Gareth Hawker, ArtWatch UK Journal’s Picture/Photography Analyst, makes his living as a portrait painter. In order to reproduce some of his paintings he has taken a keen interest in large-format inkjet (giclée) printing. To this end he has learnt both how to take photographs which are highly accurate in colour and how to print them with comparable accuracy.

Gareth Hawker writes:

Museum websites now provide an unrivalled resource for the student of paintings. Images may be called up with a speed and ease which would have been unimaginable only a few years ago. But how accurate are these images, and how far can they be relied upon? You might imagine that museum websites would all be offering an answer to these basic questions, that they would be proud to explain why the visitor should have confidence in their images, but the only museum I have come across which does so is the National Gallery,  London. It provides thorough information about its images on this page:

http://www.nationalgalleryimages.co.uk/ImageInfo.aspx

If you want to know the colour of any area of paint on a painting from the National Gallery you can find it out very simply. Download a free colour reader like the one called Pixie (http://www.nattyware.com/pixie.php). Place your cursor on the spot which interests you, and read its colour specification. (See fig.1) This specification is expressed in terms of an internationally recognised scale of colour. But what if your screen is inaccurate? For example it might make all the colours look too red or too green. No need to worry, the inaccuracy in your screen affects what you see, but it does not affect the readings. They will always be accurate no matter how wrong your screen may be. Even with a monochrome screen you can take accurate readings. If you wish, you can compare them with readings from other paintings which have also been accurately photographed.

If your screen is correctly adjusted (calibrated) you will be looking at almost exactly the same colours as if you were in the Gallery itself. However this would be an unusual situation. Even full-time colour professionals, who spend considerable sums on screen calibration, are unlikely to have a screen which is totally accurate. You still need to see the real paintings if you want to experience the colour with total accuracy.

In principle, you could also display the colour very exactly by printing the digital file on your colour printer. However setting up a printer for accurate colour printing is a very onerous task indeed – one which very few people, even professionals, do correctly. So, even though the colours may be accurately specified in the digital file, they are rarely seen as accurately when they are displayed in print or on screen. Also the colours in the digital photographic file can never be absolutely correct. They are only accurate within certain tolerances. The webpage mentioned above deals with these tolerances in detail. None of this detracts from the value of such a substantial improvement compared with the old film and plate methods of recording paintings. These readings are far more accurate.

In contrast to the National Gallery’s site, most museum websites show photographs which, regarding colour, are little more than rough approximations. Many museums seem not even to be fully aware of what such accuracy involves, and often make a beginner’s mistake. If the colour of a photograph looks wrong when a photographer examines it on his screen, he will make adjustments in Photoshop until what he sees is acceptable in his judgment. This is useless as far as objectively accurate colour is concerned. However this practice, known as ‘visual editing’, is widespread:

http://www.cis.rit.edu/museumSurvey/documents/Benchmark_Final_Report_Web.pdf (p57)

It is far preferable to adhere to international colour standards. Accuracy does not then depend on personal judgement, but on objective measurement. This means that each time you photograph the painting you need to photograph a standard colour chart next to it.

In order to check the photographed values you can read them with a colour-reader (like the Pixie mentioned above), then compare them with known values. These known values are published. They are readings taken direct from the original chart using a spectrophotometer (- a light meter which splits light into bands of different wavelength, then measures the light intensity at each wavelength band).

If the values match (within defined tolerances), this means that the photograph of the painting is as accurate as the equipment can produce (fig. 2).

There are immense practical difficulties in photographing a large collection, so perhaps it is not surprising that other Museums have not caught up with the National Gallery in this respect. Most of the pictures on the website of the Metropolitan, New York, come as a disappointment if you arrive having just seen those on the National Gallery website. Not that New York is lacking in expertise. Some of the most advanced museum photography in the world is being carried in the USA. The high standards which characterise this work are exemplified in the following pdf which is concerned with photography at the National Gallery of Art, Washington DC:

http://www.art-si.org/PDFs/TechnicalReportNGA_low.pdf

This high-quality work has not yet had a major impact on US museum websites. For example, the Metropolitan Museum, New York, offers no statement comparable to the one on the National Gallery website – no claim that its colours are accurate.

The amount of detail, too, on the National Gallery’s site is far superior. When you zoom in you see more detail, whereas, when you zoom in on pictures on the Met’s site, you often see no more detail, just the same, low-resolution image enlarged and blurred (fig. 3). (A number of museums have developed zoom systems for their websites. These merit separate discussion, but it may be worth mentioning one of the most impressive, Birmingham’s Pre-Raphaelite Online Resource. There is also Google Earth, which has produced a spectacular zoom representation of 14 paintings in the Prado. However, downloading the free Google Earth software and learning how to use it does take some time.)

The National Gallery’s full-screen zoom certainly lives up to its publicity and must count as one of the easiest to use:

http://www.culture24.org.uk/art/painting+%26+drawing/art70531

Is the National Gallery’s virtual gallery perfect? No. Some imperfections are inevitable given the limitations inherent in the medium. For example the range of colours available on a normal computer screen is not large enough to display very strong colours correctly. This question is discussed on the Gallery’s own webpage mentioned above.

Perhaps the most striking imperfection is that there is a limit on the number of pixels which may be used to display a single painting. This means that a large painting cannot be shown in as much detail as a small painting.

There are limitations, too, in the level of proof offered that the photographs are accurate. The National Gallery Company Online Picture Library will supply digital files, but these include only the Kodak Colour Control Patches and the Kodak Greyscale reference charts which are intended to help printers. More colour patches would provide a stronger demonstration of accuracy. But the webpage does assure the reader that, “Each image has been individually calibrated using the Gretag Macbeth 24-patch colour rendition chart.”

When considering these imperfections, it is easy to forget what an extraordinary achievement this virtual gallery represents. Think of all the glossy colour art reference books, all the reproductions made for hanging on walls, all the posters. Not one of them is as accurate as the images on the National Gallery website. Even if, by some remote chance, one of those reproductions elsewhere did turn out to be accurate, it would be extremely difficult to verify its accuracy. All sorts of complicated measurements would be needed – with light meters, spectrophotometers and so on. In contrast, if you wish to verify the National Gallery colours you can order a digital file which includes an image of the painting with Kodak colour patches alongside it (fig. 4). You may then place your cursor over the patches and check whether your readings match the published values. If they do, this proves that the photograph of the painting is accurate in colour (within specified tolerances). No previous photographs, from a public gallery, have come close to this level of colour objectivity nor been so available for scrutiny.

Comments may be left at: artwatch.uk@gmail.com

Printer-friendly PDF version of this article