Martin Gasser

• MSc
• Researcher at Austrian Research Institute for Artificial Intelligence

Music psychologists have long been concerned with phenomena such as repetition and tonality – both the internal representations formed by exposure to these phenomena, and how these representations vary with expertise. A question arises of whether less expert listeners can gain proficiency in perceiving a particular phenomenon by being exposed to re...

This paper describes the theory and implementation of a digital audio workstation plug-in for chord sequence generation. The plug-in is intended to encourage and inspire a composer of electronic dance music to explore loops through chord sequence pattern definition, position locking and generation into unlocked positions. A basic cyclic first-order...

In this paper we present a real-world application (the first of its kind) of machine listening in the context of a live concert in a world-famous concert hall – the Concertgebouw in Amsterdam. A real-time music tracking algorithm listens to the Royal Con-certgebouw Orchestra performing Richard Strauss' Alpensinfonie and follows the progress in the...

We present a system that we call 'The Piano Music Com-panion' and that is able to follow and understand (at least to some extent) a live piano performance. Within a few seconds this system can identify the piece that is being played, and the position within the piece. It then tracks the progress of the performer over time via a ro-bust score follow...

We present a first integrated prototype towards our long-term goal of building a complete classical music companion. When listening to a live performance of classical piano music, our system is able to almost instantly identify the piece that is being played, and the position within the piece. It then tracks the performance via an accurate and robu...

Both in interactive music listening, and in music perfor-mance research, there is a need for automatic alignment of different recordings of the same musical piece. This task is challenging, because musical pieces often contain parts that may or may not be repeated by the performer, possi-bly leading to structural differences between performances (o...

Most methods to compute content-based similarity be-tween audio samples are based on descriptors represent-ing the spectral envelope or the texture of the audio signal only. This paper describes an approach based on (i) the extraction of spectro-temporal profiles from audio and (ii) non-linear alignment of the profiles to calculate a distance measu...

We present a study on the limitations of an interactive music recommendation service based on automatic computation of audio similarity. Songs which are, according to the audio similarity function, similar to very many other songs and hence appear unwantedly often in recommendation lists keep a significant proportion of the audio collection from be...

Multivariate Gaussians are of special interest in the MIR field of automatic music recommendation. They are used as the de facto standard representation of music timbre to compute music similarity. However, standard algorithms for clustering and visualization are usually not designed to handle Gaussian distributions and their attached metrics (e.g....

In audio based music similarity, a well known effect is the existence of hubs, i.e. songs which appear similar to many other songs without showing any meaningful per-ceptual similarity. We show that this effect depends on the homogeneity of the samples under consideration. We compare three small sound collections (consisting of poly-phonic music, e...

We present an application of content-based music recom-mendation techniques within an online community platform targeted at an audience interested mainly in independent and alternative music. The web platform's goals will be de-scribed, the problems of content management approaches based on daily publishing of new music tracks will be dis-cussed, a...

A new algorithm for automatic generation of playlists with an inherent sequential order is presented. Based on a start and end song it creates a smooth transition allowing users to discover new songs in a music collection. The approach is based on audio similarity and does not require any kind of meta data. It is evaluated using both objective genr...

We propose a content-based approach to explorative visu- alization of online audio streams (e.g., web radio streams). The visualization space is defined by prototypical instances of musical concepts taken from personal music collections. Our system shows the relation of prototypes to each other and generates an animated visualization that places re...

The visualization of time-dependent simulation data, such as data sets from computational fluid dynamics (CFD) simulation, is still a very challenging task. In this paper, we present a new approach for the interactive visual analysis of flow simulation data, which is especially targeted at the analysis of time-dependent data. This supports the flex...

We propose a novel method for detecting changes in the harmonic content of musical audio signals. Our method uses a new model for Equal Tempered Pitch Class Space. This model maps 12-bin chroma vectors to the interior space of a 6-D polytope; pitch classes are mapped onto the vertices of this polytope. Close harmonic relations such as fifths and th...

This paper presents an implementation of a simple playlist generator. An audio-based music similarity measure and simple heuristics are used to create playlists given minimum user input. The ultimate goal of this work is to conduct a field study, i.e., to run the system on the users' personal col- lection and study the usage behavior over a longer...

SimVis is an interactive research implementation of a visual analysis scheme which we developed for the feature-based assessment of flow simulation data. It is based on the concept of multiple, linked views of different kinds, including views from information visualization such as adapted scatterplots and histograms. In this paper we present an ext...

Visualization of high-dimensional and time-dependent data, resulting from computational simulation, is a very challenging and resource-consuming task. Here, feature-based visualization approaches, which aim at a user-controlled reduction of the data shown at one instance of time, proof to be useful. In this paper, we present a framework for interac...

Analyzing large, time-dependent, and high-dimensional data sets resulting from CFD (computational fluid dy- namics) simulations is a difficult and often very time-consuming task. Especially in the field of the automotive engineering industry, simulations are very time-consuming themselves and have to be carried out multiple times to increase perfor...

Visualization of high-dimensional, large data sets, resulting from computational simulation, is one of the most challenging fields in scientific viualization. When visualization aims at supporting the analysis of such data sets, feature-based approches are very useful to reduce the amount of data which is shown at each instance of time and guide th...

In this paper an overview over some methods of transfer functions design for volume rendering applications will be given. First, TF design in general and multidimensional transfer functions in spe-cial will be presented. After this rather theoretic introduction a short overview over existing solutions for TF specification will be given and two rath...