Interest for Participation in ICASSP 2023 Signal Processing Grand Challenge on TTS

[Anusha Prakash]

Dear all,

We are proposing a ICASSP 2023 Signal Processing Grand Challenge on development of a lightweight, multi-speaker, multi-lingual Text to Speech (TTS) model. For this purpose, we share TTS datasets in Marathi, Hindi, Telugu, and Indian English. These studies are important for data selection in TTS, for various landscapes such as India, Europe, etc which have a large number of languages and their regional variations. These works can also enable the development of large-scale multi-speaker, multi-lingual TTS models. TTS corpora in these languages are being built as a part of the SYSPIN project at SPIRE lab, Indian Institute of Science (IISc) Bangalore, India, in which large corpora comprising 40 hours of single speaker’s speech in each of nine Indian languages are being collected.

If you are interested in participating in this challenge, please fill the following Google form: https://forms.gle/EzHQsTFPSViDcmV79

Organizing Committee:

Prasanta Kumar Ghosh, IISc, Bangalore

Hema Murthy, IIT Madras

Pranaw Kumar, CDAC Mumbai

Philipp Olbrich, GIZ, Germany

Keiichi Tokuda, Nagoya Institute of Technology

Mark Hasegawa-Johnson, University of Illinois

Heiga Zen, Google