WatsonNT/audio_data_russian_annotated

Dataset Audio Russian Annotated

This is a dataset with Russian annotated audio data, split into train for tasks like text-to-speech, speech recognition, and speaker identification.

Features

• —text: Audio transcription (string).
• —speaker_name: Speaker identifier (string).
• —audio: Audio file.
• —utterance_pitch_mean: The average pitch of the speech utterance (float64).
• —utterance_pitch_std: The standard deviation of pitch, representing variability in intonation (float64)
• —snr: Signal-to-Noise Ratio (float64).
• —c50: Clarity index, measures how much of the speech signal energy arrives within the first 50 milliseconds (float64).
• —speaking_rate: The speed of speech (string).
• —phonemes: The sequence of phonemes in the utterance, typically in phonetic transcription or string form (string).
• —stoi: Short-Time Objective Intelligibility, a metric from 0 to 1 estimating how intelligible the speech is to a listener (float64).
• —si-sdr: Scale-Invariant Signal-to-Distortion Ratio, measures the fidelity of a separated or enhanced speech signal relative to the original, invariant to signal scale (float64).
• —pesq: Perceptual Evaluation of Speech Quality, an objective measure of perceived audio quality (float64).
• —noise: A description of the background noise (string).
• —reverberation: Describes the amount or type of reverberation (string).
• —speech_monotony: A qualitative label of monotony or expressiveness of the speech (string).
• —sdr_noise: A textual description of signal-to-noise distortion (string).
• —pesq_speech_quality: A qualitative interpretation of the PESQ score (string).
• —text_description: A textual description of the speech sample (string).

Usage

Load the dataset like this:

from datasets import load_dataset
dataset = load_dataset("kijjjj/audio_data_russian_annotated", split="train")
print(dataset[0])