Music and Speech Modelling
QMUL ECS 729 Spring 2019
QMUL ECS 729 Spring 2019
Brendan O'Connor
RateMyChat receives inputs in the form of audio conversations classified by quality (classification is subjective which makes the output personalised). The algorithm learns from this data how conversational attributes behave and how this correlates with classifications. It is then able to predict the quality of conversations based on these findings.
Brendan O'Connor
RateMyChat receives inputs in the form of audio conversations classified by quality (classification is subjective which makes the output personalised). The algorithm learns from this data how conversational attributes behave and how this correlates with classifications. It is then able to predict the quality of conversations based on these findings.
Nicole Robson
Despite its anarchic and spontaneous nature, natural conversation is remarkably smooth with minimal pauses between speakers. This project compares the timing and coordination of speech in dyadic conversations between friends and strangers.
Dimitris Stoidis
The purpose of this project is to analyze the gender-specific acoustic features, that differentiate male and female speech, using examples of male actors impersonating female roles in Hollywood movies and test the classification accuracy on State-of-the-Art gender-classifiers.
Rishi Shukla
This project investigates pre-MIDI disco-era musicianship as a comparatively low-tech musical style founded on rhythmic regularity. It uses computational analysis of tempo fluctuation to identify how this was controlled through human performance alone in Rodgers’ recordings, where rhythm sections were purportedly recorded without a metronome and in uncompiled (single) takes.
Chris Yeoward
With the rise of automated playlist generation, and now automatic song mixing, this project explores the techniques used by DJs to express themselves, and put their unique spin on the music they play. Focussing on tempo, this project examines the differences between various DJs, and how an automated playlist compares.Â
Luis Zayas
This project proposes a method for annotating events of jaleos in Flamenco music, and draws connections to meter and structure by analysing a set of recordings in the soleares style.
Luis Zayas
This project proposes a method for annotating events of jaleos in Flamenco music, and draws connections to meter and structure by analysing a set of recordings in the soleares style.
Kartikeya Sharma
An analysis of relevant spectral and temporal attributes of vocal features in music, that govern the preference of a listener to listen to either live music or recorded music, whether it is for entertainment or for medical and music therapy purposes.
Jorge Zuniga
There's lots of research for bird species recognition and classification. But what about categorizing the sound by our perception or the types of sounds a bird makes. For the task of bird sound synthesis, finding suitable species for a particular synthesis model is a problem. Classifying birds by features of their sounds provides way to narrow down which birds work best for which model.