For this article, I leave the pen, or rather the keyboard, to Alain Huet.
We both share a passion for art and the desire to understand visual perception: pictorial works become playgrounds for the exploration of color.
On his website Alain offers a chromatic analysis of the paintings of great masters. This analysis will remind you of the analysis performed in challenge #4, with the colorotate tool. But his algorithm better highlights the proportions of the different colors present in a painting. Moreover, it cleverly avoids the use of a volume to represent the colors and their distribution: everything is synthesized in two images.
He explains his project here.
The painter’s palette
In front of a painting, have you ever wondered about the painter’s palette? Among all the possible colors, which ones are present in the painting? In what proportions? If this is the case, you will discover here an approach that will refine your view of works of art.
Foundations of color perception
The scientific basis of color perception has been known for a long time.
In the 17th century, Newton discovered that daylight is composed of a multitude of different colors. In the 19th century, Maxwell demonstrated that it is an electro-magnetic wave whose frequency corresponds to a specific color. His 1855 article allowed Thomas Sutton to make the first color photograph.
In the 1930s, scientific tests enabled the “Commission Internationale de l’Éclairage” color measurement to define the first colorimetrics standards, which still form the basis of today’s definitions. It is thanks to this work that the first color televisions were created in the 1940s.
A need for art lovers?
The artists did not wait for the scientists to find out that mixing yellow and blue gives green, and that arranged properly on a circle, the opposite colors harmonize well: they are called complements (see also here).
In front of a painting, the artist or color expert can easily identify the shades of color, which contributes to their appreciation of the work being observed. But what about less trained eyes? The ideal would be to have a device that maps the colors present in the image.
Manufacturers use very precise colorimeters to ensure the stability of their products (printing, paint, wallpaper, etc.). Apart from their high price, these devices are not suitable for analyzing the palette of an image. This is the justification for the website presented here.
And first of all some examples.
Claude Monet
The representation on the color circle (above, left) illustrates the use of a reduced palette with a contrast between a dominant blue-grey and a tonic orange that occupies little space on the canvas but has great expressive strength. The presentation in horizontal stripes (above, right) does not bring much benefit.
Vincent Van Gogh
You can immediately perceive the play of complementary colors, as well as the contrast between the brightness of the terrace and the darkness of the street. But do you notice that a continuity of tones runs through the work? The color circle shows this clearly. However, the multitude of colors on the circle masks certain nuances. The presentation in horizontal bands is useful in this case.
Georges Seurat
Here, too, the opposition of the complements is exploited, with greater diversity than in the previous painting, but there is little contrast (except for a few dark touches that appear as highlighting the rest of the painting): the colors occupy a larger area of the chromatic circle, but in a more restricted range of saturation.
Rik Wouters
The color circle is well occupied, but in disjointed areas, and the colors are more saturated (over-saturated with respect to the real work), especially in the reds and blues. We are dealing with a fauve !
Antoon van Dyck
In contrast, the portrait of Charles I of England uses a reduced palette and not at all complements; van Dyck plays only on the subtleties of light, which gives a feeling of serenity in harmony with the subject.
The method behind the chromatic analysis
The algorithm consists of three steps.
A good image of the work is required. It is often necessary to crop it to eliminate damaged or shadowy edges. Then the varnish of many paintings is yellowed or the picture is bad; fortunately, there are photo retouching filters that refresh the colors convincingly, even if one cannot be fully sure that the original tones will be restored.
The second step is to extract a representative palette. An image usually contains thousands of colors, and it would not be useful to display a complete list of them. It is therefore necessary to group them. The algorithm is based on the standards of the International Commission of the l’Éclairage already mentioned.
The third step is to display the two modes of presentation illustrated by the examples: on the color circle and in horizontal bands. The size of the circles and rectangles depends on the number of pixels counted in the image. Details of the algorithm are given on the site. Its development benefited from the advice of Vinciane Lacroix 😊.
It’s up to you now!
The website allows you to analyze your own images: paintings or photos, personal works or works photographed in museums and galleries. The internet is full of images just waiting to be analyzed 😉.
A Facebook page serves as a newsletter. Subscribe to it! New analyses will be published regularly.
Finally, the site offers young (and not so young) artists the opportunity to publicize their work with a short biographical note. Make the most of it!
Biography
Alain Huet is a civil engineer in telecommunications and electronics, a graduate of the Université Libre de Bruxelles. His first contact with color techniques dates back to a course on television where the standards of the International Commission on Illumination were discussed. Passionate since adolescence about the art of ancient civilizations, he has followed training courses on the subject, while pursuing a career in computer science. He has recently given lectures on engineering antique aqueducs. His new passion is extracting the palette of the great masters of painting. He has developed a chromatic analysis program, the results of which he offers on his website chromapicture.com.