SDASJun 4, 2020

PJS: phoneme-balanced Japanese singing voice corpus

arXiv:2006.02959v128 citations
Originality Synthesis-oriented
AI Analysis

This provides a freely available, phoneme-balanced dataset for researchers in singing voice synthesis, addressing legal and technical bottlenecks in the field.

The authors tackled the problems of data imbalance and copyright restrictions in singing voice corpora by constructing a phoneme-balanced Japanese singing voice corpus (PJS) with a CC BY-SA 4.0 license, enabling free and reproducible research in singing voice synthesis.

This paper presents a free Japanese singing voice corpus that can be used for highly applicable and reproducible singing voice synthesis research. A singing voice corpus helps develop singing voice synthesis, but existing corpora have two critical problems: data imbalance (singing voice corpora do not guarantee phoneme balance, unlike speaking-voice corpora) and copyright issues (cannot legally share data). As a way to avoid these problems, we constructed a PJS (phoneme-balanced Japanese singing voice) corpus that guarantees phoneme balance and is licensed with CC BY-SA 4.0, and we composed melodies using a phoneme-balanced speaking-voice corpus. This paper describes how we built the corpus.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes