Mahler: Symphony No. 5. New York Philharmonic Orchestra, Bruno Walter Feb. 10, 1947
Original 78rpm edition Columbia 12666-73D in set MM 718:
Western CD editions:
Sony Masterworks Portrait MPK 47683 (1991) 5099704768328 (European edition), 074644768329 (US edition):
Sony France Maestro SBK 64012 (1995) 5099706401223:
Bruno Walter The Edition Sony SMK 64 451 (1994) 5099706445128, 074646445129:
Collected in Bruno Walter The Edition vol. 1 Sony SX10K 66246 (10 CDs) 5099706624622 (with 64447 to 64459 – Mahler Symphonies 1, 2, 4, 9, Das Lied von der Erde, Lieder eines fahrenden Gesellen, Lieder und Gesänge aus der Jugendzeit, Wagner Orchestral Music, Violin Concerto by Beethoven with Szigeti) and Mendelssohn with Milstein):
Collected in 5 CDs Sony France 5054072 (with 83Nos. 1, 2, 4, 9, Lieder eines fahrenden Gesellen) (2002) 5099750540725:
Collected in 5 CDs Sony SM5K87988 “Columbia Legends – Bruno Walter Mahler The Great Symphonies” (with Nos. 1, 2, 4, 9, Lieder eines fahrenden Gesellen) (2003) 5099708798826:
Sony “Classic Recordings” 515301 2 (with Symphony Nos. 4 & Lieder und Gesänge aus der Jugendzeit) (2004) 5099751530121:
Collected in Sony Original Jacket Collection SX13K 92460 “Famous Mahler and Bruckner Symphonies” (13 CDs) (2004) 827969246023:
Collected in 7 CDs Sony Masters “Bruno Walter Conducts Mahler” (2012) 886919201024:
Collected in “Bruno Walter The Edition” (39 CDs) (2013) 887654895226
An invaluable historical document – and not only because it was the premiere recording of Mahler’s 5th
Originally posted on Amazon.com, 14 December 2011
The historical significance of this recording, made on February 10, 1947, cannot be overstressed: quite simply, it was the very first recording of Mahler’s 5th Symphony. Also, as one of Mahler’s closest disciples, Bruno Walter had unique legitimacy in this repertoire. Of course, there is no telling how close his interpretation in 1947 was to the way he had heard Mahler conduct it during the composer’s lifetime. Not only because any memories he had were by then something like four decades back in time, but also because it is the interpreter’s personality that shapes his interpretation, not any memory he has of someone else playing it 40 years before. Mengelberg, a close friend and early champion of the composer, had easily as much credentials as the upholder of the Mahler tradition as Walter, but it is enough to compare Walter’s strict and controlled reading of Mahler’s 4th symphony (recorded 10 May 1945, many CD reissues after CBS/Sony 32DC 578 in Japan in 1986, barcode 4988009214856, and
Sony Masterworks Portrait MPK 46450 barcode 5099704645025 in 1990 in the West), with Mengelberg’s roughly contemporary and highly charged one (1939, first reissued in the West in 1989 on Philips Legendary Classics 426 108-2, barcode 028942610828), 
to realize that we will never know who was closer to Mahler’s own interpretation, and who was doing his own thing under the pretense of doing it as Mahler did. In all likelihood they were both doing their own thing.
Still, anyone raised on the later recordings of the 5th symphony, starting with the first stereo recording of Rudolf Schwarz from 1958 (with the London Symphony Orchestra, reissued 1995 on Everest EVC 9032 – link will open new tab to my review) and continuing with the second stereo recording by Leonard Bernstein (New York Philharmonic, January 1963, Columbia now Sony), and with the exception maybe of those of Leinsdorf (Boston Symphony Orchestra, RCA, November 1963), Vaclav Neumann in Leipzig (June 1965), Solti (Chicago Symphony, Decca/London 1970) and Mehta (Los Angeles Philharmonic, Decca/London 1976, reissue 1987 Decca Ovation 417 730-2 barcode 028941773029), is in for a shock listening to Walter.
First thing to note: it is, according to the table of timings provided by Peter Fülöp in his stupendous and invaluable Mahler discography, the overall fastest-running Mahler 5th ever recorded. And this is not a figure of speech: at 61:11 (by Fülöp’s count, which is sometimes marginally different from the CD timings), it is, as of April 2010, the fastest among the phenomenal 173 versions listed in the discography (leaving out Scherchen’s three live recordings, which are cut). Second and hard on the heels of Walter comes Abravanel, in 1974, at 61:25 (Vanguard Europe 08 4008 71 barcode 3351474008717, US SVC 25 barcode 0723918002521). I haven’t drawn statistics from Fülöp’s table, but by rough eyesight most versions of the 5th are above 67 minutes, and a good number run at 70 minutes or over. For reference, Bernstein’s 1963 recording is 68:55 and his DG remake from 1987 with the Vienna Philharmonic, 74:43.
Furthermore, while Abravanel conducted the fastest ever first and second movements, that “honor” came to Walter in the Scherzo. They were both beat on the finishing line of the Adagietto by Mengelberg (1926, conveniently reissued in 2003 on Naxos Historical 8.110855, barcode 636943185520, in the transfer of Mark Obert-Thorn), and while they came again within seconds of each other in the finale, the record-bearer here is the unexpected Alain Lombard conducting the Bordeaux Orchestra in 1991 (Forlane FF 059 bc 3399240000596).
So, these are the numbers. What does it all sound like, musically?
I think it sounds great. With the Adagietto, it is probably the first movement that will be the hardest to take for those accustomed to the more brooding and ponderous approach exemplified by Bernstein’s Vienna version. They may find, as I once did, that Walter lacks a brooding, burdened quality essential to the music. With Bernstein you get the impression that you are hearing the plight of the Jewish people ever fleeing the pogroms and persecutions, like Atlas bearing the weight of a world of suffering on their shoulders. Walter’s approach is much more forceful than that, his “pesante” is powerful rather than heavy, he never lets you forget the military beneath the funeral. His march marches on in an almost grim mood of dogged implacability – the fallen soldiers of World War I returning to compel the living to stop all wars as in Abel Gance’s film “J’accuse” rather than the plight of the Jewish people. This is the kind of approach that was later followed by Leinsdorf, Solti, Neumann in Leipzig, Mehta in Los Angeles and even Sinopoli, and if you can adjust to it, it is great and entirely convincing. One originality of Walter is that, while you would have expected the approach to yield a very brisk “plötzlich schneller” section (suddenly faster) at 4:41 (and that’s the way Scherchen does it, see my review of his 1952 Westminster recording), in fact he remains surprisingly held-back, more in fact than anybody I’ve heard so far, even those like Bernstein/Vienna who adopt a basically slower tempo throughout. But it translates in no loss of passion (“leidenschaftlich” is Mahler’s marking here, and “wild”), because Walter’s lends that section an intensely despaired vehemence. One interpretive quirk that would have many followers including Bernstein (in New York) and Solti is the way Walter takes Mahler’s indication “molto espressivo” and “breiter strich” (large bowing) at 5:44 (measure 202 and after) as a prompting to perceptibly slow down the tempo. But there are also great instrumental touches, and for instance I have never heard the cello pizzicati starting at 1:02, at the “etwas gehaltener” passage (measure 35) sounding more like tolling funeral drums.
But listen to Bernstein/Vienna in the second movement, and hear how heavy and muddy it all becomes. That’s the downside of the great and appropriately ponderous approach of the first movement: it doesn’t extend so well in the next. Not so with Walter’s. Whatever one’s possible misgivings with his first movement, the second cracks like whiplashes. I don’t know who it is that is supposed to be getting hammered in such a way, the Jews or the Nazis or Mankind or the good Viennese society of Mahler’s time, but I sure am happy that it is not me. Mahler calls for the movement to be played “mit grösster Vehemenz”, the utmost vehemence, and under Walter’s baton it is, with rage even and scorching heat in the more lyrical passages. Again it is the kind of approach that was followed by Leinsdorf and Mehta, and brought to a point of perfection and maximum bite and intensity by Solti, but it’s already all there with Walter.
I already mentioned that Walter’s Scherzo was the fastest in all the discography – yet, I don’t hear it in the least. His urgency translates in no loss of charm, but provides at the same time a constant underlying intensity, as a Vienna Waltz danced on the brink of the abyss. And (like Ravel’s La Valse) what is this movement about, if it is not exactly that?
The Adagietto tends to be played nowadays as a brooding lament, something like a meditation upon the passing of things, a kind of anticipation of the finale of the 9th symphony (and Visconti’s Death in Venice didn’t help). And certainly the music can take it, and that kind of slow approach, extending the music to circa 11-12 minutes and sometimes even more, yields beautiful results. But the evidence shows that for Mahler, it was a declaration of love to Alma, and that’s how the early pioneers played it: Mengelberg’s Adagietto was just a little over 7 minutes, Walter’s is 7:45, and Rudolf Schwarz’ (reminder: the first stereo recording in 1958), 7:30. Other than them, very few versions are under the 8-minute mark. In fact, other than a few recordings of the Adagietto alone (by Kenneth Slowik, Gilbert Kaplan, Walter himself in 1937), there are only five – or rather four. I suppose the name of Heribert Brandt, conducting the Arte Sinfonica Orchestra on a Deutsche Austrophon CD published in 1999 will be as unfamiliar to any reader as it was to me. No wonder: it is a pseudonym, and looking at the movements’ timings, I bet it is Schwarz who got pirated. The others are Walter, Abravanel, and the unexpected Boulez, leading the BBC Symphony Orchestra in a live 1968 concert.
So again, as with the first movement, fully appreciating a less than 8-minute Adagietto probably requires a change of listening paradigm or expectations, not Death in Venice but a tender to passionate love song. You fully hear that passion with Walter at 3:14, on the “fliessender” passage (more flowing) leading to “etwas drängend”, somewhat pressing. That said, the love song requires the full, spacious and warm stereo sonics of Schwarz (I haven’t yet heard Abravanel); Walter’s 1947 mono sounds a bit dry and mono-chromatic. In addition, the all-important harp is often lost in the resonance of the strings.
Many versions of the finale seem to neglect Mahler’s indication, “allegro GIOCOSO”, which translates in “jocose, jocular, playful”, to which Mahler adds “frisch”, lively and in my listening experience, this sense of playfulness and joy is best conveyed by a swift and lively tempo (a slower tempo conveys other and possibly valid moods, but not a sense of elated and carefree joy). Walter provides exactly that, rounding off his recording with an enthusiastically resounding conclusion.
Noting that “tempos are quick throughout”, esteemed Mahler critic Tony Duggan from Musicweb-international commented that “though this probably reflects Walter’s more astringent approach at that time in his life, you cannot escape the impression that another determinant was the need to fit the recording on to 78rpm sides”. This is simply ridiculous, and only shows that even esteemed critics and seasoned Mahlerites can have blind – or rather, deaf listening spots when they listen with the wrong paradigm or expectations. Significantly, in his synoptic survey of the 5th, Duggan doesn’t comment on the very similar approaches of Leinsdorf, Solti and Mehta. What would he say, that it reflected their need to catch the 5 p.m. flight on the day of recording? What those tempos reflect is Walter’s very conception of the 5th, a work of considerable bite, urgency, passion. And my feeling, although nobody will ever know, is that it is close to Mahler’s own conception.
As the old saying goes: “they don’t do it that way any more”. Yet, there are lessons in interpretation to be learned from Walter’s 5th that are still valid for today.
A note on the sonics. In any reissue of 78rmps the sonic quality is going to be dependent on the state of preservation of the original records, and on the transfer made from these. I’ve been entirely satisfied with the sound of the Japanese edition on CBS/Sony 32DC 579 barcode 4988009214955, ever since it entered my collection, in 1990 (the recording was reissued in the West only in 1991 – and I couldn’t wait). It sounds great. There is a modicum of surface noice from the original records, but it is kept at a surprising minimal, and a few side joints that are slightly jarring because of the change of sonic perspective between one side and the next. But while the sound is a little dry and boxy, an incredible amount of orchestral details can vividly be heard. Other than a few menial glitches, it exposes a valiant New York Philharmonic, who play the music as if they knew it on the back of their hands (which, in 1947, cannot have been true). I bet the symphony was recorded using a reduced cello and double-bass section: they sound so incredibly lean and clear and detailed. Just listen to their staccato runs at 1:06 in the finale.
Recently I acquired the Original Jacket Collection set with Walter’s Bruckner and Mahler referenced above – not that I felt unhappy with my early Japanese edition, but because the set sold cheap on eBay and is a fine collector’s item. The sonic difference with my Japanese CD is minuscule: a touch less surface noise, but sound that is, maybe, imperceptibly less brilliant. Needless to say, this is perceptible only on A-B comparison. This will never be hi-fi anyway, it’s not for “realistic” sonics that we listen to that version. The “Original Jacket” set is a nice way to have Walter’s Mahler and Bruckner in one’s collection.