digital exposure control

[Jonathan Wienke]

The most useful digital exposure strategy is "expose to the right" or crowding non-specular highlights as close as possible to clipping without actually doing so.

Digital Exposure And Metering Strategies
Michael's original Expose To The Right article

[BJL]

Eugène [sorry, wrong name before],
   one element of exposure control is understanding the latitude of the medium: how many stops under before the image disappears into "film base plus fog" or noise, how many stops over before highlights are "blocked" or "blown out".

Phil Askey at DPReview has made his third attempt at measuring dynamic range, which seems far better than his previous ones at least, using a Stouffer Step Wedge. Results are seen for the first time in his review of the FujiFilm S3 Pro, at http://www.dpreview.com/reviews/fujifilms3pro/page18.asp

One thing I notice that fits with conventional digital wisdom: for both the S3 in normal mode (not using the special "R" pixels) and the 20D, the "latitude" is unbalanced, with about five stops below the mid-tones, only three above them. You probably know film behavior better than me, but -5 sounds better than film (good underexposure latitude), while +3 seems about as tight as reversal film, worse than negative film. (It also fits almost perfectly with the ISO definition of a sensor's base speed, which is based on placing the highlight blowout point at 170% diffuse reflectivity, or 3.2 stops above 18% gray.)

My reading would be that we should expose to control the highlights, much as with slide film, but with the reassurance that there is more latitude for underexposure when using low ISO settings, so it is usually not critical to get the highlights as far to the right as they can go.

That suggests a slight variation on Michael Reichmann's advice: expose to the right, but never go too far to the right, while being one or two stops to the left is usually safe.


P. S. Norm Koren has produced software called Imatest which tries to measure dynamic range, amongst other things, but I have not studied it yet. I trust Norm better than Phil on technical issues though!

P. P. S. The S3 Pro's "R" pixels balance the range out by adding two stops of highlight headroom, which is just as Fuji claims, so I wonder what all the complaints are about!

[Gordon Buck]

Is a variation of the Zone System likely to be useful when the sensor cannot capture the complete dynamic range of the subject with a single exposure but, say, two exposures can be combined to record all "zones"?

[Mark D Segal]

In my practical experience with digital capture, the advice from Michael and Jonathan of "exposing to the right" without blowing out highlights is extremely useful for preserving as much image information as possible.

I also find (on my Canon 1Ds) that the histogram and the flashing blow-out warnings on the LCD screen provide a very reliable guide of what to expect when you open the RAW image in Adobe Camera RAW. And once there, if by accident highlights in one or two channels have been blown-out, much detail can usually be rescued by reducing exposure in Camera RAW.

With the 1Ds (and this will vary from camera to camera) in selecting an ISO noise is pretty low on my list of considerations, because with post-capture processing programs such as Noise Ninja and PK Sharpener Pro you can really work wonders on images that you may think have been compromised before these treatments.

Another important aspect of expanding dynamic range in post-capture processing is the ability to use latent image detail for luminance blending. This consists of blending several different exposures from the same Camera RAW file to expand dynamic range as much as the latent image detail in your RAW file will allow - and the amount can be surprisingly large, again depending on camera and exposure conditions. I have recently drafted a very detailed "Aide Memoire" on implementing this approach, and if you would like to receive it as a PDF, please send me your email address. I am not set-up for accessing materials through a web-site. It is also explained in Katrin Eismann's book on Compositing and Masking.

Given the enormous potential that post-capture processing gives the photographer who creates RAW images to start with, there are simply more easily usable degrees of freedom with digital capture than ever existed in the film world.

Regards,

Mark

[BJL]

About the Zone System; a RAW file from a digital camera is not quite like a negative; it is more like an undeveloped sheet of film, that can be individually processed over and over again to get many different "negatives". From that perspective, it is not surprising that a lot of the Zone System approach is replaced by post-procesing stage. And that we can potentially spend a lot more time on digital post-processing than we ever did on film development!

What is left at exposure time is mainly "placement" to get all the relevant subject details within the gamut of the "emulsion", with the biggest emphasis seeming to be on highlight placement.

P. S. Us lazy types sometimes do more of this at exposure time, with JPEG output: adjusting exposure index, contrast settings, sharpening, color temperature and such, which emulates experimenting with film and processing choices on a frame by frame basis. If you worry about your students thinking too little with a digital camera and leaving to much to automation and default settings, there are plenty of experiments and challenges to throw at them for every step of the process!

My current learning interest is fine tuning color balance to match what I see (or what I want to see!). I prefer to do that in the field, when I can compare to the original subject, though I doubt the LCD can be trusted that much for color accuracy. Maybe I need to carry a laptop to really do it right.

[boku]

For the exclusive case where a photographer uses the digital medium by recording RAW files, we need to abandon old tenets about "getting the in-camera exposure correct". (This does not apply to JPEG shooters.)

In a perfect world where all colors are truly represented by the histogram, the "expose to the right" method will deliver the best results. It maximizes the medium. We aren't quite there yet, but you learn to know what to expect from your histogram pretty quickly. A blown out yellow flower tells you that the reds aren't represented strong enough on the histogram. Pros that live with their camera develop this sense. That makes "expose to the right" a safe and predicatable practice.

By shooting to the right, more decisions about the final print are made outside the camera as opposed the zone system. And when we talk slides, all descisions are made in camera.

This is all very rudimentary for those that speak before me, but I just wanted to strees that you need to make the paradigm shift. Digital exposure (using RAW capture) needs to be resolved in the computer for best advantage. Know what your histogram is telling you and maximize the data in the file.

All of the above advice is valid. My suggestion is that you need to make the paradigm shift. I wish you well. In my world I've dealt with unemployed (smart but stubborn) COBOL/Fortran programmers that can't move to modern Object-Oriented Programming to save their soul. You need to free yourself for the sake of your students.

[eugene appert]

I am a college photography instructor in Quebec, Canada. We have a three-year program in photography.  Next term we will have completed a full transition to digital technology.  Presently we are working out the fine points of course content and trying to standardise methods and practices which would enable our students to be as flexible and as competent as possible in the ever-evolving world of digital photography.

My question today deals exclusively with the new science of exposure, I am trying to get a feel if there is any standardised consensus about exposure control of digital capture among practicing photographers.  Given the zone system was designed for a single fixed output using relatively similar film stock that reacted in relatively similar ways to contrast manipulation, there are many who believe that any attempt to salvage the zone system as a reference for digital capture would be useless.  Insufficient consistency in dynamic range from one camera to another and the variety of output ranges possible render any reference to the zone systems defunct.  Some of my colleagues even promote abandoning the light meter in favour of the in-camera histogram with clipping preview.  

On the other hand there is an opinion that a zone-like system should be employed for digital capture even if the output range is not known, since there is no other way of precisely translating luminance to specific density.  Some cameras may produce only two or three textured zones over the grey card reading and perhaps only three or four under the grey card reading, but having mapped the densities, (whether they produce 8 or 9 or 10 zones) the photographer would know precisely where those tones would fall along the curve.  In which case, a camera-by-camera calibration would produce different exposure scales but offer the same precise control to each photographer /camera combination.  For example if a grey card exposure is calibrated to produce L*54, then a reduction of one stop is likely to produce L*36 another L*24 and still another L*12 etc. This school of thought asks how does the photographer know how a specific tone is being translated using the histogram. Some even dismiss the in-camera histogram as almost completely useless for a photographer shooting RAW since the information it provides is the result of a virtual conversion to jpeg, using gamma encoding, and icc profile incrustation according to camera default settings.  

Any feedback would be greatly appreciated, especially from educators who have successfully competed the transition.

Thank you

Eugène Appert
Montréal Canada

[mcanyes]

A few thoughts:
I am familiar with the zone system and sometimes find it useful to "think through" a shot, but don't see too much of an application to digital. For all practical purposes you are shooting slides and the zone system has never worked well for slides.

You can spot read a scene and find out if it will fit into the range of your digital camera, but if it won't fit there is only so much you can do about it. It is possible to raise the shadows in Photoshop, but you also make the noise more visible if you raise them too much.

Also, don't forget that you can see exactly what you are doing on the computer - something that the zone guys never had available.

Underexposure is a quick route to noisy images - particularly at high ISO, so don't dismiss the histogram. You can "fix" an exposure in post if you shoot raw, but it is a bad idea. As Jonathan says: expose to the right.

Personally, I find that the matrix metering in my Nikon D2H is so good that I just trust it in most cases, and sometimes do a bracket of 1 stop each way if I have time. I have been shooting for more than 30 years and I find that I am very happy to make use of the technology available to simplify my workflow.

Hope that helps,
Michael

[eugene appert]

Thank you very much for your contributions and references.

I am getting a sense from the answers you have given me that digital exposure strategies are more about getting the best or most correct or perhaps even most versatile exposure at point of source and less about manipulation of light and tone at the moment of exposure.   Most likely this is because, as Michael Canes points out, there doesn’t seem to be much we can do about it, or, because of the seemingly infinite control that we have over tone-by-tone manipulation in Photoshop.  Perhaps the argument for a zone-like system is promoted by purists who want to reduce post capture tonal manipulations to preserve a maximum of quality, though admittedly 16-bit capture is pretty robust.  The clipping preview does indicate the order of each luminance area by area along the scale of scene brightness but its only a good guess about where that luminance will fall precisely.  Although knowing that once in Photoshop, a single tone, either throughout the entire image or in a specific area can be, lifted or dropped a point or two completely independently, makes it difficult to motivate a photographer to work almost in the dark and with simple tools at the viewfinder.

Thank you for your feedback.  I think the comparison to slide film is a good one, perhaps even C 41 with its given and unalterable contrast, and the document explaining tests for true raw clipping will be helpful.

Eugene Appert

[mcanyes]

Eugene,

I guess this really comes down to whether you are studying the process as an academic exercise, or focusing on the output (print, web image, etc.) If you are focusing on the output look at http://www.outbackphoto.com/index.html
He has several e-books that are excellent.

Also,
http://www.luminous-landscape.com/
http://www.fredmiranda.com/
http://www.robgalbraith.com/bins/index.asp
are worth a look if you haven't been there yet.

As I am more interested in output, and making a living, I tend to be result oriented. I haven't touched film for about 3 years now and my work is far better. I can control the process IMHO far better than the zone guys could.

One thing I am curious about. You said "makes it difficult to motivate a photographer to work almost in the dark and with simple tools at the viewfinder." Unless you specifically wanted to work with film and a wet process why would you try to motivate a student in that direction. The zone system was "invented" to control a particular process. For what it was it worked quite well. but digital is a very different animal with a different process. It makes no sense to me to try to overlay the zone system on a digital workflow as a way of getting the correct exposure.

As an example, I was in college when calculators replaced slide rules. One of the benefits of understanding a slide rule is that you get pretty good at estimating an answer. The students that never used a slide rule made a lot of dumb mistakes on the calculator because they could not estimate. My answers tended to be more accurate. But I would never go back to the slide rule. I view the zone system the same way. I can look at a scene and think values and translate that to what I see on my monitor. But I still expose to the right for the highlights, and adjust the shadows in Photoshop. It remains to be seen exactly where the zone system will fit into all of this, but it is interesting to speculate a bit.

Michael

[eugene appert]

"...One thing I am curious about. You said "makes it difficult to motivate a photographer to work almost in the dark and with simple tools at the viewfinder."

Michael,

I think I have expressed myself poorly. I wasn’t referring to the zone system here but rather to the using the histogram to manipulate tones along the curve. The idea I was trying to express was, given the infinitely greater precision and control in Photoshop it is hard to justify attempts to do so using the histogram. “Blind” because the histogram doesn’t show the targeted density for each luminance and “simple tools” because the histogram is mostly about endpoints, and intervals of light cannot be separated and manipulated independently as one could using the zone system or to a much greater degree as one can using curves and layer masks.  But as you pointed out earlier those kinds of controls do not exist at source point of digital capture which would appear to render a system which attempts to do so (the zone system) defunct.

I think the concern about abandoning the light meter is usually expressed in terms of tonal intervals. How will the student know to place a background at a specific light value?  or, what will its relationship be to the silhouetted form before it?  or, to a shadowed middle value object etc.  I think the concern grows in situations where the photographer is more active in the manipulation of light and tone as opposed to simply trying to find the best exposure. But its strictly academic if such controls are infinitely greater in Photoshop and without too much data loss.

You say your interest is output and results; well, I am almost certain that the zone system cannot take us to digital output. If a zone-like system can be applied to digital capture I am convinced that it cannot be taken further than a series of logarithmically scaled values as points of reference in an RGB or Greyscale working space. Taking it to digital output would force us to redefine the White-Zakia definition of what a zone actually is.  Some desktop digital printers, especially the wax variety, would produce zones of less than .10 density increments which are technically not zones at all but more like 1/3 zones.


 Eugene

[mcanyes]

Eugene,

Unfortunately, most camera histograms are still lacking in the ability to look at all 3 colors so you do have to be careful when exposing an image with one bright color, flowers for example. It is easy to "blow out" a channel (color) and not catch it on the camera histogram. So expose to the right really becomes "expose almost all the way to the right." I expect we will eventually get more useful histogram readouts as more photographers push for them. This is one of the reasons I still bracket. Even though the meter is usually correct, it can't always spot a channel that is maxed out.

One of the more interesting areas of raw image conversion is how to make the most out of problem highlights. My current favorite is Bibble. Worth a look if you are not familiar with it. http://bibblelabs.com/

To address your other point about abandoning the light meter, I always use my meter when setting up lights - inside or out. Light ratios are still meaningful, and you are quite right about setting background values, etc. But I also capture to a laptop when I can so I can see exactly how the values relate to each other, and I can see the individual histograms. I haven't had a reshoot for technical reasons for a couple of years.

Michael

[Mark D Segal]

A great way (and unfortunately somewhat expensive for what it is) to get good colour balance in the field is to carry the credit-card size version of the MacBeth Color Checker. Take one photo of it in the kind of lighting conditions typical of the photo-shoot you will be doing. Then back in the digital darkroom in Levels or Curves click the grey eye dropper on the appropriate grey MacBeth patch in the image and this should exactly rebalance the colors. Compare the resulting colors of the MacBeth color patches in the image with the patches on the actual card. If OK (normally is or should be) note the recipe and apply as needed to all other images under similar lighting conditions.

Getting back to Professor Eugene's concern about luminosity, density and what the histogram shows, my statistical understanding of a histogram is that it is a density function with luminance on the X axis as the parameter and density on the Y axis. It tells us alot more than just looking at end points. The higher the "bars" of the histogram on any one luminsoty level, the more of the image's total information is concentrated at that level. Presumably with a combination of a zone-system set of grey patches ranging from black to white corresponding with levels 0 to 255 respectively, perhaps one could photograph the patches with any model digital camera, see how it distributes the patches along the luminosity scale, and measure it all using curves in Photoshop. This way you can trace the one to the other if needed say in a studio setting where one is manufacturing lighting conditions. Apart from this, while an understanding of zone-system approach is generally helpful to organize one's thinking about envisioning photographic information, it seems to me that in digital capture a combination of using the histogram, plus what one sees through the viewfinder plus the ability to manage the histogram by varying exposure to ensure capture of detail where it is needed most all taken together should generally provide a practical recipe for successful capture. Then there is all the post-capture processing latitude others and I have alluded to above.