The surprising demise of the opera film

Faced with this overwhelming contemporary fondness for the musical, from both audiences and critics, the demise of the opera film, which as I shall demonstrate below was once one of the most popular subgenres of the musical, is on the face of it puzzling. There are, I would suggest, a number of reasons for it. First and foremost is the type of music involved: opera films are by default what most people might call classical music, whereas what most people understand by the term musicals is, again by default, popular music. But this cannot be the only reason, as it begs the question of why the opera film was so popular during the 1980s, as well as the question of why opera music is frequently incorporated in film soundtracks, and why the live screening of opera in film theatres has become increasingly popular.

Live screening of staged operas, often in multiplexes as part of the specialist offer to customers, is the second reason why there might be less of a perceived market for opera films. Although live screening in a cinema does not necessarily carry the same affective weight as the ritual of going to the opera, and may appear to be more vulgarisation than democratisation, nonetheless it is attractive to some opera lovers because the experience of an opera is closely associated with liveness (the Benjaminian argument of a work of art’s ‘aura’; see Benjamin 2008) and with the whole work, whereas opera films of necessity tend to abbreviate the original. Peter Gelb, the general manager of the Metropolitan Opera of New York, pioneered live screening from 2006, incorporating behind-the-scenes material (see Affron and Affron 2014: 358–361). As he made clear in a press release, ‘this is not an opera film […]. This is live coverage of an opera event. Between the moving camera action and the behind-the-scenes coverage […] it’s more like Monday-night football than an opera movie’ (quoted in Armatage 2012: 223). As Kay Armatage points out, this has led to an expansion (and consequently democratisation) of exposure to opera:

While the multiplexes are managing to cater to old-style opera devotees, they are also attracting millions of new viewers, who neither attended opera houses before, nor usually frequent the blockbuster-oriented multiplexes. So maybe there are two kinds of new spectators: those that are new to opera and others – already opera aficionados – who are strangers to the multiplex. (220)

Partly in response to this, and to the development of online technology, by 2015 a number of opera houses were streaming operas live for free (see Cooper 2015).

Third, there is another significant development that could be considered to diminish the market for the opera film as a cinematic experience: the incorporation of multi-media for stage versions, so that staged opera becomes in some respects more ‘cinematic’. Peter Sellars’s 2005 production of Wagner’s Tristan und Isolde, for example, was dominated by Bill Viola’s work on an immense screen above the stage. Krzysztof Warlikowski regularly uses video and films in his productions, either by screening clips (Kubrick’s 2001: A Space Odyssey [1968] for Wagner’s Parsifal in 2008) or by very clear visual references, such as the combination of Marilyn Monroe in The Seven Year Itch (Billy Wilder, 1955) with images from King Kong (Merian C. Cooper and Ernest B. Schoedsack, 1933) for the 2007 stage production of Janáček’s The Makropulos Affair.

Finally, there is the gradual rise of the musical biopic. It is striking just how many singer biopics have been produced in the last decade, particularly in the case of French films, as can be seen from Table 2. The biopic has overtaken the French equivalent of the heritage film –the film patrimonial – in popularity, as many commentators have pointed out. A whole book has been devoted to the biopic genre by one of the most high-profile French researchers (Moine 2017; see also Moine 2010 and 2011). While on the one hand these films attest to audience interest in musical genres, on the other it is clear that they have supplanted to some extent not just the standard heritage film but also the opera film, in so far as it could be considered to be a variant of ‘heritage’.

All of these points suggest that there is a kind of atomisation of the opera as material for film: it is almost as if the difficulties in combining the two media have led to opera remaining in film as no more than a fragment, rather than as a work in its own right. This is even reflected in academic scholarship. The early major monographs on the opera film focus on complete films (Tambling 1987, Citron 2000), with subsequent books moving to the occasional and often fleeting presence of opera within films (Joe and Theresa 2002, Grover-Friedlander 2005, Citron 2010). This fragmentation was arguably prefigured in the 1980s group of opera films by the curious portmanteau film Aria (1987), a collection of short films by ten directors, each film illustrating an aria (see Joe 1999 and, for the relatively successful Jean-Luc Godard contribution, Cook 2000).

In the rest of this article, I want to look back at the opera film, always seen as problematic by critics, and consider how the last few opera films made in France managed to address the problems raised by the combination of the two media, anchored in the tension between realism and faithfulness: how to naturalise singing so it does not appear through the singer’s efforts to be what it really is, a performance (the solution being dubbing either by the singer-actor or by a singer for the actor); how to avoid the static nature of the aria (the solution being a focus on camera movement and editing); how to capture a sense of the real world as opposed to the enclosed space of the theatre (the solution being the incorporation of ‘natural’ rather than studio settings).

The reason for the focus on French films is that the majority of opera films made in the 1980s–1990s, the heyday of the modern opera film, were French-produced or co-produced (see Table 3), thanks in part to Daniel Toscan du Plantier, the director general of Gaumont 1975–1984, who produced the asterisked films in Table 3. As he says in a statement that echoes the title of the book in which it appears, ‘the history of opera films is key to what makes me different […] These are my films, those that I really wanted to produce, those that made me want to transmit their cultural emotion’ (1995: 181–182).² The success of the opera film in the 1980s is largely thanks to his commitment to the genre, and it is reasonable to assume that his untimely death in 2003 at the age of 61 is part of the reason why opera films have ceased to be made.

I shall first discuss the history of the French opera film and after some comments on the most successful of these, Carmen (Francesco Rosi, 1984), focus on three of the last opera films, all, like Carmen, produced by Toscan du Plantier: Boris Godounov (Andrzej Žuławski, 1989), Madame Butterfly (Frédéric Mitterrand, 1995) and Tosca (Benoît Jacquot, 2001). I have chosen these three because they are not as well-known in Anglophone film studies as the more popular opera films from the early 1980s – Trollflöjten, Ingmar Bergman’s 1975 version Mozart’s The Magic Flute (originally a TV film, but generally seen as the first of the major opera films), Don Giovanni (Joseph Losey, 1980), La traviata (Franco Zeffirelli, 1982) and Rosi’s Carmen – the first three of which attracted approximately one million French spectators, and the last over two million. By contrast the later three films I shall be considering in some detail attracted considerably fewer French spectators (see Table 3), signalling the shift in spectator interest away from the opera film.

A brief history of (French) opera films

Opera films were a staple of the silent period, with a resurgence during the 1940s and more particularly the 1950s, as can be seen in Table 4.³ As that Table demonstrates, in the 1940–1960 period Italian cinema dominated the subgenre, with the Italian director Carmine Gallone being the most prolific director. There was a lull in production during the 1970s, before the surge in the 1980s. The history of French opera films is of necessity brief, at least prior to the surge of the 1980s, because unlike the silent period during which there were many French-produced opera films,⁴ there are very few of them in the sound era, compared to the Italian or Soviet film industries as listed in Table 4. There were many musical films, especially those incorporating the popular songs of the realist singers in the 1930s (see Conway 2004), and then the popular tenors during the 1940s and 1950s, such as Tino Rossi, Georges Guétary and Luis Mariano (see Powrie 2014 for the latter); but where so-called classical music was concerned the emphasis, as is the case in contemporary musicals, was on the biopic:

• Chopin (biopic): La chanson de l’adieu (Géza von Bolváry and Albert Valentin, 1934)

• Beethoven (biopic): Un grand amour de Beethoven (Abel Gance, 1937)

• Charpentier (opera): Louise (Abel Gance, 1938)

• Schubert (biopic): Sérénade (Jean Boyer, 1940)

• Berlioz (biopic): La symphonie fantastique (Christian-Jaque, 1942)

• Lizst (biopic): Rêves d’amour (Christian Stengel, 1947)

• Offenbach (biopic): La valse de Paris (Marcel Achard, 1949)

As can be seen from the list above, Abel Gance directed two major films in this early period, one a straight biopic, the other the first major French opera film, the adaptation of a well-known opera by Gustave Charpentier (1900), Louise. Surprisingly, there were very few opera films in France in this period. La vie de bohème (Marcel L’Herbier, 1945) is based on the Henri Murger novel (1851) adapted by Puccini, and uses Puccini’s music as background,⁵ while at the opposite extreme Le barbier de Séville (Jean Loubignac, 1948) was not an adaptation but the filming of the very successful production of Rossini’s opera by the Théâtre national de l’Opéra-Comique. Neither of these can be considered an opera film in the sense of a screen adaptation involving actors and singers singing arias.

Gance is best-known for his silent films, in particular the anti-war J’accuse! (1919), the formally complex La roue (1923), and the epic six-and-half-hour-long Napoléon (1927), restored by Kevin Brownlow in 1979. It is hardly surprising, given Gance’s formal experimentation in the two latter films in particular, that he should explore a variety of techniques to integrate the music and the visual tracks in Un grand amour de Beethoven and Louise.

Gance was intent on adapting operas for the screen: a 1929 note by him suggests that it would be possible to ‘film the whole of a Boris Godunov, a Golden cockerel, or a Thaïs’ (cited in Icart 1983: 253; he is referring to operas by Mussorgsky, Rimsky-Korsakov and Massenet). His intention for Louise was to create ‘a marriage between cinema and opera’ (Gance cited in Caumartin 1938: 1068), as well as to privilege music so that it was not ‘cladding, external, a kind of filler, background sound, but as important as a character or as an image that carries more weight by what it suggests than what by what it shows’ (Gance cited in Doringe 1939: 8–9). This may partly account for the fact that Gance, following Charpentier’s recommendation, changed the hero’s profession from that of a poet to a poète-musicien (poet-musician) in the film (Niccolai 2016: 82–83). His problem, as Roger Icart points out, was how to avoid the music becoming secondary to what we see on screen, either by overloading the image track on the one hand or going for realism on the other (1984: 288); his solution was to film everything using his own invention, the painted backdrops of the Pictographe (Niccolai 2016: 84 n47)⁶ in the studio with décors by Georges Wakhevitch, studio décor being an aesthetic choice that Benoît Jacquot was later to emulate for Tosca.

One of the key challenges for the opera film, as mentioned above, is what to do with arias. Focusing on a singer singing, as opposed to the typical Hollywood film musical in which characters sing and dance at the same time, creates even more of a pause in the narrative, a potentially unnatural interlude. This can be minimised in the case of a narrative focusing on the music, such as the backstage musical; but in Louise, despite the fact that the hero is composing an opera called Louise, the musical numbers are not rehearsals. To avoid the image becoming static, especially during musical numbers, Gance used a very mobile camera with frequent travelling shots, proudly pointing out in interview that ‘all my shots are mobile’ (Caumartin 1938: 1069), an aesthetic choice emulated by Andrzej Žuławski for Boris Godounov (1989).

Henry Malherbe, a prominent music critic of the period, was impressed by the integration of music and dialogue. In his view Gance succeeded in creating a new type of film; picking up on Charpentier’s description of his opera as a ‘roman musical’ (a musical novel), he commented that ‘Louise is neither a filmed opera […] nor a film musical […]. It’s a cinematic work, quite distinct from the original musical novel’ (1939: 9). This was partly achieved by turning recitatives into dialogue, with musical numbers interspersed in the action. This solved another key problem for the opera film: what to do about realist ambient sound. But the film as a result is a standard sentimental melodrama with high-point musical numbers, even if this was unusual for the time. The male lead, Georges Thill, rather floridly explained why he thought that the rearrangement of the score worked well: ‘Song happens only when situations have reached a climax […] Then, as if words were incapable of expressing the characters’ feelings, song rises up and lifts the audience on the wings of music’ (Caumartin 1939: 1068). We can gauge just how new this might have been for spectators at the time by Thill’s evaluation of the procedure: ‘This way of concluding each scene, each dramatic situation by musical effusion is quite sensational’ (1068; emphasis in original).

Thill’s comment that the music bursts out when characters cannot contain their emotions is a standard melodramatic device. When compared with Un grand amour de Beethoven, however, the film is not melodramatic enough and appears tame in comparison, as Claude Beylie writing in the 1980s pointed out in his note on Louise: ‘Melodrama is gestured at, but Gance’s filmmaking requires him to wallow in it’ (1987: 65). This can be seen in Louise’s storm sequence (0.40–0.47), clearly an attempt by Gance to capitalise on what had been seen as a very successful piece of filmmaking in Un grand amour de Beethoven’s equivalent storm sequence when Beethoven, suddenly deaf, recovers his hearing. Louise has finally agreed to run away with Julien, and he takes her to a safe place. He returns to his apartment when a storm breaks out. Louise, frightened, runs back to him, only to find her friend Lucienne with him. Believing the worst, she runs back out into the rain (see Figure 1). We find the same sound effects and lighting effects (thunder and lightning) as in Un grand amour de Beethoven. But there is no music at all, and Gance uses somewhat hackneyed expressionist techniques (strong lighting effects, chiaroscuro, angular décor, canted frames) to mimic the Storm und Drang of the previous film. Louise comes across as less creative in its use of the music track; indeed, Beylie dismisses the film as an ‘academic pot-pourri’ (1987: 65).

Figure 1

Canted low-angle shot as Louise leaves Julien’s apartment (René Chateau).

The French opera film 1980–2000

As mentioned above, opera films in France did not appear after Gance’s Louise until the 1980s, alongside other internationally-produced opera films. Amongst the reasons for the development of this type of film, there is ‘the growth of higher education worldwide, increased prosperity and demographic factors’ (Wood 2006: 190), as well as industry-specific factors, such as the requirement to develop new markets, in this case an international market, foreign languages with or without subtitles being less of an issue for Anglophone audiences in particular (see Duault 1987: 3). Television also played an important role in the dissemination of opera on screen (Citron 2000: 40–42, 50–52), as can be seen in the case of one of the most popular operas, Carmen. There were ten film versions globally 1930–1950 (only one being French); from 1950–1980 there were 17 (again only one French) and eight TV versions (two of them French). But from 1980–2000 there were nine film versions (four of them with significant French involvement), and in addition 20 TV versions (six of them French) (see Davies and Powrie 2006). Indeed, there are directors who specialised in TV opera films, such as Jean-Pierre Ponnelle (see Citron 2000: 52–56, 2002, 2005) or Petr Weigl. Citron cautions us about the porous boundaries between theatrical and TV versions, however: ‘The Magic Flute [was] originally produced for television (1975) and subsequently shown in movie theaters. It has assumed such an independence from television that it is often considered the first modern film-opera, a status encouraged by Bergman’s position as cinéaste’ (2000: 41). The list in Table 3 is therefore not definitive, but a representative list of opera films with theatrical distribution by well-known directors.

Television may have helped disseminate opera so that it became less elitist and less tied to the opera house, but by the same token it may well also have persuaded audiences to explore the better-quality picture and sound provided by the big screen (see Citron 2000: 58–59). One of the more surprising features of this turn to filmed opera is the fact that the works, although cut, led to often unusually long films (Citron 2000: 60–61). Despite their length, these films nonetheless attracted relatively large audiences, as can be seen in Table 3 which lists the main opera films with the number of French spectators when available.

By far the most successful of the opera films was Carmen. 1983–1984 was the year of Carmen films: Carlos Saura’s flamenco version, Carmen, was released in May 1983, winning the Palme d’Or in Cannes; Peter Brook’s pared-down triptych of films using a small chamber orchestra, La tragédie de Carmen, in November 1983; Jean-Luc Godard’s Prénom Carmen, which used Beethoven string quartets as diegetic music, in January 1984, winning Venice’s Golden Lion. Rosi’s film appeared in March 1984, going on to win a number of César awards including Best Film and Best Sound. There was even a soft-porn opera film, Carmen nue (Albert López), in August 1984. Toscan du Plantier, wishing to popularise opera, had commissioned Rosi because of his track record in social-realist thrillers, as well as the fact that one of his early films, Il momento della verità (1965), focused on bullfighting. Rosi’s film was different from the other Carmens of the 1980s. Whereas the others all in one way or another questioned the myth which Carmen had become, principally by backgrounding Bizet’s music, almost to the point of non-existence in Godard’s case, or by paring the opera down until it resembled something else, as was the case with Brook, Rosi’s film, at least in appearance, presented itself as a realist illustration of the opera. It had location shooting in the Spanish towns of Seville, Ronda and Carmona (Ronda being chosen for its eighteenth-century bullfighting arena), typically Spanish scenes illustrating the entr’acte, and a ‘noise track’ with frequently obtrusive diegetic sounds.

While many reviewers found the film refreshing because of its realism and Julia Migenes-Johnson’s performance as Carmen, the film’s constant appeal to realism irked others. As one reviewer put it: ‘To sit before a live performance of Carmen in the opera house is one thing, but to sit before this exactingly slow imitation suggests listening to a concert of phonograph records while looking through a book of colored photographs of Spain’ (Curtiss 1984). If some commentators felt that the shuffling of feet, the clip-clop of horse’s hooves, the chattering of crickets, and the burbling of birds anchored the music in the ‘real’, most, even those who admired Rosi’s realism, found such features distracting. This reaction underlines the difficulty of the opera film which naturalises the body of the singer, leading to a disjunction between the melodramatic intensity of performance, and the claims of the everyday with its mundane attachment to ordinariness. The reality of flesh and blood, arguably, does not mix with vocality, which, with its links to the pre-oedipal, is ideal and transcendent; opera ‘deliberately turns its face against everyday reality – that’s its chief appeal’ (Cook 1985: 32). For Marshall Leicester, however, there is a productive interplay between Carmen’s noise-track, which approximates the characters’ awareness of being-in-the-world, and the music, which moves towards an ‘unselfconscious ecstasy’ where the world is, as it were, drowned out in a moment of epiphany (1994: 269), recalling Georges Thill’s view of the musical numbers in Louise.

The film’s image track uses – as had both Prosper Mérimée, the author of the novella, and Bizet, who adapted Mérimée’s novella – a nineteenth-century travelogue by the Baron Charles Davillier (1874), illustrated by Gustave Doré (see Powrie 2007). Many shots, and particularly those of the entr’actes, are patterned on Doré’s drawings of Spanish scenes. The historical ‘real’ thus recreated is filtered through Doré’s particular emphatic and Romantic sensitivity. Susan McClary considers that the commitment to realism ‘never lets us forget the centrality of class, imperialism and sexism to the organization of this society’ (1992: 144); however, the commitment to recreating nineteenth-century locations aligns the film with 1980s heritage films, as is the case with Madame Butterfly. In both cases, Toscan du Plantier explains how the heritage ‘real’ had to be fabricated. The locations had to be transformed because of modern developments in the case of Carmen (tarmac, plastic furniture and signs), and in the case of Madame Butterfly a complete Japanese village with lush vegetation had to be built in the Tunisian desert (Toscan du Plantier 1995: 182–189).

Madame Butterfly adopts the same attitude to location and sound as Carmen (we hear crickets when the music fades), but the camerawork is very different. The camera is rarely still, with constant crane and travelling shots, the latter winding through the Japanese house, giving the impression of a labyrinth of interlocking boxes (see Figure 2). Mitterrand’s justification for this was the importance for him of maintaining a link between interior and exterior so that the film was ‘open like a Japanese dwelling, always looking out, like Cio-Cio San as she waits’ (Mitterrand 1995: 16). Mitterrand’s attempt to adopt one of Gance’s solutions, the roving camera, is neutralised by the insistence on showing off the lovingly recreated Japan; there is a tension between the dynamism of the roving camera, and its potential to signify the instability of desire, and the constant picture-postcard refocusing on the interlocking rooms of the Japanese house that leads to an effect of stasis. The detailed and obsessive recreation of the past is all the more curious given the theatrical and performative aspects of opera. Ultimately, were it not for the music, Carmen and Madame Butterfly would be dull travelogues. This is less the case with the following films, each of which rejects the realism of Carmen and Madame Butterfly.

Figure 2

The Japanese house in Madame Butterfly (Columbia).

Boris Godounov was a project initially intended for Andrei Tarkovsky who died during early planning. Toscan du Plantier hired Žuławski, who had been Andrzej Wajda’s assistant in Poland before he settled in France. His breakthrough film was L’important c’est d’aimer (1975, 1.54 million spectators), starring Romy Schneider, while Possession (1981) starring Isabelle Adjani, and Mes nuits sont plus belles que vos jours (1989), starring Sophie Marceau, did less well. Boris Godounov is an opera about guilt and the madness it engenders. Žuławski uses labyrinthine castle spaces with dark tunnels and torchlight, as for example during the chorus ‘Slava! Slava! Slava!’ (0.17; see Figure 3).

Figure 3

Labyrinthine castle spaces in Boris Godounov (Gaumont).

But the most important feature of the film is its rejection of the kind of realism found in Carmen. This is best exemplified by the opening sequence, which begins with the premiere of the opera in 1874, the audience being in period dress, amongst them an anxious Mussorgsky. The curtain rises and we see a modern film camera on a dolly tracking across the stage, followed by the ripping of a canvas curtain that reveals realist exteriors with horses galloping across the screen (0.2; see Figures 4 and 5). As Žuławski explains in Dominique Maillet’s 2014 DVD bonus documentary, opera was not intended to be filmed or to be filmable, so that these effects are an attempt to explain the ‘fabrication’ that the opera film represents. At different points of the film, we see period décor with modern light projectors and camera teams, whether in the forest (0.26; see Figure 6), or in the shadows during the duet between Marina and the Pretender (‘O Tsarevich, I implore you’), who are naked and cavorting on the bed (1.07; see Figure 7). We also occasionally see modern-dressed Soviet-looking soldiers patrolling with their guns. There is a particularly harrowing sequence with bedraggled serfs behind modern concrete posts and barbed wire (0.11; see Figure 8). But we also see Marina playing in her transparent bath with a blue plastic duck (1.02; see Figure 9). The film ends, predictably one might argue, with a return to the theatre where the opera is being shown, the audience now in modern dress, although Mussorgsky is still amongst them (1.48).

Figure 4

The camera crew on stage at the start of Boris Godounov (Gaumont).

Figure 5

The canvas curtain ripped at the start of Boris Godounov (Gaumont).

Figure 6

The forest with camera crew (Gaumont).

Figure 7

The duet between Marina and the Pretender (Gaumont).

Figure 8

Serfs behind barbed wire (Gaumont).

Figure 9

Marina in the bath with a blue plastic duck (Gaumont).

There is a political point being made of course. Žuławski explains in the Maillet documentary that he profoundly disagrees with the conservative Orthodox ideology of the libretto, namely that an assassin can become great, even a saint, if only he repents. As he says in the documentary, his version of the opera is ‘an anti-Boris Godounov’ and that ‘Stalin’s Russia was the Russia of Boris Godounov’. The film was released in December 1989, only five weeks after the fall of the Berlin Wall; as the ‘historical’ notes at the end of the film before the credits point out, ‘this film was made in 1989’, implying that the film is an allegory for Russian imperialism, an unsurprising position to be taken by a director of Polish extraction. Much to the consternation of the French judge the conductor of the pre-recorded music, Mstislav Rostropovich, took Žuławski to court for betraying the ‘Russian soul’, and lost. One of his complaints was that at one point a village idiot urinates in a bucket in front of the church; the noise, Rostropovich claimed, made it impossible to hear the music properly, recalling the complaints made of realist sounds in Rosi’s Carmen. This explains in part the otherwise opaque first title of the film: ‘In the name of his moral rights Mstislav Rostropovitch disapproves of the sound effects superimposed on his interpretation of the opera Boris Godonouv’. As Žuławski points out in the bonus documentary, this was not true, as we hardly hear the trickle of urine at all, but he immediately explains how important the scene is from the point of view of cinema, as it is an echo of a scene from Tarkovsky’s Andrei Rublev (1966), the director originally intended for the film; it is therefore a homage as well as a recognition of the subversive nature of both Tarkovsky’s films for the Soviet State and of the anti-Russian aspect of Žuławski’s version. But this curious episode also demonstrates the key tension in the opera film between realism (ambient sounds) and fidelity (the music), while by the same token illustrating what the opera film can achieve in relation to its film-historical context, emphasising its difference, pace Warlikowski, from most staged opera.

It is not just the décor of Žuławski’s Boris Godonouv that undermines any sense of realism; as with Madame Butterfly, the camera hardly ever stops moving, giving a sense of ‘manic recklessness’, as one reviewer puts it, pointing out that ‘characters sing while rutting vigorously [1.07], and one character completes an aria while plummeting from a minaret’ [0.20] (Cairns 2016; see Figure 10); when the camera does stop, it is generally to focus briefly in extreme and harrowing close-up on the singer. The camera movement is part of a rigorous aesthetic choice, however. Žuławski complains in the bonus documentary that in opera films the visual track and the music often appear to run in parallel, rather than synergistically. He therefore decided to match musical phrases with single shots, leading to highly complex pans and tracks. This makes Boris Godounov a remarkable achievement, matched only, in my view, by the last opera film I want to analyse, Jacquot’s Tosca.

Figure 10

A character sings as he falls (Gaumont).

Like Boris Godounov, Tosca focuses on the performative and constructed nature of opera. The real locations of the three acts of the opera (the Basilica of Sant’Andrea della Valle, the Palazzo Farnese and the Castel Sant’Angelo) were reconstructed in a Cologne studio – a similar procedure to Gance’s studio set for Louise – but with vast uncluttered spaces surrounded by darkness, so that the characters emerge as if from the backstage of a theatre.⁷ The sets are inspired, according to the director of photography Romain Winding, by a variety of painters: de la Tour and Rembrandt for the lighting, and de Chirico, Delvaux and Magritte for the extraordinary use of the floors with lengthened shadows (Breteau-Skira 2010: 129; see Figure 11). In these vast spaces, the focus is also on costume and on isolated items of décor, such as Tosca’s long-trained red dress, contrasted starkly with Scarpia’s black costume (see Figure 12), or the sequence when Scarpia looks at a distorted image of himself in the knife he has been using for his supper (0.49; see Figure 13). The film makes no attempt to hide playback: we frequently see mismatches between what we hear and what the singers’ lips are doing; moreover, on several occasions the characters sing but we hear them repeating the words as ordinary dialogue over the singing (0.20, 0.26), or in a variant effect, we hear them singing but do not see them doing so, as if their song is a kind of inner monologue (0.35). This is an effect that Ponnelle had already used in his TV operas, and Marcia Citron explains its impact: ‘Interior singing allows Ponnelle to reconfigure the private element of an aria – what it means to be airing one’s thoughts, alone – and stage the number as a public spectacle that still remains private’ (2010: 123). One of its key characteristics is its mobility, echoing the mobility of the camera in both Ponnelle and Jacquot’s films; as Citron says, the interior voice ‘can recede and reappear, rearrange time, and redefine venue and agency’ (2002: 149).

Figure 11

The set of Tosca (TF1 Vidéo).

Figure 12

Costume contrasts (TF1 Vidéo).

Figure 13

Scarpia looks at a distorted image of himself (TF1 Vidéo).

Most startling of all is the three-stranded approach to what we see: the full colour studio sequences are punctuated by brief, stuttering and grainy colour sequences of the locations of the film, and by over-exposed black and white sequences of the orchestra rehearsing in the recording studio. The orchestra sequences in particular insist on the constructed nature of what we see, as if the opera can only exist in our imagination with little purchase on the real world. At the climax of the film, for example, as Tosca leaps to her death over the parapet into the darkness, her red train following her in a cloud of dust, we cut to the singer, Angela Gheorghiu, in the recording studio, sighing now that all is done, and chatting with her fellow singers (see Figures 14 and 15).

Figure 14

Tosca plunges to her death (TF1 Vidéo).

Figure 15

Gheorghiu in the recording studio (TF1 Vidéo).

One sequence in particular stands out as a remarkable moment of pure cinema in which theatre, music, camerawork and editing combine in ways that no other opera film has ever done. It occurs at the close of Act 1 when Scarpia, the evil police chief, gloats as he imagines conquering Tosca and getting rid of her lover Cavaradossi. The camera circles around Scarpia in the main film strand, circles under the cupola of the Basilica in the grainy strand, and holds the rehearsing orchestra in the black and white strand (see Table 5 and Figures 161718). The circling camera could be taken to signify Scarpia’s lust for Tosca, enfolding him to the point where he ‘takes off’ on wings of lust; but it could equally well signify the ascendency of the sacred in the battle with the profane, representing Tosca battling with Scarpia. The point here is that the film creates a moment of perfect ambiguity in a whirligig of heightened emotion. It achieves what Gance set out to do with Louise, an opera – like Tosca – dating from 1900. Both Gance and Jacquot rejected the realism that dogs so many opera films; but where Gance’s Louise strains to allow heroic melodrama its full rein, and in the end feels more like a chocolate box, Jacquot’s Tosca not only celebrates melodrama, but interrogates it in postmodern mode as emotional fantasy.

Figure 16

Scarpia in the main strand (TF1 Vidéo).

Figure 17

The cupola in the grainy strand (TF1 Vidéo).

Figure 18

The orchestra in the black and white strand (TF1 Vidéo).

Conclusion

Looking back at the development of the French opera film, it is very tempting, teleology notwithstanding, to see Louise and Tosca as the beginning and the end of what was possible for the opera film, with the various films in between as interesting but failed experiments. Both Louise and Tosca mimic the melodrama and theatricality of opera by rejecting exteriors and focusing instead on studio sets that push what we see towards melodramatic fantasy anchored in artistically reimagined sets. But then we could ask what makes a staged opera with its sets different from a filmed opera with its sets? Apart from the temptation of realism – which means the use of exteriors – that most opera films display, and which Rosi’s Carmen managed so well, the attraction of the more successful opera films lies in their fractured reality. This can be the mise en abyme we find in Boris Godounov with its references to the twentieth century. But it can be the flagrant use of the studio, which we also find in Boris Godounov, as well as in Louise and Tosca. As Beylie wrote when reviewing Louise in 1987, ‘it is the trompe-l’œil décors that charm us’ (1987: 65).

Siegfried Kracauer considered that opera film was ‘a collision of two worlds detrimental to both’ and that at best it could only be ‘an eclectic compromise between irreconcilable entities – a sham whole distorting either the opera or the film or both’ (1960: 154). He follows these comments with his view of what works and what does not work in Louise, the main element being in his view, as in the view of Gance’s contemporaries, the ‘significant camera work’. And he singles out one and only one sequence that in his view manages to achieve ‘a precarious equilibrium between music and imagery’: it is the father’s aria towards the end of the film, a kind of lullaby in which he laments Louise’s passage from childhood into a wilful independence and during which she sits on his knees as he rocks her (1.11). Kracauer writes: ‘The camera approaches the two closely, isolates their faces, and […] draws us into the very vortex of inarticulate childhood, whose memory the song itself evokes’ (154; see Figure 19).

Figure 19

The father’s lullaby in Louise (René Chateau).

I would argue that the most successful opera films combine several elements: the interrogation of realism through elaborate and obvious studio sets and/or mise en abyme; the use of mobile camerawork; and perhaps most importantly in conjunction with these, as Kracauer points out above, the use of close-up which draws us into the emotion vehicled by the song and its percolation through the body of the singer. Ultimately, the difference between a staged opera and an opera film lies in the difference between distance and closeness. The proscenium arch coupled with the ritualistic elements of ‘going to the opera’ encourage the distance of a performed rite, while the closeness afforded by the close up in most opera films breaks that distance down leading to the opposite of ritual distance: the intimacy of human emotion felt through the combination of song and facial expressions.

Competing Interests

The author has no competing interests to declare.