paodigitalhub/pao-audio-dataset

🎙️ Pa'O Audio Dataset

ပအိုဝ်ႏ အငေါဝ်း အဆင်ႏဗာႏ ရွမ်ခြွဉ်းဗူႏ

📌 Project Summary

The Pa'O Audio Dataset is an open-source initiative created to facilitate the development of speech technologies and Artificial Intelligence tools for the Pa'O language (ပအိုဝ်ႏဘာႏသာႏငေါဝ်းငွါ).

Pa'O is primarily spoken in Shan State and other regions of Myanmar. As a low-resource language in the AI landscape, this dataset provides audio recordings and corresponding transcriptions for research and development in:

• Automatic Speech Recognition (ASR)
• Text-to-Speech (TTS)
• Speech processing
• Speech analytics
• Linguistic research
• Low-resource language technology

အဆင်ႏဗာႏခြွဉ်းဗူႏ ယိုနောဝ်ꩻ အင်းမိုးလွေꩻ တဲႏဆို့ꩻမာꩻလွူꩻနွောင်ꩻအီ အေအုဲင် နယ်ꩻပညာႏတွမ်ႏ အောဝ်ႏငေါဝ်းငွါ လုဲင်ႏဖေႏ လိတ်ယာႏ ယံလွို မော်ဒယ်ဖုံႏတာႏတဲင် ထွာ ပအိုဝ်ႏဘာႏသာႏငေါဝ်းငွါအတာႏ နေးတွမ်ႏ ထာꩻနယ်ꩻမဲန်ႏတနꩻယိုထွူ ရွမ်ခြွဉ်းဗူႏခါꩻဒါႏ အွဉ်ႏစွꩻတခြင် ထာꩻခွုမ်ပဲင်ႏဖုံႏ ဗွောင်လေလင်းဖေႏ အဆင်ႏဗာႏခြွဉ်းဗူႏဒျာႏသွူ။

---

💡 How to Use

The dataset can be loaded with the Hugging Face datasets library.

Because the metadata uses file_name as a relative path to the audio file, the column can be converted to an audio feature with Audio() after loading.

1. Install Dependencies

pip install -U datasets soundfile pandas huggingface_hub

### 2. Load Dataset via Python (`datasets` Library)
```python
from datasets import load_dataset

# Load Pa'O audio dataset
dataset = load_dataset("paodigitalhub/pao-audio-dataset")

# Inspect the first sample
sample = dataset["train"][0]
print("Filename:", sample["filname"])
print("Transcript:", sample["transcript"])
print("Speaker ID:", sample["speaker_id"])
print("Age:", sample["age"])
print("Gender:", sample["gender"])
print("Region:", sample["region"])
print("Device:", sample["recording_device"])
print("Duration:", sample["duration"])

3. Load Audio Files Directly in Google Colab / Python

import os
import pandas as pd
from huggingface_hub import snapshot_download
from IPython.display import Audio, display

# Download repo snapshot
repo_dir = snapshot_download(repo_id="paodigitalhub/pao-audio-dataset", repo_type="dataset")

# Read metadata CSV
df = pd.read_csv(os.path.join(repo_dir, "metadata.csv"))

# Play first audio
sample = df.iloc[0]
audio_path = os.path.join(repo_dir, "audio", sample["filename"])

print(f"Transcript: {sample['transcript']}")
display(Audio(audio_path))

---

📊 Data Fields (အဆင်ႏဗာႏ အခရာႏဖုံႏ)

The metadata.csv file contains the following fields:

Column Name	Data Type	Description	Example
filename	string	The audio filename located inside the audio/ directory	pao_audio_0001.wav
transcript	string	Pa'O language transcript in Pa'O script (လိတ်ပအိုဝ်ႏ)	ခွဲးဟောဝ်ဒျာႏ
speaker_id	string	Unique identifier for the speaker	S01
age	int64	Age of the speaker	21
gender	string	Gender of the speaker (male / female)	male
region	string	Geographic origin / dialect region of speaker	Kyaukkachar
recording_device	string	Hardware used for recording	Smartphone
duration	float64	Duration of the audio recording in seconds	1.28

---

🎯 Intended Use Cases (ကသွုံꩻနွောင်ꩻဒါႏ ထွောင်းဖဲ့ꩻဖုံႏ)

• Automatic Speech Recognition (ASR): Training models (e.g., Whisper, MMS) to convert spoken Pa'O into text.
• Text-to-Speech (TTS): Training speech synthesis models (e.g., VITS, Piper TTS) to pronounce Pa'O text naturally.
• Linguistic Research & Dialect Analysis: Acoustic analysis and computational processing for the Pa'O language.

---

📁 Repository Structure (တဲ့ꩻဆို့ꩻဖုဲင်လာႏ အရွီးခိုႏပွုံႏ)

To ensure full compatibility with Hugging Face datasets and standard ML workflows:

pao-audio-dataset/
├── audio/                  <-- Directory containing all .wav audio files
│   ├── pao_audio_0001.wav
│   ├── pao_audio_0002.wav
│   └── ...
├── metadata.csv            <-- Mapping file containing audio metadata & transcripts
└── README.md               <-- Dataset documentation

---

👨‍💻 Key Contributors & Acknowledgments (ကေားတဲမ်းရွမ်သားဖုံႏတွမ်ႏ ဒင်ႏလိတ်ကေႏဇူꩻ)

This audio dataset has been primarily collected, curated, and maintained by:

• KhunPhanDuae (ခွန်ဖန်ဒွဲ့)
• KhunRorNa (ခွန်ရောန)

Pa'O Digital Hub is a project being organized and carried out by SuccessImprove under RYPAK.

---

🔗 Related Datasets & Resources (သွုပ်ဆွိုင်ꩻ အဆင်ႏဗာႏခြွဉ်းဖုံႏတွမ်ႏ ထောင်ထွားတာႏ ခင်ႏလမ်းဖုံႏ)

• Related Pa'O Sentence Dataset: Pa'O Sentences Dataset (paodigitalhub/pao-sentences-dataset)
• Wikimedia Commons (Pa'O Pronunciation): Category:Pa'O pronunciation
• Pa'O Language Documentation & Resource Hub: KhunPhanDuae - Pa-O-Language Website

---

📜 License

This dataset is released under the Creative Commons Attribution 4.0 International (CC-BY-4.0) license.