AI Methods in Algorithmic Composition: A Comprehensive Survey

Fernández, José‐Jesús; Vico, Francisco J.

doi:10.1613/jair.3908

Cited by 207 publications

(136 citation statements)

References 164 publications

Supporting

Mentioning

113

Contrasting

Unclassified

Order By: Relevance

“…Music generation methods can be divided into two broad categories [9,11]. On one hand are rule-based methods that use hard coded rules and constraints for style emulation and algorithmic composition.…”

Section: Introductionmentioning

confidence: 99%

A Viewpoint Approach to Symbolic Music Transformation

Bigo

Conklin

2016

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. This paper presents a general approach to the transformation of symbolic music. The method is based on viewpoints, which enable the representation of musical surfaces by sequences of abstract features. Along the transformation process, some of these sequences are conserved while some others are variable and can be replaced by generated ones. The initial piece is therefore seen as a template which is instantiated at each transformation. The method is illustrated in the paper with the particular case of transformations occurring at the harmonic level. New chord sequences are generated by sampling from a statistical model in a particular style. The pitch of the notes constituting the template piece are then transformed according to the generated chord sequence.

show abstract

Section: Introductionmentioning

confidence: 99%

A Viewpoint Approach to Symbolic Music Transformation

Bigo

Conklin

2016

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

“…Many techniques to generate music have been developed, which draw upon several approaches of artificial intelligence, such as evolutionary algorithms, machine learning and expert systems [9]. Even though some of these methods produce music which can be deemed as pleasant by human listeners, none of them is actually capable to convincingly evolve its compositional style.…”

Section: Introductionmentioning

confidence: 99%

On the Stylistic Evolution of a Society of Virtual Melody Composers

Velardo

Vallati

2015

Evolutionary and Biologically Inspired Music, Sound, Art and Design

View full text Add to dashboard Cite

Abstract. In the field of computational creativity, the area of automatic music generation deals with techniques that are able to automatically compose human-enjoyable music. Although investigations in the area started recently, numerous techniques based on artificial intelligence have been proposed. Some of them produce pleasant results, but none is able to effectively evolve the style of the musical pieces generated. In this paper, we fill this gap by proposing an evolutionary memetic system that composes melodies, exploiting a society of virtual composers. An extensive validation, performed by using both quantitative and qualitative analyses, confirms that the system is able to evolve its compositional style over time.

show abstract

“…In particular, genetic programming is very useful because it can represent music in a hierarchical manner from a note to a measure and from a measure to a syllable [3]. However, the biggest problem of the method using evolutionary algorithms is that it is difficult to evaluate the fitness how well the individuals fit [12]. This is because it takes a lot of effort and time for a person to evaluate short measures expressed as individuals.…”

mentioning

confidence: 99%

“…Because artificial neural network has learning function, it can learn and output music. Since music changes with time, we can use recurrent neural networks that can learn time series data or feedforward neural networks by representing song data recursively [12]. After learning one or more songs in the artificial neural network, the composition is performed by outputting new songs.…”

mentioning

confidence: 99%