# machine-learning
- t
The Wiggles and High School D×D
07/14/2022, 10:33 PMMuch lower loss - t
The Wiggles and High School D×D
07/14/2022, 10:33 PMDefinitely messed up somewhere - t
The Wiggles and High School D×D
07/14/2022, 10:34 PMI preferred when Rocking on the Sea was sung by Greg Dorothy the Dinosaur and Hot Potato are also solid - t
TheRoyalRuby2000
07/14/2022, 10:35 PMDude, I even remixed Dorothy the Dinosaur in FL Studio! - t
The Wiggles and High School D×D
07/14/2022, 10:35 PMUsing what? - t
TheRoyalRuby2000
07/14/2022, 10:36 PMwhat are you talking about since you said using what? Are you talking about FL Studio, the music making software? - t
The Wiggles and High School D×D
07/14/2022, 10:36 PMDid you use the Uberduck voices or were you just rearranging the music itself? - t
TheRoyalRuby2000
07/14/2022, 10:37 PMI was rearranging the music itself. But I also threw in the original vocals. - t
TheRoyalRuby2000
07/14/2022, 10:38 PMWant me to put the remix that includes the vocals in #848652201504931892? - t
The Wiggles and High School D×D
07/14/2022, 10:41 PMSure, I'll check it out later - t
TheRoyalRuby2000
07/14/2022, 10:42 PMAlright. When you check out the remix AND the Uberduck Dorothy model later, Hit me up and share your thoughts. - t
The Wiggles and High School D×D
07/14/2022, 10:45 PMSure - t
The Wiggles and High School D×D
07/15/2022, 3:44 PMLiked the remix The Dorothy voice seems to work with most voice clips like decently, though I forgot which ones still sound off. How much data did you use, considering there's room for improvement? - a
a nickname?
07/15/2022, 4:20 PMHEY - a
a nickname?
07/15/2022, 4:21 PMI Thought chris thorndyke got removed for background noises - a
a nickname?
07/15/2022, 4:21 PMwho put it back up? - a
a nickname?
07/15/2022, 4:21 PMit has background noises - a
a nickname?
07/15/2022, 4:21 PMyour can hear it to - t
TheRoyalRuby2000
07/15/2022, 4:44 PMI used 45-46 seconds of audio and 5-6 wavs. I trained her using AITCH which is a part of Pipeline. - t
The Wiggles and High School D×D
07/15/2022, 5:04 PMI thought one was supposed to use much more That clip of garbled Greg audio I sent you was trained on about 10 minutes of audio, maybe 80 files - t
The Wiggles and High School D×D
07/15/2022, 11:13 PMSample audio from one of my 100 models - t
The Wiggles and High School D×D
07/15/2022, 11:14 PMSame model input but now with Japanese text - u
{K EY1} (Kei)
07/15/2022, 11:14 PM👁️ - h
hecko
07/15/2022, 11:34 PMthat is - h
hecko
07/15/2022, 11:34 PMa new funky failure mode - h
hecko
07/15/2022, 11:34 PMdid you disable arpabet in synthesis - s
Syclon (Creator of Quandale TTS)
07/15/2022, 11:50 PMQuick Question: How do you make voice synthesis better? As in not speaking without pauses, not having seizures, not sounding robotic, etc. - s
Syclon (Creator of Quandale TTS)
07/15/2022, 11:51 PMBecause I’m thinking of revamping TickleMyTip’s voice and adding Billy Butcher’s voice as well - h
hecko
07/15/2022, 11:54 PMpauses as in it ignores punctuation? - h
hecko
07/15/2022, 11:54 PMif so then make sure to have clips with periods and commas in the middle