Goldsmiths Music Corpora

Number of corpora: 3

1. The Goldsmiths-Geerdes corpus of MIDI-transcriptions of popular music

Corpus short title: GoldPop corpus
Version: 1.0
Last updated: August 2006
Owned by: ISMS Group, Computing Department, Goldsmiths College
Owning institution: Goldsmiths College
Data: 14,063 pretty faithful MIDI transcriptions of western commercial pop songs from about 1950 to 2006
From: 1990
To: 2006
Rights: Goldsmiths has the right to use the corpus and parts of it only for non-commercial research and on in-house College computers
Availability: Available for research only within the ISMS Group

Analysis Data

Analysis description:
From: 2006
To: on-going
Analyst: Daniel Müllensiefen (Goldsmiths College)
Analyst: Jamie Forth (Goldsmiths College)
Method: Frequency analysis of melodic formulae
Method: Analysis of repeating drum patterns (SIA)
Publications: 
@incollection{MullensiefenWigginsLewis08,
      		Address = {Frankfurt},
     		Author = {M{\"u}llensiefen, Daniel AND Wiggins, Geraint AND Lewis, David},
     		Booktitle = {Systematic and Comparative Musicology: Concepts, Methods, Findings},
    		Date-Added = {2008-08-19 15:17:36 +0100},
      		Date-Modified = {2009-01-09 17:05:12 +0000},
      		Editor = {Schneider, Albrecht},
      		Pages = {133-155},
      		Publisher = {Peter Lang},
      		Series = {Hamburger Jahrbuch f{\"u}r Musikwissenschaft},
      		Title = {High-level feature descriptors and corpus-based musicology: Techniques for modelling music cognition},
      		Volume = {24},
      		Year = {2008}}
Publications: 
@incollection{FW2009,
       		Address = {London, UK},
      		Author = {Jamie Forth and Geraint A. Wiggins},
        	Booktitle = {London Algorithmics 2009: Theory and Practice},
        	Editor = {Joseph Chan and Jackie Daykin and M. Sohel Rahman},
        	Publisher = {College Publications},
        	Series = {Texts in Algorithmics},
        	Title = {Approaches for identifying salient repetition in  
		multidimensional representations of polyphonic music},
        	Year = {forthcoming}}

Documentation

Documentation: none currently available; see publications

Contact

Daniel Müllensiefen (Department of Computing, Goldsmiths College) <d.mullensiefen'at'gold.ac.uk>

2. Electronic Corpus of Lute Music

Corpus short title: ECOLM
Version: 1.0
Last updated: August 2006
Owned by: ????
Owning institution: Goldsmiths College
Data: Over 1000 fairly literal transcriptions of lute sources into TabCode format. Of these over 800 are in reasonably well-edited (though not publication-quality) forms. A database of metadata on sources and bibliography is also available.
From: 1999
To: 2006
Rights: These are academic transcriptions of out-of-copyright material and, as such, the only restrictions are that the project and corpus should get credit whenever material is used.
Availability: Can be downloaded (but not straightforwardly)

Analysis Data

Analysis description:
From: 1999
To: on-going
Analyst: Tim Crawford (Goldsmiths College)
Analyst: Michael Gale (??Current??University of Southampton)
Analyst: David Lewis (Goldsmiths College)
Method: Computer-assisted manual analysis of motives and textures
Method: Variation detection
Publications: 
@article{GaleCrawfordLute2004
				author = "{Michael GAle and Tim Crawford}",
				title = "{John Dowland’s ``Lachrimae'' At Home and Abroad}",
				journal = {The Lute: The Journal of the Lute Society},
				year = {2004},
				volume = {44},
				pages = {1-34}

Documentation

Documentation: Main documentation on website www.ecolm.org

Contact

Tim Crawford (Department of Computing, Goldsmiths College) <t.crawford'at'gold.ac.uk>

3. Artists Without a Label

Corpus short title: AWAL
Version: 1.0
Last updated: July 2008
Owned by: ISMS Group, Computing Department, Goldsmiths College
Owning institution: Goldsmiths College
Metadata: Associated XML metadata
Audio files: 48,528 audio files in m4a format
From: 1997
To: on-going
Rights: Goldsmiths has the right to use the corpus and parts of it only for non-commercial research and on in-house College computers, with 30-second broadcast rights for demonstration purposes.
Availability: Available for research only within the ISMS Group

Analysis Data

Analysis description:
From: 2008
To: on-going
Analyst: Michela Magas (Goldsmiths College)
Analyst: Benjamin Fields (Goldsmiths College)
Analyst: Christophe Rhodes (Goldsmiths College)
Analyst: Michael Casey (Goldsmiths College)
Method: Audio similarity analyses
Method: Shingle indexing
Publications: 
@inproceedings{1401641,
 author = {Magas, Michela and Casey, Michael and Rhodes, Christophe},
 title = {mHashup: fast visual music discovery via locality sensitive hashing},
 booktitle = {SIGGRAPH '08: ACM SIGGRAPH 2008 new tech demos},
 year = {2008},
 pages = {1--1},
 location = {Los Angeles, California},
 doi = {http://doi.acm.org/10.1145/1401615.1401641},
 publisher = {ACM},
 address = {New York, NY, USA},
}

Documentation

Documentation: none currently available; see publications

Contact

Benjamin Fields (Department of Computing, Goldsmiths College) <b.fields'a'gold.ac.uk>