# machine-learning
- k
KubaPL2004Gmail
02/20/2023, 3:46 PMhttps://github.com/lugia19/speechToSpeechElevenLabs speachToSpeachElevenLabs by lugia19 on github - k
KubaPL2004Gmail
02/20/2023, 3:47 PMhttps://github.com/lugia19/elevenlabslib elevenlabslib by lugia19 on github - t
The Wiggles and High School D×D
02/20/2023, 7:54 PMI seem to get some decent results from the Tortoise notebook using Japanese voice work to create English speech - t
The Wiggles and High School D×D
02/20/2023, 7:58 PMAll getting better the more audio I collect - t
The Wiggles and High School D×D
02/20/2023, 7:59 PMHere's Australian English The voice sounds right but not the emotion - t
The Wiggles and High School D×D
02/20/2023, 8:43 PMResult of Japanese speech - t
The Wiggles and High School D×D
02/20/2023, 8:44 PMI don't know how to address the robotic nature - t
TheRoyalRuby2000
02/20/2023, 10:41 PMThat actually sounds like Greg, holy shit! - i
I love the hive
02/21/2023, 1:33 PMi try to finetune stable diffusion and this is a result: - t
The Wiggles and High School D×D
02/21/2023, 10:58 PMI got Greg, Murray, and Anthony saying it on intro tracks in Wiggly Safari - j
Justin
02/22/2023, 1:01 AM - t
The Wiggles and High School D×D
02/22/2023, 2:29 AMTortoise seems fairly random when it comes to generating a specific emotion from input text for a speaker - t
The Wiggles and High School D×D
02/22/2023, 2:29 AMI've been trying quite a few times to get a good sounding audio clip from a voice I added more support for and it sounds worse than before due to weird inflictions - p
PeaNutsAreGood
02/22/2023, 2:30 AMare you using vanilla tortoise? - p
PeaNutsAreGood
02/22/2023, 2:31 AMvanilla tortoise only uses the first 4 sec of the clip - t
The Wiggles and High School D×D
02/22/2023, 2:37 AMUsing the Colab notebook - t
The Wiggles and High School D×D
02/22/2023, 2:39 AMOh that does seem to happen occasionally towards the end of the clip - t
The Wiggles and High School D×D
02/22/2023, 2:40 AMI don't know if adding more voice clips to the folder remedies that at all - t
tanooki426
02/22/2023, 2:41 AMI tried adding an equal amount of Android 17 voice clips with the Travis voice clips to make an impression of Android 17 and it got rejected from Uberduck because "it sounds too similar." - t
The Wiggles and High School D×D
02/22/2023, 2:42 AMThat's fresh - t
The Wiggles and High School D×D
02/22/2023, 2:42 AMTortoise probably isn't the best thing to use for English speakers - t
The Wiggles and High School D×D
02/22/2023, 2:42 AMBut I'm using it for a Japanese speaker and wanted to throw an English model in there for the sake of it - t
The Wiggles and High School D×D
02/22/2023, 2:43 AMDoesn't sound so Aussie now - t
The Wiggles and High School D×D
02/22/2023, 2:43 AMShe sounds great though - t
The Wiggles and High School D×D
02/22/2023, 2:46 AMRias speaks first - t
The Wiggles and High School D×D
02/22/2023, 2:46 AMIn this exchange - i
I love the hive
02/24/2023, 9:08 AMweird result from tor Tortoise TTS - g
Gosmokeless28
02/24/2023, 9:47 AMIs this better than iSTFTNet? - j
Justin
02/24/2023, 9:48 AMhmm not sure