how do i download RVC

[misty trout]

there are so many tutorials that are atleast 30 minutes long and im so confused if anyone knows a short and simple tutorial on youtube please LMK

[forest sequoia]

https://docs.ai-hub.wtf/rvc/local/mainline/

[misty trout]

thanks

[misty trout]

uh there isnt a path file

maybe there is but idk what it should be named

[forest sequoia]

You linked two different things

Path file audio is like the screenshot suggests, just the audio file of the one you want to make a cover of

voice model is something completely different, you put the voice model .pth file into the assets > weights folder and then you can select it on the webui

[misty trout]

so where can i find it

i have like 50 audio files of the character speaking

but im not sure if i need a specific one that i can download from there

[forest sequoia]

If making a cover, you can only make it one by one

So you select one of them, like idfk C:/Desktop/voices/firstaudio.mp3

convert, download, then continue with secondaudio.mp3

as for the voice model youre trying to download from the first screenshot, you download "model.pth" and "model.index"

rename those two files to whatever the models name is

put .pth file in the folder RVC1006 > assets > weights

and the .index file in the folder RVC1006 > logs

[misty trout]

like if its a character from a game like tf2 heavy or something do i just rename it tf2 heavy

[forest sequoia]

it doesnt matter what name you give it its just for you to know which one youre selecting

[misty trout]

i already did that but it said error

[forest sequoia]

Send a screenshot of the whole thing before you clicked on convert

And one of the console window

[misty trout]

i kinda dont trust sending the console window

[native nacelle]

@misty trout has reached level 2. GG!

[misty trout]

you sure?

like can i blur out the name

my name

[forest sequoia]

I just need the last lines where it gives an error

[forest sequoia]

Do you actually have a file called "audio_example.wav" located in that path or did you forget to select it

[misty trout]

no

i dont think i have it

where is it

[forest sequoia]

There is a major logic flaw you have here, let me explain

[misty trout]

i can tell, i suck with computers

i gotta smooth brain

[forest sequoia]

"Enter the path of the audio file to be processed" is where you put in an audio file that you want to convert. You said you have 50 files of someone speaking, and you want someone else to speak those lines, yes?

So in there you put in one of those 50 files

[native nacelle]

@forest sequoia has reached level 18. GG!

[misty trout]

yes

[forest sequoia]

And let me explain how to select one of those voices path

1 sec

So u go to your folder with your audios, pre select one you want like this

Then click on this

copy path

Then go back to the webui and CTRL + V (Paste) whatever it copied

[misty trout]

alright thanks

[forest sequoia]

It automatically puts " " at the start and end, usually it doesnt matter if theyre kept in or not, but remove them just in case

[misty trout]

still error

does it affect anything if there is a number

[forest sequoia]

blank spaces might be problematic

either rename to sam_50 or sam50

What GPU does your pc have btw?

[misty trout]

its a laptop but alright

[forest sequoia]

aight think 1050 is just barely ok for this

try with changed name, and screenshot me the webui again

[misty trout]

i did

but still

[forest sequoia]

I am not sure about dtype error, maybe someone with more knowledge can check

Until then, i can give you an online alternative which is fast and no need to run any files stuff

Ilaria RVC https://huggingface.co/spaces/TheStinger/Ilaria_RVC

Guide on how to use it:https://docs.google.com/document/d/1YbXcLFPaGjhOdG5NFkK3QrucCEpHZBwFUxkeMO8aB18/edit

in short
model loader tab: you upload model pth and model index

then on the inference tab, you click to upload audio, select model, convert

[misty trout]

is it like the same thing

[forest sequoia]

Yes

wouldnt link it if it wasnt xd

[misty trout]

ok

i thought RVC developers was just for the other one

like ithought the server was just dedicated to that rvc

thanks alot

[forest sequoia]

yes and no i guess

theres forks and branches that were built upon this rvc

we just help people find what they need

[misty trout]

doesnt matter as long as i can troll in tf2 as some anime girl

[forest sequoia]

Hold up

Are you looking for realtime usage?

[misty trout]

yeah

[forest sequoia]

...this is not realtime usage

[misty trout]

oh

[forest sequoia]

lol

rvc does not mean realtime voice changer btw

[misty trout]

is there any with realtime usage

[forest sequoia]

in case you thought that

well gtx 1050 is really not a strong gpu, you will have some sort of delay

could be like up to a second potentially

https://rentry.co/Forkvoicechangerguide#download-nvidia

[misty trout]

no i can run gta 5 with 60 fps

[forest sequoia]

This is realtime voice changer that will run decent with your gpu

[forest sequoia]

Doesnt matter, it depends highly on cuda cores which the 1050 doesnt have much of

[misty trout]

i literally have no idea why this 5 year old laptop with a 1050 can run gta 5 smooth as butter

this laptop is literally older than my little cousin

[forest sequoia]

gta 5 is well optimized

it doesnt take much to run it

developers did a good job

[forest sequoia]

anyway download nvidia

then download the virtual cable (this is what u need to use it ingame)

start the MMVCServerSIO.exe, it will install some files, it might take some time but this is one time only
if theres any errors just post in here ofc

then you can head over to the "uploading voice models" section and go from there

for 1050 i recommend:
f0 det: rmvpe or fcpe, try both. fcpe can reach lower delay
chunk: start with 250
extra: 1.0s

[misty trout]

i saw ''failed'' flash for a sec and then it closed itself

[native nacelle]

@misty trout has reached level 3. GG!

[misty trout]

its pretty late where i live so ima go sleep

[forest sequoia]

Fix is on this section:
https://rentry.co/Forkvoicechangerguide#press-enter-to-continue-failed-to-download-or-verify-files

[misty trout]

thanks a whole lot man

i’ll try again tomorrow

your patience is immeasurable and i love it

wish i could give a tip or something but am broke

[forest sequoia]

all ggg

[misty trout]

alright i did everything so how do i use it

[forest sequoia]

Did it load up

[misty trout]

yeah

[forest sequoia]

What was ur gpu again

[misty trout]

1050

nvidia geforce gtx 1050

[forest sequoia]

Select your gpu

F0 det: fcpe
Chunk: 250
Extra: 1.0s

Then select your model by clicking on it to load it up

Input mic
Output virtual cable
Monitor headphones if u wanna hear urself

Then look at the graph on the top left of the image

"Perf" number should be green
Let me know what number it says on perf

[misty trout]

i dont think i have a virtual cable

[forest sequoia]

Download link on the guide

[misty trout]

nvm i do

i do have it

[misty trout]

ok i think it finished its thing

now what

[forest sequoia]

Click staet

Start*

Then speak and check for the perf metric on the graph like i said

[forest sequoia]

As i said, click on the model in the bluebox to load it up haha

Where it says no image if that wasnt clear

[misty trout]

oh thanks

its kinda laggy an has a long pause between syllables

also it sounds like morty lmfao

[forest sequoia]

What number is the perf metric

The one in colors

[forest sequoia]

Owh wtf

1050 should be better than that

[misty trout]

it is

i think its a wifi problem

[forest sequoia]

You have your gpu drivers installed, and the gpu selected?

There is no internet involved here

[misty trout]

its on a website

not a program

[forest sequoia]

Its a local adress

It doesnt use internet

[misty trout]

oh ok

[forest sequoia]

Check for this

If it still doesnt go down, then i suppose try the og vanilla voice changer. It should run smoothly

sry if you have to keep bouncing back between downloads lol

https://rentry.co/VoiceChangerGuide

[misty trout]

goddamnit

[forest sequoia]

Get nvidia download. Run "start_https"

[misty trout]

i havent gotten rickrolled in over a year

[forest sequoia]

Scroll down ignore rickroll

I gtg for 20 mins

[forest sequoia]

Sorry, i meant run start_http
Without the s

[misty trout]

where

in the files?

this one in the MMVC file?

[forest sequoia]

Yes

[native nacelle]

@misty trout has reached level 4. GG!

[misty trout]

problem

[misty trout]

nvm

it work

[misty trout]

TYSM SHAD

IT WORKS

[forest sequoia]

nice

[misty trout]

thanks alot chad

[misty trout]

aw man its not working anymore

''out of memory''

[forest sequoia]

Use f0 det: rmvpe

But the out of memory error is not related to that actually, just means your GPU is really driving toward its limit 😅

[misty trout]

does it matter if there are many tabs open

like on opera gx

also still doesnt work

[forest sequoia]

gpu just trash thats the problem lol

voice changer ideally needs 6gb vram to work decently, yours has 2

[misty trout]

doesnt make sense that nearly any game i play runs at 60 fps with no lag spikes yet i try to run a simple voice changer its at its limit

[forest sequoia]

it does make sense, this is AI not a game

Im not that technical to be able to explain

[misty trout]

so what should i do

does this mean i cant use any voice changer and have to get a pc

[forest sequoia]

do you have good internet

[misty trout]

amazing most of the time but sometimes dogwater (but i just turn it on and off to make it good again)

[forest sequoia]

ehh well does it have 100 mbps down?

[misty trout]

well tbh idk how to check that but it takes about 25 minutes to download a 6 gb program so idk

[forest sequoia]

anyway you can try the online version

Google Colab:
https://colab.research.google.com/github/hinabl/voice-changer-colab/blob/master/Hina_Modified_Realtime_Voice_Changer_on_Colab.ipynb

The guide on setting it up:
https://docs.google.com/document/d/e/2PACX-1vTIceEcBfS6Zqolv_QEysrFfI_EJikPxozWptP_EjkpLVl-l-gdo-ijBonQMTviAHEYm5emmd9k9TdC/pub

Anything else, such as uploading voice models and settings, check out
https://rentry.co/VoiceChangerGuide

Skip the "[Optional] Upload a voice model (run this before the voice changer)" cell since it doesnt work, you can upload it later when the GUI opens

[misty trout]

is it compatible with discord and gams

[forest sequoia]

yes

[misty trout]

games*

ok

[forest sequoia]

limited to 4h daily thou

[misty trout]

yeah actually maybe ill just get a voice acting career

or maybe do this

yeah still doesnt work

ill go get a voice acting career

[forest sequoia]

Bruh no

No gpu in the world can reach that chunk 😂

[misty trout]

whats the best chunk for a shit gpu

[forest sequoia]

The higher the value, the more delay but the more time your gpu has to process

idk try 192 - 256

[misty trout]

neither works

yeah maybe ill just get a pc or something

[forest sequoia]

try f0 det fcpe

[misty trout]

wait what if i change the gpu to cpu

would that do anything

[forest sequoia]

your cpu probably wont run well on it

i cant stop you from using it

[misty trout]

intel i7?

do you think that could run

[forest sequoia]

full name

[forest sequoia]

Ehh

not the greatest

You can try xd

[misty trout]

i know but could it run it

[forest sequoia]

No clue

[misty trout]

alr ima try

[forest sequoia]

It would probably damage your cpu overtime

i dont recommend

[misty trout]

wish me luck i dont combust

its not like im running the voicer changer 24/7

i didnt even get to try the cpu

[forest sequoia]

it says youre already running the voice changer basically

[misty trout]

how#

[forest sequoia]

idk didnt properly close it?

check if console is closed lol

[misty trout]

oh nvm found it

its cuz i play modded ultrakill which also needs a console so i thought the open console thing was just my mods

lmfao the breathing sounds

[native nacelle]

@misty trout has reached level 5. GG!

[misty trout]

out of nowhere

im not even breathing into my mic

also i dont think the gpu is the problem its cause i wasnt doing fcpe

i was doing crepe

nvm cpu and fcpe are the only ones that work

even then really laggy

i'll just get a pc

or something

wait wtf its doing kinda better at 704

i dont care at all about delay as long as it doesnt seperate words into two

holy shit it kinda works just a long delay

okay 448 works best

but major problem

im sick

so my voice sounds weird and so does the voice changer

i dont blame the software for that

alsoi did some research

if my cpu isnt over 100 c it should last like decades

and since i may or may not get a pc soon

i dont see any problem with me using cpu

[misty trout]

so far amazing but some pauses and lag spikes

wait last thing (i hope to dear god its the last thing) @forest sequoia

how do i make the voice louder

my friend says the voice is very quiet

even when im practically screaming

[forest sequoia]

increase the volume : out