Composers do not always know what is best for their music; they may hold copyright in the dots, but not, fortunately, a monopoly on their interpretation. It is naive to presume that any recording involving the composer will automatically set a qualitative benchmark: apart from anything else, performance – especially early recorded performance – does not spring fully formed like Athene from the head of Zeus, but is always subject to the practical vagaries and contingencies of players' availability, schedules, temperament, (lack of) rehearsal time and the technical limitations of the young recording medium. Yet it is too easy to overplay the scepticism. All intelligent performers know the importance of internalising the music, and appreciate that satisfying sense of the music re-emerging from deep within them, rather than feeling their hands and body functioning as mere thoroughfares between brain and instrument. And from the evidence of many early recordings of Ravel, for the best musicians working with and around the composer – unlike us – the sheer cultural currency and ‘presence’ of the music seems largely to have evaporated the need to deploy any self-conscious ‘performance practice’, and led to a directness in that process of internalisation which we have little hope of recapturing, but from which we still have much to learn.