Tonal Curve of Leica M Monochrom

[Martin B]

Advertisement (gone after registration)

Sorry but your comment that any digital sensor only displays zones from 3-8 is contrary to my everyday experience.  You say this is "well described in the literature" and then cite: 

 

- a 2011 post by a blogger with unknown credentials on a basic learning site.  The author claims: "Cameras, however, only make instantaneous exposures, with film and expensive medium format digital cameras being able to capture 12 different stops of light variations, while most other digital cameras can only capture about 5 stops of light variations or even less."  That 5-stop claim is just not true.  The author has no idea.  

 

- an undated (says 2007 at the bottom) article by an unidentified author that talks about some unspecified CCD.  That's hardly supportive of your claim about only getting zones 3-8 from "any digital sensor".  The graph just says digital is linear, film has a curve ... obvious stuff ... nothing more, nothing about 5 stops.

 

Neither of these count as "literature" about current camera sensors.  It's just stuff some people put on the web and it's terribly out of date.  It's also so contrary to experience that it's not worth arguing about.

 

You simply say "the 5-stop claim is not true. The author has no idea". Why? Because it doesn't fit your experience? What are your credentials? The only thing I can agree with is that the information might be a bit out of date and sensor technology has caught up and increased the 5-stop range as mentioned above (BTW, 8-3=5, 5 stops).

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[zlatkob]

You simply say "the 5-stop claim is not true. The author has no idea". Why? Because it doesn't fit your experience? What are your credentials? The only thing I can agree with is that the information might be a bit out of date and sensor technology has caught up and increased the 5-stop range as mentioned above (BTW, 8-3=5, 5 stops).

 

 

It doesn't fit my experience.  But don't listen to me.  Instead, check out DxOMark.com.  They've tested the dynamic range of hundreds of cameras and all of them exceed your claimed 5-stop range.  Even cameras introduced in 2002 and 2003 exceeded that.  DxOMark's "Landscape score" is the maximum dynamic range performance and its unit is an EV.  They regard anything over 12 as excellent.  The worst camera on their list was measured with 9 stops.  The best was measured with over 14 stops:

http://www.dxomark.com/Cameras/Nikon/D810---Measurements (click Dynamic Range, tab 3 at the top).

[jaapv]

You simply say "the 5-stop claim is not true. The author has no idea". Why? Because it doesn't fit your experience? What are your credentials? The only thing I can agree with is that the information might be a bit out of date and sensor technology has caught up and increased the 5-stop range as mentioned above (BTW, 8-3=5, 5 stops).

No - he says that is blathering nonsense. That is, the sources you quote. 5 EV values? 1990...

[Martin B]

That is, the sources you quote. 5 EV values? 1990...

Another simple question: how does the absolute EV numbering system actually compare with the zone system? I am aware that for relative eV, 1 eV = 1 stop. But what does it mean for absolute eV numbers, for example when we have 14.8 eV? How does it fit into the zone system which is laid out between 0-11? I have a hard time finding a good description here how those two systems are directly related.

[jaapv]

The Zone System is not that absolute, nor was it conceived for the electronic image.  AA already postulated zones XI, XII, XIII for the highlights as film increased in dynamic range.

One could easily speak of zones -I, -II, -III for recoverable shadows in digital images. The Zone System is a tool to get tones from the subject into the print in a systematic manner, not something to tie one's hands rigidly.

It is not for nothing that the Monochrom1 (I don't know about the 2) has a Zone-type raw histogram.

As a matter of fact the histogram is the tool of choice to emulate the placement of your tonal values. Set the white point, set the black point, set middle gray and adjust the curve in between. The basic idea is the same.

[Martin B]

The Zone System is not that absolute, nor was it conceived for the electronic image.  AA already postulated zones XI, XII, XIII for the highlights as film increased in dynamic range.

One could easily speak of zones -I, -II, -III for recoverable shadows in digital images. The Zone System is a tool to get tones from the subject into the print in a systematic manner, not something to tie one's hands rigidly.

It is not for nothing that the Monochrom1 (I don't know about the 2) has a Zone-type raw histogram.

As a matter of fact the histogram is the tool of choice to emulate the placement of your tonal values. Set the white point, set the black point, set middle gray and adjust the curve in between. The basic idea is the same.

 

Yes...but when we discuss DR of modern sensors with about 14 eV, we need to understand how this fits to the zone system. Because if this means that 14 eV = 14 stops = 14 zones, then it would exceed by far what film can achieve. With such wide zone range (if this is the case) why would be there still a need to bracket or do HDR with modern sensors? I understand that it is easier to recover shadows with a wide DR - does it mean that potential 14 stops/zones are not centered around zone 5 (which is 18% grey) but more so towards the dark areas? Do you mean this by pointing out negative zones? Would this leave the highlight zones intact as up to zone 10 or 11? Or does 14 eV mean that it would go from zone -1 to +12 and centered at zone 5? 

[zlatkob]

Advertisement (gone after registration)

Yes...but when we discuss DR of modern sensors with about 14 eV, we need to understand how this fits to the zone system. Because if this means that 14 eV = 14 stops = 14 zones, then it would exceed by far what film can achieve. With such wide zone range (if this is the case) why would be there still a need to bracket or do HDR with modern sensors? I understand that it is easier to recover shadows with a wide DR - does it mean that potential 14 stops/zones are not centered around zone 5 (which is 18% grey) but more so towards the dark areas? Do you mean this by pointing out negative zones? Would this leave the highlight zones intact as up to zone 10 or 11? Or does 14 eV mean that it would go from zone -1 to +12 and centered at zone 5? 

 

Under the best circumstances for each (sensor with most dynamic range shot at low ISO vs. film with most dynamic range processed for maximum dynamic range) the two are probably not far apart, maybe within a stop of each other. 

 

The need to bracket or do HDR still arises for some photographers because the brightness range of the world far exceeds even 14 stops.  For others, it may just be easier than measuring the brightness level of everything in a scene and deciding were exactly to center the exposure — a bit of insurance in case they mis-judge.  

[otto.f]

HDR is overrated, especially with Monochromes

[jaapv]

Yes...but when we discuss DR of modern sensors with about 14 eV, we need to understand how this fits to the zone system. Because if this means that 14 eV = 14 stops = 14 zones, then it would exceed by far what film can achieve. With such wide zone range (if this is the case) why would be there still a need to bracket or do HDR with modern sensors? I understand that it is easier to recover shadows with a wide DR - does it mean that potential 14 stops/zones are not centered around zone 5 (which is 18% grey) but more so towards the dark areas? Do you mean this by pointing out negative zones? Would this leave the highlight zones intact as up to zone 10 or 11? Or does 14 eV mean that it would go from zone -1 to +12 and centered at zone 5?

I didn't/don't do HDR on film and I don't on digital. I find the result not pleasing.

 

Anyway, how are you going to get the result in a print? You must either compress the vast range you recorded in HDR to the approx 6.5 EV value DR of the print, getting a wholly unnatural result (or at least losing the fine distinctions in the midtones), or choose which range you want to render, which you basically should do when determining your original exposure.

Note that the human eye has a contrast ratio of 6.5 EV values, which closely matches prints, reason why we can take in the end result just fine.

Now why can our eye capture a vastly larger range in real life? That is because it is not a still camera, it is a video camera and is constantly adjusting exposure, leaving it to the postprocessng in our brain to create an image out of the video stream. In that sense HDR does indeed mimic our perception - but I still don't like the way it looks, as the mind can compress/decrompress seamlessly and the photograpic process cannot.

As à matter of fact, we are putting three dimensions into two when creating a photograph and that is difficult enough when we see the problems understanding perspective, with HDR we are squeezing four dimensions in. Go figure.

As for your final question: that depends entirely on your exposure. You can either decide to center your exposure, or decide to expose to the left or the right of your histogram, depending on whether you are trying to protect your highlights or whether you are going for a maximum of detail in the shadows, just like exposing and developing film.

[Manolo Laguillo]

[...] Even this still keeps me wondering why film manufacturers would do this since the shoulder is an advantage for film vs. digital. But I will try to get the actual response curves for the films which I am currently using, maybe I see the same what you described above. [...]

 

I can't agree with you, sorry: the not-so-steep shoulder (one approaching horizontality) in the film's response to light isn't an advantage, because it means less separation on film than in reality.

 

I will repeat my whole argument, but approaching it in another way.

 

The curve is the expression of the way in which 2 'things' are interconnected:

 

- light reflected from the various surfaces on the scene, and

- densities on the film.

 

An horizontal (or nearly horizontal) line means that, even when different amounts of light did hit the film, its response consisted in delivering the same (or nearly the same) density throughout for all those different inputs.

 

That is what happens,

1. left of the toe's beginning,

2. and on the shoulder when it approaches the horizontality.

 

The scientists' research in film technology since the birth of photography has been, if we look at it as photographers, not as engineers, in trying to increase its sensibility (speed), its latitude (Dynamic Range), to decrease the grain's size (the T-grain for instance), and to increase its response to the diversity of colors present in real world (panchromatic film, infrared film).

 

Part of this has been trying to compensate for the inherent tendency films have to the S shape in their characteristic curves. The way they found for compensating this problem of non-linear response was designing printing papers with a characteristic curve which is very steep.

 

When I was doing platinum printing, I had to take into account the fact that platinum is a printing out paper. This means that the image appears right during the exposure. The first parts in appearing are, of course, the blacks and shadows (the neg is less dense there). This gives time for the light grays and highlights to appear, because the blacks sort of block themselves, slowing down its emerging. Therefore, my 8x10 negatives had to be very dense in the highlights parts. I achieved this by exposing 'with generosity', and by developing more time. This whole procedure can be understood as a way of sorting out the shoulder's tendency to horizontality.

 

Let's go back to digital: In fact, when working in PP with a MM2 raw picture, I try to increase the tone separation both in the shadows and the highlights by playing with those parts of the curve which relate to these values. The way of doing it is by increasing the steepness, either by pulling up or by pushing down the line, depending on the pictures character.

 

The MMs (both 1st and 2nd generation ones) response to light overcome, so to say, this very old limitation inherent to film which is expressed in the S-shape.

 

My question now is: after all these consideration, do you still want a S-shape response, Martin?       

[Martin B]

I can't agree with you, sorry: the not-so-steep shoulder (one approaching horizontality) in the film's response to light isn't an advantage, because it means less separation on film than in reality.

 

I will repeat my whole argument, but approaching it in another way.

 

The curve is the expression of the way in which 2 'things' are interconnected:

 

- light reflected from the various surfaces on the scene, and

- densities on the film.

 

An horizontal (or nearly horizontal) line means that, even when different amounts of light did hit the film, its response consisted in delivering the same (or nearly the same) density throughout for all those different inputs.

 

That is what happens,

1. left of the toe's beginning,

2. and on the shoulder when it approaches the horizontality.

 

The scientists' research in film technology since the birth of photography has been, if we look at it as photographers, not as engineers, in trying to increase its sensibility (speed), its latitude (Dynamic Range), to decrease the grain's size (the T-grain for instance), and to increase its response to the diversity of colors present in real world (panchromatic film, infrared film).

 

Part of this has been trying to compensate for the inherent tendency films have to the S shape in their characteristic curves. The way they found for compensating this problem of non-linear response was designing printing papers with a characteristic curve which is very steep.

 

When I was doing platinum printing, I had to take into account the fact that platinum is a printing out paper. This means that the image appears right during the exposure. The first parts in appearing are, of course, the blacks and shadows (the neg is less dense there). This gives time for the light grays and highlights to appear, because the blacks sort of block themselves, slowing down its emerging. Therefore, my 8x10 negatives had to be very dense in the highlights parts. I achieved this by exposing 'with generosity', and by developing more time. This whole procedure can be understood as a way of sorting out the shoulder's tendency to horizontality.

 

Let's go back to digital: In fact, when working in PP with a MM2 raw picture, I try to increase the tone separation both in the shadows and the highlights by playing with those parts of the curve which relate to these values. The way of doing it is by increasing the steepness, either by pulling up or by pushing down the line, depending on the pictures character.

 

The MMs (both 1st and 2nd generation ones) response to light overcome, so to say, this very old limitation inherent to film which is expressed in the S-shape.

 

My question now is: after all these consideration, do you still want a S-shape response, Martin?       

 

Thanks, Manolo, again for the very good background information provided here for me! Much appreciated - a lot of learning for me in this thread! I also read yesterday in Ansel Adam's book what you mentioned earlier with the difference in films and their tonal curves. Sounds like form what you are saying in the post above that the S shape film curve is more useful and applicable for the print itself. From other comments above it sounds that a more modern digital sensor overlaps anyway with the maximum tonal range provided in film. In this case it doesn't matter if a specific film has a shoulder on either end since you can adjust the digital linear response in PP anyway as you stated above (by increasing the steepness on both sides from a virtual middle point applying an artificial S curve).

 

From all this collected information the difference I see between digital and film is likely more caused by the choice of photosensitive or inkjet paper when printing since paper (and ink) might have different tonal latitudes, too. So in an ideal case you could have a digital print and a print on photosensitive paper from the film negative look the same assuming that the contrast and tonal latitude of both print processes is very similar.

[Martin B]

I didn't/don't do HDR on film and I don't on digital. I find the result not pleasing.

 

Anyway, how are you going to get the result in a print? You must either compress the vast range you recorded in HDR to the approx 6.5 EV value DR of the print, getting a wholly unnatural result (or at least losing the fine distinctions in the midtones), or choose which range you want to render, which you basically should do when determining your original exposure.

Note that the human eye has a contrast ratio of 6.5 EV values, which closely matches prints, reason why we can take in the end result just fine.

Now why can our eye capture a vastly larger range in real life? That is because it is not a still camera, it is a video camera and is constantly adjusting exposure, leaving it to the postprocessng in our brain to create an image out of the video stream. In that sense HDR does indeed mimic our perception - but I still don't like the way it looks, as the mind can compress/decrompress seamlessly and the photograpic process cannot.

As à matter of fact, we are putting three dimensions into two when creating a photograph and that is difficult enough when we see the problems understanding perspective, with HDR we are squeezing four dimensions in. Go figure.

As for your final question: that depends entirely on your exposure. You can either decide to center your exposure, or decide to expose to the left or the right of your histogram, depending on whether you are trying to protect your highlights or whether you are going for a maximum of detail in the shadows, just like exposing and developing film.

 

Yes, this comes back to Manolo's post above that in fact it is very much related to the final print itself. Is the 6.5 eV for the print always the case or only on specific inkjet paper? Is it the same value for photosensitive paper? Or are there (slight?) differences between more and less contrast-rich papers?

[Martin B]

Under the best circumstances for each (sensor with most dynamic range shot at low ISO vs. film with most dynamic range processed for maximum dynamic range) the two are probably not far apart, maybe within a stop of each other. 

 

The need to bracket or do HDR still arises for some photographers because the brightness range of the world far exceeds even 14 stops.  For others, it may just be easier than measuring the brightness level of everything in a scene and deciding were exactly to center the exposure — a bit of insurance in case they mis-judge.  

 

Sounds like this is the case with modern digital sensors. I would love to find a graph somewhere which includes this information and overlays DR of traditional film with DR/tonal range of a modern sensor. We are always assuming what the difference is - one stop, some more or less? It is already hard to find an information out there how 14 eV translate into the traditional zone system form 1-11. I tried to find this information online without success so far showing this graphically.

[jaapv]

Yes, this comes back to Manolo's post above that in fact it is very much related to the final print itself. Is the 6.5 eV for the print always the case or only on specific inkjet paper? Is it the same value for photosensitive paper? Or are there (slight?) differences between more and less contrast-rich papers?

Of course there are differences, this is an average rule of thumb. It really seems to me that you are lost in a jungle of specifications. It is time to go out and simply do it. As I said, the Zone System is a tool, and not a religion. Expose well, postprocess properly, and experiment with printing. Find a workflow, not a theory.

[pop]

Sounds like this is the case with modern digital sensors. I would love to find a graph somewhere which includes this information and overlays DR of traditional film with DR/tonal range of a modern sensor. We are always assuming what the difference is - one stop, some more or less? It is already hard to find an information out there how 14 eV translate into the traditional zone system form 1-11. I tried to find this information online without success so far showing this graphically.

Presuming that you have access to a digital camera: nothing could be easier. Set up a stable arrangement of lighting and camera and a roughly uniform target such as a wall or cupboard door. Take a properly exposed picture of the cupboard door or wall. Take a series of shots, each with double the exposure of the preceding one. Then, starting with the properly exposed one, take a series of shots, each with half the exposure of the preceding one.

 

Open all those shots with your favorite PP software and measure the brightness of the pictures, always measuring the same spot, of course. Draw a graph. You'll see immediately the ceiling of the curve. You'll also see the floor, but this might be a bit harder to see on account of the noise.

[LeicaPassion]

Of course there are differences, this is an average rule of thumb. It really seems to me that you are lost in a jungle of specifications. It is time to go out and simply do it. As I said, the Zone System is a tool, and not a religion. Expose well, postprocess properly, and experiment with printing. Find a workflow, not a theory.

[LeicaPassion]

+1

[pico]

Yes...but when we discuss DR of modern sensors with about 14 eV, we need to understand how this fits to the zone system.

 

Do we really? For digital rendering, it might be best to move away from the traditional one-stop (or EI) difference between zones. It has not been so linear with film or silver-based paper for many years (although with T-Grain film it seems to be reverting.) 

 

The latitude of modern ink-jet printers has a greater contrast-ratio today than it had years ago, and exceeds the range of wet silver-based prints. It might be helpful to address each zone as an RGB value or a % ratio of black/white.

[Advertisement]

Hallo,

Look here: Leica M

[Martin B]

Presuming that you have access to a digital camera: nothing could be easier. Set up a stable arrangement of lighting and camera and a roughly uniform target such as a wall or cupboard door. Take a properly exposed picture of the cupboard door or wall. Take a series of shots, each with double the exposure of the preceding one. Then, starting with the properly exposed one, take a series of shots, each with half the exposure of the preceding one.

 

Open all those shots with your favorite PP software and measure the brightness of the pictures, always measuring the same spot, of course. Draw a graph. You'll see immediately the ceiling of the curve. You'll also see the floor, but this might be a bit harder to see on account of the noise.

 

Simple but great idea - I will do this! I did a similar thing when calibrating my wet printing process for prints with digital negatives. This should work - I will then be able to overlay this created graph with an existing one from a film.

[xenar]

Careful: as I said way in the beginning, if you're doing this on a Monochrom, you will trace the ACR default tonal curve that gets applied to the RAW linear data. You'll either need to get the RAW data or work out the inverse of the ACR curve to get back to the 14-bit linear space.

 

If you're doing this on other cameras, you may have other gamma corrections applied. The easiest way to get at the uncorrected data may be to use dcraw.