Pixellation v JPEG artefacts

The description of JPEG artefacts and the example image given in the course notes (Photography 1 Expanding your Vision, 32-3) differ slightly from my understanding of the term. In particular, the image appears to be exhibiting pixellation rather than artefacts.

This is an exercise to deliberately degrade an image in an attempt to explore the difference.

An enlarged portion of each image is presented below.

IMG_4141 orig extract

Portion of the original image. There is minor pixellation arising from the enlargement process.

IMG_4141 artefacts3 extract

Portion of image displaying JPEG artefacts. The image is broken into squares, each of which has simplified slightly but still displays internal detail. JPEG artefacts often manifest as ‘halos’ at hard edges. It is possible to provoke stronger artefacts by sharpening or levels adjustments between successive re-savings.

IMG_4141 pixellated extract

Portion of image displaying pixellation. Each 4×4 square is a single colour and detail has been lost.

A bit of Ruff • a review of reviews

jpegs‘, a 2008 exhibition and book by Thomas Ruff was reviewed online by David Campany (2008) and Joerg Colberg (2009). The Colberg review is the shorter of the two and concentrates on the work itself. Campany uses the exhibition as a starting point to discuss found images, archives and the difference between grain and pixels in image structure.

According to Colberg, Ruff’s interest in the nature of digital images started with the WTC attack on 11 September 2001. He had taken film images which were blank when returned from the lab, and had to rely on downloaded low-resolution visual images which he describes as ‘terribly beautiful’. Whatever processes he subsequently applies to bring out the JPEG artefacts, the results have their own beauty. They force us to pay attention to the image structure as well as the image content.

Colberg admires the beauty but wonders if the concept relies too much on the technique, “What else is there?”. In part, this questioning arises because of outside influences attempting to persuade him that there is something more significant, which is not fully explained. Ultimately, he appears to accept that being ‘just beautiful’ is sufficient.

Campany starts with a discussion on the nature of found images and the way they are arranged in archives (although he seems to have a wider definition of ‘archive’ than my understanding) which is becoming more chaotic as the amount of information and speed of gathering increases (“… there is always something wild and unpredictable about the behaviour of images …”) becoming more strained as archives are digitised and redistributed through the internet.

At this point, he introduces the concept of an ‘art of the pixel’ noting that images now exist as a mass of electronic information that takes visual form as pixels. Although pixels have replaced the grain of photographic film, Campany sees a difference between them. Grain suggests authenticity, with the photographer and equipment stretched to the limit by circumstances, while pixels are seen more as a defect (although our response is changing, presumably as a result of growing familiarity with digital images). He sees Ruff’s pixels as representing the character of modern life, switching between figuration and abstraction.

Both agree that the images in this series work better in print (book or gallery wall) than on screen. I speculate that this may be because we are used to seeing JPEG artefacts in poor images online and regard them as mistakes, whereas there is an element of surprise or abstraction when seeing them in print.

Campany, D. (2008) Thomas Ruff: Aesthetic of the Pixel [online] at: http://davidcampany.com/thomas-ruff-the-aesthetics-of-the-pixel/ (accessed 21 March 2016)

Colberg, J. (2009) Review: jpegs by Thomas Ruff [online] at: http://jmcolberg.com/weblog/2009/04/review_jpegs_by_thomas_ruff/ (accessed 21 March 2016)

Exercise 1.4 – Frame

This is, essentially, an exercise in poor composition (or at least of semi-random framing). We are asked to compose an image in one-ninth of the viewfinder and ignore the other eight-ninths at taking stage. This was difficult for the first few shots as there is a tendency to automatically compose with the entire viewfinder, but I was able to do it with a little practice and without resorting to masking-off sections of viewfinder.

On reviewing the images, I have the following comments that apply to a significant subset (although not the same subset for each comment). These are illustrated with examples below.

All images were taken with a Canon G1-X, set to fully-auto mode so the only camera function under my direct control was the focal length. These images have been resized by a Lightroom export to 1000 pixels on the longest side but are otherwise unmodified. None have been cropped.

The photographs were taken in a single session. The subject matter is a derelict village hall, awaiting demolition prior to redevelopment of the site, and its immediate environs.


Because I was composing normally, but only in one-ninth of the viewfinder, I was using a much shorter focal length than usual and wide-angle distortions are apparent, particularly when I was composing in one of the corner divisions (e.g. the daffodils and blue sign in the top left corner above)


Where I was composing in the upper part of the viewfinder, the ‘extra parts’ of the image were of mainly featureless pavings which read as negative space and can be ignored. However, composing in the lower part of the viewfinder introduces more visual clutter.


Unexpected ‘extra’ image elements, cut off by the frame edge, can introduce awkward shapes, such as the front of the car in the image above.


The occasional happy accident can occur, such as the leading lines from bottom-centre and centre-right, which help us to discover and focus on the daffodils at the top-left of the image above.

I have selected six images that are reasonably well-composed (the small piece of car is unfortunate) and work as a set to describe the subject.


The upper pair describe the building overall, two blocks, one of traditional brickwork and a prefabricated timber extension. The middle row shows building details, partly walls and partly windows of the two blocks and we see that the older brickwork has worn rather better than the newer block. The lower row deals with the surrounding pavings and I have tried to mirror the yellow tape around the guard rail with the colour of the daffodils.

In the top row, the eye is taken in a horizontal line along the kerb (the signpost and telegraph pole act as ‘stoppers’ to the left) to the basin wastes in the top-right image which leads it down into the centre row (reinforced by the two vertical white lines.

In the centre-right image, the vertical lines of the brick pier relate to the right edge of the overall frame and the perspective of the horizontal brick courses lead the eye to the left and the window. The left-hand image picks up the window motif and has a vertical emphasis to keep the eye within the set.

The bottom-left image picks up the timber cladding and both images have leading lines (angled manhole cover and the paving lines previously noted) pulling the eye back into the centre of the panel.

Framing v Cropping

I am asked to note down what I understand to be the difference between cropping and framing.

My own working definition, slightly glib and deliberately circular,  is ‘Framing is cropping done in the viewfinder. Cropping is framing done or refined in post-processing’. Both words describe the act of selecting a segment of the scene before the camera, making a decision about what to include or reject (Szarkowski, quoted in course papers, p26) and arranging the principal elements within the picture space. Szarkowski (ibid.) notes that the action forces a concentration on the picture edge and the shapes that are created by it.

Of the two dictionaries of photographic terms that I consulted, Lewinski (1987) defines neither term and Bailey (1987) fails to define framing. He defines cropping as ‘Editing of a print or image, and the rejection of unwanted areas to concentrate on the main subject of the picture.’

Michael Freeman appears to adopt the same working definition as me, and the first chapter of ‘The Photographer’s Eye‘ (2007) is an extended riff on the subject. On the page headed ‘Cropping’ (2007, 20) he tells us ‘Cropping is a way of reworking the image well after it has been shot; an option for deferring design decisions, and even of exploring new ways of organising an image.’

Freeman (2011, 156) also gives a useful description of framing, ‘Framing is how the boundaries are set; the shape of the camera frame is the given, and often the first decision is how this is going to be applied to the scene in front of the photographer … The viewpoint and the lens focal length are the two main variables, followed by decisions on what will be included and what left out.’

John Hedgecoe (1994, 154) regards both actions as cropping in his glossary and uses the term ‘framing’ (1994, 28) for the compositional technique of using foreground objects as a frame to focus attention on the main subject.

Liz Wells (2000, 283) reminds us that framing also refers to the act of placing a physical frame around a physical picture. It acts as the margin between the work and the wall and delineates the edge of the picture – which brings us neatly back to the Szarkowski quote in the course notes.

I am aware that there is a debate about whether cropping is acceptable or to be avoided, but that is a topic for another day. In my opinion, some cropping is inevitable (Film or sensor formats are often not proportional to the paper size. Optical viewfinders rarely correspond precisely to the frame captured on the film or sensor) but it is best to keep it to a minimum in order to maximise the amount of information (pixels or negative area) in the final image.


Bailey, A. (1987) The Illustrated Dictionary of Photography. Leicester: Windward

Freeman, M. (2007) The Photographer’s Eye. Lewes: Ilex Press

Freeman, M. (2011) The Photographer’s Vision. Lewes: Ilex Press

Hedgecoe, J. (1994) John Hedgecoe’s Complete Guide to Black and White Photography and Darkroom Techniques. London: Collins & Brown

Lewinsky, J. (1987) Jorge Lewinski’s Dictionary of Photography. London: Sphere Books

Wells, L. (2000) Photography: A Critical Introduction, Second Edition. London: Routledge

Exercise 1.3 – supplementary

A couple of extra exercises exploring the difference between emphasising depth and suppressing it.


These are two images of the same building, taken with similar focal length and from a similar distance. The difference is that the top image is taken square-on and the lower image is taken at a corner, creating vanishing points and perspective lines, the same technique used by Atget in many of his images of the petits métiers of Old Paris (see Jeffrey, 2008: 30)


Of the two images above, the lower is cropped from the upper. This has the same effect as increasing the focal length and shooting from the same viewpoint.

Although the dominant (yellow) lines in both images are vertical and parallel to the frame edges, the upper image has some secondary perspective lines in the scaffold planks at the top of the image and some implied perspective lines joining the pads at the base of the poles. Cropping to exclude these perspective lines has removed most of the clues to the depth of the subject and flattened the pictorial frame.

Reference: Jeffrey, I (2008) How to Read a Photograph, London: Thames & Hudson


Exercise 1.3 – Line

Part 1. A selection of images using lines to create a sense of depth.

All images captured with a Canon G1-X, at the wide end of the zoom range.


Classic use of a single vanishing point (top of the blue panel and the far-centre of the road) with diagonal lines leading to it from all of the horizontals in the scene. Both kerbs, the yellow lines, building bases and various string courses. There is also an implied line formed by the ground floor arches of the building at near-right.


The same device but with a vertical context. The vanishing point is on the lowest floor, below the camera position and all the verticals converge there. The staircase handrails and strings also lead us downward.


It is interesting that converging verticals are more acceptable when looking downward than upward. We get a sense of a high viewpoint and possibly a touch of vertigo. The vanishing point is well below the bottom of the image, so there is a risk of dragging the eye out of the image at the bottom. In this case, the handrail and the seated group act as eye-stoppers.


This image combines all of the techniques of the earlier examples. There is a horizontal vanishing point near the top-centre and the converging verticals of the ground-floor columns give us a sense of height. The staircase balustrade leads us from the camera position to the centre of the image. Fortunately, the steps and handrail are near-horizontal at the bottom-right so avoid dragging the eye strongly out of the image.


A similar effect, but using a curved line.


This is an artificial example (a detail of the pedestal of Paul Day’s statue at St Pancras station), a low-relief bronze casting which uses line to give an illusion of depth.

Part 2. A set of images using line to flatten the pictorial space.

All images captured with a Canon G1-X, in the centre or long end of the zoom range.


In this case, the subject is effectively flat and the flatness is emphasised by arranging all relevant lines parallel to the edges of the frame.


The subject in this image is deeper, but presents as a series of parallel planes. By approaching it square-on and arranging the major lines parallel with the edges of the frame I have tried to eliminate the clues about separation of the planes, thus flattening the space. Unfortunately, the shelves inside the door to the left lead to a vanishing point and compromise the illusion slightly.


A similar technique, suppressing clues about the separation of planes.


Shooting vertically with a long-focus lens and arranging the tile pattern parallel with the edges of the frame we lose the impression of depth. The converging verticals of the balustrade are reasonably minor.

Exercise 1.2 – eye-tracking

An interesting point arising from the eye-tracking exercise is that I believe I view an image slightly differently when I first encounter it, compared with later viewings. On the first encounter, I enter the image (typically near bottom-centre) and search for the key element(s). The search may take a split-second or several seconds, depending on how obvious the key element it. From there, I will explore the image (or not) before returning to the key element. With a previously-viewed image, I will skip the initial exploration and go straight to the key element.

The exercises below are an attempt to reproduce my first viewing of each image.


The red dot is the single point of focus and I go straight there. Periodically, I explore the empty space to the left and above (ignoring the space at bottom-right) and return to the red dot.


The red dot is the stronger point of focus (advancing colour) and I go straight there. There is a strong line leading to the blue dot, but I tend to explore the empty space on the way back.


source: OCA course materials, Photography 1: Expressing your Vision

The paving joint gives a strong lead in to the picture elements in the top half of the frame. I look at the concrete rounded structure but quickly lose interest and return to the chair and the light-coloured can below it, where I fixate. Further eye movements (including subsequent viewings) are trapped in the cage formed by the legs of the chair.


source: Sunday Times Magazine, 28 February 2016

Robin Williams’ eye-line gives a very strong direction to the head and torso of the Oscar. The return route is down the Oscar, through both hands, the white collar, the mouth and back to the eyes.


source: Sunday Times Magazine 28 February 2016

After initial exploration, I found the Jaguar badge and number plate and the orange radiator surround, which is strongly placed just outside the thirds point. I am led up the driver’s body to his face, then down via the orange patch on the helmet and the strong orange line on the car bonnet to the badge and number plate again. Surprisingly, it is this region, rather than the driver’s face, that I always return to.