Info about the MSc Voice Tech

Below is a high-level overview of the academic year.

Term 1A

• Introduction to Voice Technology
• Programming for Voice Technology
• Speech Sounds

Term 1B

• Speech Synthesis I
• Speech Recognition I
• Machine learning for voice technology

Term 2A

• Speech Synthesis II
• Speech Recognition II
• Thesis Design

Term 2B

There are no courses in this term. You are expected to work on and finalize your thesis.

• Preamble: The thesis project is the pinnacle of the master’s program. It is your contribution to the field and proof that you have earned the title “master of science”. However, writing a thesis is difficult and time-consuming endeavor. All the more so because this one year master’s program gives very little margin for error when it comes to delivering a completed thesis by the deadline. Furthermore, the summer plans of your supervisory team, not to mention your own holidays, plus inevitable delays make it clear that there is a benefit to starting early. Therefore, in this text, we spell out how to get to the finish line in the allotted time.
• Tips:
  1. It is useful to organize scientific texts you read in a citation management system like Zotero, which allows you to take notes, highlight, tag and organize. As you read, pay attention to the background section of the text, which in turn will give you a better idea of the state-of-the-art. Look for those texts too and save them also in your citation management system.
  2. It is a good idea to peruse the thesis repository for theses completed by previous MSc VT students and in other departments to become familiar with how a thesis is structured and topics taken up in the past. Perhaps you’ll be inspired. Pay careful attention to “future research” which may be suggested after the conclusion. You are encouraged to adopt such research questions. This creates research continuity and increases the chances that your work will have a greater impact.
  3. Speak to the Voice Tech PhD students (Xiyuan Gao, Phat Do, Li Zhu, Frank Hopwood) and instructors to see if there is any overlap with your interests and theirs. Productive alignment with PhD research can be a win-win. Link topics to PhDs and instructors, if possible. Coming up with a good research question takes time. Don’t pressure yourself.
  4. Consider unconventional approaches like taking a walk, making a mindmap, and talking it through with friends from other fields.

Thesis supervisors

There are different roles for thesis supervisors as detailed below:

• Supervisor: This is the main contact person for your thesis writing activities. The supervisor is a member of the MSc VT staff or PhD team. They are available for meetings and will review your draft and assess your final thesis using the rubric. Naturally, you can consult with any member of the MSc VT staff when it comes to support.
• Second reader: The second reader is also a member of the MSc VT staff or a PhD student who acts as a second reviewer of your thesis. They are not part of your supervision team.
• External reviewer: Partners from industry can support you in different ways. They also receive the final version of your thesis and advise your supervisor on scoring in some/all rubrics. It is the decision of the supervisor on the extent to which to take their advice into account.

Relevant teams

Program Committee 2023-2024

• Alice Vanni
• Brandi Hongell

Advisory Board for the MSc Voice Technology

• Andrea Ballista (CEO, Voiseed)
• Joris Castermans (CEO, Whispp)
• Domenico Crescenzo (CEO, Screevo)
• Kelly Davis (co-founder, Coqui.AI)
• Arjan van Hessen (Head of Imagination, Telecats)
• Simon King (Professor, University of Edinburgh)
• Esther Klabbers (Senior Scientist, ReadSpeaker)
• Javier Latorre (Apple's TTS team)
• Peter-Paul de Leeuw (CEO, Amberscript)
• Johan van der Meulen (Chief Innovation Officer & co-owner, 8D Games)
• Lorenzo Tarantino (CTO, Voiseed)
• Domenique van der Niet (CEO Humain'r)

Other

Preparatory content

This webpage has some links to online resources which prospective and current students may find helpful. Here is a similar resource from the team at Edinburgh.

For programming:

• Coursera course
• Deep learning with Python

Some grading rubrics the MSc Voice Technology

• Grading rubrics