Voice Changer kinda sound like ai

[steady latch]

The voice changer works surprisingly well, but it’s still not perfect. People can usually tell it’s AI , especially when using female voices. One of my biggest issues is that English isn’t my first language; I mainly speak Arabic, and that messes things up. Some words come out weird, sound off, or only half the word gets pronounced.

Also, I haven’t used any extra programs alongside the voice changer just Discord for the best performance.

My specs: Intel i5-11320H (11th gen) and an RTX 3050 laptop GPU (yeah, kinda mid, I know).

[late sonnet]

Try the better W-Okada from this link instead. What you use is the original version of W-Okada. https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/

[steady latch]

Is there actually a significant difference between these two versions?

[ember apex]

only speed differences

[late sonnet]

That's an expected question. The difference between fork W-Okada and the original versions is mostly performance.

[steady latch]

I'll give it a shot

[steady latch]

Does it actually improve voice quality, or is it just about generating the voice faster?

[ember apex]

its just doing the conversion faster

[steady latch]

so no difference between them just speeding up the process

[ember apex]

correct

[proven vector]

@ember apex @steady latch well technically, wokada deiteris fork also helps kinda with stability/quality, with the advanced setting force fp32 mode on, which does fp32 inference compared to the default fp16

yes it doesn't fix laughing or realism (if a model is bad, is bad), but helps with stability

[ember apex]

it doesnt increase realism

[proven vector]

also, i would really not suggest that version, you deffo followed an old tutorial with also bad settings, it's more messy, and vb audio cable creates issues on windows such as crackling (as reported by Kakaor, ex staff) and randomly stop working

you should delete the folder/zip of both, and uninstall vb audio cable from the windows app settings

People can usually tell it’s AI , especially when using female voices
RVC has limitations to things like non natural speech (ex: laughing, screams),
are you trying to do e girl trolling/catfishing or are you trans / doing roleplay?

[steady latch]

Yeah I know it doesn’t handle non-natural speech like laughs or yelling, but honestly it was even struggling with regular talking a little. I’ll try the forked version though

Just to clarify things about your question,I’m not out here catfishing or doing anything weird lol. Just messing around with voice models and trolling my friends for fun. That’s literally it.

Might use it in ai videos or cosplays/roleplaying smths like that

[proven vector]

Alr that's great

[proven vector]

Try the wokada deiteris fork

on wokada deiteris fork, you can **optionally **use more advanced settings for benefits:

• Advanced Settings -> Force FP32 mode: on (THIS IS OFF BY DEFAULT! Turning this on improves stability. Increases VRAM usage by 200 MB)
• Advanced Settings -> Disable JIT compilation: off for faster loading speed of the program, on for slightly better performance (10-15 ms) for Nvidia only)
• Reduce the delay on Windows via: https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/#reduce-more-delay

And also, try playing with the pitch and different models, you can also show a screenshot of ur settings when u installed it

[steady latch]

not sure why but i keep getting this error message when choosing a mic

[proven vector]

what browser are u using? are u sure u gave microphone perms?

[steady latch]

firefox and yes i gave perms

[proven vector]

could you please try chrome?

[steady latch]

Tried chrome and edge and they give same error

[proven vector]

did you uninstall vb audio cable from windows app settings? could you please show the recording and playback tab after installing vac ltie?

[steady latch]

Sry for late reply

But yeah I'm sure I un-installed it

[proven vector]

mm

could you please restart your pc and retry the program?

[steady latch]

Same error

[proven vector]

could you please share a screenshot of ur wokada before you touch anything related to the microphone?

[proven vector]

gpu: rtx 3050 laptop
extra: 2.7
chunk: 100
f0: rmvpe without onnx
output: line 1
monitor: headphones
input: microphone

try it like this

[steady latch]

Same issue...

[proven vector]

could you please check your microphone settings on all the browsers? this is weird

[steady latch]

Ah dumb thing but it was my system sound setting was messed up cuz of Vac but fixed it

But it works now

[proven vector]

alr, use the settings i told u

on wokada deiteris fork, you can **optionally **use more advanced settings for benefits:

• Advanced Settings -> Force FP32 mode: on (THIS IS OFF BY DEFAULT! Turning this on improves stability. Increases VRAM usage by 200 MB)
• Advanced Settings -> Disable JIT compilation: off for faster loading speed of the program, on for slightly better performance (10-15 ms) for Nvidia only)
• Reduce the delay on Windows via: https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/#reduce-more-delay

[steady latch]

works perfectly , thank you so much for your help