Susie
Deltarune · RETRAINED
https://huggingface.co/joebamnasoda69/CheckardsArrVeeSeeStuff/resolve/main/susieAIexpanded.zip?download=true
Audition Across 18 Pitch References
Click any semitone cell below to hear how this voice performs on female and male reference vocals and lock your preferred pitch.
Sonic Profile at Selected Pitch
Acoustic traits measured directly from the generated output sample. Switch pitches in the Sample Deck to see how timbre shifts.
Architecture: RVC v2 · Pitch Extractor: Mangio-crepe · Training: 275 Epochs · Source Host: Hugging Face
Voices That Sound Like Susie
Close but not quite the right fit? Audition nearby candidates matched by vocal embedding and timbre.
About Susie (Deltarune) (RETRAINED) (Mangio-crepe, RVC V2, 275 Epochs)
Model link: https://huggingface.co/joebamnasoda69/CheckardsArrVeeSeeStuff/resolve/main/susieAIexpanded.zip?download=true Dataset length: about 3 and a half minutes Trained by me on the RVC Disconnected colab (no credit needed) I unfortunately forgot whether or not I used a pretrain for this because I made this model a while ago lol Special thanks to Fenixwarrior1990 on Character AI for making the TTS I trained this on The inference audio for the last two samples were made using SynthV's Eleanor Forte AI Lite and Solaria Lite voicebanks respectively. NOTES: Ok, so the way I made this model is kinda different from the older ones I posted. Basically, in case you didn't know, Character AI has text-to-speech voice cloning now, making it easier to clone voices and stuff. Anywho, someone made a clone of Susie's voice using her little dialogue noise, and I liked the way it sounded, so I made this model out of it. It's a bit low quality, so I make a better dataset with Adobe Enhance later and retrain it. Also, if you want this thing to sound a bit better, a bit of EQ should do the trick. Recommended settings: Pitch: -12 for higher voices, 0 for lower ones Filter radius: 3 Search feature ratio: Could be anywhere between 0 and 0.65; lower numbers might mitigate some of the lo-fi-ness Volume envelope: 0 Protect voiceless consonants: 0.5 Pitch extraction algorithm: rvmpe