ekacare/denoising-impact-evaluation-dataset

【ekacare/denoising-impact-evaluation-dataset】是一个全面的基准数据集，旨在评估语音增强对医疗语音上下文中自动语音识别（ASR）系统的影响。它包括在受控声学条件下的成对噪声和去噪音频子集，以支持对去噪性能的系统分析。数据集基于500个英语音频样本，包含临床术语和药品名称。通过audiomentations库使用背景噪声、短噪声和高斯噪声三种噪声类型在不同强度级别下进行合成增强，并使用SpeechBrain的SpectralMaskEnhancement框架进行去噪处理。

The **[ekacare/denoising-impact-evaluation-dataset](https://huggingface.co/datasets/ekacare/denoising-impact-evaluation-dataset)** is a comprehensive benchmark dataset designed to evaluate the effects of speech enhancement on automatic speech recognition (ASR) systems in medical speech contexts. It includes paired noisy and denoised audio subsets under controlled acoustic conditions to support systematic analysis of denoising performance. The dataset is based on 500 English audio samples containing clinical terminology and pharmaceutical names. Synthetic augmentation was performed using the audiomentations library with three distinct noise types (background noise, short noise, Gaussian noise) at varying intensity levels, and denoising was applied using SpeechBrains SpectralMaskEnhancement framework.