away than not having it in the first place.

Lowner wrote in post #10401850
Poe,

Highly unlikely, 4 to 5 is about the limit for straightforward shots without advanced manipulation after the event. Its something that Canon are supposed to be addressing in the new 1Ds mark4, but I don't expect dramatic improvements.

This is incorrect. You can do the test exposures yourself and use Rawnalyze to confirm. My 7D has about 12 stops DR.

Bill Boehme wrote in post #10401932
From a technical perspective, you are correct because the sensor data has a high enough S/N ratio that the analog sensor data can be quantized to a resolution of 14 bits when converted to a digital word. For practical purposes, the dynamic range is somewhat less.

First, when the linear response of the RAW data has a non-linear gain applied as a part of the RAW conversion process to emulate the semi-logarithmic response of human vision, the quantization levels in the very dark areas of an image are not sufficient to support discernible details -- thus the result is "blocking" (large clumps of same-value image pixels). Next, (assume that we are still working in 16 bits) when we consider the image AFTER it has been adjusted for human visual response, quantization is still a factor in discerning detail in the very dark parts of an image so a black point is usually set at some point where this "dark noise" is masked out. When the image is finally saved, most often it is converted to 8-bit resolution. Considering that the vast majority of computer monitors only support 8-bit words, but really can only display about 6 bits (not to mention their limited gamut which is typically not much more than sRGB), it is easy to see that a lot of what the camera can record is being lost before the image finally reaches our eyes.

The great majority of printers have limitations comparable to monitor limitations. Finally, it is worth mentioning that displaying an image with a very wide dynamic range on a device such as a monitor or printer with a very limited capability for displaying dynamic range only results in a very drab lifeless looking image. That is why photographers "play" with images in post processing to boost the contrast of midtones and other "tricks" such as HDR and blended images.

The bottom line is that I would prefer to have a RAW image with a wider dynamic range that gives me more "wiggle room" for post processing than a RAW image that starts out with only about six stops of dynamic range. It is always better to have data that you can throw away than not having it in the first place.

What you say is true, that the image will be limited by the ability of the output device's own DR. However Lowner's original statement was in regards to digital sensors, which was true for sensors somewhere back in the 90s, but not today in 2010.

Poe,

I totally disagree with this pseudo "science". Using Rawanalyze to prove it only serves to cast doubt on Rawanalyze. It has moved slightly from the early days, but is still nowhere near a natural 11 or 12 stops. It can be manipulated to immitate a larger dynamic range, but I prefer to live in the real world.

I have done a series of tests on my 30D using real world exposures and see a 5 stop usable range. The tail of the curve is virtually horizontal at each end and while this might fool the science, its unusable in the real world. Film is known to have a better DR, but even that could never achieve 11 or 12 stops.

Poe wrote in post #10401400
More like 10-12 stops for the latest Canon dSLRS.

Digital sensors cannot currently record more than about 5 stops.

When they claim 10-12 stops what they mean is that in theory you can pull the highlights by a couple of stops and push the shadows by a few stops (giving you a higher 'dynamic range' overall) but the image will look terrible.

What I was referring to was the DR captured by the sensor itself. It;s about the same for transparency film, compared to about 10 for negative film.

It's easy enough to prove this, take a shot with a difference of more than 5 stops in range and watch either your shadows or highlights disappear.

When the megapixel race is over only then will manufacturers concentrate on better quality sensors delivering more DR.

curiousgeorge wrote in post #10405916
When they claim 10-12 stops what they mean is that in theory you can pull the highlights by a couple of stops and push the shadows by a few stops (giving you a higher 'dynamic range' overall) but the image will look terrible.

Pulling highlights and pushing shadows decreases, not increases, DR.

Poe wrote in post #10406111
Pulling highlights and pushing shadows decreases, not increases, DR.

I think you know what I meant

curiousgeorge wrote in post #10406136
I think you know what I meant

The tendency today is to pull shadows and push highlights. Makes for very contrasty or "punchy" photos. People don't seem to be complaining much.

So no, I don't know what you meant.

Given that we're talking about increasing DR, not increasing contrast, it should be pretty obvious what I meant.

curiousgeorge wrote in post #10406323
Given that we're talking about increasing DR, not increasing contrast, it should be pretty obvious what I meant.

Which is exactly what you're doing if all your captured tones are in the midtones and your image has the bit depth to support darker tones and brighter tones to which you spread the midtones into shadows and highlights.

Or if you've captured dark tones and bright tones and bring them into the midtones, you have reduced your DR.

Contrast is directly related to DR.

Lowner wrote in post #10405302
Poe,

I totally disagree with this pseudo "science". Using Rawanalyze to prove it only serves to cast doubt on Rawanalyze. It has moved slightly from the early days, but is still nowhere near a natural 11 or 12 stops. It can be manipulated to immitate a larger dynamic range, but I prefer to live in the real world.

I have done a series of tests on my 30D using real world exposures and see a 5 stop usable range. The tail of the curve is virtually horizontal at each end and while this might fool the science, its unusable in the real world. Film is known to have a better DR, but even that could never achieve 11 or 12 stops.

You may consider your 30D as having a "usable" 5 stop range for your photography and processing style, but that does not mean the 30D only has 5 stops of DR! You already tried to pass off the same mis-information in this thread (https://photography-on-the.net/forum/showthre​ad.php?t=750731) and it was shown to be wrong there too. What you see and what is are two different things. Folks long ago used to see the sun orbiting around the earth and believed that it did, but that didn't make it true.

Poe wrote in post #10406747
Contrast is directly related to DR.

I don't think it is.

An image that has a high DR is not automatically contrasty.

Case in point. Velvia is contrasty film, but being slide film it only has 4-5 stops DR. Whereas cheap negative film is 10-12 stops is not necessarily going to be more contrasty.

If the 7D has such high DR it would be able to cope with the first of the OP's images much better, given that it's an overcast sky (if it had been sunny the difference in stops in the scene would have been far greater, making it even harder to capture the full range).

curiousgeorge wrote in post #10406963
I don't think it is.

An image that has a high DR is not automatically contrasty....

Indeed quite the opposite,. the higher the DR the capture device is capable of, the lower the contrast will be in the resulting image ( all else being equal )
Unless we are really getting our semantics backwards somewhere along the line.

curiousgeorge wrote in post #10406963
I don't think it is.

An image that has a high DR is not automatically contrasty.

Case in point. Velvia is contrasty film, but being slide film it only has 4-5 stops DR. Whereas cheap negative film is 10-12 stops is not necessarily going to be more contrasty.

If the 7D has such high DR it would be able to cope with the first of the OP's images much better, given that it's an overcast sky (if it had been sunny the difference in stops in the scene would have been far greater, making it even harder to capture the full range).

One would need to be able to access the RAW file to more accurately determine. Depending on how the JPG was created, there may have been sacrifices made that didn't need to be made. JPGs are not a reliable measure of DR because of the compromises necessary of the format.