README.md

---
pretty_name: Kazoo (Preview)
license: cc-by-nc-4.0
size_categories: [n<1K]
tags:
  - audio
  - music
  - dataset
  - kazoo
  - membranophone
  - buzzing_instrument
  - novelty_instrument
  - folk_instruments
  - acoustic_instrument
  - sound_design
  - ai_training_data
  - sampling_96khz
  - bitdepth_24bit
  - mono
task_categories: [other]
language: [und]
thumbnail: thumbnail.png
description: "A high-fidelity acoustic dataset capturing the Kazoo, recorded for AI training, sound research, and creative development as part of the Harmonic Frontier Audio Novelty Gems Cabinet."
---

![Dataset Thumbnail](./thumbnail.png)

# Harmonic Frontier Audio – Kazoo (Preview, v0.9)

**A high-fidelity acoustic dataset designed for AI training, music research, and creative sound development.**

**Kazoo (Preview)**, created by **Harmonic Frontier Audio**, provides a compact reference set demonstrating the quality, formatting, and metadata conventions used in the *Novelty Gems Cabinet* of the Harmonic Frontier Audio catalogue.

---

### 🔎 Summary

This dataset provides high-quality, rights-cleared recordings of the **Kazoo** — a buzzing membranophone whose timbre is shaped by vocal input passing through a vibrating membrane.  
The recordings highlight the instrument’s distinct nasal resonance, rapid transient response, and expressive harmonic coloration, making it valuable for **AI music modeling**, **timbre classification**, **acoustic analysis**, and **cross-cultural sound synthesis** research.

Developed by **Harmonic Frontier Audio**, this preview follows *The Proteus Standard™* for dataset provenance, transparency, and ethical AI use.  
Learn more about the Proteus Standard → https://harmonicfrontieraudio.com/proteus-standard

Full dataset details and licensing information are available at: https://harmonicfrontieraudio.com/datasets/kazoo

It offers clean, consistent audio files that capture the Kazoo’s playful yet harmonically rich character across sustained tones, triads, glissandi, and articulation patterns.

If you find this dataset useful, please consider giving it a 🤍 on Hugging Face to help others discover it.


---

## 🎶 About the Kazoo

The **Kazoo** is a buzzing aerophone / membranophone most commonly used in folk, novelty, and participatory musical settings.  
Its sound is produced when the performer vocalizes into the instrument, causing a membrane to vibrate and impart a distinctive **buzzing, nasally-modulated character**.

Though simple in construction, the Kazoo is capable of surprising expressive range — from smooth glissandi to highly articulated tongued gestures.  
This preview dataset was designed to capture the instrument’s **core articulations neutrally**, without stylistic coloration, offering a universal training source for AI and research applications.

---

## 📂 Contents

### **Audio Files (.wav)**  
- Recorded at **96 kHz / 24-bit** WAV format  
- Exported as **mono**  
- Fade-ins and fade-outs of **3–5 ms** for transient consistency  
- DC offset checked — **none detected**  
- High-pass filtered at ~40 Hz to reduce subsonic rumble  
- No EQ, compression, saturation, or post-processing applied

### **Categories in this Preview**
1. **Sustain**  
   - Single sustained tones across multiple pitches  
   - Includes vibrato and non-vibrato variations  

2. **Glissando**  
   - Smooth pitch slides (e.g., E→E, G→G) demonstrating membrane continuity  

3. **Tonguing / Articulation**  
   - Short, percussive bursts with distinct transient character  

4. **Triads (Staccato)**  
   - Rapid three-note sequences created through vocalized changes  
   - Includes C–E–G, F–A–C, G–B–D patterns

---

## **Metadata (.csv)**  
Includes structured fields for file name, category, articulation, note pitch, vibrato status, microphone used, channel configuration, sample rate, bit depth, recording chain, and dataset version.

Metadata was hand-verified for consistency and accuracy according to Harmonic Frontier Audio standards.

---

## 🎤 Recording Notes

- Recorded in a treated studio environment with:  
  - **Microphone:** RØDE NT1-A (large-diaphragm condenser)  
  - Positioned 8–10″ from the instrument for balanced capture  
- Recording chain:  
  **RØDE NT1-A → Zoom F8n Pro (32-bit float capture)**  
  Exported as **96 kHz / 24-bit mono WAV** for release  
- No noise reduction or dynamic processing used — transient integrity preserved  
- Natural dynamics retained for realistic ML training behavior

---

## 🌈 Spectrogram Preview

Below is a spectrogram showing the Kazoo’s characteristic harmonic energy and buzzing membrane structure:

![Spectrogram Preview](Spectrogram_Preview.png)

---

## 🎧 Demonstration Audio

This repository includes **three audio examples** for listening and comparison:

**1. Dataset in musical context (mixed & mastered)**  
A brief musical excerpt demonstrating the Kazoo Preview dataset integrated into a musical arrangement, highlighting expressive timbre and real-world usability.

**Track:** *“Kazoo Example”*  
**Producer:** Blake Pullen  
**Source Audio:** Harmonic Frontier Audio – Kazoo (Preview)

<audio controls style="width:100%;">
  <source src="https://huggingface.co/datasets/Harmonic-Frontier-Audio/Kazoo_Preview/resolve/main/Demos/Kazoo_Demo_Master.mp3" type="audio/mpeg">
  Your browser does not support the audio element.
</audio>

**2. Raw dataset recording**  
An unprocessed audio excerpt taken directly from the dataset’s audio files, representing the actual licensed source material without musical context or processing.

<audio controls style="width:100%;">
  <source src="https://huggingface.co/datasets/Harmonic-Frontier-Audio/Kazoo_Preview/resolve/main/Audio/HFA_Kazoo_Melody_ShortSnippet.wav" type="audio/wav">
  Your browser does not support the audio element.
</audio>

**3. AI-generated reference (timbral substitution)**  
A fully AI-generated reference created without using any Harmonic Frontier Audio recordings.  
While the model reconstructs the melodic content of the prompt, it replaces the kazoo with a toy xylophone–like timbre, reflecting a lack of kazoo representation in typical training data. This example is included to illustrate current limitations in timbral modeling.

<audio controls style="width:100%;">
  <source src="https://huggingface.co/datasets/Harmonic-Frontier-Audio/Kazoo_Preview/resolve/main/Demos/Kazoo_AI_Generated_Reference_no_dataset_audio.mp3" type="audio/mpeg">
  Your browser does not support the audio element.
</audio>

---

## 📜 Demo & Licensing Note

The `Demos/` folder contains **demonstration material only**.  
**Only the raw recordings in the dataset audio files constitute the licensed dataset.**  
AI-generated reference tracks are **not part of the dataset** and are **not licensed for training or reuse**.

---

## 🪶 Production Notes

- Instrumental accompaniment in the contextual demo was generated using **Suno (Pro plan)** under a commercial license.  
- All dataset recordings were **performed, recorded, and produced by Blake Pullen** for Harmonic Frontier Audio.

---

## ⚡ Usage

This preview dataset is designed for:

- **Evaluation** of Harmonic Frontier Audio dataset structure and fidelity  
- **Testing** AI and DSP systems that model membranophones and vocal-modulated instruments  
- **Creative sound design**, **instrument synthesis**, and **timbre modeling**  
- **Classification and recognition research**

👉 Note: This is **not a full dataset**.

The complete Harmonic Frontier Audio **Kazoo** dataset will include:

- Full pitch range coverage  
- Extended articulations and dynamic layers  
- Additional gesture types (trills, flutter-vocalizations, rhythmic motifs)  
- Velocities and spectral variations  

---

## 💡 Full Dataset Availability

This is a **preview pack** from the *Novelty Gems Cabinet* family.  
The complete dataset — with expanded articulations — will be available for **licensing**.

For licensing inquiries:  
📩 info@harmonicfrontieraudio.com

---

## 📥 How to Use This Dataset in Python

```python
from datasets import load_dataset

dataset = load_dataset(
    "Harmonic-Frontier-Audio/Kazoo_Preview",
    split="train"
)

print(dataset)
```

> ⚙️ *Note: Parquet conversion and `load_dataset()` support will be available within 2–3 days of publication.*

---

### 🔗 Explore More from Harmonic Frontier Audio

- [Scottish Smallpipes (Preview)](https://huggingface.co/datasets/Harmonic-Frontier-Audio/Bagpipes_Scottish_Smallpipes_in_A_Preview)
- [Highland Bagpipes (Preview)](https://huggingface.co/datasets/Harmonic-Frontier-Audio/Bagpipes_Scottish_Highland_Bagpipes_Preview)
- [Irish Tin Whistle in D (Preview)](https://huggingface.co/datasets/Harmonic-Frontier-Audio/Irish_Tin_Whistle_in_D_Preview)
- [Subharmonic Phonation / Vocal Fry (Preview)](https://huggingface.co/datasets/Harmonic-Frontier-Audio/Subharmonic_Phonation_Vocal_Fry_Preview)
- [Kalimba (Preview)](https://huggingface.co/datasets/Harmonic-Frontier-Audio/Kalimba_Preview)
- [Kazoo (Preview)](https://huggingface.co/datasets/Harmonic-Frontier-Audio/Kazoo_Preview)
- [Overtone Singing (Preview)](https://huggingface.co/datasets/Harmonic-Frontier-Audio/Overtone_Singing_Preview)

*(All datasets follow The Proteus Standard™ for ethical dataset provenance and licensing.)*

---

## 📜 License

Released under **CC BY-NC 4.0**  
- Free for non-commercial use, testing, and research  
- Commercial licensing available via Harmonic Frontier Audio  
- Includes a formal rights declaration for clarity

---

## 📧 Contact

Harmonic Frontier Audio  
📩 info@harmonicfrontieraudio.com  
🌐 https://harmonicfrontieraudio.com/

---

## 🔮 Future Roadmap

This preview release is part of the **Harmonic Frontier Audio – Novelty Gems Cabinet**, within the broader Folk & World Instruments Catalogue.

Upcoming planned datasets include:

- **Kazoo (Full Dataset)**  
- **Toy Piano**  
- **Ocarina**
- **Juice Harp** 

Harmonic Frontier Audio continues expanding the world’s first unified library of ethical, rights-cleared world and vocal sound datasets for AI training, synthesis, and advanced sound design.

---

## 🗒️ Release Notes

**Version 0.9 (Nov. 2025)** – Initial Preview Pack release for Kazoo.  
See `CHANGELOG.md` for version history.

---

## Citation

If you use this dataset in your research, please cite:

Pullen, B. (2025). *Kazoo Dataset (Preview)* [Data set]. Harmonic Frontier Audio. Zenodo.  
https://doi.org/10.5281/zenodo.17717348

ORCID: https://orcid.org/0009-0003-4527-0178

### BibTeX

```bibtex
@dataset{pullen_2025_kazoo_preview,
  author       = {Blake Pullen},
  title        = {Kazoo Dataset (Preview)},
  year         = {2025},
  publisher    = {Harmonic Frontier Audio},
  version      = {0.9},
  doi          = {10.5281/zenodo.17717348},
  url          = {https://doi.org/10.5281/zenodo.17717348}
}
```