Mazinger Dubber Voice Profiles

Voice profiles for the Mazinger Dubber tool contain paired audio recordings and plain-text transcripts. The dataset, created by bakrianoo, was last updated in March 2026. Profiles are contributed by users following a specific file structure.

Use Cases

• Train a voice cloning model using the `voice.m4a` audio files and corresponding `script.txt` transcripts.
• Fine-tune a text-to-speech system on the `script.txt` and `voice.m4a` pairs from multiple speaker profiles.
• Evaluate speech synthesis quality by comparing generated audio against the provided `voice.m4a` samples.
• Build a multi-speaker TTS dataset by aggregating the `my-name/` profile folders containing audio and text files.

Strengths

• Audio files are paired with exact plain-text transcripts (`script.txt`).
• Profiles follow a standardized folder and file naming convention for consistency.
• Dataset was updated in March 2026, indicating recent maintenance.

Limitations

• The total number of profiles, audio duration, and row count are unknown.
• Audio quality and recording conditions are user-contributed and may vary.
• Geographic and linguistic diversity of speakers is unspecified.

Provenance

Source

HuggingFace user bakrianoo.

Collection Method

User-contributed voice profiles following a specified file preparation guide.

Freshness

Last updated 2026-03-25.