Single speaker
Studio recordings for TTS, speaker modelling and pronunciation coverage.
2une Voice studioSpeech datasets built around how we speak
Speech data across languages, accents, speakers, conversations and recording environments, built for training and evaluating voice models.
Feature
Voice data, ready to hear
From studio recordings to real world conversations, see the kinds of speech data we can produce.
Okay,320 ms pausedhyaan seEmphasischanges herePitch
Speech annotations
HindiEnglishDisfluencyumm...Code-switchbut eveningsNon-speech event[cough]
Priya
Wait, you actually got the tickets?
Rohan
Yeah, but —
Priya
No way!
actuallyExcitedYeahLaughterOverlap
Okay,320 ms pausedhyaan seEmphasischanges herePitch
Speech annotations
HindiEnglishDisfluencyumm...Code-switchbut eveningsNon-speech event[cough]
Priya
Wait, you actually got the tickets?
Rohan
Yeah, but —
Priya
No way!
actuallyExcitedYeahLaughterOverlap
Speech-to-Text
Speech data across Indian languages, accents and code mixed speech, with human validated transcripts from studio and real world conversations.
Okay,320 ms pausedhyaan seEmphasischanges herePitch
Speech annotations
HindiEnglishDisfluencyumm...Code-switchbut eveningsNon-speech event[cough]
Priya
Wait, you actually got the tickets?
Rohan
Yeah, but —
Priya
No way!
actuallyExcitedYeahLaughterOverlap
Text-to-Speech
Custom speech datasets with control over speakers, language, accent, emotion, speaking style, pacing and recording conditions.
Okay,320 ms pausedhyaan seEmphasischanges herePitch
Speech annotations
HindiEnglishDisfluencyumm...Code-switchbut eveningsNon-speech event[cough]
Priya
Wait, you actually got the tickets?
Rohan
Yeah, but —
Priya
No way!
actuallyExcitedYeahLaughterOverlap
Speech-to-Speech
Natural multi turn conversations that capture how people actually interact, including turn taking, interruptions, emotion and prosody.
How we work
Hear the difference
The same words can sound very different depending on the speaker, environment and way the conversation happens.
Real world
Speech recorded with background noise, distance, interruptions and environmental sound.
2une Studio
Controlled recordings with consistent acoustic conditions and high quality audio.
Conversation
Natural multi speaker speech with turn taking, overlap, pauses and interaction.
Infrastructure
Voice data, built to your requirements
Define the dataset around your model, from language and speakers to recording conditions and delivery.
2.7M+
Hours of studio clean audio
Studio recorded speech across major Indic languages, captured at 48 kHz.
4,200+
Speakers
A diverse speaker network across languages, accents and voice profiles.
968+
Studios across India
A distributed recording network for custom collection and on demand datasets.
600+
Contributors
Human transcription and quality control across voice data workflows.
Samples
Voice data, ready to hear
Explore examples across speakers, languages and recording conditions.
2une Voice studio
Two speaker conversation
Natural dialogue for conversational AI, speech recognition and turn taking evaluation.
2une Voice studio
Multi speaker & noisy
Real world speech for diarisation, speech recognition and understanding speech in noisy environments.
2une Voice studio
Single speaker
Studio recordings for TTS, speaker modelling and pronunciation coverage.
2une Voice studio
Two speaker conversation
Natural dialogue for conversational AI, speech recognition and turn taking evaluation.
2une Voice studio
Multi speaker & noisy
Real world speech for diarisation, speech recognition and understanding speech in noisy environments.
2une Voice studioSpecify a dataset or request samples
Tell us what you need.