Sneha Raman

Change your cover photo
Upload
sneha.raman
Change your cover photo
Researcher, AHOLAB Signal Processing Laboratory
This user account status is Approved

This user has not added any information to their profile yet.

Spain

I completed my undergraduate degree in engineering (Information Technology) in 2013 after which I worked at a software company for about a year. Following my interest in music and programming, I did a Masters degree in Acoustics and Music Technology from the Reid School of Music at the University of Edinburgh.

While at the MSc program, I developed a keen interest in speech signal processing as I picked Speech Synthesis as one of my optional subjects. My final MSc project was about using speech signal processing to remotely monitor Parkinson’s Disease. This final project also equipped me and led me to choose this ENRICH position titled “Enriching Disordered Speech”.

In the ENRICH project I aim to build algorithms and software to make disordered speech more intelligible. One important aspect of the project is also to measure the listening effort and cognitive load involved in listening and understanding disordered speech. We are currently focusing on oesophageal speech.

Our first results were presented in the conference proceedings 'Listening to Laryngectomees: A study of Intelligibility and Self-reported Listening Effort of Spanish Oesophageal Speech'. The study had two findings. One, healthy speech is more intelligible and less effortful compared to oesophageal speech. Two, listeners familiar (spouse, close friends etc) with oesophageal speech reported less effort in listening to oesophageal speech compared to listeners who were not.

We have also been able to use some enrichment strategies (voice conversion, duration fixing, better vocoding techniques) to make oesophageal speech more intelligible.

Recent work on the project involved evaluating disordered speech and enriched disordered speech for listening effort using electroencephalography. The aim is to find an enrichment system that results in the least listening effort and improved intelligibility.