I'll limit this review to Bartok's 3rd Piano Concerto, having no particular insight at this point on the interpretations of Bach's and Liszt's respective concertos. I'll just signal that the Liszt Piano Concerto No. 1 with Ansermet and Orchestre de la Suisse Romande (June 6, 1947) comes in terrible sound, cottony and almost entirely drowned in the surface noise of whatever shellacs it was preserved on, becoming even literally unbearable in the Quasi-adagio; imagine hearing Liszt' concerto from you neighbor's table radio next door, while frying fish on your saucepan with oil mixed with water. The Bach Concerto (in Busoni's arrangement, with the Concertgebouw Orchestra under Eduard Van Beinum, October 2 1947) is marginally better in terms of clarity, but comes also with significant surface noise and spots of annoying scratches.
A general background note on these live recordings (the Bartok is from May 30, 1948) might be useful. They surfaced publicly only at the beginning of the 1980s (Bach) and at the end of the 1990s (Liszt and Bartok). Lipatti was a shooting star of the piano scene, burnt out at 33 from Leukaemia. He was also an extremely fastidious pianist (he once said that he needed five years to prepare Beethoven's Emperor Concerto and three for Schumann's Concerto) and had time to commit very little of his repertoire to the studio. So any documentation of live performances are god-send to the music-lover and Lipatti admirer, even if he has to endure dismal sound for it.
To the best of my knowledge the Bach got its first release on a Swiss Jecklin LP, D-541, in the early 1980s (paired with a nocturne and two etudes of Chopin), and Jecklin reissued the program on CD in 1990, Piano Concerto, now complemented with Chopin's First Piano Concerto under Otto Ackermann with the Tonnhalle Zürich Orchestra from February 7, 1950 (that's Lipatti's "real" Chopin Concerto, not the one that was once attributed to him but turned out to be the recording of Polish pianist Alina Czerny-Stefanska). The same program was subsequently released on Palexa 0509, with more Chopin solo pieces, (In Concert at Zürich and Amsterdam). The Liszt Concerto and only the middle movement of Bartok's 3rd, "Adagio Religioso", were first released in 1995 on a Lipatti compilation from Archiphon, ARC 112/113 (Piano Concerto 1). So the first release of the complete Bartok (paired with the same Liszt, and Grieg's Concerto) happened only in 1999, under the auspices of Urania, Bartok, Liszt, Grieg: Piano Concertos; this EMI reissue came in 2001.
There is a small detail in Bryce Morrison's liner notes that I must take exception with: he quotes a letter from Lipatti to his teacher describing the concerto as "a true marvel of astounding emotion and purity. I read it with great emotion and accepted the proposition [that he give the premiere with Ernest Ansermet in 1947] enthusiastically". The brackets are Morrison's, but they are misleading. That couldn't have been the world premiere, since it was given in Philadelphia by Gyorgy Sandor and Eugene Ormandy on February 8 1946 (it was recorded by them two months later, Bartók Premières), and it couldn't have been the European premiere since it was given in London by Leon Kentner and Adrian Boult on February 27 of the same year (documented on Louis Kentner Plays Bartók and Liszt). So what's left? The Swiss premiere I guess, or maybe the continental premiere if you want to look at the larger scheme of things (although I'd be surprised if a German premiere hadn't taken place before).
Anyway:1948-1999, that's more than 50 years under wraps. Why?
One thing is sure: not for sonic reasons. Unlike those of Bach and Liszt, the recording of Bartok's Concerto sounds stupendously good. Compare, on the one hand, with the BBC's live, scratching and muffled recording of Kentner and Boult's European premiere from two years before, and on the other hand to the studio recordings of Katchen with Ansermet (Bartok: Pno Cto. No 3/Cto for Orch/Dance Ste) and Pennario with Golschmann (Prokofiev: Piano Concerto No. 3 in C Major, Op. 26 & Chout Ballet Suite; Bartok: Concerto for Piano No. 3), both from 1953: Lipatti's 1948 live concert sounds incomparably better, incomparably more vivid, affording much more orchestral details and extraordinary presence of the woodwinds. Jed Distler's comment on musicweb-international that "in the Bartok Third, however, EMI retains more surface noise from the original lacquers than transfers on Urania and Tahra do" leaves my mouth gaping. What "surface noise from the original lacquers"??? There may be very faint static noise like when you hear a broadcast on your radio, and some saturation in the finale, but there is no surface noise from lacquer surfaces whatsoever, and if it hadn't been for Distler's comment I would have assumed that the concerto had been recorded on tape.
So if not the sonics, then the execution? The story (recounted by Bryce Morrison, and presumably derived from the Archiphon set) is indeed that conductor Paul Sacher authorized publication of only the slow movement in his lifetime because of orchestral sloppiness in the outer ones. But I'm not sure that this is the true reason. Every critic I've read seems to sheepishly repeat this accepted truth ("You can readily perceive why Sacher didn't want us to hear the outer movements. He was probably not happy with the orchestra's poor intonation, imprecise ensemble, and sloppy entrances" said Distler of the Urania release, "the orchestra - but never Lipatti - often sounds uncertain in the outer movements" wrote Charles Timbrell in Fanfare), but in fact there isn't so much sloppiness: some moments of faulty intonation from the brass in the first movement, yes, but none of the glaring wrong entries or shamefully botched ensemble I was braced for, and nothing - given the historical importance of this recording, and its live nature - that was serious enough to justify Sacher's refusal. On the contrary, in the outer movements, Sacher conducts with considerable bite and power, underpinned by those great sonics and vivid instrumental presence. In fact and by way of paradox, it is in the opening bars of the middle movement which Sacher authorized for release that the strings sound shaky, poor in ensemble, lacking tonal refinement and way too loud to let the "religioso" atmosphere truly unfold.
But I wonder if the true reason for Sacher withholding authorization hadn't to do with his retouching of Bartok's orchestration. At 5:43 in the first movement I hear so vividly what is supposed to be the oboe and bassoon (over piano pounding and strings playing forte) that I can't help wonder if it isn't trumpet and trombones playing. Likewise at 7:06, unless my ears and the recording badly fool me it is the trumpets playing the descending sixteenth-notes that Bartok attributed to violas, clarinets and oboes. In which case, Sacher did the right thing, because at least it enables you to HEAR what Bartok wrote, and it should have been no cause for withholding. And anyway I've spotted no such thing in the finale.
All the more a pity that the recording was kept in the vaults for so long as this is, also, a significant performance. We've grown so accustomed to a gentle and lyrical approach to Bartok's 3rd Piano Concerto (at least until the finale) that we take it for granted that this character is in the very nature and fabric of the composition, as with Beethoven's 4th Piano Concerto. But in fact, most of the early generation of pianists played it as a Bartok composition, e.g. urgent and biting (the same observation can be made with Beethoven's 4th, in fact, which the pre-war and early post-war generations of pianists took as an urgent and ebullient work in the character of the "Emperor"). It was the case with Sandor-Ormandy, Kentner-Boult live, Kentner with Fricsay live again in 1950 (Ferenc Fricsay Conducts Bela Bartok - Complete RAIS Berlin Recordings 1951 - 1953), Katchen-Ansermet and Pennario Golschmann in '53, Haas-Fricsay in '54 (MONIQUE HAAS Complete Recordings on Deutsche Grammophon) and Sandor again with Gielen at the end of the 1950s (but I've now exhausted my 10 authorized product links - I'll send the rest to the comments section). The only exceptions to that rule had been Farnadi with Scherchen in 1953 for Westminster (now reissued by Tahra) and, to a lesser extent, Annie Fischer with Markevitch in 1955 (but with a swift and biting finale, while Katchen's, and to a lesser degree Sandor's in 1946 had been surprisingly held-back, with moments of gracefulness even). But it is really the famous version of Anda and Fricsay in 1959 that established the new interpretive paradigm, whose most extreme representative in the studio would be Barenboim and Boulez in 1967, with a uniquely spacious and musing first movement clocking 7:43 (compare to Sandor-Ormandy's 6:19 and Sandor-Gielen's 5:47; Farnadi-Scherchen took it in 7:33).
So the interest of Lipatti-Sacher is that they decidedly advocated the lyrical and spacious approach, in a time when it was the exception rather than the rule. It is not so much their Adagio that stands out, despite its held-back pacing, even in the faster middle section (Farnadi-Scherchen and Fischer-Markevitch were more spacious still, and Sandor-Ormandy and Kentner-Boult had proved that a much brisker tempo in that movement could be as eloquent, if not more) as their outer movements. The first unfolds in a leisurely 7:45 - that's already Barenboim's timing - and the finale is taken in a broad 7:19 (compare for instance with Pennario's 5:58 there or Fischer's 6:15; but Farnadi-Scherchen and, surprisingly, Katchen-Ansermet were even broader), which doesn't preclude moments of great muscularity, and Lipatti's touch is always admirably crisp. As a consequence, he brings great character to the Bach-like counterpoint of the piano writing in the finale, while never sounding like he lacks dynamic thrust, as could be the case with Farnadi-Scherchen or Katchen-Ansermet.
Its combination of great sonics and unique interpretive insights makes this a version that no doubt would have been considered a reference recording, had it not been kept in the vaults for 50 years.