The freshness of this set is remarkable. You do not have to listen far to be swept up by its spirit of renewal and discovery, and in Pierre-Laurent Aimard as soloist Nikolaus Harnoncourt has made an inspired choice. Theirs are not eccentric readings of these old warhorses – far from it. But they could be called idiosyncratic – from Harnoncourt would you have expected anything less? – and to the extent that the set gives a shock to received ideas it is challenging. It does not seek to banish all conventional wisdom about the pieces, but it has asked a lot of questions about them, as interpreters should, and I warm to it not only for the boldness of its answers but for finding so many of the right questions to ask.

These are modern performances which have acquired richness and some of their focus from curiosity about playing styles and sound production of the past. What can be deduced about the likely nature of mass, weight and orchestral perspectives and how the musical language was spoken from what is known of instruments and performance practices in Beethoven’s time? Harnoncourt offers some answers that will be familiar to admirers of his Teldec recording of the symphonies (11/91). He favours leaner string textures than the norm and gets his players in the excellent Chamber Orchestra of Europe to command a wide range of expressive weight and accent; this they do with an immediacy of effect that is striking. Yet there is a satisfying body to the string sound, too. From all the orchestral sections the playing is of the highest class – the woodwind and brass often pungent, the thwack of the timpani leathery and distinctive in tutti passages (and the player of them relishing the big solo moment after the cadenza in the first movement of the Concerto No 3 in C minor). The performances gain an edge from all this which has nothing to do with a ‘period’ stance but everything to do with what I take to be Harnoncourt’s objectives: to regain the freshness and force of what was once new, to recover the qualities of exhilaration and disturbance that these works possess.

I also like very much the way the playing seems to have recourse to eloquence without having to strive for it, and that is characteristic of Aimard’s contribution as well. Strong contrasts are explored and big moments encompassed as part of an unforced continuity in which nothing is hurried. Melodic values are sustained to the full, yet even when the playing appears at its most relaxed it is moving forward, alert to what may be around the next corner. The big moments do indeed stand out: one of them is the famous exchange of dramatic gestures between piano and orchestra in the development of the E flat Concerto’s first movement (at 10'55"); another the equally dramatic but very different exchange when the piano re-enters at the start of the development in the first movement of the G major Concerto (8'21"). At these junctures, conductor and pianist allow the gestures to disrupt the rhythmic continuity to a degree I don’t remember previously encountering. (And there is another instance at Aimard’s very first entry in the B flat Concerto.) Over the top? I think not, but risky maybe, and if you have strong views as to what the rubrics permit in Beethoven, or have swallowed a metronome, you may react strongly. For make no mistake, Aimard is as intrepid an explorer here as Harnoncourt – by conviction, I am sure, not simply by adoption. I find him personable and persuasive, as well as abundantly capable of firing up the orchestra to make things happen as much as they and the conductor inspire and set the scene for him.

Technically, he is superbly equipped. You notice this everywhere but perhaps especially in the finales, brimful of spontaneous touches and delight in their eventfulness and in the sheer pleasure of playing them. I need to single out the finale of the Emperor, which tingles with a continuously vital, constantly modulated dynamic life that it too rarely receives; so many players make it merely rousing. And among the first movements, I must mention that of the G major Concerto as a quite exceptional achievement, as I see it, for the way Harnoncourt and his soloist find space for the fullest characterisation of the lyricism and diversity of the solo part – Aimard begins almost as if improvising the opening statement, outside time – while integrating these qualities with the larger scheme. It is the most complex movement in the concertos and I cannot remember when I enjoyed a version of it at once so directional and free as a bird.

The first movement of the E flat Concerto is nearly as good, lacking only the all-seeing vision and authority Brendel brings to it, and perhaps a touch of Brendel’s ability to inhabit and define its remoter regions. In general, Aimard imposes himself as a personality less than Brendel – I have been revisiting his sumptuous set of the concertos with the Vienna Philharmonic and Rattle while getting to know this new one. In spite of being different exercises, their distinction touches at several points and is comparable in degree. What Aimard doesn’t match is the variety of sound and the amplitude of Brendel’s expressiveness in the first two concertos’ slow movements. Exactly how you hit largo in the C major Concerto (No 1) while conveying two slow beats in the bar, not four, is tricky; but it is immediately evident that Brendel is moving (and keeping moving) in a much richer interior world. These magnificent early achievements are no less characteristic of Beethoven than his later music; the slow movement of the B flat Concerto (No 2) is another high point of Brendel’s set, and by the side of it, inevitably perhaps, Aimard’s version seems plainer.

Balances are good, with the piano placed in a concert-hall perspective. However the balance on the piano tends to change a bit when we reach the first-movement cadenzas, and sometimes very slightly within them. The cadenzas, all Beethoven’s, are the long one in the C major Concerto (the soloist has a choice of three) and the second (less often played than the other) in the G major. Given the daring quality of the enterprise, it’s curious to find the first movement of the C major treated so sedately by Aimard, the cadenza included (his timing 19'13", as opposed to Brendel’s 17'06"). I shall return to this work less often than to the other four, I think, where I’ve found a balance of imagination and rigour that is exactly to my taste, much delight and refreshment, and where I’ve sometimes been blown away.