Datasets:
Request access for academic research
This repository is publicly accessible, but you have to accept the conditions to access its files and content.
This dataset is provided exclusively for academic research under the Qlean Academic Research License. Please apply using your institutional email address (university or research institute domain). Requests from personal email addresses (gmail, etc.) cannot be verified and may be rejected. Requests are reviewed manually within 3 business days.
本データセットは学術研究限定です。所属機関のメールアドレスでの申請をお願いします。フリーメールは身元確認ができず承認できない場合があります。審査は手動で3営業日以内に行います。
If you train a model on this dataset, any public release must prohibit commercial use and impose the same restriction on downstream users (Article 5). Many widely used open-source licenses permit commercial use and are therefore not compatible — please check the terms before publishing.
本データセットで学習したモデルを公開する場合は、商用利用を禁じ、同じ制限を下流の利用者にも課す条件とする必要があります(第5条)。広く使われているオープンソースライセンスの多くは商用利用を許諾しており、そのままでは条件を満たしません。公開前にご確認ください。
Log in or Sign Up to review the conditions and access this dataset content.
Japanese Emotional Speech Dataset
感情を込めた発話データセット
Japanese emotional speech recorded by 100 professional voice actors and voice-acting students (male and female). 6,800 studio-recorded WAV utterances (48 kHz, mono) with transcripts.
Provided by Qlean Dataset (Amanaimages Inc. / 株式会社アマナイメージズ) - rights-cleared training data from Japan, released for academic research under a gated license.
Description(概要)
日本人のプロの声優や声優科の学生男女100名による日本語の発話データセットです
Dataset details(データ仕様)
| Field | Value |
|---|---|
| Number of items(データ量) | 6,800 files |
| Total size(データ容量) | 5.90GB |
| File format(ファイル形式) | wav |
| Speakers / subjects(話者数/被写体数) | 100 |
| Sampling rate(サンプリングレート) | 48kHz |
| Channels(チャンネル) | モノラル |
| Channel separation(チャンネル分離) | ー |
| Transcripts(トランスクリプト) | あり |
| Recording environment(収録環境) | スタジオ |
| Speech category(音声カテゴリ) | 感情音声 |
| Data type(データ種別) | 音声 |
License and permitted use(ライセンスと利用範囲)
This dataset is released for academic research purposes only, under the
Qlean Academic Research License (qlean-academic-research-license).
Access is gated: submit your name, affiliation, academic status, intended use,
and contact email. Requests are reviewed manually (within 3 business days).
- License text:
LICENSE_ja.md(Japanese, prevailing) /LICENSE_en.md(English reference) - 本データセットはアカデミックな研究用途に限定して提供されます。利用条件の正本は
LICENSE_ja.mdを参照してください。
Commercial use / 商用利用: not permitted under this license. For commercial licensing, contact us via https://qleandataset.visual-bank.co.jp/contact
Citation(引用)
@misc{qleandataset_speech_emotional_speech,
title = {Japanese Emotional Speech Dataset},
author = {株式会社アマナイメージズ (Amanaimages Inc.)},
year = {2026},
publisher = {Hugging Face},
url = {https://huggingface.co/datasets/qleandataset/speech-emotional-speech}
}
DOI will be added after publication. / DOIは公開後に発行・追記します。
Management information(管理情報)
- Dataset unit ID(データ単位ID): DS-001
- Catalog: Qlean Dataset — speech datasets / EN
- Provider: Qlean Dataset, Amanaimages Inc.(株式会社アマナイメージズ・Visual Bankグループ) — https://qleandataset.visual-bank.co.jp
- Downloads last month
- 30