Scholarship pertaining to the 18th and early 19th centuries is misapplied. “Often fewer strings were used for concertos,” says Arthur Schoonderwoerd. But was it right to reduce Cristofori to only a string quintet and double bass? Anyway, the point is lost because the recording magnifies the size of the band. Two fortepianos are used, one for each work. They are miked to stretch unrealistically from centre to extreme right. Worse, an artistic disaster arises from Schoonderwoerd’s belief that with a principal violin leading the instruments and the soloist providing “a rhythmic accompaniment in the tutti passages, the orchestra could do without the gestures of a conductor”. Schoonderwoerd’s accompaniment isn’t rhythmic. It’s metrical; and with the exception of the slow movement of the Third concerto, tedious. Questionably (and damagingly in the piano transcription of the Violin Concerto, Op 61a), he ignores Beethoven’s cadenzas, substituting his own.

No need to question the beliefs of Ronald Brautigam and Andrew Parrott in the First Concerto. Parrott doesn’t make “gestures” to the Norrköping musicians. He conducts them – with distinction too, grounding the music in a strong bass-line, placing the notes where he wants them to be within an elastic pulse, and keeping the phrase structures airborne. Violins are separated, and the basic tempo for the first movement is close to Czerny’s prescribed figure of crotchet=176. It’s a fair clip but here far from hasty; and the jolting F natural on the piano (rather than the usual F sharp) at 4'19" suggests that Hans-Werner Köthen’s edition was used. Brautigam also makes Beethoven’s long third cadenza sound improvised. A deeply felt Largo (excellent clarinets) and finale that is perhaps a touch too fast – but, again, no haste – rounds off an imposing performance.

Similar tempi for the outer movements of the Third Concerto are, however, misconceived. The scherzando style so successful in the First is unsuitable for this work, its gravity diminished by a two-in-a-bar speed for the first movement (as was marked in now discredited texts) and by a lightweight scamper of the last. Timpani could be better defined but otherwise the instrumental balance is right and SACD sound very good indeed.

John O’Conor settles on four-in-a-bar, a traditional choice for the first movement of the First and an appropriate one for the corresponding movement of the Third. In fact his tempi everywhere cannot be faulted but his performances fail to ignite. They are cautious and earthbound, not least because Andreas Delfs is a negative partner. His conducting (or lack of it) mires a bored LSO in rhythmic inertia that hamstrings O’Conor too. There is little or no sense of direction, no vibrantly expectant inevitability. You’ll find those qualities in the complete sets by Aimard/Harnoncourt (Teldec, 4/03) Brendel/Rattle (Philips, 5/99) and Schiff/Haitink (Teldec, 9/97 – nla).