can someone help me set up edge TTS?

[terse latch]

what the title says

[tall pewter]

wdym set up? edge tts is only online to use

which of the online links are you using?

[terse latch]

im really ass at this kind of thing

[visual stratus]

Ayo? @terse latch level 2 !!!

[terse latch]

cuz i never did this kinda stuff before

Download the browser.

Open your Notepad & paste the following code:

the notepad as edge extension right

[tall pewter]

you are on pc right?

[terse latch]

yeah

[tall pewter]

you can do it the browser way (by installing microsoft edge on your pc)

or either use an online space

if you do it on your browser way, you will have to record with audacity the internal audio tho

on the online spaces you should be able to directly download the tts audios

you wanna do it the browser way?

[tall pewter]

no with notepad its meant the app on your windows pc

[terse latch]

which one is easier

[tall pewter]

imo the online spaces

but its your choice

[terse latch]

how come?

i plan to use it more than once anyway

so which one is easier to boot up

[tall pewter]

bc you can directly download the audios without having to use audacity

[tall pewter]

yea its all free, everytime u will have js to go to the space and use it

[terse latch]

ah thats useful

btw it works as

i give it a mp3

example of voice

and it like uses that as a baseline?

[tall pewter]

nope, thats a thing called 0 shot training, there are other tts that use that, for edge tts you can only use the premade multilingual high quality voices

[terse latch]

oh theres already existing voice models

and u can choose from that

?

[tall pewter]

yes exactly

[terse latch]

are there alot to choose from?

mainly looking for ones from dbz

[tall pewter]

yes, but those are general ai voices, that sounds human, not from characters

[terse latch]

ohh i see

damn thats sad ngl

[tall pewter]

either you can use another tts that has 0 shot training where you could inout a short audio voice of a dbz character

or either use one of the edge tts mixed with rvc, that means you can use the rvc ai voice models in #1175430844685484042 but with the tts

[terse latch]

which one is more realistic?

[tall pewter]

should be edge tts mixed to rvc

[terse latch]

is there a tutorial for it somewher

[tall pewter]

https://aihubdocs.github.io/en/rvc/cloud/ilaria-rvc/

ilaria rvc is mixed with edge tts

and you can use the rvc dbz moodels you find online on it

[terse latch]

it said to start the one below

wasnt that the right one?

[tall pewter]

click on runtime, disconnect runtime and delete, and re run everything again

[terse latch]

so i first install ilaria right

[tall pewter]

yea

it was prolly a random error

you didnt do nothing wrong dw

did you delete the runtime like i said ?

after you do that, just install ilaria and run it

[terse latch]

yeah i did

[tall pewter]

okay did you try again?

did it work?

[terse latch]

now i press start right

[visual stratus]

Ayo? @terse latch level 3 !!!

[terse latch]

the bottom 1

as gradio or cloudflare?

[tall pewter]

gradio

it doesnt really matter

[terse latch]

Same error

i think imma hit the bed

Do it tomorrow

[terse latch]

@tall pewter

Yo

U there?

[tall pewter]

did you retry using ilaria rvc colab now?

[terse latch]

yeah it gave same error

[tall pewter]

Oh rip

Sorry but i dont know at this point

You could try applio rvc colab

Which has edge tts too

Its in the docs

I talked abt this with other staff we will see

[terse latch]

@nick088

@tall pewter

And?

Anything?

[tall pewter]

you could just retry using the colab

the way you did it was correct

if you still get the error

try this instead

https://aihubdocs.github.io/en/rvc/cloud/applio-colab/

which has edge tts mixed too so dw

[terse latch]

yk what @tall pewter

ill try voice to ai

how does it work

and can i record audio on phone

i dont wanna be saying random lines in the living room lol

[tall pewter]

You wanna make an ai cover or make a voice model?

[terse latch]

i want the ai to talk

[tall pewter]

Okay

Its always the same colab

[terse latch]

ill use voice model from the

find models forum

[tall pewter]

Either you use ilaria rvc

Or applio colab

[terse latch]

which do u recommend

[tall pewter]

Its always the same guide

[terse latch]

or whats the difference

[tall pewter]

Well, the rvc fork and UI (interface)

Try ilaria rvc, if it still gives error just use applio colab

They are both good

[terse latch]

ok

[terse latch]

it worked

[tall pewter]

Ilaria rvc worked?

Good

[terse latch]

yeah

but the voice model

i put the name in?

[tall pewter]

It was prolly a random error lol

[tall pewter]

Well, you need to put the link to download the model

[terse latch]

oh i download the winrar

and in the extracted files

should be a exact name?

[tall pewter]

What ?

You dont have to extract your files locally

First, search any models

you can search:

• #1175430844685484042
• #🔍│find-models
• https://weights.gg/
• https://huggingface.co/models (but watch out cus in hugging face there arent only rvc ai voice models)
• https://applio.org/models

if there isnt one, #1159289738314919936 or make it yourself with our docs guides https://aihubdocs.github.io/en

After you found one, check from https://aihubdocs.github.io/en/rvc/cloud/ilaria-rvc/#inference-

[terse latch]

i mean this

Son Goku/Kakarot (Dragon Ball | ENG Dub) - (RVC V2) 400 Epochs

do i just put Son Goku

or the entire thing

[tall pewter]

Nope, you need to put the model download link

[terse latch]

i did that

[tall pewter]

Send the link where you got it

[terse latch]

but it wants name

[terse latch]

https://huggingface.co/AIMan2001/Goku/resolve/main/Goku.zip

[tall pewter]

Ohh you mean the name after the model link yeah, just put aby name like Goku, but remember to not use spaces

[terse latch]

so its like

SonGoku

uppercase lowercase important at all when i put?

is the name of the model just what it saves it as

or is it sum important

[tall pewter]

Nope, just out any nane you want to remember when its saved, whats important is the download link above the name

[terse latch]

okay

[terse latch]

the link is as that?

zipped?

kinda weird

just double checking

[tall pewter]

When you find a model on hugging face, be sure to right click the download button next to the .zip of the model, and copy link

Its normal its zipped

Just it should have an ?download=True or smt like that at the end

[terse latch]

i see

just for info

as i might need it

the TTS

cant i just the voice model aswell for that

on the right side?

[tall pewter]

Yea you can use the tts also with the voice model

Like on the tts thing you select a tts voice of the same gender and language that could be the most similar to the voice model you downloaded

Then input any text and click speak

And it will be used as an input audio

For you to then use the rvc model on it

[terse latch]

rvc?

[tall pewter]

Rvc is the ai of the voice models, its an speech to speech (ai voice model) conversion

But ilaria rvc and applio integrated edge tts in them

So it generates the tts output audio

And uses it as an input for the rvc to use the rvc model

[terse latch]

im confused

i put the audio in

[tall pewter]

Did you put a space when giving the model name

Never do that

[terse latch]

oh i see

so how do i remake that then

[tall pewter]

Put the same download link and put a name like "Goku"

Anything that is without spaces

[terse latch]

then i drag my audio in

[visual stratus]

Ayo? @terse latch level 4 !!!

[terse latch]

convert

done?

[tall pewter]

Yea

Or you could use the tts

If it gives you errors, refresh everything (buttons not page) and mnaully check the modelname .pth and the added.index

[terse latch]

btw

what type of male

would goku fit

cuz idk lol

[tall pewter]

I dunno tbh, you could try any english male model prolly

[terse latch]

so if i wanna reboot

i just install again or what

[tall pewter]

If you wanna do this again, you just do the same steps on the colab

But you dont have to now

I mean for like tmr or whenever you need to do ir again

Its just the same steps

[terse latch]

ok

thx

[tall pewter]

Yw

[tall pewter]

Should i mark this as solved?

[terse latch]

Yeah

[sand wigeon]

ENG Dub) - (RVC V2) 400 Epochs

[tall pewter]

There needs to be no spaces

[tall pewter]

Oke

[terse latch]

@tall pewter

yo bro

what is wrong here

i press convert

and it keeps giving me the text from before

i used to test pitch

so why does it refuse

[tall pewter]

Be sure to refresh button next to audio and change the audio to the audio you want it to say

[terse latch]

my pronounciation is to bad for the ai lol

as i thought

[terse latch]

can i make the voice

like have more energetic

energy

in the voice

[tall pewter]

you are using the tts right?

[terse latch]

Yeah

my pronunciation is bad for english

so can i like make it have more energy in the voice

Since its like

A energetic person

[tall pewter]

Well, sorry but edge tts isn't much emotional, you could try using ! But yea edge tts is like that

Bark tts has more energy but it isnt mixed with rvc and has a less quality, what you could do is use bark tta from the guide, and manually put the generated tts audio, saving it, and putting it as the audio input in ilaria rvc

[terse latch]

How come their samples are so good

But when I try he sounds constipated

Im obviously missing something

[tall pewter]

Wdym constipated?

Bad quality?

[terse latch]

lol i mean like it doesnt sound as real

Like as authentic

As their samples do

[tall pewter]

Like, not as clean as the tts audio?

[tall pewter]

what samples? On weights.gg?

[terse latch]

like theirs

Sounds more as the real character

Than mine does

Way more

[tall pewter]

The first tts audio is a typical ai voice

Could be also a problem of the model