A Tool for the Analysis and Discovery of Ottoman-Turkish Makam Music

Sertan Şentürk, Andrés Ferraro, Alastair Porter, Xavier Serra, "A Tool for The Analysis and Discovery of Ottoman-Turkish Makam Music", in Extended Abstracts for the Late Breaking Demo Session of the 16th International Society for Music Information Retrieval (ISMIR) Conference, Oct. 26th-30th 2015, Málaga (Spain).

Abstact: We present a web application for the analysis and discovery of Ottoman-Turkish makam music. The tool uses an audio-score alignment methodology developed for this music culture for the analysis. It stores the data to be analysed and also executes the analysis algorithms. In the visual interface, the extracted features and alignment results are shown in an audio player, rendering music score and audio representation synchronous to the playback. We plan to improve this prototype in the future mainly for musicological research and music education.

 

Demo Website: dunya.compmusic.upf.edu/makam

 

Recording Page

Corpora [9]:

• Music Scores: SymbTr [3]
• Audio Recording: Compmusic Turkish-makam Audio Corpus [9]
• Metadata: MusicBrainz Dunya Turkish-makam collection

 

Block Diagram of the Audio-Score Alignment Methodology

 

Code:

• Data Storage, interface and algorithms: Dunya [4, 5]
• Some of the algorithms for use in the Dunya project (e.g audio predominant melody extraction [1]): pycompmusic
• SymbTr-txt to MusicXML converter (Note: MusicXML-scores are already available in the SymbTr collection)
• Score data extractor (including score structure analysis)
• MATLAB binaries of joint tonic identification/tempo estimation [2, 6], and section linking/note-level alignment [7, 8]
• Score informed predominant melody correction 
• Note modelling