Lux Japanese Speech Corpus: Character Voice Audio for Text-to-Speech

Lux Japanese Speech Corpus is a collection of Japanese text-to-speech audio recordings by the original character 'Lux'. The dataset contains WAV files at 96kHz/16bit, provided in both raw and noise-cleaned versions, with corresponding transcriptions in a metadata file. The dataset was authored by Lami and last updated on Hugging Face on February 24,我们发现了一个问题。

Use Cases

• Train text-to-speech models based on the high-fidelity 96kHz/16bit audio files.
• Fine-tune speech synthesis systems using the character-specific voice recordings.
• Develop speech denoising algorithms by comparing the raw and cleaned audio versions.
• Create Japanese speech recognition training data using the provided transcriptions.

Strengths

• Audio files are provided in a high-quality 96kHz/16bit WAV format.
• Includes both raw and noise-cleaned versions of the same speech data.
• Transcriptions for the audio are provided in a separate metadata file.

Limitations

• Row count and total dataset size are unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• The dataset features a single character voice, which may limit generalizability for multi-speaker tasks.

Provenance

Source

Lami on Hugging Face.

Collection Method

Recordings of an original character's voice for text-to-speech.

Time Range

null

Freshness

Last updated 2025-02-24 18:02:56; freshness should be verified.

Geography

null