Home
About Level3
Search archives
Issues
- June 2007
- August 2006
- May 2005
- June 2004
- November 2003
DIT Home

Read postings about this article   |   Post a comment about this article  |  print this article [pdf]


The DiTME Project

Interdisciplinary research in music technology

 

Author - Eugene Coyle, Dan Barry, Mikel Gainza, David Dorran, Charlie Pritchard, John Feeley and Derry Fitzgerald


[<<previous]next>>]


Show/ hide article menu (click icons opposite)

5. New emerging Enterprise Ireland and European Framework projects

Following upon the success of the DiTME project, the Digital Audio Research group has been successful in winning research funding on a number of fronts, with success in an FP6 European Framework, an Enterprise Ireland Commercialisation funded project, and a DIT Abbest and a Strand 1 Ph.D. funded scholarship. There has been a significant increase in the number of researchers in the group, now standing at two postdoctoral fellows, one senior researcher (nearing completion of his Ph.D.), four full-time and two part-time Ph.D., and one full-time and one part-time MPhil research students.

5.1   Enriched access to sound archives through integration enrichment and retrieval

In 2005, the Centre for Digital Music at Queen Mary University London and the Audio Research Group at DIT established a consortium of seven partners including three companies and four academic institutes in a bid for a European Framework project. The bid was successful and the project officially started in May 2006 and will run until November 2008.

Many digital sound archives still have limited access facility to incumbent users. Materials are often in different formats, with related media in separate collections, and with non-standard, specialist, incomplete or even erroneous metadata. Thus, the end user is unable to discover the full value of the archived material. To expose the inherent value of the archived material, powerful multimedia mining techniques are needed, in combination with content extractors, meaningful descriptors, and visualisation tools. There is also a need to improve retrieval effectiveness. Existing retrieval systems often do not take into account the specific nature of the media content. The ability to search collections by speech or musical features is rare. Thus retrieval techniques are restricted and inflexible. To address this, multiple retrieval techniques need to be merged and deployed, and similarity and structure must be conceptualised in order to provide a usable service. An efficient and effective retrieval system needs to be grounded in semantic description, similarity, and structure in order to provide rich functionalities related to the exploration of sound archives.

Another issue is that of providing appropriate interaction with and presentation of material for the end-users. An archive used by musicians and music students, for instance, requires that the material can be manipulated or modified appropriately at playback. Archives of recorded broadcasts need to emphasise appropriate segmentation and interactive speech recognition features. In addition, the creation of tailored collections with customized material has been identified as a strong user need in access systems. These scenarios necessitate the development of enhanced and appropriate retrieval systems, as well as organisational structures and the means to interact with the presentation of materials. This demands appropriate metadata that can be automatically created in order to deliver, share or organise the archives.

This is the motivation for the two and a half year European project, ‘Enabling Access to Sound Archives through Integration, Enrichment and Retrieval’ (EASAIER). EASAIER allows archived materials to be accessed in different ways and at different levels. The system will be designed with libraries, museums, broadcast archives, and music schools and archives in mind. However, the tools may be used by anyone interested in accessing archived material – amateur or professional – regardless of the material involved. Furthermore, it enriches the access experience as well, since it enables the user to experiment with the materials in exciting new ways.

The focus of DIT’s research in this project is to provide a set of real-time access tools which will allow the user to process the retrieved audio in useful ways. DIT will provide the following tools: time scaling, pitch shifting, source separation, noise reduction, and equalisation and enhancement tools.

5.2  Interactive Music Archive Access System

The Interactive Music Archive Access System (IMAAS) project, funded by Enterprise Ireland, is a cross-institute collaboration with the Dublin and Cork Institutes of Technology. The Irish Traditional Music Archive (ITMA) under the direction of Nicholas Carolan is also a valued partner in the project. The collaboration arose out of mutual interests and the strong links between Dr Derry Fitzgerald of CIT and the Audio Research Group at DIT. Dr Fitzgerald, who had originally obtained his Ph.D. in the field of signal processing for audio at DIT some years previously, together with Dr Matt Cranitch, consorted with the ITMA and DIT in a bid to build an interactive music archive access system. The goal is to provide remote users with a web-based access system for music archives such as the traditional music archive in Dublin. The project aims to contribute to the emerging field of music information retrieval and efficient musical descriptors. Such descriptors include time signature, key signature, tempo and tune type, to name but a few. With this information available as metadata, remote users can query large databases of music quickly and efficiently in order to retrieve only the most relevant musical data. The end user is also provided with some powerful audio processing tools to manipulate, analyse and visualise the music which has been retrieved. The project commenced in 2006 and comes to completion in November 2009.

5.3   Audio Research Group

Whilst the individual research themes of the DiTME project – audio time-scale modification, sound source separation, and music transcription (see Sections 2 to 5) –have resulted in significant contributions in advancing knowledge in their respective fields, the combined research outcomes have resulted in an even greater contribution to the field of audio research. The emerging talents of Mikel Gainza, Dan Barry, David Dorran and team mentor Eugene Coyle of DIT, and Derry Fitzgerald who completed his Ph.D. at DIT before taking up an Irish Research Council for Science, Engineering and Technology (IRCSET) scholarship under the guidance of Matt Cranitch at CIT, have provided the core of a significant research group.

With commencement of the EASAIER, IMAAS, ABBEST and VOCAL projects, the Audio Research Group at DIT currently comprises 11 researchers. In addition to the named projects, research is underway in speech synthesis, surround sound algorithms, adaptive music for gaming, musical instrument recognition, and intelligent audio environments.          

In addition to a registered patent in sound source separation, to date the DIT Audio Research Group has published over 35 peer reviewed papers on various aspects of signal processing for audio. This has been supplemented by close on 20 publications by the CIT partner group. For more information about the group visit http://www.audioresearchgroup.com.

[<<previous]next>>]