Tight Verses and Loose Choruses: The Shaping of the Metric ‘Pocket’ Across Pop Forms


This paper analyses the construction of the metrical beat across a corpus of pop songs to show how performers’ ‘tight’ and ‘loose’ expression of the beat intersects with large-scale form and the perceived intensity of these sections. I utilise a theory of ‘pockets’, building on Danielsen’s ‘beat bins’ (2006, 2010, 2018), in which beats are spans of time during which an onset may be heard as being part of ‘the beat’. These spans are shaped so events falling at different points in the span are more or less likely to be categorized as being the beat, contra the in/out categorization of beats-as-instantaneous-points (Lerdahl & Jackendoff 1983). I also explore the qualitative effects of these ‘extended beats’, how the ‘tight’ or ‘loose’ construction of the pocket works with secondary parameters (Meyer 1989) to enhance the ‘feel’ of a section. I argue that changes in the shape of the pocket can influence our experience of musical form.I apply Music Information Retrieval techniques to the individual instrument tracks of recordings to extract timing data for all events. Plotting these timing data as distributions (pockets) rather than as microtemporal deviations from a grid offers a novel way of appreciating the nuances of expressive performances. The width of the pocket, adopting performers’ colloquial terms, can be described as tight or loose, and the inter-instrument timing relations as pushing or laying back. Through this lens, the qualitative ‘feel’ of different sections and parts of sections may be compared. For example, in Stevie Wonder’s ‘Superstition’, the pocket in the second half of all sections bar the Bridge is both ‘tighter’ (narrower distributions) and more ‘on top of the beat’ (closer to synchrony). Overall, this paper reframes questions of microtiming towards appreciating the subtle ways performers shape musical time in terms that capture the qualitative listener experience.


Fred Hosken is a PhD student at Northwestern University in Chicago, though spent the first 25 years of his life in London. His research proposes a new theory of beats as ‘pockets’ that can be utilised to provide new perspectives on rhythmic and metric structures in music, as well as new approaches to understanding patterned microtiming in performance, the phenomenon of musical ‘groove’, and musical ‘feel’ more generally.


  1. Love this!

    Of course, Anne (Danielsen) is my supervisor, so I am more than a little interested in the things you are working on, and I have some questions and thoughts.
    First of all, if you haven’t already, you should check the publication list from – Anne’s research project that works with concepts such as the beat bin, extended beats, perceptual centres etc. Many of these papers could be helpful for your analyses. I think Olivier (Lartillot) has added some other functions to MIR as well, that perhaps better models “where the beat is” in a rhythmic event than the onset locator.

    I agree with you in that pockets probably has an asymmetric shape of some sort in general, and that they vary a lot.
    -Do you think it is possible to analyse/model the probable “experienced pockets” in songs while taking different listeners’ different experiences into account?

    With a basis in categorical perception, a repeated rhythmic event belongs to a certain category/density referent/stream until it doesn’t anymore – like, we might prefer to interpret something as belonging to a very wide pocket up until a point where that interpretation is sufficiently disproved (we reinterpret the category something belongs to).
    -Will this mean that a song could be thought to have different wideness to its pockets, even if the relationship between most instrumental streams stay the same? (very hypothetical, I know, but perhaps something to philosophise about)

    I have loads more thoughts, but we can take that another time. You should definitely come visit us in Oslo!
    -Kjell Andreas

Leave a Reply