kadirnar/japanese-voice-combined

Japanese Voice Dataset Combined

This dataset combines multiple high-quality Japanese voice datasets to create a comprehensive collection of Japanese speech data. It's designed to be used for speech recognition, text-to-speech, and other speech-related machine learning tasks.

Dataset Content

This dataset contains over 86,000 audio samples in Japanese from various high-quality sources. The data comes from the following sources:

Dataset	Estimated Samples
StoryTTS	5,000
genshin-voice	50,000
japanese-anime-speech-v2	1,000
nekopara-speech	10,000
starrail-voice	20,000
Total	86,000+

Data Sources

1. starril-voice: Japanese voice lines from Star Rail game with transcriptions
2. genshin-voice: Japanese voice lines from Genshin Impact game with transcriptions
3. japanese-anime-speech-v2: Collection of Japanese speech samples from anime
4. nekopara-speech: Voice lines from Nekopara visual novels
5. StoryTTS: Japanese TTS dataset containing various narrations

Data Format

Each example in the dataset contains:

• audio: Audio data in the format of a dictionary with the following fields:
  • path: Path to the audio file
  • array: The audio data as a numpy array (when loaded)
  • sampling_rate: The sampling rate of the audio
• text: The transcription of the audio in Japanese

Usage

from datasets import load_dataset

dataset = load_dataset("kadirnar/japanese-voice-combined")

# Example usage
sample = dataset["train"][0]  # Note: access the train split first
audio = sample["audio"]
text = sample["text"]

# Play audio
from IPython.display import Audio
Audio(audio["array"], rate=audio["sampling_rate"])

License

This dataset is a compilation of data from multiple sources, each with its own license. Please refer to the original datasets for their respective licenses.

Citation

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

@misc{japanese-voice-combined,
  author = {Kadir Nar},
  title = {Japanese Voice Dataset Combined},
  year = {2025},
  publisher = {HuggingFace},
  howpublished = {\url{https://huggingface.co/datasets/kadirnar/japanese-voice-combined}}
}

Acknowledgements

Thanks to the creators of the original datasets:

• simon3000 for starrail-voice and genshin-voice datasets
• joujiboi for japanese-anime-speech-v2
• grider-withourai for nekopara-speech
• kadirnar for StoryTTS