Dataset Viewer
Auto-converted to Parquet Duplicate
audio
audioduration (s)
25
25.1
gender
stringclasses
2 values
accent
stringclasses
4 values
race
stringclasses
2 values
Female
US American (Massachusetts)
White
Male
US American (Utah)
White
Female
US American (Illinois)
White
Female
US America (Texas)
White
Female
US American (Massachusetts)
Black or African American

Voice Cloning Samples

If the samples in the dataset viewer don't load, they can also be accessed here.

Description

This dataset contains single-channel spontaneous monologue recordings from 5 individuals, all of whom are customer service representatives. Each speaker has approximately 3 hours of recorded audio. All speakers have a US American accent.

The recordings were captured using studio microphones with pop filters. Gender, accent, and race information is included as metadata for each sample.

This dataset was created for voice cloning and text-to-speech (TTS) purposes. Rights have been obtained for both use cases, and the data can be used for commercial purposes.

The recordings are rich in phonetic/NATO alphabet spelling, email addresses, phone numbers, order numbers, mailing addresses, company names, and personal names. All identifying information is fictional. Each segment includes a human-reviewed transcript with word-level timestamps.

Technical Analysis

The following metrics were computed across all samples using Silero VAD for speech detection.

Metric Value
Average sample rate 80,160 Hz
Average bit depth 22
Average mean SNR 18.95 dB
Average median RMS 0.021901
Average speech ratio 0.6717
Average spectral centroid 3.552 kHz
Average frequency content 21.497 kHz
Downloads last month
91