HOW AN IMAGE IS BUILT UP

Dante, in a letter to his patron, says, 'The work I have made is not simple; rather it is polysemous, by which I mean that it has many levels of meaning'. The intention of A TV Dante was to try to match his claim in visual terms, to have the richness of an illuminated manuscript combine with the directness and impact of a newspaper's front page. We also aimed to carry Dante's three levels of meaning (outlined in the same letter). The first level is the Literal, by means of the actors and the simple actions of events. The second is the Allegorical, by the addition of imagery that throws light on the text from different angles. The last level that Dante mentions is the Anagogical (meaning the mysterious resonance with existential truth). We hope that, when we've got things right, this unsayable essence inhabits the texture of the work itself.

As Dante built up his text with layers of meaning, we tried to use the screen in the same way with images veiled and juxtaposed. The most common such device is the making of a box within the screen so that the picture can refer to both an event and a larger context. The border itself can be split so that what is going on down below in Hell can be echoed by what is alluded to in the world above. Further reference can be made by the mixing of images and the ghosting of one upon another.

An example of how we tried to echo Dante's imagery with our own, and of how we constructed the picture on the screen, might help to give a way into the programme and how it may be 'read'. In Canto I Dante meets in succession three beasts representing the three main levels of Hell and progressively more serious groups of sin: a leopard (who symbolises the superficial sins of the flesh and vain desire), a lion (who embodies the sins of pride and violence) and a she-wolf (who represents the deepest-seated sins which come from hatred and envy).

Having first seen the leopard as a running animal, we then see merely the skin, filling the screen to emphasise the superficial and 'skin-deep' nature of the sins he represents. Other vanities may be evoked like leopard-skin coats and mock leopard-skin car upholstery. The lion appears abruptly in a head-on leap to emphasise the confrontational aspect of aggression. The action is dumbly repeated like the mindless monotony of violent acts themselves. The wolf is shown in endless hungry and desperate chase progressing nowhere.

Superimposed on this construction we see the customary numbered box (colour-coded according to the Canto) in which an expert commentary is given. In this case it is by David Attenborough, who in the series as a whole gives the naturalist's viewpoint of Dante's wildlife imagery.

The ingredients of the image, like much of the raw material for A TV Dante, were fairly basic. Our one library source did not have pictures of everything and the hiring of film footage is very expensive. We isolated some fur from our rather feeble leopard shots and, with the help of our resourceful video-editor, boosted the colour. Having despaired of finding an adequate ferociously frontal lion, I borrowed from a friend an old home-movie 'filler' shot of the sort that you could buy in the fifties to add spice to, say, a film of a family picnic. These few seconds of battered 8mm black and white film were coloured, stretched in time, and made into a loop to repeat throughout the sequence.

The wolf proved equally elusive and we decided to bring Muybridge to our assistance, again by animating one of his sets of locomotion photographs. We used the photos of a hound (appropriately distorted) coloured and looped so that the animal is literally running on the spot. Later in Canto I this loop (redistorted and more dramatically coloured) doubles as the Great Hound of Virgil's prophecy.

Meanwhile the soundtrack itself is building up its own layers, starting with the urban streets and woman's laughter that originally accompanied the leopard. On top of this a loop of martial music represents the lion. The wolf is heralded by wild cries and the scratch of claws, together with the menacing sound of a helicopter. All these strands are much less separable than those of the picture, since this general mix of sound is tempered to the dominant voices of actors and commentators.