MCFlow: A Digital Corpus of Rap Transcriptions
Keywords:rap, corpus, rhythm, rhyme, phrasing, historical trends
This paper describes a new digital corpus of rap transcriptions known as the Musical Corpus of Flow (MCFlow). MCFlow currently contains transcriptions of verses from 124 popular rap songs, performed by 86 different rappers, containing a total of 374 verses, and consisting of 5,803 measures of music. MCFlow transcriptions contain rhythmic information, encoded in musical durations, as well as prosodic information, syntactic information, and phonetic information, including the identification of rhymes. In the second part of the paper, preliminary analyses of the corpus are presented, describing the "norms" of several important features of rap deliveries. These features include speed, rhyme density, metric position of stressed syllables, metric position of rhymes, phrase length, and the metric position of phrases. Several historical trends are identified, including an increase in rhyme density and phrase variability between 1980 and 2000. In each analysis, variance between different performers is compared to variance between songs. It is found that there is generally more variability between songs than between performers.
Copyright (c) 2017 Nathaniel Condit-Schultz
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.