Thursday, 1 October 2009

‘Nocturne in Black & Gold: The Falling Rocket’ – James McNeill Whistler


So how’s that work for you then? How is this, a picture of a spent firework, different to your average “abstract” pseudo-art which smears paint randomly across a canvas?  Certainly, Whistler got it in the neck at the time he exhibited this (1877) from contemporaries like Ruskin, who accused him of “flinging a pot of paint in the public’s face.”

But compare this, as Ruskin couldn’t do, to the likes of Jackson Pollock around eighty years later, who really did fling pots of paint in the public’s face. In contrast to Pollock’s random profligacy with paint, Whistler depicts here a real scene (a fireworks display over the Thames), from which he selects a moment to stylise in paint (a falling rocket!) – and in that choice lies the difference between art and bullshit.

Think about why that matters.


  1. There is nothing random about Pollock. His paintings are very controlled. It is notable that nobody has successfully forged a Pollock, which would be a simple task if it were just a matter of flinging paint at a canvas.

  2. Paul

    What you really mean is that nobody has discovered whether a "Pollock" was made by some other person. Doing that would be a very difficult task. Why bother investing the effort?


  3. I have a question to expert artists.

    Can an art expert tell the difference between painting/artwork produced by a human and one produced by a machine using computer vision technology?

    For example, see the following images from the link below, where computer vision algorithms are being used.

    Artistic Vision

    Paul, I reckon that as computer vision technology advances over time (which I think it is still primitive at this stage), forging anyone's painting using machine can be achieved easily.

    I came across a video demo before about a robot that paints artwork, but I couldn't find the link for that, even I tried youtube by searching on phrases/terms such as computer-vision and painting.

    So, I doubt that an art expert will be able to tell the difference between machine generated paintings and ones produced by a human. I mean give the expert, say 20 different paintings of which he/she is told that they came from an amateur painter (robot) and a professional painter, and his/her task is to label them into 2 categories.

    The expert is not told a priori that the amateur is a machine, since he/she will thoroughly examine the pieces looking for a telltale sign of machine production artifacts that will make him/her bias. I bet that the result will probably come down to almost 50 : 50 , which is what is expected statistically.

  4. LGM, it would be difficult but very profitable if one could pull it off. People have tried but failed.

  5. Falafulu Fisi: judging by the results of Artistic Vision on the website, I don't think there is much danger of an expert being fooled. The technology as described is a step forward from Photoshop filters, which do no more than alter pixels, but it is not enough to imitate a painting. All it achieves is to create a generic pseudo-Impressionist style, of the sort that 'how to paint' books favour. The images are still recognisibly photographic.

    I doubt that software will be created that can fool the experts. A painting is made in an entirely different way than a photograph: it is built on the canvas, rather than being the result of a film exposed to light. That construction is the work of a distinct personality - the individual artist; a computer would have to possess consciousness (and an artistic consciousness at that) to be able to work as an artist. What is essential to works of art is that they are the result of human experience.

    Apart from that, I think your 50:50 prediction is flawed because art experts have a distinct advantage: they have seen and studied a lot of paintings. One develops skill by application and experience. This skill is similar in many fields: a horse-racing expert can pick a winner in the paddock, a car expert can spot a lemon by listening to its engine.

  6. Paul, may be what is described in the ArtVision is not near good enough, but I am convinced that machine of the future with advanced computer vision technology will be good in painting artworks in comparison to similar work done by humans. Not that the machine can think up something to draw out of its own (since they can't), but perhaps in the sense that given a scene to paint, such as the view of Auckland or Sky-tower from top of Mt. Eden for instance, the machine and the human will be required to paint that same scene. I believe that the 2 completed paintings will be very hard to be distinguished by an art expert, if say there were 20 different scenes altogether being used.

    I haven't read the following paper (since I couldn't find a free-copy on the net), but what is described in the short abstract is exactly what I meant.

    In this paper, we describe a painting robot with multi-fingered hands and stereo vision. The goal of this study is for the robot to reproduce the whole procedure involved in human painting. A painting action is divided into three phases: obtaining a 3D model, composing a picture model, and painting by a robot. In this system, various feedback techniques including computer vision and force sensors are used. As experiments, an apple and a human silhouette are painted on a canvas using this system.

    From : Painting robot with multi-fingered hands and stereo vision

  7. Only time will tell. However, fakes made by people are often detectable because, while they are stylistically accurate, they lack the individuality of the originals. The most successful fakes are usually variations on real works. As soon as the forger tries to create original works in the style of his chosen artist, the inconsistencies begin to show.


1. Commenters are welcome and invited.
2. All comments are moderated. Off-topic grandstanding, spam, and gibberish will be ignored. Tu quoque will be moderated.
3. Read the post before you comment. Challenge facts, but don't simply ignore them.
4. Use a name. If it's important enough to say, it's important enough to put a name to.
5. Above all: Act with honour. Say what you mean, and mean what you say.