Machine-Hamlet:
To be, or not to be
The film was shot in a re-dressed Robotics Lab in the School of Systems Engineering at the University of Reading, which is home to Baxter. The robot has wheels but transporting Baxter, along with his computer, attached by intravenous leads, is a laborious and expensive endeavour. The Lab is, then, a kind of prison for the robot: its concrete borders confine him but they can also be forgotten or overcome by his seeming, intangible consciousness combined with the viewer’s imagination. Constructing the film from this premise, it identifies a series of scenes and lines from Hamlet (some of which I re-order and re-assign, including, significantly, ‘Who’s there?’ which is referred to Hamlet) and conceive Shakespeare’s play in metatheatrical terms, positioning it as a play-within-a-film, improvising ‘real’ scenes with my actors who perform actors being directed in their playing of Hamlet’s characters.
In my film, Baxter-Hamlet is an alienated, lonely figure, detached from the events he observes going on around him. Frequently filming from his point of view, the film works to prompt audiences to anthropomorphise Baxter, to bring their own mental and emotional lives to bear upon his robotic point of view as an actor on the set of a film. Other methods of performance and editing are also employed to achieve this end: the human actors treat Baxter carelessly; and he is shown to have technologically different capacities for perception: his ‘memories’ are stored digitally; he can replay them in fast-forward; and his apparent perception of people disarticulates their voices and bodies into distinct, not necessarily corresponding parts. Via such means, the film encourages viewers to imagine what it might be like to experience the world as a conscious robot.
I play the part of the film's director (both for real and in the film). During filming I instructed (or ‘programmed’) my human actors to leave gaps – silences – for Hamlet’s speeches, which were recorded and inserted during post-production. At the point of filming, I had no clear idea what Baxter-Hamlet’s voice would sound like and neither did my actors, who I required to respond imaginatively to my verbal instructions about their attitudes to, and treatment of, Baxter-Hamlet in any given scene.
Post production, the challenge was to create a voice that fitted the idea of the character we had created: an entity very close to human but also technologically formed, and evoking a particular sort of character: an intelligent consciousness that is very like a man – a youthful man – but one that is alienated, divided from humanity by virtue of his difference. In the end we settled on the voice of a third-year student, Niall Mills, whose voice is relatively soft and lilting and brings an innocent quality to Baxter-Hamlet. The vocal section that opens the film is evocative of the challenge facing me: to find the right voice for, and delivery of lines of, Baxter-Hamlet. By having voices speak to us from the darkness, this section operates to stimulate the viewer's imagination as we attempt to answer the question, ‘Who’s there?’ (Click on the audio track to the right to listen to the opening section of the film, titled: ‘Who’s there?’)
For further information about the challenge of creating Baxter-Hamlet’s voice, see Katherine’s piece below, where you can click on links to hear snippets of Niall (and Oliver Dickinson) delivering lines from Hamlet’s ‘What piece of work is a man?’ soliloquy and the editing process.
Louise LePage
Director’s reflections about the film
One of the most intriguing and productive projects undertaken as part of the Baxter Project was the making of a short film, Machine-Hamlet: To be, or not to be, which casts Baxter the robot as an actor rehearsing to perform Hamlet. One of the aims of this film was to conjure a sense of subjectivity and character for the robot. The challenge was to find performative and filmic ways to manifest this.
Katherine Olive: Reflecting upon Baxter’s robot-human voice
Lola Breaux: On the set of Machine-Hamlet: To be, or not to be
Louise LePage: Director's reflections about the film
Katherine Olive
Reflecting upon Baxter’s robot-human voice
Initially, the voice of Baxter was to be sourced from mainstream text-to-speech software commonly used to aid the visually impaired by reading aloud online text and documents. Natural Reader featured a variety of voices – masculine, feminine, and even childlike – giving us a number of options for Baxter. The software is certainly helpful; I had used it for my own project, The Utopian Principle, a play which featured an omnipresent tech voice, which interacted with the cast. For most audiences, the tone of the software may be reminiscent of assistance-technology systems like Siri or Cortana, which are focused on clarity and accessibility, and relate to the sort of friendly persona we wanted to apply to Baxter.
However, Louise decided that we would source Baxter’s voice by recording human actors reading a script, and then altering their speech in post-production through software like Adobe Audition in order to create the tone we felt Baxter required.
This is a slightly different process from that in which popular interactive-voice-response technology is sourced. Voice actress Susan Bennet worked for twenty hours a week to provide the vocabulary for Siri, not including the individual vowels, consonants, syllables and diphthongs language requires. The recordings of these human sounds could then be composed and reassembled into new words and sentences through a process called ‘concatenation’, in this case using Apple’s take on A.I. technology.
Baxter, like many modern technologies, would not have been able to source his own voice from scratch, and would require at least these core human sounds in order to ‘concatenate’ his words, which is what would produce the artificial and slightly disjointed ‘robot voice’ that we have grown accustomed to in everyday life and pop-culture.
We set out to see how we could recreate this effect with our human actors, and study how the voice could possibly be altered from these conventions and reinterpreted to fit Baxter and his various characters.
This began by having our voice-actors read their scripts in a deliberately affected and disjointed way. We had them listen to snippets of the text-to-speech technology mentioned previously in order to understand typical speech patterns and recreate them. They were encouraged to approach each word separately, with no run-on pronunciation or merged words. Only small deviations were permitted in pitch between words, resulting in the tone of voice sounding more neutral. (You can listen to recorded snippets of their voices here: Niall Mills and Oliver Dickinson . No vocal effects have been added at this point.)
We then investigated different approaches to post-production sound mixing and editing to see what qualities we wanted for our robot voices. Methods included raising the pitch (while leaving in Niall’s breath - see 'Raising the pitch', top right), decreasing the quality of the recording to create a tinny effect, before layering the same voice lip over itself several times (see 'Layering', second sound-clip from the top). This resulted in an extremely stereotypical robotic voice that was reminiscent of Hollywood B movies or an early series of Doctor Who. We decided that the voice was both too dated and too malevolent-sounding for Baxter. Louise wanted a more naturally human-like mechanical voice for Baxter-Hamlet. (However, this layering proved useful for conjuring the voice of the Ghost, Hamlet’s father, who finds form as an antiquated microwave in the film - see 'Voice of the Ghost', third from the top).
We tried deepening the voice, which had the effect of making it sound friendly (see 'Deepening the voice', fourth sound-clip to the right): there is a softness to its quality. However, these deep resonances carry with them connotations of something alien as opposed to android, which, again, were unsuitable for the voice of Baxter-Hamlet.
Other measures also had to be taken for the human voice to sound appropriately artificial. Extremely small cuts and alterations had to be made to the sound clips to erase any of the usual sounds of breath or mouth-noises that would give away the identity of the human actor behind the scenes. Also, words that ran into each other in the recording booth had to be cut and separated to create the stilted speech pattern that was required.
Eventually we developed an effect that altered the length and pitch of words and phrases to a very specific degree that could then be applied to human voices to transform them into Baxter. The effect slightly distorted the original recording, resulting in a kind of warbling sound as part of the voice, which we felt accurately recreated the effect of concatenation that unconsciously notifies us that the voice we are hearing is not quite human. You can hear a clip here - see 'Final version of the voice', which is the bottom clip on the right.
The ‘Baxter Effect’ could then be applied to any voice in order for him to perform any roles he (or in some cases, she) was given.
Lola Breaux:
On the set of Machine-Hamlet: To be, or not to be
As a member of the Baxter Project team and therefore present at different stages of the production of Machine-Hamlet, I was in a position to witness some enlightening moments that I would like to share here.
The opening of the film appeals to a concept that will be central throughout the ensuing sequences, and which is certainly pertinent when referring to the physical distribution of the set during principal photography; the “off-screen”. The off-screen is a notoriously contentious concept since its limits, both in terms of what it excludes and what it includes, are difficult to determine. A comprehensive definition would have to accommodate both the metaphysical and the more tangible elements involved but, since this is not the place to attempt to clarify this, for the purposes of the present task I will just say that the off-screen here, I believe, exists in between Noël Burch’s definition as a ‘purely imaginary space’[1] and Deleuze’s notion of the ‘out-of-field’[2], as something which is definitely concrete and has the possibility of being included in the frame at a later point.
Against a black screen we hear a mechanical voice pronounce the words “Who’s there?”, which the spectator familiar with Hamlet is probably expecting, and a warm human female voice instructing the source to modify the delivery of the line. So here, without a visual aid to guide us, we must rely solely on what the soundtrack is giving us. Roland Barthes says that the ‘grain’ is what gives the voice a tangible presence. ‘The “grain” is the body as it sings, the hand as it writes, the limb as it performs. If we perceive the “grain” […] I am determined to listen to my relation with the body of the man or woman singing or playing and that relation is erotic.’[3] The mechanical quality of the voice as well as the sound of Baxter’s robotic body which we hear on the soundtrack are the grain, and are all we have to work with initially. We are intrigued and our imagination starts to conjecture on the possible appearance of the source. The sense of anticipation is heightened by the decision of not showing the robot’s appearance straight away; instead of Baxter, the first image we see is that of the human film crew via 'his' point of view. This creates a sense of otherness. But who is the other? Are the crew the other? Is Baxter the other? Do we identify with Baxter by sharing his point of view, or is the action we see alienating him from us or us from him? The film quite cleverly invites us to reflect on this and doesn’t impose a definitive answer.
It is interesting that the voice which is so prominent and is achieving so much here, was conspicuous by its absence during the filming. In her reflections on the film Louise writes about how she instructed the human actors to leave gaps for Hamlet’s speeches, which would be later recorded and inserted at the post-production stage. Observing this in situ, these intervals of silence as performed by the team were awkward, funny, interesting and a whole array of adjectives which are too numerous to mention. My attention was concentrated on the humans instead of the robot, which is what one would expect to attract the most interest. It seemed to me that this unusual interaction was making the actors more aware of their condition as humans and creating the need to perform accordingly. It was at that precise moment that I realized how Louise was answering the question “if we think about the robot as a performer, what does this teach us about performance, theatre, and ourselves as human beings in a posthuman world?” which she had posed initially, for in the presence of a robot all I could notice was how people were behaving.
The logistics involved in observing the filming merit some space here. The sequence in which the crew set up the minimalist décor, allows the audience to see the appearance of the lab where the footage was photographed prior to its makeover. The three black curtains reduce the dimensions of the pro-filmic space, yet the full size of the room is still perceivable through our view of the ceiling and the gaps between the curtains. Those who were not to be involved in the action had to stand outside the curtains; another embodiment of off-screen space. Those of us who were outside could see the action taking place by means of a computer screen which was projecting and recording the activity taking place before Baxter’s 'eyes'/camera. This screen was Josh Oliver’s (the puppeteer operating Baxter’s head and arms) means to know when to [re]act to the speeches by other characters via movement (e.g. the nodding of the head during “The Actors Arrive” passage, or Baxter’s arm holding Yorick’s skull).
The experience created a sense of displacement for many involved in the production. For the actors, the displaced/absent voice made them more alert, perhaps even tense, and certainly self-conscious. For us observing from the outside, the information we were receiving of what was happening inside was coming from different sources. We could see the image on the screen, but the voices coming from the inside of the curtains seemed at some points to attract more attention. I noticed for example, that when the actors and crew on the other side of the curtains directly addressed someone who was outside, this person would stop looking at the computer screen, where they could see the image of whoever was speaking to them, and they turned their gaze to the curtain. From time to time Louise would step out of the curtains and check with the people observing the screen to see how things looked from the outside. The attention of those outside would shift from the screen to Louise and then back to the screen. In some cases people looked confused. I am sure film crews working in productions involving CGI, or circumstances in which interaction between elements is mediated in some way, would be able to share similar anecdotes, and would probably find all this quite commonplace. For me, however, it was an extraordinary experience which was illuminating in many ways, not so much for what I observed in Baxter, but for what I observed in the humans. It felt as if the robot was a mirror reflecting our humanity straight back at us, and to be honest, it looked weird.
[1] Noël Burch Theory of Film Practice (Translated by Helen R. Lane; introduction by Annette Michelson. London: Secker and Warburg, 1973), p.21.
[2] Giles Deleuze Cinema 1: The Movement-Image (Translated by Hugh Tomlinson and Barbara Habberjam. London: Athlone, 1986), pp.16-18.
[3] Roland Barthes Image, Music, Text (Translated by Stephen Heath. London: Fontana, 1972), p.188.