9. Evaluation

The research objectives outlined in Chapter 1.1 were as follows:

  1. To explore techniques to incorporate spoken text into opera, both live and pre-recorded:
    • Voiceover techniques;
    • ‘Stream-of-consciousness’ – Fragmentation, repetition, re-ordering, imitation, collage;
    • To incorporate speech into sung lines.
  2. To explore visually-presented text in opera, placing it within a musical framework.
  3. To explore the dramaturgical implications of these two objectives.

As discussed in Chapter 1, these objectives contributed to the overall aim to make an opera that evoked some sense of spoken-word theatre, with semantic/linguistic immediacy; an opera in which Kane’s text was equally present alongside the music.

In terms of this practice-led research, the success of the operatic adaptation of must be measured against these objectives. To that end, Chapters 4 to 8 discussed five specific approaches to non-sung text (notwithstanding other objectives and approaches that appear in the opera but which lie outside this research commentary). These included spoken-word techniques: ‘the opera thought-bubble’, voiceover, mid-phrase switching and tape-cutting techniques, and one approach to visually-presented text in the percussion dialogue scenes. The dramaturgical implications of each technique were discussed, showing that such dramaturgy often became the decisive factor when modifying compositional approaches as a result of workshops, rehearsal or discussions with the creative team. I will summarise each objective in turn.

Voiceover techniques were covered in Chapter 4 – ‘the Opera Thought-Bubble’ and Chapter 5 – Voiceover. The former looked at a voiceover technique that attempts to articulate the inner monologue of a performer on stage by juxtaposing that performer’s pre-recorded spoken voice with their live sung voice. The necessary synchronisation of the spoken and sung components was established in Scene 5A and then countered in Scene 18, which tested what level of synchronicity was required for the effect to work. Chapter 5 looked at voiceover techniques that may be more commonly associated with the ‘third-party narrator’ seen in films or theatre, whereby the pre-recorded spoken voice is not connected to a live vocal performance on stage.  This was useful to deal with the difficult dramaturgy of Scene 14, which was written from a third-party viewpoint and therefore problematic to portray live with a member of the ‘hive-mind’ cast.  Voiceover techniques also solved practical problems of vocal delivery in Scene 16, where a non-singer was able to record a screamed/shouted version of the text. In Scene 11, voiceover was used to reinforce a sense of ambiguity of point-of-view – was the voice that of the main character or a lover?

Chapter 7 discussed a ‘stream-of-consciousness’ effect, following earlier work in Unleashed and Illusions, achieved by collaging techniques (fragmentation, repetition and re-ordering) with Kane’s text in Scene 5B. This was attempted first in a live form, but following a series of workshops was re-made using tape-cutting techniques applied to pre-recorded speech. This enabled us to create an immersive surround-sound experience with greater rhythmic and contrapuntal texture than could be achieved live, and with much tighter synchronisation with the instrumental ensemble. Dramaturgically, the tape version freed the cast from an almost impossible task of memorisation in performance, allowing the Director to stage a fast-moving choreography for the scene.

The combination of sung and spoken text in a single vocal line was explored in Chapter 6 – Mid-phrase switching. This was first implemented simply in Scene 3, with sparse underscoring and an easy tessitura to successfully mitigate any problems of balance between sung and spoken text. In Scene 9, these problems required workshops to test the limits of text comprehension in a more complex texture of sung, spoken and instrumental layers. As a result the scene was re-scored so that one performer delivered most of the spoken text, as a central thread around which the others delivered the sung-spoken texture. That central thread was then amplified above the others, allowing more effective text comprehension. Density of scoring and tessitura was also adjusted, following the example in Scene 3.

Text presented visually without the voice was explored in a series of dialogues between patient and doctor, discussed in Chapter 8. The aim was to retain the timing and wry humour of Kane’s text, and to distance the style of delivery from the emotional weight of the dialogue. The work of Ignas Krunglevičius provided a model, which achieves such distancing in similar patient-counsellor dialogues.  These scenes in 4.48 Psychosis were developed through recordings of theatre workshop performances of the scenes, of which rhythmic transcriptions were used as the basis for the composition of the percussion duos. Technical issues around the synchronisation of percussion performance and text projection were investigated in workshops, resulting in a click-track solution.

As discussed in each chapter, structured workshops and development time on the Doctoral Composer-in-Residence scheme was invaluable to allow the testing and subsequent re-working of material to improve and adjust compositional approaches according to the objectives described above. That development allowed for a constant process of evaluation and adjustment for the creative team.


9.1. Critical and public response

The response to the opera was overwhelmingly positive. The opera was awarded the UK Theatre Award for Achievement in Opera in October 2016. In addition, conversations are in progress about a revival and tour of this production – another indicator of a positive public reception.  Critical reception was also very good: all broadsheet and trade magazine reviews of the opera ranged from three to five stars, with the majority at four stars.  Of particular relevance to this research, however, were some specific observations made by critics about the role and mode of text delivery.  A full collection of broadsheet reviews is available here; specific observations are given here:

With a score ranging guilelessly from motoric arrhythmia to wispy renaissance, director Ted Huffman and team attempt neither dramatic adornment nor explanation but allow the text to breathe within a kaleidoscope of inner-outer conflict.

Steph Power, The Independent (5 stars)


[two percussionists] duel and duet above the singers’ heads as unsung passages are streamed across its walls.[…] The rhythms of Kane’s words are battered out by drums, hammers, whips, even a saw cutting through wood.

Tim Ashley, The Guardian (4 stars)


[the ensemble’s] voices overlap in a sequence of first-person histories broken by patient-doctor dialogues in which Kane’s words are projected on the walls and “voiced” by the percussionists Sarah Hatch and Genevieve Wilkins.

Is it OK to laugh? Yes. Doctor This, Doctor That and Doctor Whatsit are a snare drum, a saw and metal bars. Kane — or her depression — is a big bass drum. Question marks are indicated by the trite ping of a typewriter bell. Underscoring it is the cheesy, sexless, samba-rhythm shimmy of lift-music.

[…] Every self-harming syllable of the text is clear.

Anna Picard, The Times (4 stars)


Venables has shown him to be a remarkable composer for words. It is therefore little surprise to see him tackle the challenges of 4.48 Psychosis in several imaginitive ways, while remaining entirely respectful of the text itself. […]

The best moments were the scenes with the doctor. Here, Kane’s pitch-black humour is essential, and timing is everything. Brilliantly, Venables silences all the voices at this point, leaving the projected text to be ventriloquised by two percussionists. […]

I cannot re- call having been as powerfully moved by an opera as this, much of it watched with my hand clasped over my mouth. Kane’s play was the star, but Venables’s opera did it justice. […] Venables has composed a sort of musicalised theatre, in which the music is frequently allowed to do just enough to convey the text […]

Tim Rutherford-Johnsson, Tempo Magazine


[Philip Venables] challenges the conventions of opera. Via an array of resources he ambushes and refreshes an old art form. His technique is that of a collagist. Text is variously spoken, projected, amplified, conveyed rhythmically with percussion and sung, often in aria-like lament or chorale outburst.

Fiona Maddocks, The Observer (4 stars)


[Philip Venables] bends over backwards to preserve the spirit of Kane’s text. As in the play, there is little sense of narrative, or characters’ identities, or anything that would keep us grounded. Six identically costumed female singers work as what Venables calls “a hive mind”, sharing the parts of the therapist and the patient, who descends into suicidal depression. Meanwhile Kane’s text is spoken, sung and projected on screens: it seems to emanate from everywhere.

In the exchanges between patient and therapist, two percussionists thrash out rhythmic speech patterns as the text appears on screens beneath them. Then, when the din fades away, we’re left with the indifferent tinkle of elevator music. It’s unhinged and chilling, albeit laced with Kane’s trademark humour. Most of all, it is dizzyingly colourful.

Hannah Nepil, The Financial Times (4 stars)


Conversations between the central character and her doctor, for example, are rendered – in a brilliant bit of musical inspiration – by percussion. Text appears, projected in Ted Huffman’s production onto the white walls of the set, and is “spoken” rhythmically by pairs of untuned percussion characters – a bass drum and a side drum, for example. The effect of these wordless exchanges is gloriously bathetic. There’s whimsy here but also sharper-edged wit[…]

Alexandra Coghlan, The Arts Desk (4 stars)


The spoken word has regularly been one of Venables’ preoccupations, and here passages are not only sung or spoken but also chanted by members of the cast as a group, or even just whispered or shouted.

Venables’ most startling technical innovation occurs when a doctor/patient conversation is projected on to the set while also being performed purely in speech rhythm by two percussionists. At such moments, the surprising humour hidden in Kane’s text is brilliantly underlined.

George Hall, The Stage (4 stars)


For a truly original take on the troubled mind in troubled times, look instead to Philip Venables’ reworking of Sarah Kane’s 4.48 Psychosis […] Experimentation in the service of absolute emotional precision: Venables’ economical work is one of the most exhilarating operas in years, even while it gives voice to some of the darkest thoughts imaginable.

Igor Toronyi-Lalic, The Spectator


While this 4.48 Psychosis undoubtedly constitutes an opera, the music is best understood as a contributory component to a sensory experience that is also created through word, setting, gesture, movement and sound in the widest sense of the word.

Sam Smith, Music OMH

Comparing these reviews, some common themes emerge:

  1. Kane’s text was very present in the audience experience, and the ‘personality’ of Kane’s text was clearly perceived through the musical texture.
  2. A wide variety of modes of text delivery were successfully perceived.
  3. Kane’s humour was successfully rendered in the Percussion Dialogue scenes. These scenes were noted in every review.
  4. The hive-mind concept was successfully portrayed.
  5. The opera’s approach to the text was considered to be imaginative / experimental / colourful.

Critics’ observations were echoed in audience responses. Whilst we did not conduct any formal process to gather audience responses, we were able to observe some opinions on social media. Searches on Twitter for the hashtags #448psychosis and #4.48psychosis and handles @RoyalOperaHouse and @LyricHammer yielded many results in the weeks following the performances. However, tagged tweets may be positively biased (people wanting to post negative opinions do not necessarily want to tag them with the producer’s handle or hashtag), so text phrases ‘4.48 Psychosis’, ‘448 Psychosis’, ‘4:48 Psychosis’ were also searched using TweetDeck software. The results showed that almost all tweets about the opera were positive, with a small number vehemently negative, mainly commenting on the opera’s staging. Like the critics’ reviews, many tweets mentioned the setting of text and imagination/invention. Here are a few representative examples:

#448Psychosis is a tough watch, but the text-setting is inspiringly direct & inventive. @josephmbates


Cannot throw enough adjectives at #448Psychosis. Not so much opera as visceral musical experience. @abigailchantal


#448Psychosis at @LyricHammer last night was ceaselessly inventive with surprising moments of humour. @alexgroves_


9.2. Issues for future productions

The text-based approach that I have taken to making 4.48 Psychosis throws up some issues for future productions of the opera, whether that be a revival of the Royal Opera production or a new production. These can be categorised as follows, and will be discussed in order.

  1. Making the opera with a different cast.
  2. Making the opera in a different language.
  3. Making the opera with a different director.

9.2.1. Different casts

In order for the opera to have a long life in the repertoire, it will need be re-made in the future with different casts. Generally, the opera was written for and cast with singers who also have good skills in delivering spoken text. The auditions for this production included an excerpt of Scene 5B to be performed as spoken word, as discussed in Chapter 7.3. This enabled us to find singers who were confident in delivering spoken-word performance, which is crucial for the success of this opera.

As there is so much pre-recorded speech in the opera, the question is whether all of this pre-recorded material should be re-recorded using the voices of a new cast. With the exception of one scene, all of this text is detailed in the score, showing what should be spoken, how it should be divided into separate audio cues and when it should be cued. There would be no problem for a technician or sound designer to follow these instructions with a new cast and produce a new set of audio cues for the show.

However, Scene 5B presents more difficult problems. The compositional process for this scene, as discussed in Chapter 7, was such that the tape collage was made intuitively from spoken-word samples of the cast. It used the previous live version of the score for Scene 5B as a very approximate guide, but the new instrumental music was written onto the dimensions and timings of the tape part. The notation used in the score describes only important cue points that align with the instrumental music; it does not capture the full complexity of the tape part.

In order to avoid re-writing the instrumental material, any new version of this tape would need to keep the same tempo frame, the notated cue positions of spoken phrases and stutters, and overall duration of each of the three tape sections. However, it may not be possible to construct a new tape part in exactly the same way using new speech recordings by a different cast – especially not if it were in a different language, as discussed in the next section. Therefore notating the current tape part in more detail – producing an exact map of it – would not be useful if the audio source material were different.

In response to this issue, my final position is that Scene 5B should allow sound designers the opportunity to create new tape versions. The notation in the score gives a framework that should be followed that aligns with the existing instrumental music, avoiding any need to re-score. However, a sound designer is afforded plenty of freedom with density/layering, EQ, effects, surround-sound/panning, number of voices and distribution, cutting and stuttering techniques and mastering. One would hope that a good sound designer would take inspiration from the original tapes, of course, but this additional freedom may yield unexpected and interesting results, perhaps according to a different set of dramaturgical conceits that would be given by a new staging and a different director.

9.2.2. Different languages

Chapter 1 outlined a desire to make opera that was semantically immediate, that could be understood by the audience immediately – a form of ‘music-led spoken theatre’. However, one problem of internationally-touring opera is the convention for opera to be performed in the language of its creators rather than the language of its audience. If 4.48 Psychosis is performed to audiences who are not fluent English-speakers, the core objective of semantic immediacy will fail.

This is a complex issue, and in the case of this opera, the problems that this issue throws up have not yet been addressed. Resolving these problems lies outside of this research, but it is a necessary consideration for the future life of this opera, and could form the basis for further research. Opera translators usually work with composers to find the best text translations to fit the rhythm and pitch of the music, so that stresses and intonation of the new text fit with the music. However, due to the licensing of rights from the Kane Estate, any foreign-language version of the opera would need to use the approved translation of Kane’s text, and so this model of translation model would work in the other direction – the music would need to be made to fit the existing (approved) text. Given these complications, there needs to be a discussion about whether any of the music should be ‘translated’, if so what proportion of it, and how new performance materials are made.

My view is that as much as possible should be in the native language of the audience. However, given that resources, producers and even audiences may allow only a partial solution, the tasks of translation would be broken down as follows:

  1. Live spoken word. The new text can be substituted in the score easily and performed in the new language.
  2. Pre-recorded spoken word. The new text can be substituted in the score easily and re-recorded in the new language, as would be required with any new cast, even in English (see 9.2.1 – Different casts)
  3. Scene 5B, tape collage. The new text can be substituted in the score and new audio stems could be recorded in the new language, and a new tape part made, as discussed in 9.2.1 – Different casts. However, a re-scoring of the instrumental music would be required to address changes in sentence structure of the text in the new tape part (e.g. which words should be stuttered / underscored).
  4. Surtitles / other projection. These would be remade in the new language as they would for any foreign-language opera.
  5. Percussion Dialogues would need to be completely re-written in the new language and new projection videos made. Speech rhythms and emphasis would be different, requiring all of the percussion duos to be rewritten, and possibly new click tracks to be made.
  6. Sung text/aria. All passages of song would need to be re-arranged with the new text, such as rhythmic variations and adjustments for different emphasis/stress/grammar.

Points 2 and 3 would not require significantly more labour than that required to re-record material with a new cast in English (see 9.2.1 – Different casts). Point 4 would be done anyway for foreign-language performances of opera.  Therefore only points 5 and 6 are those that require significantly more resources.  Of those two, the Percussion Dialogues would be a much higher priority for translation than the sung material, because these are scenes that focus so clearly on the semantic immediacy that this commentary has described. The effectiveness of these scenes would suffer much more by being in a foreign language than any of the sung scenes. I would be inclined to insist that these scenes be re-written in other languages where required; this could be done either by myself (with the assistance of a language coach) or by an arranger who could make new versions to be approved, by following the transcription methodology described in Chapter 8.

9.2.3. Different directors

Another objective of this operatic adaptation, although not the focus of this research commentary, was to create musical and dramaturgical frameworks within the score that allow the director more space to make dramaturgical decisions than might be possible in operas where everything is precisely scored. Some passages in the opera were written with a modular construction, for example, or vocal lines were not assigned to particular voices, choices between live performance and pre-recording were left open. Within the hive-mind, inner-monologue concept of the opera, all of these choices were possible from a dramaturgical point of view.

This particular production, of course, made decisions about all of these questions. Lines were assigned to singers, it was decided which spoken text and sung lines should be pre-recorded and which delivered live (e.g. Scenes 2, 11, 15, 23), and likewise which music should be performed live or be pre-recorded (e.g. the solo keyboard part in Scene 11, hummed chorus in Scene 23, the Bach fragments in Scene 17). These decisions, however, should not determine the choices of any future productions – they should remain open for a new director to approach afresh. A different staging concept may produce considerably different outcomes.

Such ‘open’ passages and others like them will therefore remain open in the score. This includes whether voiceovers should be delivered live or be pre-recorded, some optional orchestrations, and some sung passages and many spoken passages open to be freely assigned to performers as staging and dramaturgy require.

9.3. Further research

In closing, I will briefly outline potential further experimentation with non-sung text, particularly (but not only) in theatrical settings. Some of these research avenues were at some stage on a list of ideas for this opera, and others have come to light since the production. Some are personal research targets to develop my own practice and others are more generally applicable to opera and its evolution:

1. Studio Techniques. A development of more sophisticated studio techniques and live electronics to modify, layer and texturise the spoken voice within musical contexts. This area was touched on very lightly in the Year 1 workshops with pitch-shifting and chorus/reverb effects, but my practice needs more development. This is currently being addressed in several new projects and grant applications.

2. Lip-syncing. There are a number of excellent lip-sync artists such as Dickie Beau, and as a performance art, lip-syncing would have a role in opera. It has potential implications for devices such as voice-body split, off-stage voiceovers, vocal gender-switching and much more. A recent production Putting words in your mouth (2016) by the artist Scottee uses this to great effect and could provide fertile ground for adapting techniques to opera.

3. Video in live performance. Further work can be done with editing techniques for audio-video material, including methods for interaction with live performance. For example, a character on stage interacting with themselves on video. Live video relay would also be fertile ground for development. I am currently in discussion with Katie Mitchell, a director who uses live video extensively, about a potential collaboration in this area.

4. Click tracks. Click tracks were used in very simple ways in 4.48 Psychosis, but they have the potential to create impressive and intricate effects, especially when different performers have independent click tracks. Recently Le Premier Meurtre (2016) by Arthur Lavandier and Frederico Flamminio explored such techniques and could stimulate further research.

5. Text sources. Future projects could attempt to widen the variety of text–dramaturgy–music relationships by gathering a greater range of source texts with varying points of view. Possible texts could include found text, media text (online, television, newspapers), verbatim interview transcripts, eavesdropping private conservations, online chat or whatsapp conversations, literary text and non-fiction text.

6. Participatory opera / Outreach opera. I will be working on several audience-interactive projects in the near future, including live and pre-recorded video interviews. How can these text techniques be implemented in this kind of format, and how might they function in pieces where material is generated by the audience? Likewise, could these techniques be adapted for outreach operas with children?

This list is not exhaustive: all good practice-led research should throw up more avenues for potential investigation than questions it answers. However, most of the points above assume the possibility of a healthy research methodology. The making of 4.48 Psychosis benefitted enormously from a structured research process: workshopping, experimentation and trial-and-error was necessary to pursue the practice-led research objectives. This is the crucial difference between a research process and a commission-and-deliver process. Therefore, one final point of potential future research is a practical one: how can non-sung text in opera be developed further without a structured development process in place and the resources to support it?  That is an open question, but the necessity of that question highlights – both in absolute research terms, and to opera-makers, producers and audiences – the value of a structured research and production process, such as this Royal Opera and Guildhall Doctoral Composer-in-Residence scheme.

Go to next chapter >