This model aims to replicate the smooth, expressive tone of Michael Jackson’s voice during the Dangerous era (1991–1993), focusing on his softer and more emotional vocal style. It was trained using the RVC V2 framework with RVMPE and KLM HiFiGAN for high-quality timbre reproduction. The model is optimized over 500 epochs to preserve clarity, dynamics, and signature vocal nuances such as breathiness, vibrato, and phrasing typical of MJ’s Dangerous-era performances.

🎶 Training Dataset (Audio Sources):
The dataset was carefully curated and cleaned from studio recordings, live performances, and acapella tracks from the Dangerous album and era, including:

“Remember the Time”

“In the Closet”

“Who Is It”

“Heal the World”

“Will You Be There”

“Give In to Me”

“Keep the Faith”

“Gone Too Soon”
https://huggingface.co/JEFFERSONMUSIC/MJDangerousEraSoft40K/resolve/main/MJDangerousEraSoft40K.zip?download=true

Additional Details:

Tags: No tags available

Download Link: https://huggingface.co/JEFFERSONMUSIC/MJDangerousEraSoft40K/resolve/main/MJDangerousEraSoft40K.zip?download=true Hosted on Hugging Face