Home

Session 3

Session 3: 4.30 - 6.30 pm
Session chair: Ju-Chiang Wang (National Taiwan University)


4.30 - 4.55 pm
Analyzing Sound Tracings - A Multimodal Approach to Music Information Retrieval
Kristian Nymoen (University of Oslo); Baptiste Caramiaux (IRCAM/CNRS); Mariusz Kozak (University of Chicago); Jim Torresen (University of Oslo)

Abstract
This paper investigates differences in the gestures people relate to pitched and non-pitched sounds respectively. An experiment has been carried out where participants were asked to move a rod in the air, pretending that moving it would create the sound they heard. By applying and interpreting the results from Canonical Correlation Analysis we are able to determine both simple and more complex correspondences between features of motion and features of sound in our data set. Particularly, the presence of a distinct pitch seems to influence how people relate gesture to sound. This identification of salient relationships between sounds and gestures contributes as a multi-modal approach to music information retrieval.


4.55 - 5.20 pm
Advantages of nonstationary Gabor transforms in beat tracking
Andre Holzapfel (INESC Porto); Gino Angelo Velasco (University of Vienna); Nicki Holighaus (University of Vienna); Monika Dörfler (University of Vienna); Arthur Flexer (ÖFAI)

Abstract
In this paper the potential of using nonstationary Gabor transform for beat tracking in music is examined. Nonstationary Gabor transforms are a generalization of the shorttime Fourier transform, which allow
exibility in choosing the number of bins per octave, while retaining a perfect inverse transform. In this paper, it is evaluated if these properties can lead to an improved beat tracking in music signals, thus presenting an approach that introduces recent fi ndings in mathematics to music information retrieval. For this, both nonstationary Gabor transforms and short-time Fourier transform are integrated into a simple beat tracking framework. Statistically signi cant improvements are observed on a large dataset, which motivates to integrate the nonstationary Gabor transform into state of the art approaches for beat tracking and tempo estimation.


5.20 - 5.45 pm
A Musical Mood Trajectory Estimation Method Using Lyrics and Acoustic Features
Naoki Nishikawa (Kyoto University); Katsutoshi Itoyama (Kyoto University); Hiromasa Fujihara (National Institute of Advanced Industrial Science & Technology & Queen Mary, University of London); Masataka Goto (National Institute of Advanced Industrial Science & Technology); Tetsuya Ogata (Kyoto University); Hiroshi G. Okuno (Kyoto University)

Abstract
In this paper, we present a new method that represents an overall musical time-varying impression of a song by a pair of mood trajectories estimated from lyrics and audio signals. The mood trajectory of the lyrics is obtained by using the probabilistic latent semantic analysis (PLSA) to estimate topics (representing impressions) from words in the lyrics. The mood trajectory of the audio signals is estimated from acoustic features by using the multiple linear regression analysis. In our experiments, the mood trajectories of 100 songs in Last.fm’s Best of 2010 were estimated. The detailed analysis of the 100 songs confirms that acoustic features provide more accurate mood trajectory and the 21% resulting mood trajectories are matched to realistic musical mood available at Last.fm.


5.45 - 6.10 pm
What is a 'musical world'? An affinity propagation approach
Eugenio Tacchini (Università degli Studi di Milano); Ernesto Damiani (Università degli Studi di Milano)

Abstract
This work proposes a method based on the affinity propagation clustering technique to classify artists and find representative artists for each musical category ("musical world") using only the listening history log of a music service.

Two variants of the proposed method are compared with a classic k-means clustering approach and an evaluation based on folksonomy analysis is provided. The results suggest that affinity propagation is highly effective in the music domain, allowing for better classification of artists than classic clustering techniques.

Furthermore, an analysis of the results indicates that classifying music by genres, even using more than one genre for each artist, is sometimes an oversimplification of the dynamics that govern the music ecosystem. While most of the slusters found have a strict relationship with a music genre, the characterization of some of the emerged "musical worlds" is related to other aspects like the geographic origin of the artists, the prominent themes in the lyrics, the evocative potential and the association with a culture/lifestyle or the context in which the music has been used.


6.10 - 6.30 pm
CLOSING REMARKS

Theme by Danetsoft and Danang Probo Sayekti inspired by Maksimer