What does the MM 'raw histogram' say about the camera's dynamic range?

[photomeme]

Advertisement (gone after registration)

It seems that one could do a simple test to see how many stops of exposure move an illuminated object from "blue" to "red", or from one end to the other of the so-called 'raw histogram'.

 

Anyone with an MM have a chance to give it a go?

[Advertisement]

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

[algrove]

Since there are 11 EV divisions on the histogram and the tones recorded stretch all the way from left to right covering all these 11 EV, then IMHO the camera's dynamic range is 11 stops. It will not show more.

[SpiritShooter]

I am not a technogeek when it comes to sensors and stuff, but I would think that the since the histogram is a graphic representation of the exposure, Leica could have divided the zones into 8, 10, 12, or whatever that desired.

 

My gut feeling is that they may have been trying to relate the MM histogram to the Zone System and not necessarily relate it to the actual dynamic range of the camera.

 

Mike

[johnbuckley]

I am not a technogeek when it comes to sensors and stuff, but I would think that the since the histogram is a graphic representation of the exposure, Leica could have divided the zones into 8, 10, 12, or whatever that desired.

 

My gut feeling is that they may have been trying to relate the MM histogram to the Zone System and not necessarily relate it to the actual dynamic range of the camera.

 

Mike

 

I have wondered the same thing, re: Zone System.

 

A note to those going to Photokina -- Lou, or David Farkas, if he is reading this: when talking to Leica folks, it really would be nice to know what they were thinking as they designed the histogram. Lou may be exactly right - that there are 11 stops of range. But how those lines match to the Zone System has made me wonder, and since Mike gave voice to it, perhaps others are wondering about the relationship, too. JB

[photomeme]

manual:

 

•The histogram appears in the lower half of the picture.

When using DNG format (see p. 115) the histogram

is divided into 11 sections, with the graduations

corresponding to a 1EV difference in brightness.

If clipping indicators are also activated (see next

section), the left (blue) and/or right (red) ends of the

histogram flash.

Notes:

•The division of the DNG histogram is comparable to

the zone system. This system was used in analog

B&W photography to adjust the exposure so that

the contrast range present in the subject can be

reproduced as required in the developed picture.

 

I have wondered the same thing, re: Zone System.

 

A note to those going to Photokina -- Lou, or David Farkas, if he is reading this: when talking to Leica folks, it really would be nice to know what they were thinking as they designed the histogram. Lou may be exactly right - that there are 11 stops of range. But how those lines match to the Zone System has made me wonder, and since Mike gave voice to it, perhaps others are wondering about the relationship, too. JB

[jaapv]

I am not a technogeek when it comes to sensors and stuff, but I would think that the since the histogram is a graphic representation of the exposure, Leica could have divided the zones into 8, 10, 12, or whatever that desired.

 

My gut feeling is that they may have been trying to relate the MM histogram to the Zone System and not necessarily relate it to the actual dynamic range of the camera.

 

Mike

YThe Zone system has 10 zones ( or 8, depending on interpretation), but not 11.

[mjh]

Advertisement (gone after registration)

YThe Zone system has 10 zones ( or 8, depending on interpretation), but not 11.

I suppose the correspondence to the zone system is that one zone equals 1 EV. Since the dynamic range of the M Monochrom (or the M9 for that matter) is somewhere between 11 and 12 EV there are 11 subdivisions.

[tele_player]

In 'The Negative', Ansel Adams numbers the zones from 0 to X, which is eleven zones, and discards the highest and lowest to refer to the 'dynamic range' as zones I to IX, which is 9 zones.

So, according to Adams, The Zone system has 11 zones ( or 9, depending on interpretation).

 

-Robert

 

YThe Zone system has 10 zones ( or 8, depending on interpretation), but not 11.

[photomeme]

Two strategies I see for the 'harsh clipping' issue.

 

The first is simply to bring down the maximum brightness of the last 'zone' of the histogram without changing the rest of the histogram. This improves images that look good over most of the image but have a small area much brighter than the rest, like a bit of sky breaking through an image naturally mid-tone, such that there appears on a CRT a very harsh and unnatural white blotch that is discordant with the image. It won't make the transition any less abrupt, but less harsh, and the image should be more pleasing without sacrificing dynamic range (i.e., you can effectively expose to the right).

 

The second is to expose a bit more conservatively and compress the last 1-2 'zones' by reducing contrast for that portion to create a more smooth transition to any blown out area. This will sacrifice dynamic range at the right end of the histogram, introduce more noise on the left side, but create a more film-like transition to white. For a very close approximation to film like shoulder in the brighter tones, you may lose several stops of dynamic range and/or raise noise in the shadows significantly.

 

Nothing I'm describing brings back blown highlights, of course.

 

The most direct way to achieve these effects, for now, is manual manipulation of curves. i also foresee applying noise reduction selectively (or on a gradient) to darker areas of images where I've applied above technique #2.

 

I see a lot of MM shots in these early days not taking advantage of either strategy.

 

In 'The Negative', Ansel Adams numbers the zones from 0 to X, which is eleven zones, and discards the highest and lowest to refer to the 'dynamic range' as zones I to IX, which is 9 zones.

So, according to Adams, The Zone system has 11 zones ( or 9, depending on interpretation).

 

-Robert

[photomeme]

Thanks for reminding us about the specifics! Although Adams spoke of zone 0 and zone X as areas with no information, just full black or full white, so it would be like the area to the left of the histogram and to the right of the histogram, respectively, in the MM case, would it not? So using his terminology we have zone 0, full black, 11 zones with information content, and zone 12 of full white. Suffering, of course, the abrupt transitions on the right of a linear sensor without Bayer data for algorithmic trickery.

 

In 'The Negative', Ansel Adams numbers the zones from 0 to X, which is eleven zones, and discards the highest and lowest to refer to the 'dynamic range' as zones I to IX, which is 9 zones.

So, according to Adams, The Zone system has 11 zones ( or 9, depending on interpretation).

 

-Robert

[SpiritShooter]

I suppose the correspondence to the zone system is that one zone equals 1 EV. Since the dynamic range of the M Monochrom (or the M9 for that matter) is somewhere between 11 and 12 EV there are 11 subdivisions.

 

Interesting in that Edwin Putts says in his most recent expose that he is finding that the MM has 6-1/2 stops of usable dynamic range and when you include pure black and pure white (clear transparency) the camera has 8-9 stops of actual dynamic range.

 

Regardless of what the dynamic range is, I would be a beer or three that Leica takes that exposure range and expands it to fill the graphic width of their histogram. The 11 divisions represent the Zone System steps. A nice touch to say the least.

 

Mike

[algrove]

@john-yes I will ask the question and any other s within reason at the show.

 

As regards my histograms.

 

When I shoot a dark image say at night I find the vertical zone delineations are bunched up over to the left side of the histogram which seems correct to me.

 

However, when I shoot an image having a wider tonal range I get those same vertical zone lines more spread out from the left side to the right side of the histogram.

 

These vertical zone lines are never the same for all images, but appear different with each image giving me the impression they are reacting to my images' dynamics.

 

Any others experiencing this same thing?

[photomeme]

Leica is betting a beer against you in it's manual.

 

It claims unambiguously that there is usable data in each zone, one zone is 1EV.

 

Can't account for EP's comment on dynamic range. Not sure he factors in the possibilities of a full linear raw workflow on a production release MM. As I recall his number for dynamic range on the M9 itself was (absurdly?) low.

 

Much of the commentary on the M9's sensor used metrics more appropriate to CMOS than CCD. CMOS builds noise reduction into the hardware layer (as it must, because CMOS technology is inherently noisier than CCD technology); still, that noise reduction has advanced to quite sophisticated and better than CCD.

 

Not sure I like the look of CMOS over CCD technology, there's an ever-present plasticky element; although less pronounced in recent CMOS sensors, it's still there. To some of us, CCD seems to look noticably better at base ISO.

 

To the extent a measure of 'usable' dynamic range relates to a mathematical measurement of luminence 'noise', the much more fine and less patterned grain of the MM (because Bayerless and uninterpolated) surely adds a stop or two or recoverable range in the shadows over whatever EP is doing.

 

Interesting in that Edwin Putts says in his most recent expose that he is finding that the MM has 6-1/2 stops of usable dynamic range and when you include pure black and pure white (clear transparency) the camera has 8-9 stops of actual dynamic range.

 

Regardless of what the dynamic range is, I would be a beer or three that Leica takes that exposure range and expands it to fill the graphic width of their histogram. The 11 divisions represent the Zone System steps. A nice touch to say the least.

 

Mike

[AlanG]

These are from my 1969 version of the book Photographic Sensitometry by Todd and Zakia.

I always thought these examples were interesting and I feel if you are going to think about how to expose MM files you should be comparing the approach to shooting slide film not to negative film. Negative films may have a coating of high speed emulsion over another coating of lower speed emulsion.... sort of like shooting two images for HDR.

 

I figure usage of these images for discussion here alone falls under "fair use." Horizontal streaks are from my scanner.

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[chrism]

Alan,

I hadn't heard of the two layers of emulsion with differing sensitivities. Are any of the films we use today made that way?

Otherwise, I confess to some confusion. Most of the digital world has been about 'expose to the right' followed by the comment that this is the opposite approach to that used for film (not that any film users ever had a histogram unless they scanned their negatives, but I do have enough darkroom experience to understand one might expose for the shadows and develop/print for the highlights). Then I learn that the MM is particularly prone to unrecoverable highlights (OK, not phrased fairly - blown highlights, the fault of the photographer and not the camera, will not be recoverable) and that we should now expose for the shadows again. To me, in my simplicity, that means don't point the camera down and lock the exposure then reframe, but simply take what comes as the bright sky ought to be allowed to affect the meter. But then I ask myself, with only a luminance channel, surely software is no better at recovering fully black shadows?

 

So what's a girl to do?

 

Chris

[AlanG]

Alan,

I hadn't heard of the two layers of emulsion with differing sensitivities. Are any of the films we use today made that way?

 

 

I can't say I know the exact preparation methods of any film and this is probably proprietary. I remember being told in school that Plus X was such an example and this was a reason you could get a printable negative from it with many stops of overexposure. These were the kinds of things we measured after exposing film in sensitometers in my early classes. (42 years ago!) Any film that still has a good slope in the densest sections of its characteristic curve would handle overexposure pretty well. The book "The Science of Photography" by Baines and Bomback has a section that describes coating machines and how they work. It says this, "Many modern fast negative materials are double-coated, the film first receiving a coating of a slower emulsion before the extremely rapid emulsion is applied." My edition is from 1970 and the first edition was 1958. So I'm not sure what was meant by "modern."

 

It also helps explain how very primitive snapshot cameras could get acceptable b/w photos in various conditions. And it also explains why an MM does not work the way b/w film works.

 

As far as digital cameras are concerned, the way they are set up to measure the "exposure" and how they apply a curve on the back LCD for judging the image are factors. As well as how they process and display their histograms. I doubt if the histogram from two different brands of cameras would look the same even if they shot the same subject with the identical amount of light hitting the sensor.

 

I think one has to get used to using a given camera and I have never tried out the MM so I have no first hand knowledge about exposing with one. But I'd venture an opinion that it is much different than b/w film and using the zone system as applied to negative film may not be the best approach other than as a general method to understand the limited brightness range of paper and how that applies to the range in the subject.

 

To give you advice, if highlight detail is important to you then you must make sure it is not blown out. (I am sure you already know this.) Then you can either add light into the shadows or try to pull them up in the raw processing stage. (I carry at least 7 strobe heads with me on each shoot and even light up exteriors of homes at dusk to get the look that I want.)

 

Here are histograms from DXO, ACDSee, and C1 from the enclosed image that looked the same on screen in each program. This sample clearly shows the need for supplementary lighting in order to hold highlights and get nice mid tones and shadows. And the yellow lighting was part of the design and needed to be preserved.

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[IkarusJohn]

These are from my 1969 version of the book Photographic Sensitometry by Todd and Zakia.

I always thought these examples were interesting and I feel if you are going to think about how to expose MM files you should be comparing the approach to shooting slide film not to negative film. Negative films may have a coating of high speed emulsion over another coating of lower speed emulsion.... sort of like shooting two images for HDR.

 

I figure usage of these images for discussion here alone falls under "fair use." Horizontal streaks are from my scanner.

 

Thanks for this Alan. Unless I have misunderstood, this suggests that for negative B&W film of that era, it is better to expose for the shadows on the basis that over-exposure is easier to recover than under-exposure. The 6 stop over-exposed image (which suggests exposing for the shadows) can be recovered in printing, whereas the 6 stop under-exposed image (exposed for highlights) cannot.

 

Conversely, much like slide film (where highlights are easily, irretrievably lost),* the Monochrom should be exposed for highlights. This raises an interesting issue with regard to Adams's zone approach to assessing whites, darks and greys, and whether it can be a useful if modified to the dynamic range of the Monochrom sensor, and exposure metering.

 

I guess only time and experience, and careful assessment of the histogram, will give us the answer on how to expose well. I am used to taking ambient readings as, for a relative novice when it comes to Adams's zones and such ideas, an incident reading keeps everything simple; unless there is a significant difference between where you take the reading and what's lighting your subject, an incident reading cares not a jot about reflectivity or shadows. You are therefore guaranteed true, dense blacks and clean whites, provided you take a proper reading and the range is within the capability of your sensor.

 

The trick (for me, anyway) will be understanding any idiosyncrasies of my incident light meter, and the centre-weighted meter in the camera, with the additional complication of which filter to use when and what impact that has on my incident reading.

 

Lots of fun ahead. I'm very confident that by the time I have assessed all that and come to a decision, the subject will have wandered off in disgust!

 

Cheers

John

 

* In my relative youth, I spent a life changing few months backpacking through Asia on my way to Europe. I spent two months in Nepal, trekking through the mountains with a Nikon FE, a couple of lenses and a bag of Kodachrome. I was shocked when I got it developed in London to find that in many of the images, the mountains had vanished completely. A hard lesson ...

[AlanG]

Thanks for this Alan. Unless I have misunderstood, this suggests that for negative B&W film of that era, it is better to expose for the shadows on the basis that over-exposure is easier to recover than under-exposure. The 6 stop over-exposed image (which suggests exposing for the shadows) can be recovered in printing, whereas the 6 stop under-exposed image (exposed for highlights) cannot.

 

Conversely, much like slide film (where highlights are easily, irretrievably lost),* the Monochrom should be exposed for highlights. This raises an interesting issue with regard to Adams's zone approach to assessing whites, darks and greys, and whether it can be a useful if modified to the dynamic range of the Monochrom sensor, and exposure metering.

 

I guess only time and experience, and careful assessment of the histogram, will give us the answer on how to expose well. I am used to taking ambient readings as, for a relative novice when it comes to Adams's zones and such ideas, an incident reading keeps everything simple; unless there is a significant difference between where you take the reading and what's lighting your subject, an incident reading cares not a jot about reflectivity or shadows. You are therefore guaranteed true, dense blacks and clean whites, provided you take a proper reading and the range is within the capability of your sensor.

 

The trick (for me, anyway) will be understanding any idiosyncrasies of my incident light meter, and the centre-weighted meter in the camera, with the additional complication of which filter to use when and what impact that has on my incident reading.

 

 

Yes for b/w negatives you expose for the shadows in that era and this one. You either adjusted the processing for the highlights (as with the zone system) or burned them in. (Or made contrast masks.)

 

As for incident readings , that is probably a good way to start assuming there is no light emanating from your subject and that your darkest subjects that you want detail in are being illuminated at the same value.

 

I will tell you that I shot interiors on film (with hot lights or strobes) for about 25 years and rarely used a meter at all to compute exposure. I proofed on Polaroid and through experience and a test book I carried, I knew how that related to the characteristics of my film. I did use both incident and spot meters outside or to get initial values or lighting ratios. When I had to deal with filter factors, many different leaf shutters and lenses, shutter efficiency, reciprocity failure, etc. the actual number on the meter was not very useful.

 

Today I rely on shooting tethered or the LCD and find that live view is great. If the image is within the histogram (no clipping) I know I'm good to go in post processing and I really don't have to think about exposure beyond that unless I'm adding lights or fill.

[Advertisement]

Hallo,

Look here: Leica M

[IkarusJohn]

I must confess, I had come to the same conclusion.

 

I have been opposed to chimping, but I can see more studying of the histogram when my new camera arrives in about 8 weeks time.

 

Thanks - most helpful.

John