Psst.....Wanna Buy an Algorithm?

The Proliferation of Generative Music Systems throughout

Everyday Life and its Effects on the Music Industry


Paul Brown M.A.

Research Student, Anglia Ruskin University, Cambridge, United Kingdom






Generative music is created within a system as a result of the rules implemented by its creator. In computer based systems these rules, and therefore the music produced, may be further influenced by external sources converted into digital data. This combination enables generative music systems to create music for a specific environment or purpose. The adaptability of music produced by generative music systems could substitute many instances where linear music is currently used in everyday situations.  After examining the issues surrounding the music produced by generative music systems, this paper reviews instances where generative music systems could be used to replace linear music systems in everyday settings and enhance the effects of music in those settings. Such settings include healthcare, personal fitness, other commercial setting such as restaurants, retail outlets, offices and factories, in the motor industry, in domestic settings and within the telecommunication industry. Since the “pure”[1], “unfixed” music produced by generative music systems is, in the UK at least, copyright free and therefore exempt from the usual public performance licensing procedures, after looking at linear forms of “environmental” music, the paper then goes on to examine the importance and nature of the algorithm within generative music systems and considers how the generative music composer could be remunerated each time music generated by his or her algorithm is performed in public. The title of this paper makes inference to the likely attempts that the established music industry will make to firstly ignore the idea of generative music systems and then to suppress their proliferation by attempting to outlaw such systems despite their legality. The paper therefore concludes by discussing some of the implications of generative music systems on the music industry.


This paper follows on from, and includes elements of, my MA thesis “Is the Future of Music Generative?” and acts as a precursor to my research degree at ARU in Cambridge where I propose to examine the feasibility of using generative music systems as tools to assist music therapists and patient carers in health settings.




1. Introduction


“Generative music is commonly agreed to describe music in which a system or process is composed to generate music rather than the composition of the direct musical event which will result from that system.  The generative composer has only indirect control the final musical result, and the creativity of the compositional process is found in the decisions about how the system will operate and the rules inside the system”(2). 


As generative music is created by the composition of music within a system this means that musicians, recording artists, producers, and to some extent the composer, can be absent from the creation process. The generative music composer, besides defining the musical parameters within the piece, essentially separates him or herself from the creation of the final piece of music. Advances in computer technology have enabled generative music systems to be created where the generative music composer has more control over the rules he or she creates within the system or process and therefore the final music produced. The rules within the generative music system may be further influenced by external environmental changes converted into digital data. This combination allows generative music composers to create systems that can produce music for a specific environment or purpose.  The adaptability of music produced by generative music systems could substitute many instances where linear music is currently used in everyday situations. 


It is my belief, that the intention of creating a generative music system is to produce a unique piece of music each time the system is initiated or reset and therefore means generative music in its purest form is not recorded. As it is not recorded, although copyright subsists in the generative music system itself, no copyright subsists in the music that it produces. Since no music copyright exists within pure generative music the role of traditional music companies in relation to generative music such as collection societies, record labels and publishers, which rely on copyright subsisting within music to allow them to carry out their day-to-day business, is brought into question. The absence of a business model within the traditional music industry to support the proliferation of generative music systems means that a new model must be sought and developed from elsewhere.


2. I Am Not a Lawyer But…


The copyright status of computer related generative music systems and the pure generative works that they create seems, in the UK at least, to be fairly straight forward. For the purposes of UK Copyright Law a computer program is considered as a “literary work” which, according to Section 3 (1) of the Copyright Designs and Patents Act of 1988 (3), “means any work, other than a dramatic or musical work, which is written, spoken or sung”. Section 3 (2) of the Act states that “Copyright does not subsist in a literary, dramatic or musical work unless and until it is recorded, in writing or otherwise”. In the case of a musical work this would include making a video, tape, CD or digital recording or copy of the work as well as a musical score and in the case of a computer program a digital copy of that program.  A computer based generative music system, then, is protected under UK Copyright Law. Section 12 (2) of The Duration of Copyright and Rights in Performances Regulations 1995 (4) states that “Copyright expires at the end of the period of 70 years from the end of the calendar year in which the author dies”. If the generative work that it produced were “recorded, in writing or otherwise” then as a “computer generated” work copyright would, according to Section 12 (3), expire “at the end of the period of 50 years from the end of the calendar year in which the work was made”. The pure generative music produced by the generative music system, however, is not fixed and therefore copyright does not subsist in it.


This does not mean that the creator of pure generative music is not deemed its author. Section 9 (3) tells us that “in the case of a literary, dramatic, musical or artistic work which is computer-generated, the author shall be taken to be the person by whom the arrangements necessary for the creation of the work are undertaken”. The author, however, is unable to exert his or her Moral Rights as set out in Section 77 of the Act because, according to Section 79 (2), “The right does not apply in relation to the following descriptions of work—  (a) a computer program; (b) the design of a typeface; (c) any computer-generated work””.


In the US, the copyright position in relation to pure generative music is not so apparent. According to Section 102 a) of the US Copyright Act (5) "Copyright protection subsists, in accordance with this title, in original works of authorship fixed in any tangible medium of expression, now known or later developed, from which they can be perceived, reproduced, or otherwise communicated, either directly or with the aid of a machine or device." In the preceding definitions in Section 101 it states that "A work is “fixed” in a tangible medium of expression when its embodiment in a copy or phonorecord, by or under the authority of the author, is sufficiently permanent or stable to permit it to be perceived, reproduced, or otherwise communicated for a period of more than transitory duration."


In the case generative music systems, the nature of the music that the systems produce is often dependent on random choices made within the system. The question that arises, therefore, is are the random choices within the system “sufficiently permanent or stable to permit it to be perceived, reproduced, or otherwise communicated for a period of more than transitory duration” to be considered ““fixed” in a tangible medium”. My own feeling is no. The random choices are never fixed within the system as the intention is for them, and therefore the nature of the music created, to change each time that the system is reset.[2] But I am not a lawyer.

3. Applications


A number of generative music composition tools exist for general public consumption[3]. These include MuSoft's a Musical Generator (6), Digital Expressions' ArtSong (7), Nyr Sound's Chaosynth (8), FractMus (9), Algorithmic Arts' SoftStep (10), Sseyo’s Koan Pro (11)[4] and Madwave’s Madplayer (13).


With the exception of the telecommunications industry where Sseyo’s generative music engine enables generative ringtones to be created in real time within Tao’s intent Sound System (iSS) and Madwave’s Madmixer, which incorporates its MadTone Generator for creating generative ringtones, no other commercial applications of generative music systems, to date, have been found other than those that are used as music composition tools.


Generative music systems have two distinct advantages over conventional linear music systems. The immediate advantage to companies using a generative music system is that, in the UK at least, neither a PRS (Performing Right Society) (14) nor PPL (Phonographic Performance Limited) (15) license fee is payable as the pure generative music played is not fixed and therefore copyright does not subsist in it. The second advantage is that generative music systems are able to adapt and vary the music that they produce in real-time and therefore give the end-user greater control over the characteristics over the music produced.


Another possible application for generative music systems within the telecommunications industry could be “on hold” music. On-Hold music is an important tool in keeping customers on the line while waiting for an operator or customer services assistant.  Studies have found that callers are likely to think that they have been forgotten if there is no music on the line and that music reduces the perceived waiting time (16) with callers staying on the line up to 20% longer if they heard music that they liked (17). Studies have also shown that callers tended to hold on longer if jazz, country or classical music was playing rather than pop or relaxation music (18) and that the choice of music played influenced the callers’ image of the company (16).


PRS recommends that companies that use on-hold music change the music regularly, match music to customer profile and choose music to represent the desired company image. They also recommend that companies be unique in their choice of music and to focus on the value to the company by using the music and not the cost (19). Whatever genre of music is played there should also be an air of familiarity about it as “familiar music produces more discrete events in the memory than unfamiliar music thus increasing a person’s estimate of time” (16). In other words music that is familiar to the listener makes time seem to pass quicker than when unfamiliar is played.


The detection of a caller’s telephone number could instruct the generative music system to create appropriate music, based caller’s preferences, while the caller is waiting for assistance ensuring that the music played is varied but of a style that is preferred by the caller so that the caller is willing to wait longer but the perceived waiting time is shorter.


In commercial environments such as restaurants and bars the style of music forms an image of the establishment in the mind of the consumer. While classical music creates an up market image for an establishment it is thought that jazz music attracts a more affluent customer (16). Customers in a cafeteria were found to be more willing to pay a higher price for food when classical music was playing in the background than other types of music (20). In the same study it was found that both classical and pop music may have increased the sales in the cafeteria as compared with easy listening and silence. Classical music was also found to appease rowdy or aggressive behaviour in customers whereas heavy metal was found to have the reverse effect (16).


The tempo of the music played is also another important consideration in commercial settings. In the 1940s, an American company, Muzak®, developed Stimulus Progression programming, an "elaborate system that arranges songs according to tempo and time of day, taking into account the typical lulls that hit workers mid-morning and mid-afternoon" (21). In one study fast music was found to significantly increase the time spent drinking a can of soft drink (22) and managers of restaurants tend to play faster music at lunchtime and in the evening to clear tables more quickly as diners tend to eat quicker when faster music is played (16). Another study found that music tempo variation can significantly affect the pace of in-store traffic flow and dollar sales volume (23). Fast music was found to increase both productivity and morale of employees occupied with monotonous work. In a study at a voucher processing centre for a bank, fast music lead to 22.3% more vouchers being processed than when slow music was played (24).


Another important factor is the volume of the music played in the restaurant or bar. Music played at low volume tends to increase expenditure drinks (25) and tends to be most effective when it is quiet enough to be discrete but loud enough to suppress background noise (16). Music also reduces perceived waiting time, can facilitate conversation and alleviate uncomfortable silences (16).


In commercials settings such as restaurants and bars changes in local environmental readings could instruct a generative music system to create appropriate music that was faster at lunchtimes and in the evening in restaurants to encourage customers to eat more quickly and drink more and automatically adjust the volume of music so that it remained discrete no matter how many people were in the establishment. Additionally, if sensors detected that the restaurant was not full it could over rule the instruction to play faster music by playing slower music so that customers felt more inclined to linger longer and therefore give the appearance that the restaurant was always busy and therefore popular. Other possible applications include automatically playing more stimulating music in offices and factories at morning or mid afternoon lulls or if room temperatures rose to a level that were in danger of making the workers lethargic or in retail environments where appropriate styles, tempos and volumes of music are played depending on the time of day, the temperature within the shop or the number of customers within the shop. Booming music, for example, may not be appropriate in a shop on a hot Monday morning when customers are merely there to browse rather than buy.


Another possible application for generative music systems could be within the motor industry. One study found that for many people the car is the only place that they can listen to loud music without annoying other people (26). Another study has shown that the type of music played within the car can have a marked effect on driving behaviour (27). A driver playing loud music with a high tempo can either knowingly or unwittingly increase the speed of the vehicle thereby breaking the speed limit and potentially endangering the lives of over road users. Similarly, a driver playing the same music in a traffic jam could become irritated and stressed by being held up thus increasing the risk of “road rage” related incidents. An appropriate generative music system could automatically play music that both discouraged speeding by encouraging the driver to maintain a safe, sensible speed and also relieve the stress associated with being caught in traffic jams.


A generative music system, influenced by physical data gathered from within a vehicle and its surroundings, could create appropriate music with a tempo and volume that encourages driving at a safe and sensible speeds and when detecting the close proximity of other vehicles when travelling at low speeds play appropriate relaxing music to subdue driver stress levels.


In the home many people use music to accompany domestic or solitary tasks including housework, studying or resting (26). When music was used as an accompaniment to routine activities such as washing or cooking the effect of the music was found to increase positivity, present mindedness and arousal in participants of one study (28). The same study concluded that one of the primary functions of music in contemporary everyday life appeared to include the enhancement or distraction of attention from a mundane domestic task and that music was used to enhance states of relaxation. A generative music system could play music that was both appropriate for the activity that the individual was undertaking within the domestic environment and also play appropriate music based on other data such as the time of day, temperature within rooms or even the presence of a certain individual and based on pre-determined information regarding the individual’s musical preferences. The application of such a system has been previously been presented and discussed by John Eacott and Mark d’Inverno in their paper “Embedded Intelligent Music or iHiFi – The Intelligent HiFi” (29).


Within the area of sport and personal fitness studies have shown that synchronisation of exercise to music has increased work output, reduced the rate of perceived exertion during exercise, increased enjoyment levels and enhanced affective states at both medium and high levels of work intensity (30).


Using physiological readings from the athlete a generative music system could create music within a work-out session with characteristics, such as tempo and rhythm, which were attuned to the ability of the athlete in order to sustain endurance and participation levels throughout the work-out session without the danger of overexertion.


Extensive research has been carried out within the health sector on the effects of music on people. These include the effects of music on physiological parameters such as heart rate, respiration rate, skin temperature, blood pressure, skin conductance, salivary cortisol levels and levels of Immunoglobulin A. The scope of this paper, however, prevents a detailed presentation of those studies[5].


Overall, however, studies have found that people’s music tastes are highly individual with different types of music eliciting different moods. Music also counteracts stress and anxiety and promotes relaxation in patients. The type of music played can also influence physiological elements such as body temperature, heart rate, blood pressure, respiration rate and production of salivary Iga in patients (cited in 31).


Decreased stress and anxiety in patients could have positive outcomes for both patients and their carers. The feelings of stress and anxiety have a number of contributing factors including fear of pain or of the treatment outcome and also the patients’ feelings of loss of control over their body and immediate environment (32). Physical reactions to stress and anxiety include restlessness, trembling, muscle tension, fatigue, shortness of breath, increased heart rate and elevated blood pressure that could lead to medical complications (33).  


Decreased levels in stress and anxiety could also result in an improvement in the feeling of well being in the patient, a decreased length of hospital stay, decreased medications such as pain relievers and sedatives and an overall increased perception of the patients’ stay in hospital (34).


A generative music system that is adaptable, autonomous, non-invasive and holistic could help the music therapist to create and sustain a dialogue between with patients and also help the health carer to sustain patients’ comfort without the need for constant carer-patient monitoring or the over use of pain relieving drugs. My intention is to commence developing such a generative music system during the course of my research degree.

4. Is it “Musak®”?


Environmental music is not a 20th century phenomenon. It origins, in fact, lie within the church of the 17th century where improvised organ music was played to “maintain mood and concentration during inaudible portions of the service taking place at the altar and during quiet segments such as communion” (35). Much of the music played at garden and birthday parties and other celebrations in the late 18th century was specifically composed as background music “intended to blend in with the general ambience of the event rather than be listened to directly” (35). 19th century upmarket restaurants in Europe, and then the US, also used music in the background to enhance the client’s dining experience. During WW II in England music was also played in defence plants with the specific intention of reducing fatigue in its workers (35).


The provider of environmental music that has dominated the latter half of the 20th century is Musak®. Musak®’s intention is to produce “music that is functional and to be heard but is not supposed to listened to directly” (35). Indeed, it considers the music making process to be more science that art creating music that functions as an integrated physical and psychosocial part of the overall music environment and is purposely non-entertaining, non-distracting and unobtrusive in nature (35).


Each fifteen minute segment of Musak®’s programming contains five or six recordings that have been analysed to give a “stimulus value”, each value being “based on a set of numerical values derived from such variables as tempo, rhythm, instrumentation and size of orchestra” (35). The average stimulus value increases gradually over the fifteen minute period and is followed by a short period of silence because Musak® have found that constant sound can induce fatigue.


The purpose of presenting Musak® and its predecessors here is not solely used to justify the possible application of generative music systems in similar and other environments. Its presence is also used to illustrate that a generative music system could bring even greater benefits to the end user by delivering music in a more effective and organic way than a Musak®-type system could.  The linear nature of the Musak® system means that it is unable to change the characteristics of the music it plays in real-time in the same way that a generative music system could because of its inability to take into account changes in the end user’s environment such as fluctuations in room temperature, brightness, ambient noise and the number  of people in a room.

5. Pay per Play


The delivery of algorithms, whether as the programme itself or separate files enhancing the function of the generative music system, and the payment thereof are important factors when considering the proliferation of generative music systems.


When looking for a model on which to base a delivery and performance accounting structure for generative music systems traditional music industry organisations do not provide very good examples. Both PPL and PRS, for example, have been lacklustre in their efforts to embrace advances in technology in order to create systems that can identify individual public performances of works by their members, preferring instead to refer to other data such as chart information to help them allocate license revenue. The consequence of this is that many members of PPL and PRS fail to receive royalties to which they are entitled.


Instead it may be necessary to turn to the mobile communications industry for a more suitable model. Here many products can now be delivered to mobile phones and other mobile devices. These, of course, include ringtones, trutones, pictures games, music, and videos both as downloads and together with television, streamed live. Payment methods include “pay as you go” or monthly with itemised bills delivered to your door. The development of a similar system could enable the passage of algorithms to be tracked around a network of generative music systems allowing a more accurate and reliable distribution of royalties.

6. A Crack in Ivory Towers?


In the UK, and in fact anywhere else, non-copyright works cannot be licensed. Copyright and the ownership of copyright are the essential building blocks of the music industry. A proliferation of sophisticated generative music systems, installed in both public and private places and producing pure generative music, could have a significant impact on various organisations within the music industry with many of them finding it difficult, if not impossible, to function. The immediate impact on PPL, PRS, record companies and music publishers would be a decrease in licensing income for public performance in commercial environments. Other areas that could suffer include licensing revenue collected from “on hold” music services. Record companies, MCPS (Mechanical Copyright Protection Society Ltd) (36) and music publishers would also find revenue from the sale and manufacturing of CDs decreasing as both commercial premises and the music buying public gradually turn to generative music systems. The erosion of MCPS and PRS revenue collected by licensing ringtones may have, in fact, already started because pure generative ringtones are copyright free.

7. Conclusion


The intention of this paper is not to herald the imminent collapse of the music industry due the development and dissemination of generative music systems. Its intention is merely to highlight the advantages that sophisticated generative music systems could have over linear music systems; the absence of copyright in pure generative music and the ability to adapt the music they produce in real-time. Both these advantages could, in time, give generative music systems a real commercial edge over conventional linear music systems enabling music to be delivered more effectively, and with greater benefits to, the end user.


The intention of this paper is neither to de-value the work of generative music composers by suggesting that they now turn their attention to simply designing a newer sort of wallpaper music to adorn the walls of McDonalds, Starbucks and Gap. The development of generative music systems will require the composer to acquire knowledge and skills from other disciplines in order to develop appropriate generative music systems. Nor is it the intention of this paper to proclaim the end of the creation of new music. In fact the generative music systems suggested may even be a catalyst for new works.


There is, however, one important factor that may impede the progress of the dissemination of generative music systems: people. Until people come to realise how generative music systems might benefit themselves and even their businesses, the music industry machine will continue to spoon feed the majority of the population with the manufactured, publicity and marketing led music it believes they demand.

8. References


 (1) Brown, Paul (2005) “Is the Future of Music Generative?” Music Therapy Today (online) Vol VI (2) 215-274 Available from


(2) Rich, Owain (2003) “The Evolution of the Score and Generative Music: To What Extent Can Computer Code Ever Be Considered a Musical Language?” (email to author) [online]


(3) Great Britain (1988) Copyright, Designs and Patents Act 1988, HMSO, London


(4) Great Britain (1995) The Duration of Copyright and Rights in Performances Regulations 1995, HMSO, London


(5) USA (2003) Copyright Law of the United States of America and Related Laws Contained in Title 17 of the United States Code [online] (Accessed 1st November 2005) Available from

(6) Musoft [online] (last accessed 30th October 2005) Available from

(7) Digital Expressions [online] (last accessed 30th October 2005) Available from

(8) Nyr Sound [online] (last accessed 30th October 2005) Available from

(9) FractMus [online] (last accessed 30th October 2005) Available from

(10) Algorithmic Arts [online] (last accessed 30th October 2005) Available from

(11) SSEYO [online] (last accessed 30 October 2005) Available from


(12) Miller, David (2003) “Game of Chance” [online] (last accessed 30th October 2005) Available from


(13) MadWaves [online] (last accessed 30 October 2005) Available from


(14) Performing Rights Society [online] (last accessed 31 October 2005) Available from


(15) Phonographic Performance Limited [online] (accessed 31 October 2005) Available from


(16) Areni, Charles (2003) “Exploring Managers’ Implicit Theories of Atmospheric Music: Comparing Academic Analysis to Industry Insight”, Journal of Services Marketing, Vol 17 No 2 pp 161-184

(17) PRS (1997) “Music Proves a Valuable Asset to Business” [online] (Accessed 3 November 2005) Available from

(18) Ramos, L.V. (1993). “The Effects of On-Hold Telephone Music on the Number of Premature Disconnections to a Statewide Protective Services Abuse Hot Line”, Journal of Music Therapy, XXX (2), 119-129


(19) PRS (2004) “Music on Hold. Is it Really Worth the Wait?” [online] (Accessed 3 November 2005) Available from


(20) North, A C & Hargreaves, D J (1998) “The Effect of Music on Atmosphere and Purchasing Intentions in a Cafeteria” Journal of Applied Social Psychology, 28(24), pp 2254-2273.


(21) Garton, A  (1996) “Lost Time Accidents – A Journey Towards Self-Evolving , Generative Music” [online] (Accessed 3 November 2005) Available from


(22) McElrea, H & Standing, L (1992). “Fast Music Causes Fast Drinking” Perceptual and Motor Skills, 75, p 362.


(23) Milliman, R.E. (1982). “Using Background Music to Affect the Behaviour of Supermarket Shoppers”, Journal of Marketing, 46, 86-91.


(24) North, A C & MacKenzie, LC (2000) “Musical Tempo, Productivity and Morale” Performing Right Society, London


(25) Sullivan, Malcolm (2002) “The Impact of Pitch, Volume and Tempo On the Atmospheric Effects of Music”, International Journal of Retail And Distribution Management, Volume 30, Number 6, pp 323-330


(26) Sloboda, J. A. (1999). “Everyday Uses of Music Listening: A Preliminary Study”. In Suk

Won Yi (ed) Music, mind and science. Seoul: Western Music Institute


27) Oblad, C. (2000). “On Using Music - About the Car as a Concert Hall”. Proceedings of

the 6th International Conference on Music Perception and Cognition, August 5-10th

2000, Keele University, Keele.


(28) Sloboda, J.A., O'Neill, S. & Ivafdi, V. (2000). “Everyday Experience of Music: An

Experience-Sampling Study”, Proceedings of the 6th International Conference on Music

Perception and Cognition, 5-10th August, Keele University, Keele.


(29) Eacott, J and d'Inverno, M (2002) “Embedded Intelligent Music or iHiFI – The Intelligent HiFi” [online] (Accessed 2 November 2005) Available from


(30) Karageorghis, C.I. & Terry, P.C. (1997). “The Psychophysical Effects of Music in

Sport and Exercise: A Review”, Journal of Sport Behaviour, 20(1), 54-68.


(31) Brown, Paul (2005) “Generative Music Systems as Tools to Assist Music Therapists and Patient Carers in Health Settings (Provisional Title)” Draft Research Degree Proposal


(32) Cowan, D. S. (1991). “Music Therapy in the Surgical Arena”. Music Therapy Perspectives, 9, 42-45.


(33) Robb, S.L., Nichols, R.J., Rutan, R.L., Bishop, B.L. & Parker, J.C. (1995). “The Effects of Music Assisted Relaxation on Pre-Operative Anxiety”, Journal of Music Therapy, XXXII (1), 2-21.


(34) Walworth, Darcy DeLoach (2003) “The Effect of Preferred Music Genre Selection Versus Preferred Song Selection on Experimentally Induced Anxiety Levels” [online] (Last Accessed 2 November 2005) Available from

(35) Fink, Michael (1989) “Inside the Music Business: Music in Contemporary Life” Schirmer Books, New York

(36) Mechanical Copyright Protection Society [online] (accessed 6 November 2005) Available from

[1] The term ‘pure’ generative music is used in this paper to describe music that has been produced by a generative music system or process and not subsequently recorded.


[2] At the time of finishing writing this paper I had yet to receive a response clarifying the situation in the States from the Legal Departments of both ASCAP and BMI.

[3] The absence of both Max/MSP and Supercollider here is deliberate because they require greater technical skill to operate

[4] All these applications and others are discussed in David Miller’s article “Game of Chance” (12)

[5] A summary of some of those studies relating to the effect of music on various physiological parameters is set out in Appendix 1 of the research degree proposal, entitled “Generative Music Systems as Tools to Assist Music Therapists and Patient Carers in Health Settings” (27).