Jerome F. Weber
Classic Record Collector
, June 2007
When this performance and recording commemorated the sesquicentenary of Mozart's death, there had been only three recordings, the Joseph Messner of 1930 on Christschall (now Orfeo C396 951 B), the Harl McDonald on RCA Victor and HMV (once on Camden CAL276), and the Bruno Kittel on Deutsche Grammophon (now <D 459004-2). HMV captured Bruno Walter's interpretation in Paris but it was rejected by him and issued only much later (Orfeo 7 63912-2.) As with the Kittel, wider circulation of the De Sabata was delayed by wartime conditions. It was issued on Deutsche Grammophon and on HMV in Britain after the war, and it was the first CetraSoria LP in America in June 1949 (Camden 1001). I believe it has been continuously available ever since. The original Cetra album was unusually lavish with text and photos, claiming 300 singers and 160 players in a public performance. It has been clarified (ICRC Summer & Autumn 2000, CRC, Winter 2000) that the concert performance took place on 3 December with Gigli, Caniglia, Stignani and Pasero, while for the recording sessions Tagliavini, Tassinari and Tajo were substituted for contractual reasons. In 1968, one of many LP reissues (Heliodor Camden 88005) printed an impossible date and place of April 1939 in Berlin, confusing it with De Sabata's BPO sessions for Deutsche Grammophon in that month (reissued on Heliodor at the same time). Typically the wrong date was repeated on Cetra's own 1991 CD reissue (Orfeo CD01): that label's presentation is often at fault.
It was noted in CRC, Winter 2000, that the Heliodor LP is different to the Cetra recording, but no one seems to have compared it with the earlier issue (DG 68422/9). That set, which can be found at the Berlin Musikarchiv, has the same soloists but Grammophon's own series of matrix numbers from 1716GS to 1 731 GS were assigned (three of the sides have take 4/4), but not in sequential order as the Cetra numbers were. (I thank Dieter Lerch for this valuable clarification.) The Cetra matrix numbers were 2-70664 to 2-70668, with take II added to eight of the 16 side numbers. Comparing the Grammophon shellacs with the Heliodor LP should establish that the reported differences noted in Mr Bernard d'Abrera's letter are explained by alternate takes made during the same sessions.
This basilica, a large church in the shape of a Greek cross built on the site of the baths of Diocletian, has been used frequently for concerts, though its shape makes it appear to be an unlikely place for that purpose. The acoustics are resonant (but not as 'unfortunate' as the notes here suggest), and the huge forces produce some powerful climaxes. Some movements end with a huge ritard, yet the tempos are not as broad as might be expected, the timing closely matching most of the traditional performances dating back more than a few years. In the notes, G. Paolo Zeccara speculates without foundation that the choruses were recorded first, then the solos. The matrix numbers indicate that the work was recorded straight through, with half of the 16 sides requiring a take 2.
While Cetra's CD reissue was an abysmal transfer that showed how badly the Cedar process can be handled, this one is not as good as expected – I have not heard the issue recommended by Kenneth Morgan in ICRC, Summer 2000 (Musica Classica MC2019/20.) At worst, 'Confutatis' and 'Lacrymosa' sound as if the pressing was offcentre or someone was dragging irregularly on the turntable. Overall, timing both discs carefully and omitting the variable gaps between sides, the Naxos plays 1:28 slower than the Cetra CD, but the two movements cited betray the greatest differences, 5 to 6 per cent slower. Listening through the obstacles, the performance still impresses as a powerful interpretation of its time. It always was a landmark. I would like to hear what the finest source materials might yield in another transfer.