I downloaded w-okada and I want to know if there is any way to make my voice much more realistic, because when I download english rvc models and speak in russian, I can hear an english accent, the way they pronounce some letters and etc. The same situation happens with russian models too, although the voice was sampled from a russian-speaking person, the voice sounds with an english accent for some reason. I’m really curious if there is some new app that would allow me to sound realistic, or maybe w-okada is the best - but in that case, what should I do?
#How to sound more realistic, is this a app problem or just me x)
desert spire
<@&1159293204038955078> <:cat_seriously:1366692376575213599> with all love askin...
Make sure to read help guidelines before start asking, but when you provide a very long query, I'm not sure which one to summarize. By the way. What is your PC GPU? There are newer voice changer versions you'd like to try.
manic nacelle
my gpu nvidea geforce rtx 2060, sorry for not reading the guidelines
but my CP is Intel Core i7-10750H
manic nacelle
The Intel Core i7 is a "CP**U**".
ma bad bro, i will try what u sent me and gonna comeback with review prob tommorow
thx anyway!
quartz sigil
but my CP is Intel Core i7-10750H
Your what is Intel...
manic nacelle
Your what is Intel...
what can I say, shi happens
arctic ember
Your what is Intel...
he said the gpu is rtx 2060
arctic ember
I downloaded w-okada and I want to know if there is any way to make my voice muc...
the best pretrain so far to use by the model makers was trained from english speaking dataset
some russian one is rather old and has quality issue
manic nacelle
I was trying to find russian ones with good quality online but couldn't find one unfortunately
even the russian models here, being russian they sound like english
with that hard accent on letter R and shi
arctic ember
with that hard accent on letter R and shi
it kinda struggles on hard trills because the pretrain model was not quite trained on that
manic nacelle
it kinda struggles on hard trills because the pretrain model was not quite train...
that sad broski
manic nacelle
Try Tg Develop's W-Okada fork. https://docs.aihub.gg/realtime-voice-changer/loca...
yo broski, I downloaded the files from ypur link: voice-changer-windows-amd64-cuda.zip.001 and voice-changer-windows-amd64-cuda.zip.002 from the release.
When I try to extract the archive using 7-Zip by opening the .001 file, I get this error: ! C:\Users\Legion\Downloads\voice-changer-windows-amd64-cuda.zip.001: Unexpected end of archive
Both files are downloaded completely. I have tried re-downloading the files several times, but the error still appears.
Could you please help me understand what might be causing this issue?
chatgpt trying to help, but it either me hardly silly, or chatgpt trying to sabotage me
@desert spire, sorry for pinging you in advance, but i really could use some help
lapis bay
Don't use an index file. Find a suitable model, or better yet, create one yourself. Personally, I haven't had success with either 1 or 2 yet. But models that say R or sh are generally freely available. Whispers, laughter—impossible tasks. The only thing I know is that it's possible, but achieving it is a challenge.