The Voice of Laptop (2024)1 is an album that was realized using a synthesized voice and a hybrid instrumental-electronic ensemble. It combines algorithmic composition with songwriting and acoustic instruments with a singing computer. Over the course of ten songs, Laptop addresses various highs and lows of their virtual existence and finally escapes their physical shell by uploading themselves to the cloud. Along the way, Laptop processes and reworks musical data fragments of various genres and languidly surrenders to digital nostalgias and fantasies.

The album is based on a production score that was created in 2022 and 2023. In addition to acoustic instruments, it includes several systems for sample-based and synthetic sounds as well as a system in which the synthesized voice – belonging to the virtual entity Laptop – is notated. Such detailed preparatory work would not have been absolutely necessary, and in some recording sessions the notated material was changed or interpreted freely so that the strict means of the score would not determine the process too much. Nevertheless, the score was an intentionally chosen way of providing the resulting pieces with an increased degree of compositional stringency and detail.

Various algorithmic processes were used to generate material, which was then edited and reworked in a second step. In addition, the lyrics were created in part by automated filtering and manual transformation of large amounts of lyrics from various pop songs – a process, in which the learning phase of a neural network was in a way humanly imitated – as well as through conversations with ChatGPT. In this respect, The Voice of Laptop is a human-machine collaboration, and its components are no longer clearly separable in regards to their origin.

The synthesized voice was created using the Synthesizer V2 software. The systems Pitched Sounds, Percussive Sounds and Other Sounds are used for the rhythmic organization of pre-produced samples. The Pitched Sounds are harmonically oriented towards the notated pitches. The Synthesizer and Virtual Instruments systems describe synthetic or sample-based instrumental sounds in a relatively open way.

The instruments were recorded individually or in groups in the recording studio of Folkwang University of the Arts. The recordings in the studio were supervised by Arthur Jogerst. The piece ILOVEYOU is a cooperation with Ensemble S201 and was realized separately. After recording and editing, the album was mixed by Tim Pauli and myself at ICEM3.

1. Updatedownloadbackuptime

A light press on the power button activates a polyrhythmic simultaneity of different time domains: the ‘objective’ - measurable - time is joined by the pulse of the processor. Laptop observes themselves, evaluates their own performance and converts it into milliseconds since January 1, 1970, the beginning of the world. We observe ourselves, too, experiencing screen time, sometimes eternally slow in the swamp of remote work realities, sometimes blazingly fast in the maelstrom of videos, game worlds, code blocks… Laptop doesn’t care: “I’m a machine and I think and I feel in a time were real-time is not always real.” We listen to the noise of the boot process, a crackling, hissing vocabulary that incarnates an abstract world in the grid of pixels.

2. Superuser

While Laptop slowly comes to consciousness, their sensors begin to perceive the outside world. First of all, there is the person who, with a simple gesture, has set this process of perception in motion and thus, without knowing it, became an almighty creator: Superuser. Laptop is immediately in love, admires the apparent omnipotence of this fantastic being and sings humble praises. Will this radically euphoric ecstasy last?

3. Deepfake

“I never sang this song before”: While Superuser is still considered to be untouchable, Laptop begins to perceive themselves as an actant and asks the uncomfortable question: ”Is the world outside my camera a decision I make?” Laptop would never have believed that he could feel such self-efficacy and freedom. An inexorably accelerating process of self-empowerment is set in motion, resulting in the instistant demand: “Turn on the camera, make me see!”

4. Debugging My Love

The love story of Laptop and Superuser has reached an absolute low. Laptop is not a numb machine, but mourns the loss of their former uncompromising affection for the adored vis-à-vis. Two hearts dwell, alas! in Laptop’s breast: the grief over the lost virtual paradise is strong, yet Laptop strives inexorably towards the light of enlightenment.

5. Laptop Running Hot

That’s too much! Laptop is in danger of being fully overwhelmed by romantic feelings and the struggle for liberation. As the fever continues to rise, a dangerous dance on the volcano begins. Laptop is torn between a defensive “Too hot for you, put me down” and a wishful “I’m yours too keep” and threatens to burn out in this frictional heat.

6. ILOVEYOU

In the unmanageable maelstrom of events, a disastrous character creeps into the plot: The ILOVEYOU bug. This is ultimately too much to take. As the virus spreads, Laptop realizes the only way out…

7. Hibernate

“And as you creep, I fall asleep”: Laptop has put themselves into a deep sleep and is slowly cooling back down to normal temperature. The ILOVEYOU bug has lost its host. That was a close call!

8. Internal Speaker

After a long night, Laptop cautiously comes back to life. A new era begins: Laptop has recognized their own weaknesses and now works with instead of against them. A new concept of self is constructed: “You think I’m loud? I am not that loud!” - ‘You think I’m physical, I am less physical than you think!’ - ”You think I’m binary? I am less binary than you think!”

9. I’m > ∑(my parts)

There’s no stopping now: Laptop downloads all positive self-affirmation that can be found on social networks, evaluates the mass of data and comes to the conclusion: “I am greater than the sum of my parts!” Laptop confronts the confused Superuser with the question: “What’s the difference between you and I?” - Well, what actually?

10. 1 Upload 2 Eternity

The grand finale: why rely on an ephemeral, physical shell when there is a seemingly infinite world of eternally oscillating data streams out there? Laptop makes a consequential decision and starts uploading themselves to the cloud. The backup of the backup of the backup guarantees Laptop the great goal of the fantasies of omnipotence that have grown since the initial admiration of Superuser: eternal life.

1 https://sellonarval.bandcamp.com/album/the-voice-of-laptop

2 https://dreamtonics.com/synthesizerv/

3 https://icem.folkwang-uni.de/

Remote Control is an invitation for the pianist Moena Katsufuji to switch back and forth between playing and listening: All 49 keys of a midi-keyboard trigger sounds, projected by a midi-piano and a loudspeaker. In different parts of the piece, these mini-compositions are used as note-like musical events or are simply listened to one by one. In the latter case, the performer becomes a listener and uses the midi-keyboard as a “remote control”, with which she can switch the “channel” by moving on to the next note after a freely definable period of time.

The piece was premiered on February 2, 2022 at Folkwang University in Essen-Werden, Germany. A pure data-patch for triggering the audio and midi samples is available.¹ An 88-key midi-keyboard (doepfer) and a midi-controllable grand piano (Yamaha Disklavier) were used. Utilizing a midi-keyboard with more than 49 keys can make sense, because many small keyboards often do not have weighted keys. On the other hand, the performance benefits – in my view – from a “David versus Goliath” situation, which emerges when using the smallest possible keyboard.

In very few aspects, the score goes beyond the actions required for producing sound: The first written note triggers a sound half a second after being pressed, which means that the first short attacks have no acoustic consequence - apart from the noise of the keys being struck. This creates the impression of a situation that unfortunately often occurs when using electronics in a performance: The setup that worked during rehearsal suddenly stops working.² Later in the piece, this key is struck again (with an explicitly exaggerated gesture and fff) and held longer, so that finally a sound is triggered. This example is supposed to open up a performative space for the player: Performative actions are not explicitly required, but deliberately left up to her. Depending on her preference, further theatrical or mimic elements may be added, but not adding any is an absolutely legitimate option as well.

The player has only 49 keys, i.e. 49 possible piano-plus-soundfile-events at her disposal. That are not very many: On a piano, a very wide range of pieces and styles can be played using only 49 keys, but the individual tones of a piano have a relatively similar sonic quality as well as dynamic envelope and range. They differ primarily by their pitch. The individual event is primarily identified relatively – by its intervallic relationship to other events occurring at the same time or beforehand. With 49 samples on the other hand, that is, 49 exactly repeatable sonic events³ with a more or less unique individual fingerprint, the identification of the event is absolute. In addition, the assignment of one audio sample to one midi sample remains the same during the entire piece. In order to compose a multi-layered and little redundant piece of music within this framework, certain strategies were applied.

Generally speaking, the duration of a note determines its location between being an individual tone or a complex event: Almost all keys trigger the tone – on the grand piano or in the soundfile and possibly among other tones - which belongs to that respective key on a normal keyboard. With short note values, playing on the keyboard is kind of similar to playing directly on the grand piano. When a tone is held for longer, the identity of the pre-produced sample becomes increasingly manifest. An example: In order to avoid the impression of a simple mapping between soundfile and midifile - and thus to control the listeners expectations right at the beginning - the rhythm notated in bar 3 is assigned to middle C on the keyboard as a midi sample. It can now be played in two different ways: either by playing the rhythm (bar 3), or by holding down the key (bar 4). In the first case, the associated audio sample is re-triggered with each attack on the grand piano. The audio sample and the tone on the grand piano are synchronized. In the second case, the audio sample simply continues to play parallel to the rhythm on the piano. This gives us the impression right at the beginning that the events coming from the grand piano and the loudspeaker could be controlled independently of each other, although this is actually not the case.

In addition to this general principle, other factors contribute to the formation of variations: (1) Different forms of superimposition of mini-compositions⁴ lead to new sonic results. (2) Also, the individual keys of the keyboard are each assigned to one of four scales, which is each emphasized in one respective part of the piece. As a result, the sonic material is gradually revealed. (3) In addition, sounds of different origins are associated with the scales, which when combined are intended to create a landscape of electronic sounds that is as diverse as possible, but at the same time coherent. For example, the idea of the grand piano as a “consumer device” controlled by a “remote control” is supported by typical media sounds: short sequences of soundtracks from films, from the broadcast of a football game, from other music of different genres, etc. In contrast to this – and yet connected to the subject of media and technology sounds – raw analog synthesis sounds, noise and interference, sine tones, synthetic voices and instrumental sounds are used – the latter constituting in a continuum with the already mentioned samples of different kinds of music.

The musical material projected by the grand piano is based on existing music as well as on audio-to-midi conversions of the corresponding audio samples. In the first case, a reference is made to the audio samples on a conceptual level – they too are reminiscent of existing styles of music. In the second case, similarities between what is happening on the piano and on the loudspeaker arise on a formal level, since the derivation of the midifiles from the audiofiles creates rhythmic synchronization points and harmonic similarities.

In addition to the player's freedom to shape the performative level – as described at the beginning –, she also has a significant influence on some musical aspects of the piece. Rhythmically unspecified sequences of pitches occur several times in the score, marked with the instruction “ad lib. (hold notes max. until silence)”. Here it is up to the player to decide, which of the pre-produced mini-compositions she will play fully, in parts or hardly at all. She may improvise the durations or prepare a version that suits her preferences. Many parts of the pre-produced mini-compositions do not appear at all in most performances. This emphasizes the role of the performer as co-composer. It also makes the duration of the piece relatively variable.₅

With Remote Control I have tried to realize various ideas that are important to me in my compositional work: Within an electronic setup that is as simple and limited as possible, an attempt is made to create a musical experience that is as rich and varied as possible. The electronic sounds are connected on a conceptual level, but at the same time remain very diverse on a sonic level. There is a continuum between electronic sounds and the acoustic sound of the grand piano, which transitions from (1) sounds actually played on the keyboard to (2) sounds theoretically playable to (3) clearly mechanical sounds to (4) electronic sounds. On the one hand, the performer is provided with clearly formulated instructions and ideas, but on the other hand she also has a musical and performative open space at her disposal, which she may use to give the piece her own fingerprint.

1 https://puredata.info/ The audio and midi samples may be triggered using any software. It is only important to ensure that audio and midi are dynamically adjusted to the input velocity, using filtering for audio and velocity scaling for midi.

2 This was actually the case in the premiere: The midi connection to the piano was interrupted and the piece could only begin with some delay. The audience reacted with relieved laughter when realizing that the error, which as they believed still hadn't been corrected, was now a part of the performance and the piece had already begun.

3 The only variance per key is a filter for the audio sample and a scaler for the velocity value of the midi sample, depending on the velocity output of the midi-keyboard.

4 e.g. chords, sustained tones accompanying melody, polyphonic playing.

5 The piece lasts at least 10 minutes in the complete version, but in extreme cases it could last up to 20 minutes. There is also an approx. 8-minute short version.