What makes music groovy?

©20140417 Volkskrant

Today PLOS ONE will publish an interesting study on rhythm, groove and syncopation that uses an often criticized methodology: questionnaire and web-based research (cf. Honing & Reips, 2008). However, this study is a good example of how an unspectacular method (compared to brain imaging techniques, elegantly controlled experiments or advanced computational modelling) can still be quite informative.

The paper takes advantage of people's intrinsic interest in rhythm, timing and what's often called 'groove', in combination with the sheer fun of participating in an online listening experiment that has to do with music (cf. Honing & Ladinig, 2008).

Based on sixty-six responses, the authors were able to extract an inverted U-curve for a music-theoretic measure of syncopation, a shape that was absent for an alternative, information-theoretic measure based on the acoustic quality of the soundexamples used (i.e. entropy). As such the study provides evidence that the theoretical notion of syncopation, as defined by Longuet-Higgins' L-model in the 1980s, might be an important component of 'groovyness': pleasure and dance-inducing aspects of many musics ranging from James Brown to Marvin Gay and from Funkadelic to Stevie Wonder.

The first author, Maria Witek (Aarhus University), encouraged me to mention that the questionnaire and all soundexamples are online. Feel free to take part.

[See also article in de Volkskrant and interview on Radio 1; both in Dutch]

Witek, M., Clarke, E., Wallentin, M., Kringelbach, M., & Vuust, P. (2014). Syncopation, Body-Movement and Pleasure in Groove Music PLoS ONE, 9 (4) DOI: 10.1371/journal.pone.0094446

Honing, H., & Reips, U.-D. (2008). Web-based versus lab-based studies: a response to Kendall (2008). Empirical Musicology Review, 3 (2), 73-77.

Honing, H., & Ladinig, O. (2008). The potential of the Internet for music perception research: A comment on lab-based versus Web-based studies. Empirical Musicology Review, 3 (1), 4-7.

Art meets science?

Below a video impression of an evening that was organized this week by the Studium Generale of the University of Groningen. The idea of the lecture/concert was to explore tempo and timing, swing and groove from the perspective of both the performer and the listener (an idea that turned out not always to be a success ;-) See for a longer fragment here.

How well would you do as an expert?

In the Netherlands (and I’m sure there are versions of it in the UK and the US as well) there is a weekly radio show containing a returning item in which music experts are asked to compare and judge two or three CD recordings of the same piece, without knowing who the musicians are. They have to guess the performers and describe why they do (or don’t) like that particular performance.

How well would you do in such a test? The common hypothesis is that experts do this much better, e.g. under the assumption that they are more sensitive in their listening skills. But do experts indeed hear more detali and more nuances when compared to a 'common listener'? Or do they just have more terminology available to verbalize these differences?

Two years ago our group did a large-scale online listening experiment with a similar task. Participants were asked to compare several pairs of recordings of well-known musicians. One of the recordings was taken directly from a CD, but the other was originally performed at another tempo (faster or slower) and then scaled to be similar in tempo to the former recording. The task was to judge which recording was real and which one was manipulated, by focusing on the timing used by the performer.

To give you an idea of the difficulty of the task, below an example.

A
B
(See answer at the bottom.)

The results were recently published in the Journal of Experimental Psychology, with a surprising outcome: the judgments seem to be largely influenced by exposure to music (listening a lot to one’s favorite music) and not (at all) by the level of expertise (amount of formal musical training). One seems to learn a lot by simply listening.

Honing, H., & Ladinig, O. (2009). Exposure influences expressive timing judgments in music. Journal of Experimental Psychology: Human Perception and Performance, 35 (1), 281-288 DOI: 10.1037/a0012732

* The first recording is the original. It is Glenn Gould performing English Suite No. 4 by J.S. Bach. The second recording is Sviatoslav Richter performing the same piece. However, this recording was sped up from 70 to 87 bpm making his use of tempo rubato 'unnatural'.

Final ritards in Balinese music?

This week an interesting article appeared in Empirical Musicology Review, an open peer-reviewed journal on music. Andrew McGraw (University of Richmond) discusses the use of tempo-change in Balinese music.

The most common kind of tempo-change is often referred to as the ‘final ritard’: the typical slowing down at the end of a music performance, apparent in Javanese and Balinese gamelan music, music from the Western Baroque and Romantic period, but also in quite some pop and jazz genres.

An important contribution to this topic is made by a family of computational theories, so-called ‘kinematic models’, that propose an explicit relation between the laws of physical motion (elementary mechanics) in the real world and chaneg of tempo (so-called "expressive timing") in music performance. These models were shown to produce a good fit with a variety of empirical performance data, suggesting that the final ritard alludes to human movement: the pattern of runners’ deceleration.

Unfortunately, the McGraw study is yet another example of a study that takes ‘tempo curves’ too seriously as a potential description, or even mental representation, of tempo-change in music (A hobby horse of mine that I shouldn't bring in once more; cf. here).

Furthermore, the author seems to be unaware of the notorious mistake made by Feldman, Epstein and Richards (MIT, Cambridge, Mass.) in their 1992 study. In there the authors propose a theory of tempo change (or rubato) based on the laws of physical motion, but in the end fit the empirical data to models unrelated to these laws. So indeed, the conclusion that "previous idealized models are too simplistic to describe Balinese music" is correct. In fact, it has been shown for both music performance and music perception. The challenge is still to model the regularity and structure that can characterize this particular use of tempo rubato.

Nevertheless, the paper is a much needed contribution to music perception and cognition research by studying other than Western classical music, a genre that is still dominating the literature.

McGraw, A.C. (2008). The Perception and Cognition of Time in Balinese Music. Empirical Musicology Review, 3(2), 38-54.

HONING, H. (2005). IS THERE A PERCEPTION-BASED ALTERNATIVE TO KINEMATIC MODELS OF TEMPO RUBATO?. Music Perception, 23(1), 79-85. DOI: 10.1525/mp.2005.23.1.79

Can musical expression be understood?

In 1985 the composer György Ligeti published some wonderful Etudes for piano. I regularly listen to them ever since. They were also the material (or at least a few bars of “Cordes Vides”) to test an idea for an expression editor: a system that was aimed to facilitate editing operations that are musically meaningful while at the same time amounting to perceptually sound results. The editor, named Expresso, was conceived as a calculus of expression. Quite an ambitious project that aimed to formally describe how different types of expression are linked to different types of musical structure.

The research was inspired by several researchers in the field of “musical expression”: Henry Shaffer, Eric F. Clarke, John Sloboda,Christhoper Longuet-Higgins, Johan Sundberg, Alf Gabrielsson (just to name a few European researchers), all of them contributed, along with quite a few fellow researchers all over the world, to an understanding on how "expressive nuances" in music performance make the difference between one and another performance. A recent paper by Luke Windsor (University of Leeds) summarizes and elaborates on this research. You might like it.

Windsor, W.L. et al. (2006). A structurally guided method for the decomposition of expression in music performance. The Journal of the Acoustical Society of America, 119(2), 1182. DOI: 10.1121/1.2146091

A 2006 recording of Glenn Gould?

Sony Music recently released a new recording (made in 2006) of Glenn Gould performing the Goldberg Variations. Curious, not? The recording was made using measurements of the old recordings and then regenerating the performance on a computer-controlled grand piano, a modern pianola.

This technology dates from the early nineties, a time when several piano companies (including Yamaha and Bösendorfer) combined MIDI and modern solenoid technology with the older idea of a pianola. Old paper piano rolls with recordings of Rachmaninoff, Bartok, Stravinsky and others were translated to MIDI and could be reproduced ‘live’ on modern instruments like the Yamaha Disklavier. Until now, the only left challenge was to be able to do this for recordings of which no piano-rolls were available.

Besides the technicalties of all this, for most people the real surprise —or perhaps disillusion— might well be the realization that a piano performance can be reduced to the ‘when’, ‘what’ and ‘how fast’ the piano keys are pressed. Three numbers per note can fully capture a piano performance, and it allows for replicating any performance on a grand piano(-la). The moment a pianist hits the key with a certain velocity, the hammer releases, and any gesture that is made after that can be considered merely dramatic: it will have no effect on the sound. This realization puts all theories about the magic of touché in a different perspective.

Nevertheless, while it is relatively easy to make the translation from audio (say a recording from Glenn Gould from 1955) to the what (which notes), and the when (timing) in a MIDI-like representation, the problem is in the ‘reverse engineering’ of key velocity. What was the speed of Gould’s finger presses on the specific piano he used? The Zenph Studios claim to have solved it for at least a few recordings. Only trust your ears :-)

Piano touch unraveled?

A few postings ago I mentioned a remake of Glenn Gould’s Goldberg variations. It was related to the topic of piano touch (or touché), a notion pianists and music lovers often talk about, and that is, nevertheless, surrounded with a lot of magic.

Several researchers are researching this topic, including Werner Goebl and Caroline Palmer at McGill University, Canada. They presented their recent findings at the SMPC conference on music perception and cognition in Montreal. Using a movement tracking device it was possible to track a pianist’s finger movements on a digital piano keyboard (Apparently a grand piano could not be used because of the need to film/measure these movements from the piano towards the hands; see photo).

By analyzing the performances of twelve professional pianists, they found that different finger movements did not lead to differences in timing precision or in tone intensity. That is a novel finding. However, the actual relation between the finger movements and the resulting velocity of the piano key after contact was not studied as yet (a replication of this study on a modern pianola —like the Yamaha Disklavier or Bösendorfer— seems a logical next step).

My hunch is that the finger dynamics will not matter so much (as was in part suggested by this study). The gestures made by a pianist, including finger movements and what is generally referred to as piano touch, have more to do with habit and a sense of control, then that they actually have an influence on the key velocity that, next to the timing, effectively contributes to the sound and musical quality of the performance. This type of research will find out soon …

Why doesn’t it groove?

Jazz and pop musicians spent a lot of time trying to work out ‘the feel’, ‘the groove’, or how to ‘time’ a particular piece of music. It is everything but arbitrary, and even the smallest detail counts. All to get the right timing at the right tempo. It clearly matters!

Music performance studies have looked at these timing details a lot. While often focusing on classical music, more and more studies are now looking at jazz, pop and world music. Tomorrow Bas de Haas (studying at the University of Utrecht) hopes to graduate on a MSc thesis on groove and swing. He asked three well known Dutch drummers —Joost Lijbaart, Joost Kroon and Marcel Seriese— to play a fragment of Swing, the famous break from Funky Drummer by James Brown, and a so-called Shuffle. And had them do this at different tempi.

As always, the relation between timing and tempo turns out to be more complicated than thought of previously. A straightforward model would be that all timing scales proportionally with tempo. It is like making a particular movement: when performing it at a different speed, people generally do it faster (or slower) by doing every part of the movement faster (or slower) in proportion. This apparently works for computer models that imitate human movement, but does not work for music, let alone for our ears. If you slow down a recording you probably immediately hear that something is wrong. Not because the tempo is wrong, but because the timing sounds awkward.

The challenge is to make a model of timing and tempo that, when, for instance, Funky Drummer is scaled to a different tempo, it still sounds groovy. Bas de Haas hopes to show his first attempts at a conference in Montreal this summer.

* See here for a related paper.