Despite all the technical stuff I write on this blog, my interest remains mostly in photography, and by extension, printing. I don’t often show what I make, but I’m going to try to break that pattern. This entry will be about composition – a subject that’s evidently important, but I admit having difficulty dealing with it much of the time. Practice hopefully helps, at least in part.
The thing with composition is that on the one hand, it makes or breaks an image – it’s crucial. At the same time, it’s hellishly difficult to put into words what makes a composition ‘work’. Lots of attempts have been made, but the ones I’m aware of mostly fall short in trying to smash a complicated topic into ‘rules’ like the ‘rule of thirds‘, ‘rule of odds’ or the ‘golden ratio‘. These never helped me much, as composition evidently is a more complex topic.
Much more useful I find breaking down what composition actually consists of, and in those terms, this Wikipedia entry is actually a nice start. It lists line, shape, color, texture, value, form and space as elements of any composition in the visual arts. We could probably debate this, but for all intents and purposes, it’s a pretty comprehensive of the visual aspects of a photograph, painting, drawing or even a sculpture, garden or building. Perhaps I’d add the boundary, limits or edge to this list, as particularly in photographs, paintings etc. how the image content relate to the edges turns out to be essential.
The other part to composition, actually already hinted at when I mentioned the edges, is how these elements relate to each other. And frankly, I think there are no meaningful ‘rules’ for that. At least not as far as I can tell with what I know now – so ask me again in 20 years or so. There’s hope yet. For now, I think that every image (let’s stick to images for a bit) has its own inherent logic and set of ‘rules’ if you will. As a result, in crafting an image, the decisions on how to make the different elements of the composition relate to each other will have to be made from the ground up. The only way I see to facilitate this process is to practice, practice, practice – and to reflect (alone, or together) on examples.
Right, all the lofty talk aside, let’s look at the damn pictures alright. Expect no particularly meaningful artworks and certainly no photos with a conceptual layer underneath them. They’re just a set of digital images shot within a few hours at the same location by means of practice. My main intent here was to deliberately go through the following process:
- Identify a ‘thing’ of interest; i.e. a subject.
- Figure out what’s part of it and what not – i.e., what’s inside or outside the image.
- Determine how the various compositional elements relate to each other so that they ‘work’.
- Try and resolve any problems that arise along the way.
I was doing this together with a friend who is (exceedingly) well-trained in photographic composition. This meant I had the opportunity to bounce ideas off of him a lot and reflect together with him. As a result, I think it’s inevitable that some of his way of seeing and composing has crept into the images.
The stair shadow was the first thing that grabbed my attention. Together with my friend, we identified four or five different ways of framing the spiral stair and its shadow. This presented two challenges: (1) a lot of different and in themselves interesting things happened in and around that staircase. It’s not just one subject, but a whole slew of subjects. (2) There were some objects close by that presented problems in composition, due to dissonant shapes or colors.
In the end, I isolated the part of the shadow where it’s kinked by the angle of the wall it projected onto, placing emphasis on the open structure of the steps and the light dots this created within the shadow. I decided to make a small part of the stair part of the composition, mainly because of the holes in the steps that associate with their counterparts in the shadow. For the left vertical edge, I chose to let it coincide with the tangent of the arc of the shadow.
The shadow of the straight, central spine of the staircase formed a natural boundary on the right side, effectively making the image frame itself. Vertical positioning was somewhat less controlled; I wanted to include enough of the staircase to show some of the steps, and I wanted to keep the off-center corner of the wall a true vertical. The net result is a pragmatic compromise.
In terms of exposure, retaining sufficient detail in the surrounding white wall around the shadow was crucial. I let the shadows underneath the steps fall where they may; any texture there I did not consider as relevant to the image.
This was a problematic one. The object of interest is the boundary between the sunlit and shadow area of the image, which zigzags from the bottom center of the frame to the top right. It’s this zig-zag shape that caught my attention and I attempted to isolate it.
In framing this image, I intended to include as much as possible of the bottom section of the zig-zag shape; the gradient part that runs over the concrete floor. This turned out to be impossible without either including too much extraneous stuff at the top side of the frame, and/or without pushing the other parts of the zig-zag too far to the background, making them smaller and less significant.
Another problem was the shadow area on the left, of which I ended up having to include more than necessary. This area in the present image has too much weight and does not balance optimally with the lighter fields on the right-hand side. I suppose I could crop the left side of the image slightly to obtain a 3:4 ratio.
Exposure, again, was a matter of not blowing out the highlights on the white wall, as the detail and color nuances there are essential. The stark blue section on the left is a nice ‘gift’ to contrast the warmer tones on the right-hand side, but as said, this area has too much weight as it is. In post processing, I even slightly burned in the shadow area to make the grated well cover less distracting, as it has a tendency to compete with the subject of the image.
What grabbed my attention here was how the shadow of the fence to the left interacted with the architecture. However, framing was problematic and I did not take enough time to get it right. I ended up with the portrait orientation, but in hindsight I doubt it was the optimal solution. I suspect it might have worked better to get in closer and use a wider angle perspective on a landscape format.
What does work, is the somewhat unsettling effect of the conflicting angles in the image, which was a large part of the appeal of the scene. Nothing is quite nicely vertical or horizontal, the diagonals are at odds with each other – and yet, there is this odd compatibility between the shadow lines and the lines of the structure they project onto.
A fatal problem in my view is the large white section top right. This has no clear function, and it’s difficult to get rid of. Cropping the top would be an option, if not for the tan diagonal band pointing into (out of) the top left corner. I feel this has an important role to play, as it balances out the bottom right shadow lines. Try and crop it off, and the image starts to ‘limp’.
So this one ultimately doesn’t work quite well. I was too quick in calling it good as I fired the shutter.
This one I can hardly take credit for, because the essential choices were pointed out by my friend. These were: (1) the rhythm of the tree trunks, with the double row of trees in the foreground and the background. (2) The base of the image, i.e. the nearly solid black line, which is actually a wall that extends down to the waterline of a sluice/lock.
Initially I imagined a framing that allowed in a much wider band at the base, since the wall of the lock had a nice texture to it. However, it would have only distracted from the main theme, i.e. the alternating verticals of the trees. Also, I shot a version with more air, including the tree tops. This empty space however did nothing for the composition as a whole and only served as a distraction.
In the end, the closer crop works the best, but is still not optimal. I think this image would have worked better in a panoramic format with a higher aspect ratio. At present, the sky visible through the branches still competes with the rhythm of the trunks, which is quite low in value and subtle in contrast to begin with. Indeed, cropping this image by taking roughly 1/3 from the top part makes it work much better.
A further improvement might have been the use of a longer lens, reducing the difference in length between the foreground and background trees a bit.
I love textures like these, even if they’re a bit gratuitous, since it’s the kind of thing ‘nature’ throws at you and all you have to do is catch it. Textures like these in my experience take little effort to make an appealing image from. My friend calls things like these ‘happy little accidents’, and I see his point. Making a good photograph from them, is another matter.
Especially in terms of composition, this is a totally different ballgame than the preceding ones. Different, but not necessarily less interesting, in my opinion. Here, there are no clear lines that form an interplay. More essential in this case are elements like texture and color.
For me, there are three textures that create an interplay here. Firstly, there is the bluish-grey reptile-skin-like texture of the dried oil patches in the bottom half of the image. I chose a part of this texture that had an interesting, meandering line to it, and a usable contrast between the irregularly shaped patches and the ‘canals’ between them.
The second elements is the ochre rusty or weathered paint texture that reminds me a bit of some types of pink granite. It quite literally interacts with the first pattern by forming the canals between its patches. Moreover, it creates a nice split-color contrast: if you map out the dominant hues of the first and the second patterns, they end up nearly opposite each other, but not quite.
Finally, there is the element of the blue-black solid tar patches on the left. They aid the image by creating a visual anchor, or weight, due to their low value. By offering a visual black point, they give ‘punch’ to the image as a whole. It helps (happy little accident) that on the right half of the image, the first texture leans a bit in the direction of these more solid patches, and that the meandering line in that pattern emanates from the group of black patches on the left.
The beauty of digital of course is that I was able to boost the contrast (a lot) and saturation (a bit) in post processing to emphasize the effect, as the lighting was sort of bleak and suboptimal for the nature of this image. Getting this just right in the analog darkroom on RA4 paper would have required some serious trickery, either by pushing the film during development, or by means of supplementary masking during printing. Digital makes it so much easier to just get what you envision, sometimes.
The nearly straight angle of the waterline and the turquoise fence was what grabbed my attention here. I had my doubts about this one and it took me a while to decide to actually make this photo. We were chatting at this point, so I had plenty of time to observe the scene. My friend ended up making a photo at the exact same spot, but his photo is fundamentally different. There’s overlap in content, but the visual logic is almost literally perpendicular to my image.
The main choice here was if I thought it feasible to go with this overly obvious and bold given of the perpendicular angle. Obviously I ended up giving it a try, and in doing so, the choices were:
1: Do I place the vertical (water)line dead center and make it a true vertical? Or not? I decided against it because that felt a little too bold, and it would introduce a problem with the next choice. Also, how much of the line to include? I wanted to include as much as possible because that would help me to preserve a decent amount of the fence at the bottom as well, and it would retain more of the jagged nature of the vertical line, which in itself I found interesting.
2: How do the four fields of color relate to each other? Placing the waterline in the center, I would have ended up with a little triangle of green grass in the top right corner, and I felt that would have been problematic. If the grass was to be part of this to begin with, it might as well actually play a significant part. The brown mud and the blue water create a nice contrast; the color of the grass creates some friction. The turquoise of the fence ties it together in a way – also geometrically, by actually connecting the three vertical fields together.
In hindsight, the horizontal band of brickwork at the bottom is an unnecessary distraction. It doesn’t hurt the image too much, but it might as well be cropped off.
My friend is a sucker for Piet Mondrian’s work. We discussed this scene a bit before I made the shot. He only carried a wide angle, with which this image wasn’t possible. There wasn’t much discussion about how the image was to be framed. The top red fence is a natural end point, as is the bottom horizontal shadow. A shadow impinges the frame on the left side, and I kept its influence at a bare minimum while still allowing in as much as possible from the fences’ verticals which form the divisions of the little Mondrian-like fields.
The aluminum tear plate service lid in the bottom right corner was a problem that was just impossible to work around in composing the image. In post-processing, I might have subdued its value a bit so that it doesn’t draw as much attention. It’s easy to overdo it, creating an unnatural look, though.
Three elements interact here: the rounded shape of the quay, the trees and their sharply angled shadows, and the color contrast between the grass and the water. There was an asymmetry in the scene that kept me from framing this as a dead-centered composition. Doing so would have included more of the shaded and otherwise empty area in the bottom left corner, weighing it down and making the entire image lean heavily to the right. The image also appears to do so slightly anyway, which is mostly due to the angle of tree trunks and their shadows.
The bottom edge of the image is a bit of a compromise; the horizontal tiled path that now acts as a base is not necessary, and in fact I find the image stronger without it. However, this leaves very little space between the shadows of the lower trees and the edge of the frame. At the top of the image, I wanted to have some headspace above the arc form of the quay. What happens in the tree tops is irrelevant and I felt they could be cropped at will.
In post processing I subdued the saturation of the green grass a little. Sunlit grass in this time of year tends to produce a hue that I find obnoxious and off-putting. The saturation of the grass as it was in real life did not balance out with the more subdued blue hues in the water, so I used digital trickery to restore this balance.
This is another color and texture composition. The afternoon sun tends to bring out the texture in pavements and in this case, there’s in inherently attractive (to my eye) color contrast between the rusty iron of the expansion seam and the associated bit of loose rust at the bottom of the image, and the more bluish-grey hue of the asphalt.
In framing, I really wanted to include the right-hand interface between the iron and the asphalt with the smooth tar edge that runs vertically. On the left side, I wanted to include part of the black blob top left as a balance to this vertical shape of similar hue and value. This puts the ‘zipper’ off-center, which prevents the image from becoming too ‘monumental’. I was careless at the top edge and clipped the two bolts on the left.
I’m divided on the bottom of the image. I included the horizontal line mostly because of the shade cast by the sidewalk, creating an inverted-T shape with the verticals. But looking at it now, I doubt if that’s really such a good thing. It grounds the composition, and perhaps it works better if it’s allowed to float free from this artificial horizon.
This one is primarily about the turquoise paint job on the bridge and the waterworks in the background, and of course the combination of horizontals and diagonals. Initially I started composing this with the V-shape off-center, placing more emphasis on the water surface and the background. However, it became immediately clear that this made little sense, given the dominant interplay of lines in the foreground. The top and bottom edges are straightforward, using available elements as a frame.
The duplicate shadows are what caught my eye, in combination with the upright tree trunks. Visually this is similar to ‘blue and green lobe’ above, but without additional elements. The rhythm of the trunks in the upper two-thirds of the image is essential and I tried to optimize for this. The overlapping trunks near the left edge are somewhat unfortunate. I included no horizon in the image; it would have been superfluous. Moreover, there were too many distracting elements up there.
I mentioned the oppressive nature of green grass in this season earlier. In this image, it was so prominent that the only viable option was to get rid of the color altogether. I feel that’s perfectly fine, too, since the essence of the image is in the lines. The colors played no role whatsoever in composing the image. I think I used a blue filter in post processing to do the B&W conversion. It’s a quick & dirty job; note the flat tonality in the upper section of the grass. This seems mostly due to much of the image information being at the edge of the image’s gamut – perhaps I should have given slightly less exposure, recovering more of the shadow detail in the tree trunks in post processing. In all likelihood, some more texture could be salvaged through a multi-layer edit, but this went beyond the scope of the composition exercise I was doing here.
So, that’s it. Just some practice shots with reflection on how they were made. Hopefully you didn’t actually read all of the above. In all honesty, I mainly wrote it as a means of forcing an explicit reflection process onto myself. Doing that for an (invisible, probably non-existent) audience just helps in making it more concrete. In case you’re still here – you’re probably even more of an oddball than me, congratulations 😉