voa-engines/w-v-rp-dataset

VAD Test Dataset - Comparing Different VAD Configurations

Dataset Description

This dataset contains audio segments processed with different Voice Activity Detection (VAD) configurations to compare their effectiveness. Each segment has been transcribed using Whisper.

VAD Configurations Included:

• LongPauseThreshold
• LongPause
• LowSensitivity
• Old
• Default

Dataset Structure

• audio/: Contains the segmented audio files organized by VAD configuration
  • audio/LongPauseThreshold/: Audio segments from the LongPauseThreshold configuration
  • audio/LongPause/: Audio segments from the LongPause configuration
  • audio/LowSensitivity/: Audio segments from the LowSensitivity configuration
  • audio/Old/: Audio segments from the Old configuration
  • audio/Default/: Audio segments from the Default configuration
• vad_test_dataset.csv: Original metadata for all segments
• vad_test_dataset_with_relative_paths.csv: Same as above but with relative paths
• dataset_dict.json: Dataset in HuggingFace compatible format

Loading the Dataset

You can load this dataset using the HuggingFace Datasets library:

from datasets import load_dataset

# Load from the Hub
dataset = load_dataset("voa-engines/w-v-rp-dataset")

# Access the first example
example = dataset["train"][0]
print(f"Audio file: {example['audio']}")
print(f"Text: {example['text']}")
print(f"VAD config: {example['config_name']}")

VAD Parameters Used

LongPauseThreshold

LongPause

LowSensitivity

Old

Default

Created on: 2025-04-17