We begin by examining the most salient feature of the GMdm: its scoring. Berlioz has asked for quite an array of performers; the Schlesinger full score of 1838 calls for the following: four flutes, two oboes, two English horns, four clarinets, eight bassoons, twelve horns, eight pair of timpani, two bass drums (one apparently a small bass drum in B Flat), one tamtam, three pairs of cymbals, fifty violins, twenty violas, twenty celli and eighteen double basses; the chorus consists of seventy soprani (which presumably subdivide into parts where appropriate), sixty tenors, and seventy basses; and the four brass bands: North, four cornets, four trombones, and one bass ophicleide; East, two first trumpets, two second trumpets, and four trombones; West, four trumpets and four trombones; South, four trumpets, four trombones, and four ophicleides; and of course in addition to all of this the solo tenor for the Sanctus.1 Berlioz further allows that "the Chorus may be doubled or tripled and the orchestra be proportionately increased. But in the event of an exceptionally large chorus say 700 to 800 voices, the entire chorus should only be used for the Dies Irae, the Tuba Mirum and the Lacrymosa, the rest of the movements being restricted to 400 voices."2
Obviously one of the first decisions to be undertaken is the size of the ensemble. Usually this will be dictated by economic and practical considerations, though it might be hoped that any musical organization which undertakes to perform the GMdm will attempt to follow Berlioz' directions as far as possible. Unfortunately it has not been possible to determine the size of the ensemble for each of the recorded performances, though in some cases available information will be related in the next chapter, under the separate discussion of each recording.
Next might be considered the matter of the "unusual" instruments. The bass of the ophicleide family was not an uncommon instrument in outdoor wind bands, particularly in France in the 19th Century, but it is quite obsolete now, and parts written for it in the standard literature (such as Felix Mendelssohn's Overture to A Midsummer Night's Dream) are commonly played on the bass tuba. The ophicleide has a coarser tone than the more sophisticated tuba, so it could be argued that Berlioz' tone colors work only when the proper instrument is used. On the other hand, modern tuba technique is such that the suitably-minded player may probably be able to achieve the proper tonal effect, and in any event it is surely easier to find competent tuba players and instruments in working order than it would be to find ophicleidists and ophicleides. It is not definitely known that any of the recordings utilize ophicleides, nor is it known if any performances of the GMdm in recent times have used ophicleides.
Next comes the question of the cornets in the first (North) brass band. Since the other brass bands have trumpets, it is evident that Berlioz did in fact have this particular lighter, more pointed tonal quality in mind for the first band. In the available recordings, there are some instances in which cornets have apparently been used, and many instances in which the lack of difference in tonal quality suggests that trumpets have been substituted. The problem of instruments and players is nowhere near that in the case of the ophicleide; but trumpets have become a common substitution for the cornets.
The size of the chorus is another major consideration, particularly since many performances of choral works are given by unpaid choral ensembles, and the problem of paying the additional performers essentially does not exist. There do remain the problems of organization, training, and space for the singers on their risers (a terrible problem in the 1972 Weiss performance, preserved as WE, in which the present author was a participant). Given the right circumstances and the proper training, a fairly large chorus (in excess of two hundred voices) will be able to carry Berlioz' imaginative vocal writing quite well.
The question of locale is not merely a practical consideration, but an artistic one as well. Two hundred each of instrumentalists and singers need to be situated in such a fashion that the physical demands of Berlioz' music be met. Perhaps the most practical situation is to have the chorus on risers in back of the orchestra, with the smaller ensemble (if such is being used for the more intimate movements) concentrated toward the center of the group to facilitate ensemble. The four brass bands are ideally located at the four corners of the ensemble--just what Berlioz meant by his compass points if the ensemble is not a round or square one is not certain--but in modern performances, particularly those given in concert auditoriums rather than in cathedrals, balconies and organ lofts will sometimes provide suitable vantages, particularly if the antiphonal effect is to be emphasized. (See the next chapter for comments on quadraphonic recordings.) A performance by an amateur ensemble from Rochester, Minnesota under Jere Lantz was given at Orchestra Hall in Minneapolis on 7 March 1982, and the brass ensembles (half-sized to begin with) were scattered across the stage along with the orchestra and chorus, minimizing the antiphonal effects, so it may be seen that there are also pitfalls if Berlioz' concept is not followed at least in spirit. When it has been possible to determine the direction and placement of the brass bands in the twenty recordings, this has been mentioned; however, it must also be considered that technical considerations such as microphone placement, tape mixing, and so forth have been used in the case of the commercial recordings in order to optimize the composer's effects.
The purchase, or rental, of parts should not pose any particular problem insofar as variant editions are concerned. The edition most commonly in print is based on the Malherbe-Weingartner, or OBE, which in turn is based on the posthumous third edition. There are numerous differences in phrasing and accents but few variations of actual note readings, save for the discrepancies in the ophicleide parts (chiefly in matters of octave displacement in the second "Hosanna" fugue), noted by Barzun.3 These matters may easily be set aright by consulting NBE. Fortunately in this work there are no decision-making problems such as there are, say, in Händel's Messiah, where alternate versions of arias, choruses, and number ordering abound.
The Tenor Solo
One major consideration in producing a performance of the GMdm, and one which must be dealt with subjectively, is the selection of a tenor soloist to sing the Sanctus. Since the twenty recordings offer a sort of an overview of the types of singers who have been engaged to sing the part in recent years, they will be examined critically here.
The question must first be addressed: What sort of a voice did Berlioz have in mind for the part? Bearing in mind the ecstatic nature of the music, from the background string haloes to the fairly high solo tessitura, the impression might be of a floated tone, with plenty of voix mixte or outright head tone in the upper reaches. Then too, Berlioz was engaged in the composition of Benvenuto Cellini when he received the commission for the GMdm, and Cellini is a lyric tenor. This solo is not for a Faust, nor certainly an Aeneas!
The tenor who sang the first performance in 1837 was Gilbert-Louis Duprez (1806-1896), of whom Cairns says the following:
Originally a rather light lyric tenor, he developed darker tones and considerable power in Italy, where he sang Arnold in the Italian première of William Tell (1831) and created the role of Edgardo in Lucia di Lammermoor (1835). He was reputedly the first tenor to sing top C as a chest note [though see below], with electrifying effect--though Rossini likened the sound to "the squawk of a capon having its throat cut" and considered that the note in question (in Arnold's aria in Act IV) should be sung in head voice, as [Adolph] Nourrit, the first Arnold, had done. Berlioz admired Duprez' developed "Italian" quality, preferring it to Nourrit's (Correspondance inédite, 97-8), but he soon came to doubt his sensitivity and artistic humility.4
Since Berlioz was working on the GMdm--and Cellini, for that matter--in the period immediately preceding Duprez' Opéra debut on 17 April 1837, it is possible that he had that sort of voice, with chest voice still used above the staff, in mind both for the Sanctus and the role of Cellini. Indeed, while it is the 17 April date which is given for Duprez' debut, and therefore traditionally the beginning of the popularization of high chest voice singing in tenors, Stefan Zucker points out that Hippolyte Prévost, in the Revue de Théâtre for 3 February 1835, reported that "[Giovanni Battista] Rubini emitted a high C from the chest in a rendition he gave, during an intermission of a performance of I Puritani, of a cavatina from Pacini's Niobe--this two years before Duprez' sensational debut at the Opera."5 Thus it is evident that Duprez' high C from the chest, and presumably his B Flat and A as well, were not entirely new and original sounds to the French public, and Berlioz was almost certainly aware of the vocal possibilities available to him.
It is, however, very difficult to sing quietly in a high range, and so one may suppose for the sake of argument that some sort of voix mixte, that is to say, a mixture of the floated head voice and the full, resonant chest voice, is the most appropriate tonal quality for the solo in the Sanctus.
With this in mind, the tenors in the twenty recordings may be examined, with a brief subjective description of their vocal qualities.
Georges Jouatte (FO), the first recorded exponent of the part, has a light, clarion voice which exhibits no strain in the upper registers. Ray de Voll (HO) has a small, wavery, but lyric voice, the sort typically found in English oratorio performances. Léopold Simoneau in Salzburg (MI) has a full, lyric sound, with however very little dynamic differentiation. David Lloyd (MH) has an adenoidal tone, somewhere between lyric and heroic, but with a pronounced wobble. Jean Giraudeau (SC) has a thin, small voice, somewhat reedy in character; he utilizes a very marked voix mixte on top, as well as some portamento.
Simoneau, this time in Boston (MX) is still warm and lyric, but now shows more brilliance on top, as well as some straining at, and some scooping into, the high notes. Richard Lewis (BE) has a very full and heroic voice, with pronounced wobble. Cesare Valletti (OR) has a warm, thin, lyric voice, sounding aged but still somewhat elegant, and with a heroic ring on the top notes. Peter Schreier (MZ) has a small, focused, almost Baroque tone, with a slight bleat on top.
The Soviet performance (CH) is a special case and will be dealt with separately. Charles Bressler (AB) has a very sweet, lyric voice, with almost the flutelike quality of a countertenor. Ronald Dowd (DA) is focused in tone, with some pronounced adenoidal qualities, along with glottal sounds, some sloppy phrasing, and a sign of strain on top. William Johns (BU) has a weightier voice than most, its thick tone mostly well-controlled, but with a heroic ring on top. Kenneth Westrick (WE) has a small, tight voice, very bleaty in the upper register. Jon Gilbertson (AV) is robust, lyric, and well-controlled, and forceful on top.
Robert Tear's first recording (FR) shows him burly and tight, heavy on top, with pronounced glottal sounds. Stuart Burrows (BS) is extremely sweet and lyric, with a lovely light voix mixte in the upper notes. Kenneth Riegel (MA) has a small yet lyric voice, with some glottal sounds and a slight squally spread on top. Placido Domingo (BA) has a very full, ripe, lyric voice, extremely well-controlled, somewhat heroic in quality throughout its register, and plangent on top. Tear's second recording (PR) is again burly and tight, with the same glottal sounds (made worse by the close microphoning and the excellent recording quality); this time he shows a fine but precarious voix mixte on top, although the voice is now somewhat strained and requires scooping up to some of the higher notes.
The special case of CH is its exercising of a performance option apparently rarely taken. On the first page of the Sanctus, Berlioz allows: "This Solo may be sung by 10 tenor-voices in unison."6 The anonymous annotator of the Soviet album has gotten this option completely turned around, and apparently considers it rather a recommendation: "In the only number ('Sanctus') which Berlioz singles out for a tenor solo, he proposes immediately that ten voices perform it; thus, again, emphasizing his monumental intention."7 This appears to be what in fact has happened in CH, as a rather diffusely-recorded group of tenors, timid-sounding, lacking in ensemble and short on breath, seem to lunge for the top notes. It is a remarkably disappointing version.
Some of the solo tenors take liberties with the score, beyond the expected variations of phrasing, dynamics and tone. The most notable of these is the reshaping of the very opening word, "Sanctus." In measure 6, Berlioz calls for a half-note D Flat, phrased to a dotted quarter F, on the syllable "Sanc-," followed by an eighth-note E Flat "-tus." The tenors in MI, BE, AB and PR perform this with the "-tus" phrased through the F and E Flat, with all performances save PR showing the same change at measure 93, the repeat of essentially the same material. This reshaping can be traced to corrupt editions of the score, such as Leopold Damrosch's 1880 vocal score for Schirmer, and an undated Brandus vocal score. Tenors in MI, DA, and BU extend the "u" vowel in "tua" in measure 43 for the next one-and-a-half beats, presumably to enable them to keep singing the more focused vowel through the minor-seventh leap.
With some of these subjective procedures out of the way, we proceed now to a full-fledged aesthetic criticism of the twenty whole performances.
Copyright © 1983, 1995 by Matthew B. Tepper
Table of Contents
Back to The Hector Berlioz Page
Back to Matthew B. Tepper's Home Page