« Mine That Bird! (OT) | Main | Epic Curtis Up for Sale »

Sunday, 03 May 2009


Gorgeous photos, wonderful article. Thank you.

A very significant contribution! Thanks a lot!

However, the claim in the caption of the 1st picture is wrong. The brightness of a shadow is not influenced by the brightness of the medium it falls on!

You can produce a completely black shadow on a totally white background. Try this: take a white paper, put a non-transparent object in front of it, wrap the whole setup with a black cloth and take a photo with a flash a little bit off the camera axis. You will get a wonderful black shadow on white paper.

Actually, the second picture shows two shadows on the same light-gray surface and the one in front is distinctly brighter than the one in the back. A clear indication, that the surface is unimportant.

What really decides about how deep the shadow is, is the fact, whether there are objects reflecting light into the shadows or not. In a city you usually have enough walls around, acting as reflectors, in order to get reasonably open shadows.

To make a long story short, the shadow in the first picture is not medium tone due to the light surface but due to objects like the wall on the right side, bouncing light into the shadow.



excellent article!!!

I 'grew up' shooting K64, then Velvia and Provia, so I got used to their unforgiving handling of highlights. It has taken me a surprisingly long time to wrap my head around digital capture's handling of sunlit scenes, and to overcome an intuitive dread of mid-day sun cultivated by working with chrome films. After more than six years shooting digital I still have to get past that initial shudder when I see stark shadows. Yet it's so liberating to get past the 'Velvia aesthetic' of subdued subject contrast range and neon color. Oddly enough, looking at Edward Hopper's paintings has been the best antidote for me.

Carl, thanks a lot for this understandable explanation of light intensities and the effect on film and sensor. I thought I had a fair understanding of this topic, but your description made me correct my perception especially on the highlight topic.

"In effect, the established standardized rating (known first as ASA and then ISO) was really about a one stop push."

Gawd, I wish somebody had f***ing told me then. I did everything by the book (many books), and I always had underexposed negatives.

That's a very lucid description, Carl. Well done.

My related anecdote is about the disappointment I had to swallow when I got back some transparencies of an important portrait shoot some years ago. I thought I was being clever by using my client's new Nikon D100 as a substitute for Polaroid preview shots. I set up and adjusted the lighting and exposure until everything was perfect in the D100's playback picture and histogram. Then I plugged the strobes into my Hasselblad and took the "real" shots on Astia.

What a disaster. The trannies looked like there had been no fill light of any kind. What you point out in this article about the different behaviour of the various recording media - neg versus pos versus digital - was a serious trap for the unwary.

Thanks to all for reactions.

Wolfgang, there are many factors that influence the brightness of a shadow, including environmental reflections. But the brightness of the surface is important: click on the first illustration for a closer look and follow the shadow of the stop sign post as it transitions from the middle gray asphalt to the bright gray concrete sidewalk.

For some reason, I spent my whole film life tied slavishly to rated speeds for normal development (though I very frequently pushed TRI-X). It was, after all, an official standard! Seems kinda dumb in hindsight. I'd read about establishing personal ASAs in the 1970s, but somehow rejected the concept. Maybe it was as simple as not feeling I could afford to give up the speed (I guess this is the photographic equivalent of "we need the eggs!").

I did eventually figure out that my Kodak Plus-X negatives printed better than my Ilford FP-4 negatives because the slightly purple base tone of the Plus-X increased the contrast of the Ilford variable-contrast paper.

Great article. This is how we learned exposure back in the day, and it's surprising how much you sometimes need a refresher.

Wow, did I need that! Thank you, Carl Weese.

Wolfgang, Carl: apples, oranges. Let me see if I understand your exchange about shadow tonality, or lack thereof:

When there is absolutely no ambient light or secondary source, shadows are simply not illuminated--the surface is literally invisible. Let's call this "absolute shadow" (which is, btw and AFAIK, an exceedingly unlikely occurrence outdoors in daylight (on Earth, anway)). On the other hand, any ambient light or secondary direct source falling on the shadowed surface will fill-light it, in which case it is just another lit surface, albeit more dimly lit than other parts of the scene.

Please correct me if I've misunderstood. And thank you both for forcing me to puzzle this out.

Nice to see a lucid account, with explanation, of things it took me a fair amount of trial and error to figure out.

Carl, very interesting. Would you rate current colour neg film like Fuji 400H likewise at say 200 or 160ASA?

Robert, I read somewhere that there's no "open shade" on the moon because there is no atmosphere to scatter sunlight. Sounds like your idea of absolute shadow. Light scatter in the atmosphere is what limits open shade to a three stop differential from sunlight. If the shade isn't "open" (like the underpass shot) then it can be more than three stops darker. Or bounce from surrounding objects or clouds in the sky can make the difference less than three stops.

There are other factors too. A vertical surface like a wall is in a better position to catch bounce light from another wall than a horizontal surface like a road. With the camera pointing down, a nearby shadow on the ground will look lighter than one farther away because of the viewing angle (that's happening in the Hot Bar Keep shot).

Guy, I haven't shot color negative in many years now, so I don't know about current films. Give it a try by bracketing some typical subjects. See if the lower speed gives richer shadow detail, or just unnecessary added density.

Photos taken on the surface of the Moon show plenty of light in the shadows. While its true that there's no atmosphere, the surface (and spacesuits)will bounce light around.

Unless of course it was all done in a sound stage with aliens looking over our shoulder.

Its not the sun I fear, its those few clouds in the sky on a sunny day that drive me crazy. ch

Deep, black shadows give pictures a meaning. Your clean technique turns every scene into a dry and unemotional record of so-called reality.

Carl, that's a very succinct explanation and I appreciate that it covers both film and digital, for comparison sake.

Guy: Yes, in my relatively short experience, 400H loves being exposed at 200, or even 100 in the right light, esp. if you meter the right thang. (I got this info from a interview of a pretty well-respected wedding photographer who shoots 400H almost exclusively -- maybe Jonathon Canlas - his work is nice). Sometimes, with 400H you can just let the highlights blow a bit and can get some lovely effects (a look not available in digital without a whole heck of a lot of computer time, if at all).

Thanks for the further clarifications.

To be fair (and obvious), perceptual or photographic "absolute shadow" (zero light or tone) is not uncommon, where a surface is merely dark enough that its luminance doesn't register (clipping). But this is an artifact rather than a representation of nature. If we could somehow render this in a truly nonreflective medium, it would, I'm sure, look as unnatural as it is.

To put this in some perspective: the nature, perception and representation of shadows challenged painters for centuries. “No shadow is black. It always has a color. Nature knows only colors … white and black are not colors.” (Renoir)

i certainly take your point about the range of tonal values we generally encounter in sunlit scenes and the overall sufficiency of our media, but i think you are kind of sidestepping one of the issues which makes a lot of people yearn for more dr in digital sensors (discounting another one of the issues, which is that too many people expose poorly). in your scenes the point of the photographs is, largely, the contrasts of light and shadow. while the scene may have 'really' looked like your rendering, in our perception we probably recall more detail in the shadows, because we can easily adjust and see into them with our eyes in real life. so when a lot of us take a sunlit picture with a person or people in it, shadows (especially on the face) are something we are trying to overcome, not something we want to emphasize for dramatic effect. that is, we mainly just want to pull detail out of the dark pits where the eyes reside, without making it look really nasty. yeah, sure, we could have used fill flash, but that isn't always possible or even desireable. with bw neg film, this was pretty easy; so far i've found it more difficult with digital sensors, though one thing i think helps is to not increase iso from the camera's base, eg, my 5d handles sunlit shadows better at iso 100 than at iso 200 (it seems like most users have taken a cursory look at frames made at the two isos and concluded they look equally good, and then proceed to use 200 as their everyday base iso. but doing so sacrifices some ability to retain sunlit highlights and still bring shadows up in a natural-looking way).

overall i think that the discussions about how much dr we need to photograph a scene have been talking past each other a bit. you and ctein are talking about using the full range of tones from white to grey to black to represent the values in the scene as lit, but other people want to be able to hold detail in clouds and still go down into deep shadows as if they were the middle grey. that is, they want to be able to view their photograph as if it were a real scene they are looking at with their eyes. the difference is a little bit like the way many people like their photos to be 100% in sharp focus front to back, rather than accepting a single plane of focus as the camera (and their eyes, but not necessarily their whole perceptual system) see the scene in a particular moment.

personally i like to see what the camera sees--it is more interesting to me to get exposed to a specific point of view--and likewise, i think the arguments for dr sufficiency are premised on representing a camera's eye view of the world. but it's no surprise that people still want to see what their mind's eye did.

Robert, on the Renoir quote, it happens my wife is a painter--classic academy training and a lifetime of experience--and she wouldn't be caught dead owning a tube of black paint. All shadow values in her work are mixed from deep colors, just as all grays are mixed from pigments with intense but complementary hues.

cb, I find deep shadows that retain descriptive power vastly more interesting than a print area that looks like spilled India ink.

Chris, there are plenty of daylight situations that exceed the ability of film as well as sensors, but sunlight from front or side doesn't get there. Many subjects in overcast light present a much more difficult dynamic range than sunlight...but that will be the subject of a future article.

Dear Chris,

We're not talking past each other. I'm not sure you got the import of my last column. Recall what I said in my last column about having difficulty coming up with photographs that even used up the full exposure range of my camera. There were very few of them, and none of them exceeded the 11 stop range. I could have produced exactly the look you're after from the two photographs in my column; I just don't happen to think it's an attractive look for those photographs.

The problem we are concerned with is mostly an aesthetic one; it is rarely a technical one. A good digital camera can record all the visible tones in almost every scene without clipping. The problem is rendering those tones in a way you'd find attractive. People who want a photograph that "looks like what they see" don't realize that this requires making aesthetic choices, not technical ones. Simply capturing ever-increasing exposure ranges doesn't address that.

There are many artistic solutions to the problem: the classic S-shaped characteristic curve, HDR-like tonal mappings, and even a straight linear rendition, ala Stephen Johnson's Digital Parks project from the 1990s. All of them present problems with "realism" because all of them are mapping a subject luminance range to a much shorter display/print range.

In summary, the arguments for which you referred to as "DR sufficiency" don't depend on accepting the camera's view of the world. My camera has sufficient range to capture everything in the scene for 98% of what I photograph. Producing an appealing photograph from that is a whole different problem.

There are technical problems associated with making such photographs, but they're not in the sensor; they're in the optics! To record a 12 stop subject luminance range with good tonal separation in the shadows means that flare and light scattering have to be held down to 0.01%! That is much, much harder for lens and camera designers to achieve than you might imagine. And the longer the exposure range you want to capture, the worse the problem becomes

~ pax \ Ctein
[ Please excuse any word-salad. MacSpeech in training! ]
-- Ctein's Online Gallery http://ctein.com 
-- Digital Restorations http://photo-repair.com 

well, i guess we're still talking past each other then. it looks like robert e was taking the point though.

i'll try once more. consider a subject on the street. lots of patches of bright sun, reflective cars, etc, but also lots of shade, from open shade to the deep shadow of doorways, etc. you expose close to the sunny 16 rule to retain proper color and detail in sky and highlight areas. but your main subject, let's say, is a face, and that face is facing away from the sun (or even facing away, and already in shade as well). sure, it is no problem to represent the face as shadowed, but since that is the main subject, we really want to locally raise it a bit. now the skin, which might have been just above grey had it been sunlit, is at best that minus 3 stops (using your own estimate for shadows). in other words, the region you hope is about in the middle is now about 5 stops below the highlights, but we'd like to raise it back up 2~3 stops. but wait--that's just the cheeks. now go into the eyes, possibly quite dark eyes--god help you if the subject is wearing a hat--and we're talking about regions where the image quality is falling apart. if you were to raise them up along with the rest of the face, they would look terrible and unnatural.

hypothetically, if we had say two stops more range over what a good sensor (5d would be my standard, but a lot are as good or better) gives us now, raising those shadows selectively to craft a picture more representative of our perceptions would be much easier. i am not saying i 'need' two more stops to do photography, obviously. but i would have plently of uses for it if i had it. and we can perhaps agree that sunlit scenes would only be a start, and not the most significant gain.

Chris, consider the work of Alex Webb and Constantin Manos. They work with short-scale Kodachrome in harsh sunlight, but construct their pictures to read entirely from the highlights. If a person's face is turned out of the sun, they wait. Then shoot when the architecture of the picture fits in the range of K64. They don't do spot news.

There's a reason news photographers have a flash mounted on their cameras. They aren't making art, they can't wait, they're making recognizable shots of politicians or celebrities, and that calls for exactly what you've described: fill flash to open the micro shadows of three dimensional objects--like people--in harsh sunlight. It looks awful, but it gets published. Back in the twenties, the guys shooting for the Daily News with Speed Graphics carried as many giant flashbulbs as sheets of film in the 4x5 holders...for the same reason.

Carl, I saw Alex Webb doing a "show and tell" recently, and he's been using his dwindling "stash" of his favored, discontinued Kodachrome 200, though he admits that he did also shoot some K64 recently. Costa Manos and David Alan Harvey are now shooting with the Leica M8 and 28mm (35 equivalent) lenses. Have a look at Costa's wonderful "Magnum in Motion" essay and also click on "Leica M8" and have a look at his digital work, it looks just like his previous work with K64!

ctein: i'll take your post (evidently written before my second post, which i likewise wrote before yours was up) as proof that, in fact, we're still talking past each other. how was it unclear that i was referring precisely to aesthetic choices? and the point is, lots of people would like to have a technical base which could provide more freedom to make new, or at least their own, aesthetic choices. as for the argument about how we're at the limits determined by lens flare--clever, but considering your current camera isn't straining under the effects of flare, i think we will be able to handle that 12th stop just fine.

and carl, i admire the photos of webb and manos (greatly). i was careful to say that i personally don't need more stops to make good photos. but really, those guys' work is more a demonstration of my point than a counterargument. they have chosen to stick with the (distinctive) way their camera system (including film/sensor) saw the scene, or at least the spirit of that vision--well and good. they choose not to raise shaded faces, and in fact frequently push them to solid black. all i am saying is that for other people who sometimes want to raise important subject details veiled in deep shadow--and it would be specious to suggest that there aren't plenty of great artworks where this has occurred, without resort to fill flash--the greater dynamic range in capture would be a real advantage.

can you really not bring yourself to agree with this? even if it is not what your current work is all about, you can't see how it might be put to good use by somebody?

Outstanding article ! This is the stuff that doesn't appear in books, but you'd pay 300 USD to go on a day's training course and the Famous Artist wouldn't be able to tell you this clearly and succinctly...

In fact it's "Great Briefing in the Sky" material: (Richard Bach, but in his period writing about aviation, before he got into soppy sloppy stuff about seagulls and things...) It's stuff that when you hear it, you assume that everybody else must have been told it at the Great Briefing in the Sky which you somehow missed...


Chris, I haven't said that more range wouldn't be useful, in fact I'd love for sensors to read up into the highlights as much as negative film. I've spent years shooting large format and ULF to make contact prints in platinum/palladium specifically because of the glorious tonal range available. A textured gray sky can fall on Zone XII and print with beautiful, delicate detail. I've been delighted to find in recent years that nearly all of that rich data in the negative is accessible to scanning. It would be great for a DSLR sensor to do this, but that's a ways off. In the meantime, my point here has been that the sensors we've already got can handle strong sunlight with aplomb.

Excellent, excellent article Carl! Thank you!

The comments to this entry are closed.



Blog powered by Typepad
Member since 06/2007