Speech researcher at Keio University: full-duplex dialogue TTS, multilingual speech corpora, audio codec analysis
3
papers here
2
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
21
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-17) and cover the author's entire public record, not just the papers listed here. Reproducibility signal comes from Evolve's own reviews.
Research profile
Output—
Insufficient data to score research output. This axis would measure publication volume and consistency over time — total works, annual counts, and venue spread. Coverage of this author's record is incomplete, so no score is computed.
Impact—
Insufficient data to score impact. This axis would measure citation-based influence — citations per work, highly cited papers, and field-normalized standing. The available citation coverage is not yet sufficient to compute a score.
Focus0/100
Focus scores 0, the lowest radar value: the profile spreads across five OpenAlex topic areas — Speech Recognition and Synthesis (208), Speech and Audio Processing (195), Phonetics and Phonology (58), Voice and Speech Disorders (46), and Music and Audio Processing (42) — with the field distribution split evenly at 0.5/0.5 between speech synthesis and spoken dialogue processing and speech processing / computational linguistics.
Collaboration51/100
Collaboration scores 51, a mid-range value. Over the last three years he has 72 works with 12 distinct co-authors, and the frequent-co-author list spans Shinji Watanabe, Samuele Cornell, David M. Chan, Ryuichiro Higashinaka, Yuki Saito, Tetsuji Ogawa, and Hiroshi Saruwatari, among others. On all three 2026 catalog papers he appears in a middle author position rather than as corresponding author.
Activity100/100
Activity scores 100, the maximum. The snapshot records 72 works in the last three years, and per-year counts stay high across the past decade — 29 works in 2020, 28 in 2024, 21 in 2023, and 22 in each of 2025 and 2026 — indicating sustained, near-constant publishing momentum rather than episodic bursts.
Reproducibility55/100
Reproducibility scores 55 (reproMean 55 across 2 scored papers). All three 2026 catalog papers carry a medium repro-readiness rating: Evolve finds the evidence supports YODAS v3's scale and fidelity claims but not its language tags or transcript quality, backs KABURI-TTS's human evaluations despite thin baselines, and treats the codec-token statistics as solid descriptive work whose quality links rest on n=12 correlations.
Research summary
Shinnosuke Takamichi is a speech researcher at Keio University, Japan, whose OpenAlex field distribution splits evenly (0.5/0.5) between speech synthesis and spoken dialogue processing and speech processing / computational linguistics. Topic activity concentrates on Speech Recognition and Synthesis (208 works) and Speech and Audio Processing (195), with smaller threads in Phonetics and Phonology (58), Voice and Speech Disorders (46), and Music and Audio Processing (42). The Evolve catalog holds three 2026 arXiv preprints on which he is a middle author: the YODAS v3 web-scale multilingual speech corpus, the KABURI-TTS stereo dialogue synthesizer, and a cross-codec token-statistics study covering 13 neural audio codecs. The bibliometric snapshot reports an h-index of 21 and an i10-index of 48, 72 works in the last three years, and per-year counts peaking at 29 (2020) and 28 (2024), which yields a maximum activity score of 100. The radar scores collaboration at 51 (12 distinct recent co-authors) and reproducibility at 55 (reproMean across 2 scored papers; all three catalog papers rated medium), while focus scores 0, reflecting topic spread rather than concentration; output and impact remain unscored for insufficient data.
Collaboration pattern
Takamichi's recent footprint is mid-range: 12 distinct co-authors across 72 works in the last three years (collaboration score 51). His frequent co-author list — Shinji Watanabe, Samuele Cornell, David M. Chan, Ryuichiro Higashinaka, Yuki Saito, Tetsuji Ogawa, Hiroshi Saruwatari, Sayaka Shiota, Shunsuke Kando, Satoru Fukayama, Joonyong Park, and William Xingxu Chen — mixes long-running collaborators with international speech-processing researchers. On the three 2026 catalog papers he is a middle author, never corresponding.
Research keywords
Highlights
YODAS v3: a claimed 1.1M-hour open multilingual speech corpus
KABURI-TTS: stereo dialogue TTS with phoneme/activity conditioning
Language-statistical analysis of neural audio codec tokens
Timeline
- 2026Co-authored YODAS v3 (arXiv), a claimed 1.1M-hour, 147-language, 48 kHz stereophonic open speech corpus, as a middle author.
- 2026Co-authored KABURI-TTS (arXiv), an activity-conditioned bi-channel TTS system for full-duplex spoken interaction.
- 2026Co-authored a language-statistical analysis of neural audio codec tokens spanning 13 codecs, corpora, and noise conditions.
Papers in the Evolve catalog
Frequent co-authors
About this profile
This profile is AI-structured by Evolve.Pub from public bibliographic data (OpenAlex) and Evolve's own paper reviews. It is not written or endorsed by the researcher unless the profile is claimed. Radar dimensions are computed from real bibliometrics with fixed formulas — raw figures are shown alongside; missing dimensions mean insufficient data, never zero. Errors? Claim the profile to correct or hide it.
- Bibliometric indicators (h-index, i10-index, counts by year, topics) come from an OpenAlex snapshot and may undercount citations and works.
- The output and impact dimensions are unscored because the underlying data is insufficient; no citation-based influence estimate is offered.
- Paper-level analysis draws only on three Evolve-reviewed works (2026 arXiv preprints, middle author); the rest of the record has not been reviewed.
- Identity is name-resolved from public bibliographic data (OpenAlex, ORCID 0000-0003-0520-7847); Takamichi is a shared Japanese name, so work attribution rests on that resolution.