Scripted Monologue

Our offering of speech data comprised of people reading selected phrases on mobile devices, recorded in 16Kh 16 bit. Each recording is validated by multiple human annotators to ensure a perfect match between text and audio. These datasets are ideal to create base-line models.

Arabic Scripted Monologue

716 hours of Arabic recorded by speakers from the Egypt, Jordan, MSA, and Yemen.

Spontaneous IVR

French Scripted Monologue

826 hours of French Canadian and 844 hours of French (France)

German Scripted Monologue

Russian Scripted Monologue

Portuguese Scripted Monologue

720 hours of Portuguese (Portugal), 371 hours of Portuguese (Brazil)

Japanese Scripted Monologue

Italian Scripted Monologue

English Scripted Monologue

988 hours recorded by native speakers from the UK, the US, and Australia

Telecommunication

Accented English Scripted Monologue

434 hours recorded by speakers of French, Arabic, Spanish, and other languages

Scripted Speech

Dutch Scripted Monologue

1045 hours of Dutch (Belgium), 1265 hours of Dutch (Netherlands)

Spanish Scripted Monologue

280 hours of Spanish (Mexico), 278 hours of Spanish (Spain)

Mandarin Chinese (PRC) Scripted Monologue

Czech Scripted Monologue