Bryan Duggan

A system for automatically annotating traditional Irish music field recordings

This paper presents MATT2 (machine annotation of traditional tunes). MATT2 is a novel system which can automatically annotate field recordings of traditional Irish music with useful metadata such as tune name, key signature, time signature, composer and discography. MATT2 works by using a number of algorithms to automatically transcribe digital audio to be annotated to the ABC music notation language. It then compares these transcriptions against a corpus of 860 human made transcriptions in ABC using a variation of the edit distance algorithm. Results using MATT2 to annotate fifty recordings of flute and fiddle tunes demonstrate a high success rate at annotating recordings made by different musicians. Additionally, several of the recordings successfully annotated in testing MATT2 were recorded in imperfect conditions, with badly degraded audio.

Tunepal: searching a digital library of traditional music scores

Purpose – This paper aims to describe the Tunepal project as an example of a music information retrieval (MIR) system that is having an impact on how musicians access, learn and play traditional Irish music around the world.Design/methodology/approach – This paper describes the functionality of the Tunepal system: consisting of the tune corpus, the web site tunepal.org and mobile apps supporting iOS and Android OS. Tunepal facilitates query‐by‐title and query‐by‐playing music (QBP) searches and allows a musician to retrieve and playback scores amongst other supported functions.Findings – Tunepal has been favorably received and musicians report that the system is being used in a variety of scenarios including archiving and the preparation of sleeve notes for commercial recordings. Tunepal has a growing user base in 25 countries.Originality/value – The comprehensive tune corpus (over 16,000 compositions), the query‐by‐playing technology and the fact that the mobile apps provide access to the corpus in situ ...

Rhythm Inference From Audio Recordings of Irish Traditional Music

A new method is proposed to infer rhythmic information from audio recordings of Irish traditional tunes. The method relies on the repetitive nature of this musical genre. Low-level spectral features and autocorrelation are used to obtain a low-dimensional representation, on which logistic regression models are trained. Two experiments are conducted to predict rhythmic information at different levels of precision. The method is tested on a collection of session recordings, and high accuracy scores are reported.

VR 3D Visualization Interface

Human activities produce large amounts of data, from online transactions to scientific experiments, this vast trove of information is being collected on a daily basis and great efforts are being made to produce tools to analyze it, in an attempt to gather insight and knowledge from this inexhaustible source. The use of 3D techniques to display information, and sorting techniques to arrange it are not new. Node-edge graphs predate computers, but their visualization in real-time, three-dimensional virtual reality is quite recent. Some efforts have been conducted by visualization researchers to verify the validity of these techniques, and some HCI experiments have been carried out in navigating 3D spaces. This paper describes a HCI framework to facilitate engagement in an immersive 3D virtual reality (VR) visualization, by means of a compound control system consisting of existing technologies, for navigation and highlighting, and a purpose built controller for selection and manipulation of data; within a software environment that was built using an existing real time 3D game engine.