In 1989 there appeared, in Roger Norrington's EMI version with the London Classical Players, the first performance of the Symphonie fantastique in modern times that came close to the sounds of Berlioz's orchestra. Though this did not, of course, invalidate performances by our own contemporary orchestras, it did reveal a huge amount that could otherwise only be guessed at. In the brass section alone, it was revelatory to hear the effect of correctly stopped horns, differently crooked instruments, valveless trumpets adjacent to cornets a pistons, and narrow-bore trombones, not to mention the fearful ophicleide, together substituting (as Norrington wrote) for the habitual blare of a Berlioz brass section ''a kaleidoscope of colours''. Norrington's performance was vital, sensitive and exciting, building an interpretation on the sounds of his players and not resting on them. Now he is joined by another fine Berliozian, John Eliot Gardiner.
In his preface, Gardiner sets out the issues again, and claims to ''recreate as closely as the available documentation permits the sound and atmosphere of the first performance''. The essential documentation is contained in Nicholas Temperley's edition of the score for the New Berlioz Edition, where he points out that a definitive version is not possible since different sources show changes. One such change is the cornet obbligato in the waltz, which certainly did not feature in the first performance: Gardiner includes this surely impermanent afterthought of Berlioz's, Norrington not. One may also point out that the instruments of Gardiner's orchestra, wind and brass very properly identified in the booklet, are often of later date than the symphony, even when they are not modern reproductions.
Nevertheless, even if we have here only an approximation to the sounds of the first performance, that is a good starting point. Gardiner's performance is in some ways sharper than Norrington's, and more insistent on detail. This can lead to over-phrasing, though he does almost nothing that cannot be justified from Berlioz's intricately, often oddly, marked score (a small misjudgement is the anticipation of a rallentando in the Ball, at 4'11''). He can also test audibility with his pianissimos: the final G in the double-bass phrase in the work's twelfth bar has all but vanished (1'08''). Both conductors have trouble with the natural horn in E flat later at bar 50: the written F (sounding A flat) which opens the phrase cannot be heard, even though Berlioz has anticipated trouble by marking it p to the woodwind pp and strings ppp (the note has to be stopped down from F sharp by hand, with consequent muffling). Norrington produced a splendidly raucous sound in the Dies irae with his ophicleide borborygms, which Gardiner caps with the use of a serpent (Berlioz feared it might be out of tune: not here). Gardiner uses more sonorous bells, effectively, and also more precise timpani in the ''Scene aux champs'', rather less so; however, Gardiner's placing of the distant oboe at the start of the movement is exquisitely judged.
Both performances are of endless fascination and enjoyment. Norrington is, in general, more concerned to use his recovered instrumental sounds to shape a performance of the kind with which we are familiar, cleaning everything up and presenting it afresh. He is thus more attentive to a smooth line for the idee fixe, to a steady, menacing tread for the ''Marche au supplice'', to mystery and atmosphere in the Ball and the opening of the Witches' Sabbath. Gardiner is perhaps more interested in the kind of performance with which Berlioz might have startled his audience that December night in 1830. So he plays the music with an extra emphasis on sudden flicks of phrasing, an extra abruptness in the stamp of a rhythm or the snap of an interrupting chord, a concern for the extreme.
Who can tell what the instruments really sounded like (for the whole question of playing traditions and techniques is so elusive of answer), and who knows how Habeneck really conducted the first performance, ''pas irreprochable'' as Berlioz found it? What matters is that we have, to set beside other well-loved performances with a modern orchestra, two that take us very close to the sound-world out of which Berlioz created a completely new kind of music.'