The dataset is currently empty. Upload or create new data files. Then, you will be able to explore them in the Dataset Viewer.

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

Melody: Speech Enhancement Pipeline

📑 Paper    |    🌐 Project Page    |    💾 Released Resources    |    📦 Repo

This is the resource page of the our resources collection on Huggingface.

Dataset

Dataset Link
Melody-Speech-En 🤗
Please also check the raw data after our processing if you are interested: [toolevalxm/Melody-Speech-En-Raw](https://huggingface.co/datasets/toolevalxm/Melody-Speech-En-Raw).

Introduction

While having high-quality audio recordings theoretically allows us to generate reliable transcriptions as ground truth, two challenges arise: 1) Obtaining noise-free recordings is impractical in real-world scenarios; 2) Manual transcriptions are expensive and time-consuming. Thus, we adopt a fully model-based approach for synthesizing all the enhanced transcriptions using OpenAI Whisper-large-v3, as it has state-of-the-art performance and robust handling of diverse audio conditions.

*Due to our collaborators' compliance requirements, we only release the Speech-En subset (this page) of full dataset.

License

The license for this dataset is Apache-2.0.

Downloads last month
6