Beethoven: Violin Concerto, Sonata for Violin & Piano No. 5 “Spring”. Joseph Szigeti, Philharmonic-Symphony Orchestra of New York, Bruno Walter, Mieczyslaw Horszowski. Sony Masterworks Portrait MPK 52536 (1992) barcode 074645253626 (American edition), 5099705253625 (European edition)
Reissue (Concerto only) Bruno Walter The Edition SMK 64 459 (1994) 074646445921, 5099706445920:
Recorded 5 April 1947 (Concerto), 1 December 1953 (Sonata)
Whatever you think of the interpretive approach, this 1947 recording of the VC is marred by Szigeti’s intonational problems
Originally posted on Amazon.com, 28 December 2009
When I heard this very recording of Beethoven’s Violin Concerto by Joseph Szigeti and Bruno Walter, made in New York in 1947 and their second, after the one they did in London in 1932, in Dante’s apparently excellent transfer from 1998, Lys 397 (link will open a new tab to my review), and considering its glowing reputation, I couldn’t believe my ears. Was I the only one to hear Szigeti’s plain (at best) to sour (frequently) tone, his numerous little finger slips and blemishes? Or could it be the Dante transfer? So I bought this Sony Masterworks Portrait 1992 reissue, just to check (and to get the “Spring Sonata” as a bonus).
It ain’t my ears. Granted, the Sony transfer is marginally better, giving a more brilliant sonic image of the Concerto (and running at a faster speed and higher pitch than Dante, but here, it’s hard to tell who is right, since I don’t suppose anybody has a cue of what precise pitch was used on that recording day 60+ years ago). But all the problems of Szigeti’s intonation and playing stand out as clearly as on the Dante transfer.
So what is it with all those glowing reviews that the recording gets on the big online retailers (either under the entry for this edition, or the one of the 1994 reissue on the Bruno Walter edition, SMK 64 459, where it is paired with Mendelssohn’s Concerto played by Nathan Milstein)? Is it the other listeners’ ears? Or is it that they attribute Szigeti’s timbral deficiencies to the recording rather than to the fiddler, and make do with them? But Szigeti in 1932, again in Dante’s transfer (contained in a valuable companion set to this Walter Beethoven series, 
volume 7, Dante Lys 518-519 , barcode 3421710425186, with the rest of Walter’s pre-war, pre-US Beethoven recordings, including a Pastorale with Vienna and an Emperor Concerto with the same orchestra and Walter Gieseking) is free of all these problems. So I think it is the fiddler. Szigeti is (in)famous for the marked deterioration of his faculties, due to arthritis, that marred his last recordings from the 1960s, but judging from this it seems to have established has early as the late 1940s.
So is it then the overall conception, so convincing perhaps to some that it leads them either to turn a deaf ear on the violinist’s tonal production, or excuse it? Walter and Szigeti’s conception hadn’t changed much since 1932; Walter is here slightly preferable to his former self, more muscular and disciplined (in 1932 he tended to play accordion with the tempo in the first movement). Overall it is a gentle and lyrical approach, with plenty of time to enjoy the view in the more lyrical concertante passages in the first movement (but as mentioned, Walter does whip up the orchestra to fine muscularity in the tuttis), a middle Larghetto that sounds much more like a Largo and a none too rushed allegro finale. It is a far cry from the forward-moving, muscular but no less lyrical approach of Heifetz-Toscanini in 1940, then (to say nothing of Heifetz’ cold but stellar tone), which I personally much prefer. Even in that approach, Szigeti and Walter’s 1932 recording is preferable, all the more so as the sound, in Dante’s transfer, isn’t vastly inferior to the 1947 recording. Still, it is Walter’s third and stereo recording, made in 1961 with Zino Francescatti and the Columbia Symphony Orchestra, that is (by a wide margin) his best, and not only for sonic reasons (multiple reissues, I have it on Odyssey MBK 45637, 5099704563725).
By the way, Szigeti plays Joachim’s cadenzas (with a cut at the end of the first one), which had more currency in the early days of recording than today. They’re not very Beethovenian, but they are great and original, the kind you might have expected from Ysaye in fact. Szigeti there is great, free of the intonational problems present in the rest of the recording. They are definitely something to cherish from the 1947 recording.
I don’t hear the same intonational problems in Szigeti’s Spring Sonata, made with Horszowski in 1953; what I hear is a reading that goes for charm but is somewhat short in tension, energy and dynamism. Still this will be my reason to keep the CD. The Sonata by the way originally came on LP paired with the 6th, now on a Szigeti recital collated by Sony on a companion Masterworks Portraits, MPK 52569 5099705256923 or 074645256924 (see my review on Amazon.com pending repost here).