• Home
• About CHARM
• Discography
• Sound Files
• Research Projects
  • Analysing Motif
  • Gesture in Schubert
  • Mazurkas Project
  • Recording Business
• Outputs
• Analysing Recordings
• History of Recordings
• Download Area

Purely coincidental? Joyce Hatto and Chopin's Mazurkas

Nicholas Cook and Craig Sapp

The story of Joyce Hatto is one of the most inspirational in recent musical history. Born in 1928, she was known through the 1950s and 60s as a concert and recording pianist specializing in twentieth-century British music, Liszt, and Chopin. In 1970, however, she was diagnosed with cancer, which prompted her withdrawal from the concert hall by the end of the decade. For twenty years she was little heard of. But then came the extraordinary final phase of Hatto's career, in the form of an outpouring of recordings that encompass many of the summits of the pianistic repertory from Scarlatti to Messiaen. The Presto Classical website currently lists a total of 104 CDs issued on the Concert Artist/Fidelio label owned by Hatto's husband and (in Ateş Orga's words) 'unsung producer', William Barrington-Coupe; most of these have been issued since 2003, but the recording dates range from 1989 up to shortly before Hatto's death at the age of 77 in June 2006. Critical response to the recordings varies, as critical response always does, but Richard Dyer of the Boston Globe spoke for many when in 2005 he described her as 'the greatest living pianist that almost no one has ever heard of'.

'Comparatively little of the Joyce of those days', Orga writes of the period before Hatto's illness, 'was to give any indication of the flood, the phoenix-like re-invention of herself, to come twenty, thirty years later'. But what makes the story of the Joyce Hatto recordings the more astonishing is the contrast between the acute difficulties of her everyday life and the scale of her output. Dyer's obituary tells us that 'Miss Hatto underwent five major surgeries and for the last 10 years of her life was hospitalized for several days every seven weeks—68 scheduled trips to the hospital in all'. (He also tells us that she recorded materials for about 50 further CDs, now awaiting production.) No wonder then that Orga's MusicWeb obituary refers to Hatto's 'superhuman energy'; elsewhere he details one of her marathon recording sessions ('assuming correct documentation', he notes), and adds, 'impossible, many cynics would uphold'. Bryce Morrison's Gramophone review of the Liszt and Chopin-Godowsky studies (2006) similarly refers to the 'doubting Thomases, of which there are apparently many', who 'may well wonder how Joyce Hatto achieved such unalloyed mastery and musicianship when tragically beset with ill-health. But others will surely celebrate an awe-inspiring triumph of mind over matter, of the indomitable nature of the human spirit'.

Readers of Gramophone will have picked up Morrison's reference to the controversy following Jeremy Nicholas' article 'Piano Dreams' in the March 2006 issue of that journal: a number of readers wrote in to question various aspects of the story, for example whether Hatto was really ill or even whether the recordings issued in her name were hers at all, and in the July 2006 issue Nicholas challenged the doubters to substantiate their accusations by providing evidence that would 'stand up in a court of law'. There were no takers, but the mutterings continued. Indeed they were given a new lease of life in what one contributor called 'one of the most bizarre threads in the history of this newsgroup' ('Hatto hoax?', on rec.music.classical.recordings). It started when, in January 2007, a contributor to the group who had recently purchased some of Hatto's CDs reported, 'I have noticed something eerie: that the pianist playing the Mozart sonatas cannot be the pianist playing Prokoviev or the pianist playing Albeniz. I have the distinct feeling of being the victim of some sort of hoax. Does anyone else share these feelings?'

A few contributors offered reasoned responses. One attacked the basic premise of stylistic consistency: 'Many pianists have personalities that vary greatly with repertoire. I have heard pianists seemingly transform personalities within the space of one recital on countless occasions'. Another questioned what the motivation of any such hoax could possibly be, asking 'why on earth any pianist would choose to go into a studio and record literally hundreds of works in the sure and certain advance knowledge that they would all be issued under someone else's name'. Other contributors, however, resorted to simple counter-assertion: the first response to the original posting read 'You are wrong. I suggest that you trade in your ears for another set.' But nobody had any hard evidence, and the discussion increasingly floated free from reality, at one point questioning whether such a person as Joyce Hatto had ever existed. The thread ended, as such threads usually do, in a series of flames, although one of the later contributions summed up the feeling that perhaps lay behind the whole controversy: 'It is hard to believe that one pianist unknown to us suddenly plays every composition in the repertoire better than any other pianist ever did'. Precisely those qualities that elicited some listeners' admiration, it seems, provoked disbelief in others.

The AHRC Research Centre for the History and Analysis of Recorded Music (CHARM) was established in 2004 at Royal Holloway, University of London, in partnership with King's College London and the University of Sheffield. Its mission is to promote the study of recordings within musicology, and its portfolio of activities include a project on recordings of Chopin's mazurkas. We have to date built up an archive of over 1500 recordings of individual mazurkas, including nearly 30 complete sets, and the purpose of the project is to devise computational methods for extracting and analysing musically significant information from recordings. Our focus is principally on timing and dynamic information, and on this basis we hope to draw conclusions regarding, for example, the discovery of historical trends in Chopin performance and the characterisation of different national styles of playing. But as will become clear, other kinds of conclusion can also be drawn from the techniques we have developed.

We make considerable use of different forms of visualisation, including what we call 'timescapes'. As the name implies, these are based on timing information, specifically the duration of each beat in the performance: what they show is where in a particular recording one pianist's rubato is most like another's. A particular advantage is that they compare relative timings, in other words they show relationships that would otherwise be hidden by the fact that one recording is globally faster or slower than another. (In general it's relative rather than absolute timing that matters most in characterising pianistic style.) A typical example is Figure 1, which shows Arthur Rubinstein's 1939 recording of the Mazurka Op. 68 No. 3; the colours indicate which other recording Rubinstein's most resembles at each point. (Each recording is represented by a different colour.) The horizontal axis represents musical time, from the beginning to the end of the piece, while the vertical axis shows how far the similarities persist into the higher-level structure of the piece. What Figure 1 is saying is that on the large scale, the first third or so of Rubinstein's 1939 recording is most like his own 1966 recording (represented by the large patch of orange), but thereafter it is closer to the grand average of all the recordings for which we have data (that's the black). Elsewhere, in particular close to the bottom of the triangle, there are flecks of different colours, which represent more fleeting similarities to a number of other recordings, without any very obvious patterns emerging.

Figure 1 is typical of the way different recordings relate to one another; this is how we expect timescapes to look. Occasionally, however, we find a quite different sort of picture, as in Figure 2. This shows the recording of Op. 68 No. 3 by Jerzy Śmidowicz on the compilation Fryderyk Chopin, Complete Works: The Golden Age of Polish Pianists (Muza PNCD 300), and it looks the way it does because of its extreme similarity to the Śmidowicz recording on The Great Polish Chopin Tradition V: Śmidowicz (Selene CD-s 9905.50). Actually they are more than similar: they are reissues of the same original recording, first released in 1948 (Muza 1345), and the small markings at the bottom of Figure 2 merely reflect limitations in the accuracy of data capture.

Normally such duplications of the same recording on different releases or reissues occasion no surprise. We were, however, taken aback when we saw the very similar image at Figure 3. This shows the Joyce Hatto recording of Op. 68 No. 3, taken from the Concert Artist/Fidelio box set Chopin: The Mazurkas, which was released in 2006 (CACD 20012). According to the booklet accompaying the CDs, the recordings were made at Concert Artist Studios, Cambridge, on 27 April 1997 and 19 March 2004, though the slipcase changes the latter date to December 2005; the Concert Artists website adds the information that the recording was 'revised by Joyce Hatto shortly before she died' and 'completely remastered'. But all this is puzzling, because what Figure 3 shows is that this recording of Op. 68 No. 3 is virtually indistinguishable from that on a commercial recording credited to Eugen Indjic. Currently available on the Calliope label (Intégrale des mazurkas: Frédéric Chopin, 3321), this recording was first issued in 1988 by Claves (Chopin: 57 Mazurkas, CD 50-8812/13). Nor is it just Op. 68 No. 3 which exhibits this apparent identity: Figure 4 shows the Mazurka Op. 17 No. 4, and again the Hatto and Indjic recordings are virtually indistinguishable.

Timescapes are one way of expressing the relationship between the timing of different recorded performances; another is statistical (Pearson) correlation, where 1 represents total identity and 0 the lack of any relationship. The following table shows correlations for both Op. 17 No. 4 and Op. 68 No. 3, in each case comparing the relationship between the Hatto and Indjic recordings with relationships between different recordings by the same perfomer (Rubinstein 1939 and 1966), and recordings by different pianists (Indjic and Rubinstein 1939).

Op. 17 No. 4

Op. 68 No. 3

The figure of 0.996 for the Hatto and Indjic recordings of Op. 68 No. 4, as against 0.993 for the two reissues of the same Śmidowicz recording, is saying that the Hatto and Indjic recordings of Op. 17 No. 4 resemble one another just as closely as do the two releases of the same Śmidowicz performance. (The figures are not 1.00 because of limitations in the accuracy of data capture.) And it's the same story for the complete set of mazurkas. Both the Hatto and Indjic recordings contain the same 57 mazurkas, although this is not obvious at first sight, since the works without opus number appear in different places, and for some reason Op. 41 No. 1 appears twice on the Hatto set (at the end of the first CD and again at the beginning of the second). In addition the durations of the tracks vary slightly. This is partly because they incorporate varying amounts of silence at the beginning or end of the track, but it is also because they play at slightly different speeds; in the case of Op. 17 No. 4 the Hatto version plays about 0.7% slower, whereas in the case of Op. 68 No. 3 it is 2.8% slower—and in the case of another mazurka, Op. 24 No. 2, it is 1.2% faster. (Because our timescapes are based on relative timing, they show the similarities between the recordings despite these changes.) Even with the timing changes, however, the correlation for the entire set of track durations is far closer for the Hatto and Indjic sets than for the others:

These graphic and statistical analyses are valuable in that they bring a degree of objectivity to the comparison between the recordings. For many readers, however, the most convincing demonstration of the relationship between them may simply be to listen to them side by side. While we would recommend purchasing the original CDs for this purpose (both are readily available), we have made a demonstration available here: this is a version of Op. 17 No. 4 in which the left track is taken from CACD 20012 and the right track from Calliope 3321, with the timing of the latter slowed down by 0.7% so that the two recordings synchronize. (In each case we have used the left track of the original recording.) We believe that for most readers this will settle the matter.

The primary purpose of this article has to provide an insight into the potential of the computer-based methods of analysing recordings we are developing at CHARM: approaches which might be considered of purely academic interest turn out to have very practical applications. But there is only so much that analysis can achieve. We have documented the similarities between CACD 20012 and Calliope 3321, which in our opinion demonstrate beyond reasonable doubt that, despite the packaging, both recordings are of the same performance (or set of performances). We have no basis on which to speculate as to how this situation might have come about. It would, however, be very desirable for those who are in a position to clarify this to do so. In the obituary of Hatto which he published in The Independent, Ateş Orga cited the judgement of the American pianist and Horowitz pupil, Ivan Davis, that Hatto 'will have extraordinary posthumous acclaim'. There is a danger that, in the absence of such clarification, uncertainty as to the nature and extent of Hatto's recorded legacy may undermine the lasting reputation of a remarkable musician. That would be a sad end to an inspirational career.

Discography

Bibliography