Pioneering Haydn On Record
by Christopher Howell
An occasional series dedicated to commercial and live recordings of Haydn symphonies up to and including those of the first two complete cycles by Ernst Märzendorfer and Antal Dorati
Symphony no. 104 in D major Hob. I/104 (1795)
Part 4: Live and Broadcast Recordings
A pdf version of this article can be downloaded for offline reading and printing here.
In this final part of the article, I discuss a number of live and/or studio recordings made for broadcasting that have survived – some readily available on CD, some on YouTube and others circulating privately. I daresay this list could be very long indeed and there may well be many more interesting performances awaiting discovery. In the earlier parts of this article I discussed live performances by Furtwängler, Rosbaud, Scherchen, Munch, Rowicki, Barbirolli and Kempe, comparing them with their studio recordings – or their purported studio recording in the case of the first of these[1]. Here I deal with conductors who did not take the London Symphony into the studio: Toscanini, Celibidache (2), Schuricht (4), Sawallisch, Monteux, Willcocks, Giulini, Monteux, Komives, Martinon, Ančerl and Dixon.
Harvey Sachs’s Performance Chronology of Arturo Toscanini shows that he conducted this symphony just once, on 31 October 1943, with the NBC Symphony Orchestra in Studio 8H[2]. The programme began with Mozart’s Die Zauberflöte Overture and concluded with Beethoven’s 8th Symphony. This performance can be found on You Tube in very decent sound and I am a little puzzled that I have been unable to trace any CD issue of it – perhaps I have missed something. The dry sound typical of Studio 8H and the general characteristics of the performance leave little doubt that it is authentic.
Toscanini maximizes the dramatic contrasts in the introduction, while finding time to give a beseeching, almost operatic, tone to the violins’ piano phrases. The Allegro is fast but manages not to be breathless. Toscanini generates terrific tension as the music proceeds, but there are touches of affection too. No repeat. The Andante is fairly swift, though there have been swifter ones. Here Toscanini adopts a tender, lyrical manner – except in the dramatic fortissimo outbursts and, unlike practically anyone else who took a similar tempo, he succeeds in making the later stages not sound busy as the music breaks into semiquaver triplets. The final bars are very tenderly treated, without slackening the tempo. First repeat only. Toscaninian nerves begin to show in the minuet. When discussing Furtwängler’s performance I doubted whether this survey would throw up a faster minuet, but I should have remembered that Toscanini was awaiting and he shaves it by a further seven seconds. It sounds a bit rabid, with a scurrying trio, but the soft passages have an attractive lilt, so I wonder if the forte passages would have created a different impression in a more amenable acoustic. So, too, the finale, which has tremendous drive, with gunshot sforzati and amazing articulation from the strings. Surprisingly, Toscanini relaxes the tempo towards the end of the chorale-like passages. No repeat. In spite of these reservations, if this performance had been issued in 1943 it should have held a high position for many years afterwards – if nothing else, here is no-holds-barred Haydn, dramatic and passionate.
Sergiu Celibidache returned to this symphony several times throughout his career. His Stuttgart and Munich traversals from his later phases are beyond my cut-off date. Here I deal with a performance with the Berlin Philharmonic, from 20 January 1946, and another from 7 December 1959 with the Orchestra Alessandro Scarlatti di Napoli della RAI.
Even in 1950, Celibidache took measured tempi in three movements out of four. He begins with a grandly Brucknerian introduction and follows this with an Allegro that unfolds in a dignified manner. Textures are full and there is strong conviction. It must be said, too, that is on the slow side of the norm rather than sui generis and nothing like so slow as Horenstein. No repeat. The stopwatch tells us that he stretches the tempo in the Andante, with first repeat only, almost to the same extent as Annovazzi, whom I described as “almighty slow”. It is to Celibidache’s credit that, while listening, this seemed, like the first movement, just on the slow side of normal. I did not feel, in this case, that there were passages where there did not seem to be enough music to fill the tempo. Evidently Celibidache was able to draw the ear onward and keep the music afloat so that we adjust to his time-frame. He needed this ability in the minuet, which really is ceremoniously slow, yet he does give the music an agreeable lilt and here, too, I did not realize, as I listened, that this was actually substantially slower than the slowest I had heard till now. The finale is a surprise – one of the fastest, full of wild excitement. In the context, I found it puzzling. It seems to belong to a different performance. No repeat.
Celibidache has often been quoted as opposing the playing of repeats in classical symphonies (except in minuets and scherzos) since the literal repetition of a section of the music was contrary to his conception of music as a temporal phenomenon in constant development[3]. But his path towards this final view seems not to have been entirely linear. He omitted most repeats in this symphony in 1950, but the 1959 Naples traversal is a rare case of a Celibidache performance with all repeats. Moreover, his tempi on this occasion were by no means outside the norm. More importantly, there is a greater sense of involvement in the music. He begins much as before, but the Allegro unfolds in a manner that combines grandeur and vitality. The Andante is gently and gracefully caressed, with a somewhat processional feeling in the fortissimo interjections. The minuet is just sufficiently faster than before to achieve an attractive lilt. The finale, on the other hand, is now slightly slower, replacing wild excitement with joyful exuberance. The chorale-like passages are warmly expressed without loss of momentum. This performance is among the finest and it is to be hoped that it can be heard one day in the best possible sound. As so often with these RAI performances that circulate from hand to hand, the version I have is boxy and opaque, sounding less well than the Berlin recording of nine years before.
No fewer than four recordings survive by Carl Schuricht. The earliest, with the Stuttgart Radio SO, dates from 10 September 1952. From 1955, we have two with the Orchestre National de l’ORF (the French National Radio Orchestra) – the first during their regular season in Paris in January, the second on 21 September as part of the Montreux Festival[4]. Lastly comes a 1959 performance with the Monte-Carlo Opera Orchestra. The Stuttgart performance seems to be a studio broadcast, without audience but also without retakes – witness the early timpani entry at bar 83 of the Andante. The sound is more than decent. The two with the ORF orchestra have a not very intrusive audience. Although the stopwatch shows that Schuricht was marginally swifter in Montreux, they sound almost identical and I shall not try to discuss them separately. They are well recorded though a certain steeliness to the upper strings in the Montreux version needs taming. The Monte-Carlo version may be audience recording, with dynamic compression, tape rumble and a baby crying during the opening of the Andante.
The two ORF performances are the go-to Schuricht versions, so I shall discuss them first, then note the differences with the other two. There are, I should add, sufficient similarities to leave no doubt that the same conductor is responsible for all of them – you can never be quite sure what you are getting with these live retrievals unless an announcement is included.
The introduction has great breadth in 1955, possibly seeming even broader than it actually is. During the piano passages, Schuricht is continually grading the dynamics and it may be noted that he has the knack, given to only a few, of making long-held chords not sound static – something seems to be happening during these chords. The Allegro begins serenely, even prayerfully. The forte passages, while unhurried, move a little more swiftly. Schuricht shows that the old-school method of playing the forte passages a little faster than the piano ones can work if the conductor is able move from one tempo to another with such ease that it sounds entirely natural. The beginning of the development finds him going pretty slowly, but with a sense of suppressed tension. This is a ruggedly impressive account. No repeat. The Andante goes at what seems to me an ideal tempo, leisurely enough not to seem busy in its later stages while somehow maintaining a two-in-a-bar feeling and, above all, heartfelt and songful. Schuricht presses ahead a little in the first fortissimo interjection. First repeat only. The minuet has a brisk lilt – notice how Schuricht has the repeated notes sing as a phrase rather than stab them out equally. In the trio, he plays the rising third motive much slower each time. This is a ploy we have found in several older performances, going back to Barbirolli’s 1928 version, making me wonder if this was an old tradition. I still find it odd, but Schuricht plays the rest of the trio – up to tempo – very beautifully indeed. The finale goes full tilt, a riotous village festival with no slackening in the chorale-like passages. How entrancingly Schuricht brings in the recapitulation.
Compared with this (these), the Stuttgart performance is more classical, plain-speaking, with less tempo variation in the first movement and a relatively cautious finale. The slowed-down beginning of the trio is present but slightly less, and this is the one movement I could prefer here.
The Monte-Carlo performance seems less nuanced. This may be the effect of the recording, or of a less responsive orchestra. Schuricht seems a little impatient in the introduction and does not overdo the tempo changes in the Allegro. The Andante appears somewhat unsettled, doubtless the baby crying did not help, and alone of these performances, becomes busy towards the end. The slowing at the start of the trio is back with a vengeance. Maybe I would have been more appreciative of these two performances, especially the Stuttgart one, if they had been the only ones, but the ORF traversals are something approaching wonderful while the other two are not quite that.
RAI’s recording of Wolfgang Sawallisch conducting this symphony with their Rome orchestra has a more open sound than the Celibidache recording discussed above. This has been posted on YouTube by Keith Goodman, who gives the date as 14 December 1955 and states that this was Sawallisch’s debut in Italy. However, a little research in the Radio Corriere (RAI’s now-defunct equivalent to the Radio Times) shows that a concert by this orchestra under Sawallisch was broadcast on 10 December 1955, with works by Borodin, Dvořák and Britten. The Haydn would seem, from the absence of applause, to be a studio recording, though live in the sense of a one-off without editing – or at least, I presume that the flute’s entry half a bar early at bar 131 of the second movement would have been corrected if retakes were allowed. It was not broadcast until 12 April 1956. It was not unusual for RAI to engage a conductor while in Italy for concerts with their orchestras to set down a few extra pieces in the studio.
Sawallisch’s official discography of Haydn symphonies seems limited to nos. 92, 94, 100 and 101 for Philips with the Vienna SO in the early 1960s. The present traversal is a solid professional job, and slightly more than that in places. The introduction suggested this might be the sort of nameless and blameless performance that Sawallisch could sometimes turn in, but the Allegro and the finale remind us that he was able to make a steady-as-she-goes tempo convincing through gutsy rhythmic resilience. There is vitality to the playing in this first movement and a sense of lusty enjoyment in the finale, though the chorale-like passages sag. Rhythmic resilience keeps the ceremonious minuet afloat, but tension is lost in the trio. The Andante, on the other hand, is rather brisk and chirpy, even moving ahead in the fortissimo interjections. The later stages sound uncomfortably busy. This movement has its first repeat, there no repeats in the outer movements.
Somm have issued a performance recorded for broadcast by the BBC on 25 December 1960 by the Royal PO conducted by Pierre Monteux (Ariadne 5028-2). For all the expert restoration by Paul Baily, there is a degree of distortion above a certain decibel level that makes this one for Monteux completists. Monteux is strictly classical in style, with brisk and unfussy tempi in the outer movements. Both these movements seem a little muted at the outset but Monteux soon has the orchestra responding with great energy and vitality. He joins those who find a tempo in the Andante that manages a two-in-the-bar feeling while remaining sufficiently expansive to avoid business later on. His minuet is quite swift and he allows a slight slowing of the rising-third phrases that introduce the trio, though far less than Schuricht and several others. The Andante has its first repeat, the outer movements have none. Though very nice, I do not find this performance particularly distinctive – the best of Monteux is perhaps to be sought elsewhere.
The reference books describe Sir David Willcocks as a choral conductor but, unlike many such, he had no difficulty in handling the orchestra, whether as part of a choral-orchestral work or, when the occasion arose, on its own. He recorded orchestral works by Vaughan Williams and Dyson on LPs otherwise dedicated to choral music by these composers. On 4 September 1964 he conducted the larger part of a Prom with the London SO and the BBC Singers dedicated entirely to Haydn. The concert began with the “Emperor” Quartet played by the Amadeus Quartet, after which Willcocks conducted the “Nelson” Mass (soloists: Elisabeth Robinson, Sybil Michelow, Alexander Young and John Shirley-Quirk), followed by the Horn Concerto in D (soloist: Barry Tuckwell) and the London Symphony. The Royal Albert Hall had not yet acquired the “flying saucers” that ameliorated the acoustics and some uncomfortable effects can be heard when a sharp chord is followed by a rest. The BBC Proms Archive describes all these works except the Symphony as Proms premières. I find it incredible that the “Nelson” Mass had not been given before, but I take their word for it.
Willcocks gives a grave and well-shaded account of the introduction, after which the Allegro is lively and festive even though the tempo is not especially fast. No repeat. The Andante is on the slow side, but is well sprung and among the more convincing of the slower performances. The tempo allows for affectionate handling of the later stages. First repeat only. The minuet is ceremonial but buoyant with a delicate trio. The finale, without repeat, seemed a little slow as it began but soon sounds quite rowdy in this acoustic – it has something of an English country dance to it – and draws a few cheers at the end. Not a historic retrieval, maybe, but nice to hear.
A perusal of the Carlo Maria Giulini Concert Register[5] shows that he conducted Symphony 94 incessantly all round the globe, no. 99 from time to time and a handful of performances of no. 104 in the early 1960s, of which that with the Orchestra Sinfonica di Milano della RAI on 2 March 1965 was the last. There was also an isolated performance of no. 49 in 1951. The Milan performance circulates in what may derive from a fairly recent re-broadcast, since the sound is basically good but suffers from the appalling dynamic compression for which RAI has lately become notorious. The fortissimo and piano phrases in the introduction come out at exactly the same dynamic level, so one listens with ears of faith, supported by the belief that Giulini himself would never have levelled out the dynamics in this way. This performance deserves issue in a proper transfer from the original tapes, assuming the dynamic compression did not start there, for Giulini is in fine form and obtains splendid playing from the orchestra.
In an earlier article on Symphony 49 I commented on a performance by the RAI’s Naples orchestra conducted by Bruno Aprea, noting that, with an Italian orchestra under an Italian conductor, the string playing had that vibrancy we used to associate with groups like I Musici or I Virtuosi di Roma. Whereas the RAI orchestras did not display this feature under Celibidache or Sawallisch, the Milan band certainly does so under Giulini. Another remarkable feature of the outer movements is that Giulini, at least in the 1960s, could obtain intense, even fierce articulation recalling Toscanini, though at tempi which, if not really slow, are on the broad side of the norm. The first movement impresses by its grandeur and conviction but the finale is a revelation – I don’t think any other performance heard till now has brought the symphony to such a triumphant conclusion. The Andante has a heartfelt flow at a tempo that is actually not so very slow – but without becoming busy later on. The minuet and trio lilt nicely. First repeat in the second movement, none in the first movement and finale. One of Giulini’s previous performances was with the Philharmonia Orchestra, so what a pity a recording did not follow.
A public performance survives of Yevgeny Mravinsky conducting this symphony with the Leningrad PO on 3 October 1965. I heard this on YouTube but it was once available on Russian Disc (RD CD 11 163). After an introduction where the great Shostakovich interpreter finds moments of glacial stillness in the piano passages, Mravinsky proves the perfect classicist. The orchestral textures are lean and transparent, the dynamics finely controlled. The Allegro is driven, but within strictly classical lines and the softer phrases elegantly turned. The repeat is played. The Andante is quite slow but again, elegance and controlled grace are the keywords. Even the fortissimo interjections kept free of romantic passion. First repeat only. The minuet is quite swift with a pleasing lilt and a delicate trio. I particularly appreciated the cool flute dominating the first piano section (bars 9-16) of the minuet. The finale, with repeat, is taken at a tempo that might have defeated a lesser band, yet even this is never allowed to become rowdy or rambunctious. The sheer virtuosity impresses and you would expect it to bring the house down, but the applause that follows seems no more than polite. Mravinsky’s rigid control was perhaps self-defeating.
The London Symphony opened a programme on 8 April 1967 in which the Orchestra Alessandro Scarlatti di Napoli della RAI was conducted by the composer and conductor János Komives. Komives (1932-2005) was born in Budapest, where he studied composition with Kodály and conducting with Somogyi. He fled Hungary in 1956 after the Soviet Invasion and completed his studies at the Paris Conservatoire with Milhaud. He took French nationality and his career was mainly French-based. As a conductor, we are told[6] that he specialized particularly in French 20th century music, but his few recordings also include Haydn’s complete works for wind ensemble.
Perhaps I would have reacted more enthusiastically if I had heard this performance – available on YouTube – earlier on in the survey. Komives gets generally good and attentive playing – not a given with this orchestra – and chooses lively but not extreme tempi. I thought the music sagged here and there in the first two movements, but I must say the minuet and trio are sufficiently well sprung that I did not realize till I saw the timing at the end that this is actually among the slower ones. The finale is clean and crisp but I get the idea Komives did not aim to challenge his players. I am sure the Neapolitans found this enjoyable enough, but those who had attended the Celibidache performance eight years earlier will have rightly found Komives undistinctive. First repeat in the Andante, none in the outer movements.
A broadcast from the Orchestra Hall, Chicago on 9 May 1967 by the Chicago SO under Jean Martinon sounds well on YouTube if you tame the treble. The performance has that something extra that was lacking in Komives. Martinon conducts a relatively mobile, but well-shaped, introduction – more four-in-a-bar than eight, followed by an Allegro that is swift and fleet. Martinon obtains light, transparent textures from the potentially mighty forces in front of him and, thus far, has much in common with his compatriot Monteux. No repeat. Martinon is a little slower than Monteux in the Andante, I think for the better, since he nevertheless takes a graceful, serenade-like view. Even the fortissimo passages are songful rather than dramatic. First repeat only. The minuet goes with a joyful spring, without turning it into a scherzo. The trio is marginally slower, but it is delicately shaped and does not sag. The finale is one of the fastest and the comparison with Mravinsky is interesting. Both orchestras are clearly well able to play the notes at this tempo, but under Martinon there is a sense of joyful exuberance that I did not find in Mravinsky’s over-controlled version. No repeat. It sounds as if some Haydn records from Martinon during his CSO years would not have been a bad idea.
Two performances of this symphony by Karel Ančerl seem to survive[7]. The 1959 Dresden performance has eluded me and indeed, it is not very clear from the source consulted whether it has ever been issued. A performance with the Netherlands Radio PO on 6 July 1970, once available on Tahra, can be heard on YouTube in good sound. Ančerl does not hang around in the introduction and his Allegro beats all records, at least of the repeatless versions, except the 1962 Krips. But, while Krips presented a frothy Haydn on the lines of early Mozart, Ančerl uses more generous bow-strokes and generates the sort of euphoric effect we know from his Dvořák performances. This is thrilling, but a few of the softer passages sound a little uncomfortable at this speed. His Andante is also fairly swift, but hardly seems so, so delicately is it phrased. The opening bars are beautifully handled and Ančerl is almost alone among purveyors of faster tempi in totally avoiding any hint of business later on. First repeat only. The minuet is, by contrast, rather slow, but it has a gorgeous lilt, suggesting a sousedská (a sort of slow Bohemian waltz). This lilt carries into the trio. The finale goes at about Martinon’s tempo and Ančerl is at least as successful as Martinon in creating a joyfully exuberant atmosphere. The “Bartered Bride” seems only just round the corner, but who cares? An unusual but very attractive performance.
Dean Dixon’s performance with the Frankfurt Radio SO comes from a studio session dated 6-7 December 1973. Theoretically this is just outside my cut-off date, but I since this conductor interests me I shall stretch a point. So far in this last part of my survey, I have not mentioned editions since they have all used the old Breitkopf text – and only Ančerl could have used the Robbins Landon edition anyway (but did not). So the first question with Dixon, as with Märzendorfer, is, which edition is he using? At first, I thought the old Breitkopf, since the string phrases in the first forte passage of the Allegro are bowed separately, not legato as in Robbins Landon, but not in the holograph manuscript viewable at IMSLP. However, as already remarked concerning Märzendorfer’s performance, they are slurred legato in the holograph when the passage returns in the recapitulation, which Märzendorfer does but Dixon does not. Hence my initial assumption that Dixon was not using an Urtext edition. But in the Andante, the extra notes for the second violins at bar 123, present in the Breitkopf edition but absent in the holograph and the Robbins Landon edition, are not played, so here at least Dixon is clearly is not using the Breitkopf edition. Moreover, in the minuet and trio he very deliberately plays with separate bows various quavers that are slurred in Robbins Landon but not in the holograph (so it would be nice to know Robbins Landon’s authority for slurring them – maybe the contemporary orchestral parts he discovered in the 1960s). I note that Märzendorfer has these notes staccato in the trio – at his much faster tempo, this did not stand out in the same way when I first heard it. As for the quavers in the minuet, at Märzendorfer’s faster tempo I was not quite sure if they are separated or slurred but, if separated, they are not punched out as by Dixon. Dixon also has the flute solo towards the end of the second movement entirely staccato, as does Märzendorfer (but as Robbins Landon does not). It would seem that Dixon, too, is using an independently researched Urtext, based on the holograph but not always reaching the same conclusions as Märzendorfer’s editor.
But what of the performance? It is perhaps a pity that Dixon, having chosen some sort of Urtext edition, is short on repeats – he gives that in the first movement, only the first in the Andante and omits that in the finale. Nonetheless, he is magnificent after his fashion. Dixon was now in his late phase when the firebrand of earlier days had given way to a Kapellmeister who seemingly drew inspiration from Klemperer. Or, in this symphony, the parallel may be Sanderling. After a dignified introduction, the first movement Allegro unfolds grandly at a steady pace, enlivened with vital articulation. The Andante is slow, though there have been slower ones and Dixon holds the attention with delicate phrasing and well sprung rhythms. His minuet just misses being the slowest of all, a big-hearted ceremonial affair that stays afloat by the resilience of its rhythms. And, as pointed out above, Dixon uses his tempo to point incisively certain staccatos that are in the holograph but not in the printed editions. Indeed, he may have argued that these staccatos and absence of slurs indicate that Haydn wanted a tempo in which they are heard properly. The finale is again steady, big and satisfying, though here I have to say that Giulini, at a tempo only marginally faster, showed more completely how a broad tempo for this movement can provide a triumphant conclusion to the symphony. All the same, this is a powerful and impressive performance of the old, big band type.
So how do these live/broadcast recordings change the picture. In the first place, for an Urtext score with all repeats, Märzendorfer remains on his own (within my time-frame, obviously). Jump back and forth from Dixon to Märzendorfer, furthermore, and the latter, with his hard timpani sticks and forward, almost rasping wind, almost sounds like an original-instruments production. Still, if you want a performance that demonstrates the sheer grandeur of Haydn, Dixon goes one better than Sanderling in playing an Urtext (but Klemperer played all the repeats). Leaving aside textual considerations, the 1955 Schuricht, the 1959 Celibidache and the Giulini are particularly fine, with Willcocks, Martinon and Ančerl interesting outsiders.
Timings below are, as far as possible, net of applause and shuffling/tuning between movements, to make them comparable to those of the commercial recordings.
| I | II | III | IV | |
| Toscanini | 06:33 repeat omitted | 06:58 1st repeat only | 03:51 | 04:33 repeat omitted |
| Celibidache 1950 | 08:06 repeat omitted | 09:36 1st repeat only | 06:18 | 04:46 repeat omitted |
| Celibidache 1959 | 08:44 with repeat | 10:07 with both repeats | 05:37 | 06:13 with repeat |
| Schuricht 1952 | 07:24 repeat omitted | 06:49 1st repeat only | 04:56 | 05:08 repeat omitted |
| Schuricht 1.1955 | 08:12 repeat omitted | 07:23 1st repeat only | 04:52 | 04:47 repeat omitted |
| Schuricht 9.1955 | 08:09 repeat omitted | 07:16 1st repeat only | 04:52 | 04:45 repeat omitted |
| Schuricht 1959 | 06:55 repeat omitted | 07:05 1st repeat only | 04:45 | 04:32 repeat omitted |
| Sawallisch | 07:14 repeat omitted | 06:06 1st repeat only | 05:32 | 04:59 repeat omitted |
| Monteux | 06:54 repeat omitted | 06:46 1st repeat only | 05:05 | 04:48 repeat omitted |
| Willcocks | 07:20 repeat omitted | 08:33 1st repeat only | 05:22 | 05:14 repeat omitted |
| Giulini | 07:24 repeat omitted | 07:40 1st repeat only | 05:17 | 05:19 repeat omitted |
| Mravinsky | 08:36 with repeat | 07:50 1st repeat only | 05:11 | 06:07 with repeat |
| Komives | 07:33 repeat omitted | 07:28 1st repeat only | 05:35 | 05:16 repeat omitted |
| Martinon | 06:47 repeat omitted | 07:36 1st repeat only | 04:58 | 04:30 repeat omitted |
| Ančerl | 06:10 repeat omitted | 06:28 1st repeat only | 05:35 | 04:29 repeat omitted |
| Dixon | 09:52 with repeat | 07:35 1st repeat only | 06:07 | 05:28 repeat omitted |
Christopher Howell © 2026
[1] Rowicki’s live performance was discussed in lieu of his studio recording, which I was unable to hear.
[2] https://it.scribd.com/document/342450854/TOSCANINI-by-Harvey-Sachs-Supplementary-Chapters, viewed 19.7.2026.
[3] Though I wish I could find a hard and fast quotation in which Celibidache says this, as opposed to philosophical outings by people who claim to know what Celibidache believed.
[4] They were both available on René Gagnuaux’s site “Mon Musée Musicale” but at the time of writing the links do not work. There seem to be several YouTube options, though it is not clear which of the two performances is offered.
[5] https://www.abruckner.com/down/editorsnote/listsanddata/giulinicarlomariac/giulini_-_concert_register.pdf, viewed 23.7.2026
[6] A fairly detailed biography can be found here: http://perctek.com/index.php?title=Komives,_Janos, viewed 25.7.2026.
[7] http://www.karel-ancerl.com/pagb21.php?ID_COMPOSITEUR=53&comp=Haydn, viewed 25.7.2026
