OPERA / The pictures are better: Is the best Wagner staging the one in your mind's eye? Bayan Northcott reflects on changing perceptions of opera

Shortly after 6.30pm this evening, the expectant silence of Radio 3 will be infiltrated by a hushed A minor cello phrase arching up to F before falling through a strangely plangent woodwind progression to an unresolved chord, followed by a pregnant pause. But then, instead of finally settling back on the key note, the phrase will proceed to repeat itself on a higher degree of the scale - and then again, and again: at once a technical device for postponing the completion of a cadence, and a sonic emblem for the unassuaged longing that will run through Tristan und Isolde until its heroine finally expires of fulfilment in a phosphorescent B major some four hours later.

So tightly interwoven are the concerns of text and composition in Wagner's most revolutionary music drama, so obsessive and ramified their development, that it may be wondered how many listeners, once drawn in to its all-enveloping world, will feel deprived, or even notice for the duration, that they are not actually in an opera house. Some, encountering the piece for the first time, may wish they had the visual memories of at least one staging to furnish a theatre of the mind as they follow the narrative. On the other hand, those who have seen the work often enough may feel closer to the essential experience, undistracted by the theatrical perils of hard seats and noisy neighbours, undisturbed by the sight of principal singers so at variance with the romantic ideal; undepressed by that dreary scenic tradition of symbolic Modernismus that seems to have been the lot of most Tristan stagings since the war; uninfuriated by this producer's Marxist axe-grinding, or that one's determination to show that the real love tangle is between Tristan, Melot and King Mark; and so on, and on.

To suggest that operas, or at any rate some of them, may only communicate fully when liberated from the theatre might sound like armchair elitism or sheer naughtiness. After all, opera-going, at least in this country, commands wider social acceptance than ever before and the prevailing view among theorists and producers remains that opera is and must be, above all, theatre: that the exercise of theatricality alone is capable of fusing and projecting the musical, verbal and visual components of a perilously hybrid art. The trouble with this argument is that it rests on an implicit equation of the theatrical with the dramatic. Yet many Baroque operas are lavishly theatrical in their demand for spectacle without generating much in the way of drama as we now understand it, while a great deal of symphonic music sounds dramatic with no reference to the theatre.

Then, too, there have been undeniable, perhaps fundamental, changes over the last century in the way opera is apprehended. Concert programmes of the late-18th and 19th centuries may have been crammed with operatic arias and ensembles - more so than today - but if one wanted to hear a complete score, one virtually had to attend a theatrical performance. There is a famous story of Bruckner sitting, sunken-headed throughout a Bayreuth presentation of Gotterdammerung, evidently absorbed in the glories of Wagner's harmony and orchestration and suddenly looking up near the end to enquire why the female lead was on fire. Plenty of 19th-century music-lovers entertained aesthetic or moral doubts about the theatre and one wonders how many attended operas for the music, or the drama through music, rather than the staging.

Their descendants have certainly had much to gain from the 20th-century media explosion; doubtless a substantial proportion of the opera public today derives its primary enjoyment from recordings and television relays with comparatively infrequent visits to the theatre. But television tends to miniaturise; to substitute a factitious sensation of audience feedback for the live reality. And recordings negate that feedback altogether, focusing attention the more firmly on a truism doubtless unwelcome to many proponents of total theatre: that while much of the repertory has survived decade after decade despite the most ludicrous plots and theatrical inconsistencies, there is no single great opera that is not self-sufficient as a musical structure, or sequence of structures. And such perceptions raise in turn the naughtiest thought of all. Could it be that the increasing exhibitionism of certain opera producers - about which we have heard so boringly much in the musical press over the last couple of decades at the expense of intelligent comment on the works themselves - reflects an unconscious fear that their very function could now be in question?

It was a belated function in the first place. Until well into the 19th century, opera stagings were mostly knocked up informally between librettists and stage managers and doubtless depended heavily on the innate theatrical sense of their singers. In Munich and at Bayreuth, Wagner co-opted designers and machinists but essentially produced his works himself. And 40 years on, Mahler was still mounting the standard repertoire at the Vienna Opera in the same way. The idea of the producer as one brought in to subject music, text and design alike to a specific theatrical interpretation is new; newer still, the notion that such an interpretation ought to challenge all received views of an opera and to present it 'critically'.

Where a stage work demands an elaborate mise en scene or a cast of thousands, somebody has got to sort it out, of course. And apart from a few authenticists seeking to reproduce 18th-century theatrical conditions, no one would probably argue at this late stage in our culture that the operatic producer should, or even could, be abolished - any more than they might claim that a recording could ever quite reproduce that special collective feeling of sitting in an audience totally absorbed by a live theatrical event. And where recent producers have been prepared to acknowledge the paramountcy of operatic scores - to work through, rather than against, the music - memorable presentations have continued to happen. One thinks, at random, of Peter Hall's magical Glyndebourne vision of Britten's A Midsummer Night's Dream, the uncanny audio- visual interplay of bright and dark in Peter Stein's Welsh National Opera staging of Verdi's Otello, or Graham Vick's marvellous handling of the complex ebb and flow of Berio's Un re in ascolto at Covent Garden.

But it is rather difficult to recall Wagner productions of comparable quality. And, as one anticipates tonight's Tristan from this year's Bayreuth Festival - with Siegfried Jerusalem and Waltraud Meier under Daniel Barenboim - maybe it is not so difficult to work out why. The little girl who famously remarked that she preferred radio to television because the pictures are better could hardly have imagined a more apposite example. Not only does the score of Tristan seem to subsume in precise technical procedures every emotional gesture and psychological shift of its principal characters, but to contain the entire temporal and spatial dimensions of the drama, as instanced in the extraordinary pacing, for instance, which compresses the few violent actions into mere seconds, or the strange way this most internal of dramas is constantly counterpointed - through off-stage voices, hunting horns, sea sounds - by intimations of vast horizons.

Consummate man of the theatre though he was, even Wagner seems to have harboured doubts whether any mere theatrical presentation could remotely match his powers of evocation. For, in the middle of composing Parsifal, he suddenly remarked: 'Oh, my heart sinks at the thought of everything to do with costumes and make-up, and when I think that characters like Kundry are going to be personified on a stage, it immediately puts me in mind of those dreadful artists' balls, and now I've invented the invisible orchestra, I'd like to invent invisible staging, too.' He was joking, of course. Or was he?

'Tristan und Isolde', Radio 3, 6.30 pm today