Image Quality? Image Depth.

In the previous post I pointed you to Mike's remarks over on ToP, which, if you have not, you should still go read.

This may or may not be an accurate paraphrase, but it's what I think he said, and what I want to think about now:

Image Quality isn't a real thing, as an objective measure of goodness in any meaningful way. Photographs simply have properties, which are neither good nor bad, they simply are. There is no use to be gained by assigning subjective qualities to these properties, such as sharpness, dynamic range, contrast, saturation, and so on and so forth. These are certainly measurable things, but we should not use them to measure the quality of the image in any way.

This is a point I have harped on at some length, although not quite in these words. The image exists as it is, and these technical properties either serve and support the image, or they do not. If they serve and support the image, we could describe them as good, to be sure. However, the exactly opposite quality might serve and support a different image. Simply measuring something like sharpness yields no insight into the goodness of the image. Sharpness is neither good nor bad by itself, it is good or bad only in the context of a specific image.

So far, so good. This has all been said before. In fact, this next part has been said before as well, but perhaps not quite in this context.

Photographic properties come in three flavors, for my purpose here:

• Properties which can be altered in post, but only "uni-directionally" for lack of a better word. I can remove sharpness, I can remove color, but it is difficult to add sharpness or color when it it not present.
• Properties which can be altered "bi-directionally" in post. I can alter apparent contrast or saturation in either direction.
• Properties which cannot be meaningfully altered in post. Things like the relationship of the various objects in the frame, the expression on the subject's face, the position of the key light.

Of course what I am saying here is not strictly true. With enough effort you can do pretty much anything to a photograph in post. Further, as I have noted, the line between what can and cannot be done in post is constantly moving (for example the light field camera allows us to focus in post). In broad strokes, though, there are these three families of properties.

What does make sense is to lump the first family of properties together into some sort of named bucket. It is a fact that a camera which produces higher resolution images is a more flexible tool than one which produces lower resolution images. Is the phrase "image quality" a good one to use here? I agree with Mike, essentially, that it is not. The phrase sounds too much like "good photograph" which is a completely different thing almost entirely concerned with the third family of photographic properties.

Perhaps we can measure things like sharpness, dynamic range, color gamut, and so on (in general, the things we can alter only in one direction in post, given the tools of today, whenever that is) into something called "Image Depth" or "Imaging Potential" or similar.

The Online Photographer

Once again, Mike nails it. Read it, know it, love it. It's everything I believe in, it's everything I think you should, too.

You Gotta Know The Rules To Break 'Em

You have to know the rules to break them. I've addressed this bit of conventional wisdom before, but my thoughts have been refined on this point.

You do not, strictly speaking, have to know the rules first. What you need are two things: a grasp on the problems that are solved by the rule, and another solution to those problems.

The rule of thirds is truly about placing things off center, to create a livelier visual result. Various rules about portraiture are, mostly, about creating a flattering look, avoiding making the subject's nose look too big, and so forth. Perhaps your approach to portraiture plunges the nose into darkness, so it's OK if it's too big. We focus on the eyes, perhaps, because as animals we're most interested in the eyes of other animals. By creating or observing an equally interesting element, we might focus on that element instead of the eyes -- the main thing here is to not focus on something random and uninteresting. And so on.

Knowing the rule, therefore, is perhaps nice, but knowing what problem the rules solves is essential. You might well grasp the problem without knowing the rule in the first place.

Solve the problem, don't worry about the rule.

FSA Photo of the Week

Something a little more timeless?

The Histogram

Another favorite tool of the Internet Expert is the histogram. They like to show you the histogram of your photograph and point out features of it to explain to you why your photograph sucks.

Ignore these people.

In fact, ignore the histogram, for the most part. It can be a useful learning tool, but it's really not showing you anything that is not blindingly obvious in the image, so learn to see the image instead. By all means, compare with the histogram a few times, until you can pretty much visualize what it will look like for any image. This should take a very short time.

The only other time the histogram could be useful if for checking your images as you shoot. The histogram just tells you what your photograph looks like, after all, but sometimes it's pretty hard to tell what your photograph looks like when you're peering at the back of your digital camera. So, consider looking at the histogram instead when it's hard to see the image.

Anyone who tells you that the histogram has to start at the bottom and go all the way to the top is a twit. In the first place, they're gabbling away in technical terms about something that's extremely obvious visually -- a full range of tones, from blacks to whites. Presumably they're being technical to make themselves sound smart, or because having finally understood the histogram they think it is important. In the second place, it is simply not the case that every image needs black blacks and white whites. The tonal range has a profound effect on the mood of the image, and sometimes you want one of the other ones.

Of course you should think about the tonal range of your photograph, and what the emotional impact of it is. Adjust to taste. There is no "right answer" here.

FSA Photo of the Week

I may as well just define a week to be 8 days long.

Technology and New Ways of Seeing

In my HDR Manifesto, I proposed that a relatively recent technology, HDR and tonemapping, could be applied to create a new aesthetic vision. This has happened before.

Street photography was enabled by two things, fast film emulsions and the Leica camera. Without the fast emulsion you couldn't freeze motion in the necessary way, and without the small 35mm camera first available in the form of the Leica, you could not be discreet enough to accomplish the "street" way of envisioning the world. The essential new idea in "street" is the notion of an image which is present only for an instant, the momentary coming-together of things in motion (usually people) in a fortuitous way to be an image worth shooting.

Prior to "street" as "street" we have many urban scenes. Some of the earliest photographs made were of city streets, but they were not "street", envisioning the world instead in a painterly way: here is a street, it is picturesque, or otherwise worth recording. I will wait until the light is so, and then I will set my camera so, and I will record this street. This is a fine way of seeing and working, it produced many excellent photographs. It is how landscape done. These photographs were not "street" in the way we understand it today.

I maintain that the very idea of "street" was a new and essential thing -- before this, it had never really occurred to artists that there could exist these fleeting compositions. Things either were or were not good compositions. It is this essential idea of the instantaneous composition of people going about their business, a composition here and in the the next moment gone, that defines and motivates "street". This idea would have been difficult to even conceive without the Leica and without the fast film emulsions of the early to mid 20th century.

In the same way, creating an image without a strong key light is largely inconceivably to photographers of my era, photographers who learned prior to the techniques of HDR. Most practitioners of HDR are simply in love with the process and the look, they are creating images which please them are their cadre. This is fine. They are, I think, missing what the truly have in their hands, which is a technology that truly enables a new way of seeing, a truly new aesthetic, in the same way that the Leica enabled the idea of the instantaneous composition that is "street."

A Manifesto for HDR

I was pondering my remarks in the previous post, talking about new ideas and new ways of photographing old things. Naturally one cannot really give a firm example of such an idea, since it is by definition something that hasn't been thought of yet. You can talk about past ideas, such as street photography, which perhaps I will in the future.

In any case, what came out of my fevered brain was the idea of using HDR to create a truly new but undeniably photographic vision. It's not something I want to do, and I strongly suspect that I would hate the results. I am part of the old guard, this is inevitable.

I submit to you a manifesto for a new aesthetic built around HDR.

While you read it, I want you to imagine an HDR image of Half Dome, that reveals it completely, without deep shadows or strong highlights. An image with the strong contrasts of Ansel Adams, but without the shadows. An image that renders Half Dome as if it were a hyper-detailed architect's pencil drawing of a mountain shaped structure. An image which conveys the hugeness of Half Dome, its power and majesty, by revealing every detail of its forms and textures completely. In particular, I want you to imagine that, somehow, the usual muddiness and visual confusion of such HDR images is not present, because of an extremely carefully selected point of view, and very careful processing.

MANIFESTO

We propose to do nothing less than to destroy light in photography. Not to literally eliminate it, but to eliminate its tyranny over the photographic image. We choose to reveal subjects, not to conceal. We choose to strip away, as far as possible, the shadows and the highlights, to nakedly reveal the structure and form of our subjects. To reveal form by placing texture against texture, color against color, tone against tone, rather than through the modeling effects of a strongly directional light.

Our ideals are the engineering drawing, the blueprint, the exploded view, the architectural plan. Ours is an era of technology, we choose to embrace the visual idioms of technology. We choose to fully reveal the structure and form of our subject, and by doing so, to fully reveal the idea of the subject, and our relationship to the subject. The form without the idea is of no interest to us.

The obliteration of light should never obliterate form, instead it should reveal and clarify form. This is no easy task. It requires careful attention to every detail, it requires a new way of seeing and thinking about imaging. It requires careful application of technology. The techniques of HDR are one way to realize this aesthetic, but there are other ways. One might also choose to reveal structure and form with many light sources, or with very long exposures, perhaps.

In all cases, the dominance of the directional light shall be crushed.

Have all the photographs been taken?

Have we come to a world in which all photographs have been, in some meaningful way, taken? Certainly we have a lot of pictures of the Eiffel Tower and the Golden Gate Bridge. Since the moon is tidally locked, there really only are a couple dozen pictures possible of the moon. Do we need more Ansel Adams inspired mountains?

I don't mean to tell you to stop taking redundant pictures for your own pleasure. Indeed, I think it a fine pedagogical device. Making it your mountain rather than Adams' certainly may satisfy some personal desire. There is nothing wrong with that. Decorating your home with imitation Avedons isn't a bad thing to do at all, nor is selling them. These are fine things, but what they are not is new or original.

The question is, though, what kinds of things can we do that are actually new work? Can we, in fact, do fundamentally new work? The camera is hampered by its perfection, here. Any painting of the Eiffel Tower is arguably a new work, but another photograph of it is most likely indistinguishable from at least one other that has been taken. Obviously if we can take photographs of new objects, these might be new work. Can we take photographs of old objects in new ways?

New objects certainly include people. We keep making them, so there will always be the possibility of a new portrait. People invent new fashions, new fads, new buildings and cars, so probably "street" will always have the potential for fresh and interesting work as it both records and comments on the current state of mankind (or something similarly fatuous). These works of man, buildings, streets, factories, homes, bicycles, and toasters continue to be created, to age, to decay and rot. There will always be something to be shot here, surely. Imitating an existing photograph with new things in it isn't particularly new, but perhaps there's some novelty there. Creating a new photograph around a new idea having to do with the new things, well now, that would be something, wouldn't it?

What about old objects? Can we take a fundamentally new photograph of Half Dome, of the Empire State Building, of the Grand Canyon? Perhaps. Our attitudes and ideas change and evolve over time. It may be that a photograph of Half Dome might be able to express and clarify a more modern idea of Half Dome, or Yosemite Valley, or the planet we live on, somehow. It will not succeed by merely aping Ansel Adams, although it will surely fail if it does not acknowledge Adams work. It will probably only work when a photographer with a genuinely new attitude toward Half Dome photographs it. This in itself is a job and a half, since Adams has long told us what to think about Half Dome, hasn't he?

What isn't new is imitations of old work. The river of photographs in which we now live makes the old work extremely thick on the ground, avoiding imitation is increasingly difficult.

In the early days of photography you could stick your lens anywhere and create something new. Later, we had arguments and rebellions about sharpness, color, lighting. Artists with various ideas formed cliques and created art based on specific technical and aesthetic ideas. These, in the large, in broad and general terms, have been mined out. Subjects have, again in the large, been mined out.

To create new work now requires a greater clarity of vision than ever, whether that vision be conscious or unconscious. To be new, a photograph must be motivated by a laser focused idea, must be fully and powerfully expressive of that motive.

Painting, Polaroid, and Digital

Edwin Land (the Polaroid guy) gave a presentation to the Royal Photographic Society in London, in 1949, in which he said on the topic of painting and drawing:

... the artist initiates his activity by observing his subject matter and then and then responds, as he proceeds, to a two-fold stimulus: the original subject matter and his own growing but uncompleted work.

The important point here is that the incomplete drawing or painting is part of the process. Being able to see the work as it is created is fundamentally and powerfully different.

He then went on to talk about his Polaroid process, and how it allowed a similar mode of working for photographers. You shoot, you see, you shoot more. Interestingly, he is completely unconcerned with technical details here. In Land's mind, you're working for aesthetic perfection, not technical. Each image is a reaction to the previous images taken, as well as to the subject. You're working to simultaneously develop and realize an artist impluse.

This struck me like a bolt a lightning, albeit a very small one. I am as opposed to chimping as anyone, but I still do it when I'm out shooting digital. It's pretty hard to avoid. However, I find that I do it as much for artistic reasons as technical. I pretty much trust my camera to get the exposure right, and you can't really see focus on the back of the thing anyways. Sometimes I play with exposure compensation a bit.

Mostly, I chimp to see what I've got, and what I should do next. Mr. Land has pointed the way, here, and I intend to do this more, and with more intent.

The next time some joker tells you tape over the screen on the back of your camera, tell said joker to get stuffed. I do suggest you turn off the histogram display and all the little technical details, though, and just look at your picture.

What is Photography?

Another post on this topic, from a slightly different angle.

If we try to define a photographer by any specific act, we get in to trouble. If the photographer must press the button, are photograms then not photography? If not, then Fox Talbot and Man Ray are abruptly no longer photographers, which seems problematic. Photograms are made by placing objects on light sensitive materials, exposing to light, and then developing. So perhaps the fact that they developed makes them photographers. In this case, we circle around to photographers who pressed the button, but did not develop, or did not print. Is Henri Cartier-Bresson not a photographer, now?

For any action or collection of actions we can set down, we can probably find or at any rate imagine a bona fide photographer who fails to comply with our list.

I suppose one might make some headway with a list of a dozen actions or so, and claim that anyone who performs three or more of these is a photographer. Possibly. Such a wide net strikes me as likely to ensnare some people who are certainly not photographers, or at any rate not the photographer of a specific image we have in mind.

The point is, as usual, that photography is not an easy thing to get our arms around. There simply isn't a pat definition. We know it when we see it, perhaps? Except that we don't, what with the occasional spates of people who make photorealistic pencil drawings and whatnot.

FSA Photo of the week

Here ya go:

Correct White Balance

Another in my ongoing quest to stomp on conventional wisdom where it is wrong.

A canard one sees from time to time online and, I suppose, in camera clubs and other places where the amateur with a little knowledge congregates, is the idea of "correct" white balance. This generally means that white objects are rendered as white on the screen or print. This is a classic case of a technical detail that is easy to understand and, in this age of digital, easy to adjust. The amateur with a little knowledge loves these things, they are handy stand-ins for actual knowledge and understanding, like shutter speeds, EVs, and so on. Since they can see it and you cannot, they get to feel superior and correct you. They can easily edit a copy of your picture, and show you. It's one click in most photo editors.

White whites are not, however, "accurate" white balance. The normal reason white objects fail to render as white is because the light falling on them is colored. A white shirt in shade will be a bit blue. The same white shirt under incandescent lights will be a little yellowy orangeish. The camera sees and faithfully records this. The white balance settings in the camera will try to make some effort to work out the color of the ambient light and make white things look whiter, but this is a somewhat haphazard affair. White whites are not always even pleasing, or flattering, white balance. Getting the whites white is, at best, a reasonable starting point and at worst a purely arbitrary measurement that weenies can use to feel smart.

Our eyes and the massive amount of brain we devote to seeing don't really care. We see white objects as white, almost no matter what the ambient lighting is. We mentally subtract the blueish tones of shade, and the yellowy tones of incandescent light. We do this when we're looking at photographs as well, which is why the angry yammering about "incorrect white balance!" is so confusing the new neophyte. We literally have to train ourselves to see white balance in photographs.

In reality, white balance is simply an effect. It can no more be correct than contrast, saturation, or any of a number of other effects we can apply to an image. It can certainly be ugly, it can fail to support the image, it can be visually confusing. It certainly affects the way we feel when we see a photograph. Like contrast and saturation, when a set of images are presented as a collection of related frames, we should try to make the white balance more or less match image to image, to help support the coherency of the set as a set. We can make the white balance more blue or more yellow, to subtly alter the way the viewer feels when the view the image.

A photograph obviously taken in shade, with the white balance adjusted to render white things a pure white arguably looks a little off. Our visual cortex sees the shade and tries to subtract the blue, and then things are just all wrong. We expect a slight blue tone to the whites, and miss it when it is gone.

When someone tells you your white balance is incorrect, by all means take a look. If all the people look orange, they might have a point. Your white balance isn't wrong, but it is ugly. If, having taken the time to examine the colors, they are where you want them, the correct response is "go pound sand, weenie."

FSA photo of the week

Here it is, a day late:

What is Photography?

Mike J over on ToP has a wonderful series of blog posts discussing Doug Rickard's book of photographs drawn from Google's Street View and some surrounding ideas. The most important one, I think, is here. If that one interests you, then read the posts before and after for a little ways as well, there's some more commentary and ideas there.

This all swirls around a single central issue which is "what is photography?" I don't propose to answer that question, but instead to talk around it a bit. These are interesting ideas, I think it's important to wrestle with the question but it is not important to devise a definitive answer. Indeed, I think there is no definitive answer, and trying to produce one is a waste of effort.

There are things we certainly recognize as photography. If I load some film into a camera, point that camera at something, press the shutter release, develop and print that film onto a piece of photographic paper, I have pretty much undeniably made a photograph. What I have done is photography. There are, similarly, things we recognize as certainly not photography. Fishing, for instance, is definitely not photography. Painting a picture, applying pigments to canvas with brush and trowel, is a more interesting case. It is certainly not photography, although it does produce images and contains some of the same elements (selecting a frame, objects to place in the frame, and so on).

Between the two there appears to be a spectrum, a continuum of photographic techniques combined with techniques of drawing and painting. We probably all have ideas about individual works in here: that's not a photograph, it's a digital painting; that's a photogram and I consider that photography. Any two people will probably disagree on some works, and agree on others. This is ok, I can see no particular benefit to having clear and universally agreed-upon lines that delineate photography from digital art from painting from whatever else you might imagine. What is important is that we understand that there is a continuum in which we might need to agree to disagree.

Just as there is a spectrum of possibility between painting and photography, there is a spectrum of possibility between photography and editing. If I edit a set of 227 photographs into a portfolio of 7 images, that is pretty much just editing. If I select from a single person's flickr stream their 10 best images, that is probably pretty much editing too. If I select 10 outstanding images from all the photographs posted on the web, it begins to get fuzzy. As in the previous case, you may well have a personal opinion to the effect that this is, or is not, photography or editing. That's ok. What we need to recognize is that at this point we're entering an area where opinions may differ.

Someone who thinks that photography is about "selecting an image" whether it be from the real world, from Street View, or from the web, may reasonably consider this editing-like activity to be "photography" and there are legitimate reasons for considering it so. If you, personally, think that a shutter release button-press is necessary for photography, so be it. You disagree.

Walker Evans: American Photographs

In keeping with my track record of reviewing books only decades after they are published, I will now record my thoughts on the 75th anniversary edition of Walker Evans' seminal book, American Photographs.

This book contains 87 photographs, in two sections. They are presented with no accompanying text whatsoever, only a number referring to a very brief descriptive text for each photograph at the end of each of the two sections. The reproductions are quite nice, and the book is incredibly cheap - less than $US25 at the moment. It might be one of the best values out there.

There are two essays included with the current edition. The first, on the making of the current edition, is a fascinating glimpse into the problems inherent in re-making a book of photographs after 25 years. Technology advances, and photographs are lost or harder to obtain. The second essay was included in the original edition, and is a discussion of the work and the photographer, written by a friend of Evans, Lincoln Kerstein. I found it somewhat labored and not particularly interesting. It is probably more of interest for the snapshot it gives us of how Evans was viewed in his time, than of much current interest.

I have mostly been familiar with Evans through is FSA work, and the work he did nominally for the FSA but in reality for the book Let Us Now Praise Famous Men (which was a collaboration with James Agee). The current volume gives us a much broader view of the photographer's work. These photographs were, for the most part, selected for the book by Evans himself to accompany an exhibit of his photographs in 1938, so we're probably still not looking at a fully representative view of his work.

The first thing that struck me, looking through this work, was how relatively poor it was. This is not to say the work is bad, it certainly is not. However, it fares poorly when stacked up next to the towering, gigantic, monumental work he did for the book with Agee. The work is still excellent and well worth your time to look at carefully. The second thing that struck me was the breadth of Evans' work. There are at least two different photographers in Walker Evans, the view camera man and the 35mm man. While it's not a perfect correlation, the work with the smaller camera is much looser, in some sense. There's a number of images that flirt with "street" while being, ultimately, documentary. The view camera work is generally more formal, each image having a very deliberate feeling of structure.

The subject matter is all over the place. The title of the book is apt, Evans work here seems to have been an honest stab at capturing the breadth of the nation, in some interesting sense. There are people, street scenes, architecture, still lifes, and strange things drawn, painted, or glued to walls. There are very few landscapes, which is telling in a book of this title. Evans seems to have felt that America is in her people, their belongings, their buildings, their streets.

Buy it, keep it near. It's superb.