kronk / README.md
russdill's picture
initial commit
9403433

Piper TTS voice model trained samples of Kronk from the Emporer's New Groove.

Script and subtitles were used to pull audio samples. Audio samples with excessive noise and cross-talk were dropped. Remaining samples were passed through MVSep DnR v3 to remove background noise a music. A second trimming step was performed to remove unsatisfactory samples or portions of samples as well as trim silence. A final step was performed to normalize volume levels of all samples. TextyMcSpeechy was used to train the model.