edwixx/karaoke_songs_long
Karaoke Songs Long Dataset (karaoke_songs_long) A comprehensive collection of instrumental karaoke versions of popular songs in WAV format, spanning multiple genres and artists. Includes a captions.csv file with metadata for each track. Dataset Description This dataset contains hundreds of instrumental karaoke tracks (vocals removed/downmixed) in high-quality WAV format. The songs cover a wide range of artists — from Adele and Taylor Swift to Queen, Elvis Presley… See the full description on the dataset page: https://huggingface.co/datasets/edwixx/karaoke_songs_long.
Karaoke Songs Long Dataset (karaokesongslong)
A comprehensive collection of instrumental karaoke versions of popular songs in WAV format, spanning multiple genres and artists. Includes a captions.csv file with metadata for each track.
Dataset Description
This dataset contains hundreds of instrumental karaoke tracks (vocals removed/downmixed) in high-quality WAV format. The songs cover a wide range of artists — from Adele and Taylor Swift to Queen, Elvis Presley, and international artists — making it suitable for karaoke applications, singing voice analysis, and music information retrieval research.
Use Cases
- Karaoke Applications: Build karaoke apps with a pre-curated song library
- Singing Voice Separation Research: Use as reference instrumental tracks for source separation evaluation
- Music Information Retrieval: Analyze song structure, chord progressions, and genre classification
- Audio Augmentation: Add instrumental tracks as background for audio dataset augmentation
- Music Education: Provide backing tracks for vocal practice and performance
File Inventory
Audio Files
The dataset contains approximately 300+ instrumental karaoke WAV files in the root directory, named by artist and song title. Examples include:
Plus many more from artists including Billie Eilish, Maroon 5, Sam Smith, Rihanna, Lady Gaga, Bruno Mars, Coldplay, Sia, One Direction, Dua Lipa, Lana Del Rey, and numerous international artists (Tagalog, Indonesian, and other language karaoke tracks).
Metadata File
Usage
Loading the Dataset
from datasets import load_dataset
import pandas as pd
# Load dataset
dataset = load_dataset("edwixx/karaoke_songs_long", trust_remote_code=True)
# Or read captions directly
captions = pd.read_csv("captions.csv")
print(f"Total tracks: {len(captions)}")Working with Audio
import librosa
# Load a specific karaoke track
audio, sr = librosa.load("Adele_-_Hello_(Karaoke_Version).wav", sr=22050)
print(f"Duration: {len(audio)/sr:.2f} seconds")Searching for Songs
# List all Adele karaoke tracks
ls *Adele*.wav
# Search captions for a specific artist
grep -i "queen" captions.csvMetadata
- Author: Anurag Kanade
- Language: English (primarily), with some international tracks
- License: MIT
- Hosted by: Hugging Face Datasets
- Created: 2025-11-21
- Last Updated: 2026-06-25
- Storage Size: ~25.6 GB
- File Format: WAV (PCM, multi-channel instrumental)
- Downloads: 10
Citation
@misc{edwixx-karaoke_songs_long,
author = {Anurag Kanade},
title = {karaoke_songs_long},
year = {2026},
publisher = {Hugging Face},
journal = {Hugging Face Hub},
howpublished = {\url{https://huggingface.co/datasets/edwixx/karaoke_songs_long}}
}