We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

eess.AS

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Electrical Engineering and Systems Science > Audio and Speech Processing

Title: A Survey on Neural Speech Synthesis

Abstract: Text to speech (TTS), or speech synthesis, which aims to synthesize intelligible and natural speech given text, is a hot research topic in speech, language, and machine learning communities and has broad applications in the industry. As the development of deep learning and artificial intelligence, neural network-based TTS has significantly improved the quality of synthesized speech in recent years. In this paper, we conduct a comprehensive survey on neural TTS, aiming to provide a good understanding of current research and future trends. We focus on the key components in neural TTS, including text analysis, acoustic models and vocoders, and several advanced topics, including fast TTS, low-resource TTS, robust TTS, expressive TTS, and adaptive TTS, etc. We further summarize resources related to TTS (e.g., datasets, opensource implementations) and discuss future research directions. This survey can serve both academic researchers and industry practitioners working on TTS.
Comments: A comprehensive survey on TTS, 63 pages, 18 tables, 7 figures, 457 references
Subjects: Audio and Speech Processing (eess.AS); Computation and Language (cs.CL); Machine Learning (cs.LG); Multimedia (cs.MM); Sound (cs.SD)
Cite as: arXiv:2106.15561 [eess.AS]
  (or arXiv:2106.15561v3 [eess.AS] for this version)

Submission history

From: Xu Tan [view email]
[v1] Tue, 29 Jun 2021 16:50:51 GMT (377kb,D)
[v2] Wed, 30 Jun 2021 15:58:54 GMT (384kb,D)
[v3] Fri, 23 Jul 2021 12:32:52 GMT (385kb,D)

Link back to: arXiv, form interface, contact.