** The Music and Science list is managed by the Institute of Musical Research (www.music.sas.ac.uk) as a bulletin board and discussion forum for researchers working at the shared boundaries of science and music. **
MESSAGE FOLLOWS:
Dear all,
Next Wednesday 2 February, at 3pm, Matthias Mauch will present the
seminar 'Lyrics-to-Audio Alignment: Methods of Integrating Textual
Chord Labels and an Application'.
The seminar will take place in room 105 in the Electronic
Engineering building, Queen Mary University of London, Mile End
Road, London E1 4NS. Directions of how to get to Queen Mary and
details of future seminars can be found at
http://www.eecs.qmul.ac.uk/newsevents/researchgroupevents.php?i=12.
The room is under access control, so people from outside QM will
need to contact C4DM to get in - the lab phone number is +44 (0)20
7882 7480 and if I'm not available, anyone else in the lab should
be able to help. If you are coming from outside Queen Mary, please
let me know, so I can make sure no-one's stuck outside the doors.
All are welcome to attend. For those unable to do so, a video
recording of the seminar will be streamed live and also made
available online after a few days. Please see the above website
for details.
If you wish to be added to / removed from our mailing list, please
send me an email and I'll be happy to do so.
Wednesday's seminar (2 February, 3pm):
Title:
Lyrics-to-Audio Alignment: Methods of Integrating Textual
Chord Labels and an Application
Speaker:
Matthias Mauch (National Institute of Advanced Industrial
Science and Technology, Japan)
Abstract:
Aligning lyrics to audio has a wide range of applications such as
the automatic generation karaoke of scores, song-browsing by
lyrics, and the generation of audio thumbnails. Existing methods
are restricted to using only lyrics and match them to phoneme
features extracted from the audio (usually mel-frequency cepstral
coefficients). Our novel idea is to integrate the textual chord
information provided in the paired chords-lyrics format known from
song books and Internet sites into the inference procedure. We
propose two novel methods that implement this idea: firstly,
assuming that all chords of a song are known, we extend a hidden
Markov model (HMM) framework by including chord changes in the
Markov chain and an additional audio feature (chroma) in the
emission vector; secondly, for the more realistic case in which
some chord information is missing, we present a method that
recovers the missing chord information by exploiting repetition in
the song. We conducted experiments with five changing parameters
and show that with accuracies of 87.5% and 76.0%, respectively,
both methods perform better than the baseline with statistical
significance. We will demonstrate Song Prompter, a software system
that acts as a performance assistant by showing horizontally
scrolling lyrics and chords in a graphical user interface,
together with an audio accompaniment consisting of bass and MIDI
drums. The application shows that the automatic alignment is
accurate enough to be used in a musical performance.
Bio:
Matthias Mauch received the Diplom degree in mathematics from the
University of Rostock, Germany, in collaboration with the
Max-Planck Institute for Demographic Research. He received his
Ph.D. degree from the Centre for Digital Music at Queen Mary
University of London, U.K. Currently he is a post-doc research
scientist at the Media Interaction Group of the National Institute
of Advanced Industrial Science and Technology (AIST), Japan. His
research focuses on the automatic extraction of high-level musical
features from audio, with an emphasis on harmonic progressions and
repetitions. He is songwriter in the band Zweieck. Website:
http://matthiasmauch.net
Emmanouil Benetos
--
Centre for Digital Music (C4DM)
School of Electronic Engineering and Computer Science
Queen Mary, University of London
[log in to unmask]
Tel: +44 (0)20 7882 7480
Fax: +44 (0)20 7882 7997
C4DM Web-site : http://www.elec.qmul.ac.uk/digitalmusic/index.html