First let me pose a question regarding the function of oddball concerto cadenzas, when and where to use them. Outstretching, experimental forays may work in concert, where you can take in their range of colours and leave the hall either stimulated by invention or appalled by vulgarity, or whatever effect the ‘new’ music has on you, but I would respectfully suggest that including such a cadenza on a recording is far more problematic – unless it’s included as a separately tracked stand-alone ‘fill-up’, that is.

The history of recorded music has witnessed some fairly weird interpolations into recognised classics, Schnabel’s Mozart being an obvious case in point and more recently (and more provocatively) Schnittke’s cadenzas for Beethoven’s Violin Concerto. Earlier Gramophone reaction to this novel invasion (as presented by Gidon Kremer) has been largely negative; and much as I’d like to be an exception, I can’t. Go to 18'27" in the current version of the concerto’s first movement and instead of having Beethoven ruminating on what has just passed, you have Schnittke leaping to the late 19th and 20th centuries (though there is a side reference to Beethoven’s Seventh Symphony). Otherwise, we visit Brahms, Bartók and Shostakovich before the timps, walloped by hard sticks, call some sort of order and at 22'22" the main theme returns in a state of shock, as if having suddenly woken from a bad nightmare. This unsettling onslaught is made all the more conspicuous by Vadim Gluzman’s pure-toned playing of the Larghetto before the Rondo delivers another Schnittke ‘treat’ (6'34") where the timps return (why?) and beyond a swarm of rising trills from the orchestra’s strings the rondo theme reappears, less uncomfortably than was the case for the first movement’s second theme, it has to be admitted.

Lovely playing from Gluzman elsewhere and a fine account of the orchestral score by the Lucerne Symphony Orchestra under James Gaffigan (the recording dates from December 2019 to January 2020). Not one for this old curmudgeon, I’m afraid, though I’m not suggesting that Kreisler and Joachim should share a default position cadenza-wise between them.

Having come clean about ‘Schnittke confronts Beethoven’, I happily admit that I do very much admire Schnittke as a composer, including the Concerto No 3 (1978) programmed here. Defying tradition, Schnittke follows a slow-fast-slow pattern (rather than fast-slow-fast) and the often beautiful finale includes music that has a Dvořákian mellifluousness about it as well as a touch of Stravinskian austerity. To my ears this final homecoming rather resembles the Berg Concerto, though it doesn’t lay to rest in quite the same way. Again, the performance is extremely fine; I just wish its disc-mate had been another Schnittke piece (or three). Still, you may well disagree. If you don’t and would rather something a little more conventional, in the best sense of that term, then Midori with the Festival Strings Lucerne under Daniel Dodds (a fine recording that post-dates Gluzman and Gaffigan by a few months) offer a beautiful performance, tenderly expressed, thoughtfully balanced and weaving Kreisler’s cadenzas into the fabric of the score as if they’d always been there. To hear the way Midori re-enters the main body of the first movement after the cadenza has ended is to share in a sublime dream. With her, you can gratefully open the curtains and gaze at the stars.