DAFx-14 / Program 17th International conference on Digital Audio Effects DAFx-14 September 1-5, 2014 Erlangen, Germany
Home Register Authors Venues Program Proceedings Pictures Committees Travel Guide Links

DAFx-14 Program

Tutorials:

At DAFx-14 we are very excited to be able to offer the following tutorials:

Anssi Klapuri - Multipitch Analysis of Music Signals.

Abstract:

Pitch analysis is an essential part of making sense of music signals. Whereas skilled human musicians perform the task seemingly easily, computational extraction of the note pitches and expressive nuances from polyphonic music signals has turned out to be hard. This tutorial starts from the fundamentals of pitch estimation, explaining the basic challenges of the task (robustness to different sound sources, robustness to polyphony and additive noise, octave ambiguity, inharmonicity, missing data, time-frequency resolution) and the processing principles and sources of information that can be used to tackle those challenges. Among the processing principles, we will discuss why autocorrelation-type estimators (as used in speech processing) do not work for polyphonic data and how they can be amended; how phase information can be utilized; how timbral information must be either explicitly modeled or normalized away; etc. Example pictures and sounds will be presented in order to illustrate what kind of data we are dealing with and to develop intuition. Towards the end of the talk, I will describe some state-of-the-art systems by different researchers, and from my own experience, mention some of the practical challenges that I have encountered when developing real-time multipitch estimation on mobile devices in last few years.

Bio:

Anssi Klapuri received his Ph.D. degree from Tampere University of Technology (TUT), Tampere, Finland. He visited as a post-doctoral researcher at Ecole Centrale de Lille, France, and Cambridge University, UK, in 2005 and 2006, respectively. He worked until 2009 as a professor (pro term) at TUT. In 2009 he joined Queen Mary, University of London as a lecturer in Sound and Music Processing. In September 2011 he joined Ovelin Ltd to develop game-based musical instrument learning applications, while continuing part-time at TUT. His research interests include audio signal processing, auditory modeling, and machine learning.

Meinard Müller - Audio Structure Analysis of Music.

Abstract:

One of the attributes distinguishing music from other sound sources is the hierarchical structure in which music is organized. Individual sound events corresponding to individual notes form larger structures such as motives, phrases, and chords, and these elements again form larger constructs that determine the overall layout of the composition. One important goal of audio structure analysis is to divide up a given music recording into temporal segments that correspond to musical parts and to group these segments into musically meaningful categories. One challenge is that there are many different criteria for segmenting and structuring music. This results in conceptually different approaches, which may be loosely categorized in repetition-based, novelty-based, and homogeneity-based approaches. Furthermore, one has to account for different musical dimensions such as melody, harmony, rhythm, and timbre. In this tutorial, I will give an overview of current approaches for the computational analysis of the structure of music recordings, which has been a very active research problem within the area of music information retrieval. As one example, I present a novel audio thumbnailing procedure to determine the audio segment that best represents a given music recording. Furthermore, I show how path and block structures of self-similarity matrices, the most important tool used in automated structure analysis, can be enhanced and transformed. Finally, I report on a recent novelty-based segmentation approach that combines homogeneity and repetition principles in a single representation referred to as structure feature.

Bio:

Meinard Müller studied mathematics (Diplom) and computer science (Ph.D.) at the University of Bonn, Germany. In 2002/2003, he conducted postdoctoral research in combinatorics at the Mathematical Department of Keio University, Japan. In 2007, he finished his Habilitation at Bonn University in the field of multimedia retrieval writing a book titled "Information Retrieval for Music and Motion," which appeared as Springer monograph. From 2007 to 2012, he was a member of the Saarland University and the Max-Planck Institut für Informatik leading the research group "Multimedia Information Retrieval and Music Processing" within the Cluster of Excellence on Multimodal Computing and Interaction. Since September 2012, Meinard Müller holds a professorship for Semantic Audio Processing at the International Audio Laboratories Erlangen, which is a joint institution of the Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU) and the Fraunhofer-Institut für Integrierte Schaltungen IIS. His recent research interests include content-based multimedia retrieval, audio signal processing, music processing, music information retrieval, and motion processing.

Jürgen Herre, Bernd Edler, Sascha Disch - Perceptual Audio Coding.

Abstract:

Perceptual audio has been a key ingredient in the multimedia revolution, enabling the availability of high-quality audio over channels with limited channel capacity, such as the Internet, broadcasting or wireless services. Today, mp3 and other perceptual audio coding technologies are ubiquitous in devices, such as CD/DVD players, computers, portable music players and cellular phones. This tutorial covers the basics of perceptual audio coding, starting with what it means to operate according to psychoacoustic principles rather than Mean Square Error (MSE). The most relevant psychoacoustic effects will be briefly reviewed. From the modules of a perceptual audio coder, the filterbank and strategies for quantization and coding are examined in some detail. Furthermore, we discuss tools for joint stereo coding of two channels. Alongside, the most common coding artefacts that originate from violating perceptual transparency criteria will be demonstrated and explained. Beyond these concepts, modern perceptual audio coders feature tools that can significantly boost their performance further at low bitrates, for example, audio bandwidth extension, parametric stereo or unified speech and audio coding. Some sound examples will be given to illustrate these new advanced tools. Finally, an overview of today's state of the art in compression efficiency is given as well as an outlook of some currently ongoing coding developments.

Bios:

Jürgen Herre joined the Fraunhofer Institute for Integrated Circuits in 1989. He contributed to many perceptual coding algorithms for high quality audio, including MPEG-1 Layer 3 ("MP3") and - during a PostDoctoral term at Bell Laboratories - MPEG-2 Advanced Audio Coding (AAC). Working on more advanced multimedia technologies including MPEG-4, MPEG-7 and MPEG-D, Dr. Herre is currently the Chief Executive Scientist for the Audio/Multimedia activities at Fraunhofer IIS, Erlangen, Germany. Since September 2010, he is a professor at the Friedrich-Alexander University of Erlangen-Nürnberg and the International Audio Laboratories Erlangen.
Bernd Edler obtained his Dipl.-Ing. degree from the Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU) in 1985 and his Ph.D. from the University of Hannover in 1994. There he continued his research in the field of audio coding with a focus on transforms, filter banks, and perception. Since October 2010 he is a professor at the International Audio Laboratories Erlangen, which is a joint institution of FAU and Fraunhofer IIS.
Sascha Disch received his Dipl.-Ing. degree in electrical engineering from the Technical University Hamburg-Harburg (TUHH) in 1999 and joined the Fraunhofer Institute for Integrated Circuits (IIS) the same year. Ever since he has been working in research and development of perceptual audio coding and audio processing. From 2007 to 2010 he was a researcher at the Laboratory of Information Technology, Leibniz University Hannover (LUH), receiving his Doctoral Degree (Dr.-Ing.) in 2011. He contributed to the standardization of MPEG Surround and the MPEG Unified Speech and Audio Coding (USAC). His research interests as a Senior Scientist at Fraunhofer include waveform and parametric audio coding, audio bandwidth extension and digital audio effects.

Sascha Spors, Matthias Geier, Max Schäfer - Sound Field Synthesis with the SoundScape Renderer.

Abstract:

Sound field synthesis with massive-multichannel loudspeaker arrays has been an active research field for the last few decades. Several rendering methods for multiple loudspeakers have been developed including Wave Field Synthesis, Ambisonics, and Vector Base Amplitude Panning. Different loudspeaker installations exist at many institutions throughout Europe. While the their operating software is often home-made and specific to the particular loudspeaker set-up, there exists also a versatile open-source software tool for real-time spatial audio reproduction, the SoundScape Renderer (SSR). It can be adapted to various loudspeaker configurations and provides modules for the most common rendering methods. For headphone use also spatial sound by binaural synthesis is supported. The tutorial gives an introduction to the most common sound field rendering methods, presents the SoundScape Renderer and some of its rendering methods, and allows hands-on experience for a limited number of participants using the 128 loudspeaker array at the Chair of Multimedia Communications and Signal Processing (LMS).

Bios:

Sascha Spors is a Professor and heads the group for Signal Processing and Virtual Acoustics at the Institute of Communications Engineering, Universität Rostock. From 2005 to 2012, he was heading the audio technology group at the Quality and Usability Lab, Deutsche Telekom Laboratories, Technische Universität Berlin, as a senior research scientist. From the Electrical, Electronic and Communication Engineering Faculty of the University Erlangen-Nuremberg, he obtained his Doctoral Degree (Dr.-Ing.) with distinction in January 2006, as a result of his work as a research scientist at the Chair of Multimedia Communications and Signal Processing. The topic of his doctoral thesis was the active compensation of the listening room characteristics for sound reproduction systems. During his thesis work from 2001 to 2006, he conducted research on wavefield analysis, wavefield synthesis and massive multichannel adaptation problems. Sascha Spors is member of the IEEE Signal Processing Society, the "Deutsche Gesellschaft für Akustik (DEGA)" and the Audio Engineering Society (AES). In 2011 he has received the Lothar-Cremer-Preis of the DEGA. He is member of the IEEE Audio and Acoustic Signal Processing Technical Committee and chair of the AES Technical Committee on Spatial Audio.
Matthias Geier is currently working as a research assistant at the Institute of Communications Engineering, University of Rostock. From 2007 to 2012, he was working as a research assistant at the Quality and Usability Lab of Deutsche Telekom Laboratories, TU Berlin. He studied Electrical Engineering/Sound Engineering at University of Technology and University of Music and Dramatic Arts in Graz, Austria, where he received his diploma degree (Diplom-Ingenieur) in 2006.
Maximilian Schäfer studies Electrical Engineering at the Faculty of Engineering of the Friedrich-Alexander-University Erlangen-Nuremberg (FAU). He is a performing musician and acts as consultant in recording, arranging and management in the music business.

Keynote Speakers:

We are very happy to introduce our three keynote speakers for DAFx-14:

Christof Faller - Improving Time-Frequency Upmix through Time-Domain Processing.

Abstract:

Upmix has been broadly used in the professional (broadcast) and consumer (home cinema) domains, to convert stereo signals to 5.1 surround. Our motivation to add time-domain methods (such as reverberators, early reflections, equalisers, exciters, and compressors) came originally from the desire of scalability. At the advent of 3D multi-channel surround, we wanted an upmix that would be scalable to almost any number of output channels. It became quickly clear, for instance, that ambience signals to be reproduced with many loudspeakers need to be generated very differently than in a 5.1 upmix. The initial efforts in adding reverberators were frustrating: while one could hear the potential (amazing envelopment), difficult items sounded too often too bad. Ultimately, time-domain processing improved the quality of our upmix beyond scalability. Specifically, I”ll describe: early reflections for depth in three dimensions, reverberators for generation of multi-channel ambience signals, equalisation of the center channel, and the use of exciters to enhance room signals.

Bio:

Christof Faller received an Dipl. Ing. degree in electrical engineering from ETH Zurich, Switzerland, in 2000, and a Ph.D. degree for his work on parametric multichannel audio coding from EPFL Lausanne, Switzerland, in 2004. From 2000 to 2004 he worked in the Speech and Acoustics Research Department at Bell Labs Lucent and its spin-off Agere Systems, where he worked on audio coding for satellite radio, MP3 Surround, and the MPEG Surround international standard. Dr. Faller is currently managing director at Illusonic, a company he founded in 2006, and teaches at the Swiss Federal Institute of Technology (EPFL) in Lausanne. He has won a number of awards for his contributions and inventions in spatial audio.

Geoffroy Peeters - Audio Indexing for Music Analysis and Music Creativity.

Abstract:

Since the end of the 90's, audio signal analysis has been used more and more in connection with machine learning for the development of audio indexing. One of the specific types of audio content targeted by this indexing technologies is music and the corresponding research field named Music Information Retrieval (MIR). MIR attempt to develop tools for the automatic analysis of music (score, tempo, chord, key, instrumentation, genre, mood, tag classification).
In this talk, I will review the development of this research field, its connection with other research fields and the motivation for its development: from the initial search and navigation over large music collections paradigm (music search engine) to the more recent computational musicology, ethnomusicology and the use of MIR for music creativity.

Bio:

Geoffroy Peeters is a senior researcher at IRCAM where he is leading audio and music indexing activities. He received his Ph.D degree in 2001 and Habilitation degree (accreditation to supervise research) in 2013 from the University-Paris-VI. He has developed new algorithms for timbre description, sound classification, audio identification, rhythm description, music structure discovery, audio summary, music genre/ mood recognition. He owns several patents in these fields. He is co-author of the ISO MPEG-7 audio standard.

Christian Hoyer - The Beatles and Erlangen? Bubenreuth near Erlangen - the Place where the World-Famous Instruments are Made.

Abstract:

Legendary bands, orchestras, stars and virtuosos like Yehudi Menuhin, the Bavarian Radio Orchestra, Peter Kraus, Elvis, the Beatles and the Rolling Stones - they all were playing Bubenreuth instruments. In the post-war years, displaced persons from the Sudeten region (Czechoslovakia) brought musical instrument manufacturing and related industries to the region around Erlangen. The region is home to one of every tenth German musical instrument manufacturers. Bubenreuth in particular was transformed from a farming village to a metropolis of German string instrument making. The community council of Bubenreuth - then a small village of less than 500 inhabitants - decided by 1949 that more than 2.000 people would be resettled there in the following years to come.
Whether it is those learning to play an instrument, musicians in philharmonic orchestras or rock stars - they all appreciate Franconian violins and guitars. Both small artisan workshops and semi-industrial manufacturers produce quality products for the home market, but mainly for export. The viola da gamba-shaped electric bass designed by Walter Höfner in 1956 and played by Sir Paul McCartney exemplifies the story of Bubenreuth's roots in the instrument making tradition of the 17th century and how it extends to the manufacturing of electric guitars today. A museum was formed in 2009 in order to maintain the cultural heritage of Bubenreuth.

Bio:

Christian Hoyer (born 1976) studied History, Political Science and East European Studies at the Universities of Marburg, Keele (UK), London and Erlangen. He received his PhD degree in history in 2007 ("Lord Salisbury and 19th Century English Foreign Policy"). He worked as a lecturer at the University of Erlangen and as a research assistant at the Centre for British Studies at the University of Bamberg. From 2007 till 2011 he had been working as manager of "history communication" for the manufacturer of electric guitars Warwick/Framus. This position included the curatorship of the Framus Museum and the corporate archives in Markneukirchen (Saxony). One of his main tasks was to build up a company museum. Successfully having achieved this, the award for "History Communication of the Association of German Business Archivists" (Vereinigung deutscher Wirtschaftsarchivare VDW) had been awarded to the Framus Museum in 2009. Since 2012, Dr. Hoyer is working with the Erlangen based publishing house Palm & Enke and is head of the Bubenreutheum museum association.

DAFx-14 Program:

DAFx-14 is generously sponsored by the following leading audio engineering companies: