thejorseman/CloneHeroCharts

---
{}
---
# Dataset Card for Clone Hero Charts

## Dataset Description

This dataset contains tokenized Clone Hero charts with audio conditioning features for automatic chart generation.

### Dataset Summary

- **Total examples**: 6407
- **Instrument distribution**:
- guitar: 4175 examples
- bass: 944 examples
- drums: 1288 examples

### Supported Tasks

- **Chart generation from audio**: Generate Clone Hero charts from audio input
- **Music transcription**: Transcribe audio to guitar/bass/drums notation
- **Beat detection and alignment**: Align audio features with beat timings

### Languages

Metadata is in English. Music is language-agnostic.

## Dataset Structure

### Data Instances

Each row represents one instrument track from one song. A song with guitar, bass, and drums generates three separate rows.

### Data Fields

Key fields include:
- `tokens`: Integer token sequence (vocabulary size: 187)
- `mert_embeddings`: MERT audio embeddings [num_beats, 768] (optional)
- `logmel_frames`: Log-mel spectrogram frames [num_beats, 32, 128] (optional)
- `instrument`: Instrument name ("guitar", "bass", "drums")
- `beat_times_s`: Beat timing in seconds
- `song_name`, `artist`: Song metadata

### Data Splits

- `train`: 6407 examples

## Dataset Creation

### Source Data

Clone Hero chart files (.chart, .mid) with optional audio stems.

### Personal and Sensitive Information

No personal information included.

## Considerations for Using the Data

### Social Impact of Dataset

Enables research in music generation, automatic charting, and audio-to-symbolic transcription.

### Discussion of Biases

Charts may reflect popularity biases in music selection. Instrument distribution may be skewed toward guitar.

## Additional Information

### Dataset Curators

Auto-generated by auto-charter toolkit.

### Licensing Information

MIT License

### Citation Information

If you use this dataset, please cite the auto-charter project:
```
@software{auto_charter,
title = {Auto-Charter: Clone Hero Dataset Builder},
author = {TheJorseman},
year = {2025},
url = {https://github.com/thejorseman/auto-charter}
}
```