• 多學科研究/DoS
  • 風格與流派/G&C
  • 電聲音樂學/MEM
  • 表演與展示/PPP
  • 製作與處理/SPM
  • 音樂與結構/Str
  • 術語表
    Glossary
  • 參考文獻
    Bibliography
  • 資料作者
    Authors
  • 研究人員
    Researchers
  • 新聞
    News
  • 鏈接
    Links
  • 講座
    Lecture
  • 音樂會
    Concert
  • 共同體
    Community
  • EARS 2
  • 樹擴展
    Tree Expansion

搜索
Search

Home Page

树状结构导航/Treeview Navigation

Markov chains

马尔科夫链/Ma3 Er3 Ke3 Fu1 Lian4

n/a

CHEARS: False
EARS: False
CMT: True
EARS2 Encyclopedia: False


Figure 19.13


Figure 19.14

CHEARS:

源自《计算机音乐教程》、罗玆、人民音乐出版社、2011。

An important subclass of stochastic algorithms are the family of Markov chain techniques. The Markov chain is one of the earliest and most popular strategies for algorithmic composition of music. First formulated in 1906 by the Russian mathematician A. A. Markov (1856 1922), a Markov chain is a probability system in which the likelihood of a future event is determined by the state of one or more events in the immediate past.

The probabilities in a Markov chain can be laid out in a state-transition matrix. The state-transition matrix in figure 19.13 is a Markov chain for a simple melody composition, using the notes of a pentatonic scale: G, A, C, D, and E. The cells show the probability of a next pitch, given a current pitch shown at the left of the row.

Figure 19.14 shows a possible sequence of pitches generated by this matrix. One can increase the sophistication of this method by increasing the order of the chain. The order of a Markov chain indicates the number of prior states that are taken into consideration.

- Events in a zeroth-order chain (equivalent to a regular probability table with no looking back at previous states) are independent of one another.

- Events in a first-order chain have one predecessor (such as the example in figure 19.13).

- Events in a second-order chain look back two states, and so on.

Advanced Markov Techniques for Music: an interesting extension involves turning the Markov chain into an interactive composition and performance game. Playing with the control parameters in real time, musicians can change the state of the Markovian process on a moment-to-moment basis, as music is generated.

Hierarchical Markov Chains: one chain that generates high-level structure could select one of a number of types of sections, for example, a fast movement, a slow movement, a movement in a particular key, or a movement that relied on a particular pitch row. Within a section intermediate-level chains might select the sequence of phrases contained in the section. The details of each phrase type might be filled in by a low-level chain.

(Source: Curtis Roads, The Computer Music Tutorial. MIT Press. 1995)

EARS:

EARS 2:

参看其它/See Also

术语翻译/Terms Translator

术语校对/Terms Proofreader

术语顾问/Consultant to terminology

参考文献/Bibliography

标题Title副标题Subtitle

讲座/Lecture

题目Title时间/Start Time

评论/Comments

评论/Comment ID用户名/User Name内容/Content时间/Input Date
241mungo

Source of CHEARS comes from Chapter 19 (Page 878 to 880).

4/11/2015 3:49:49 AM

发表评论(预先注册)/Add comment on this term(members ONLY)