How do you guys get those covers in such a good quality? If I do it I get something fun but the quality is terrible. I use audio of 5 minutes to train my voice.

I was even using 10 minutes audio and the voice was fucking horrible, even I was setting up the pitch that it is okay with my normal voice, and training 300 times.

Cause we do. I honestly dk.

I say it always, I can sing really good and I am a english god. okay jk

My next question is not about kits.ai or any other service which can do AI covers. However, since everybody in this topic are interested in that AI voice processing stuff and my question is very near to that, i thought some of you might help me with my problem. Does any of you know about any service, app, vst plugin or anything which can analyze speech audio recording with more than one speaker and split it into separate tracks for each speaking person? I found some services but they try to describe speech into text and then label the speakers in the text. I didn't test these out, so cannot express any opinions on these. In my case it is better to keep original audio but just have separate file for each speaker. In order to keep this topic cleaner, you may answer me in private message. Many thanks in advance.

OK so update. Kits is getting super shitty. Apparently you can only do 3 minutes of audio a month of youtube conversions, and audio files from what I'm being told

the fuck

might as well just use voicemy.ai at this point instead

oh my gosh how dumb!
-- (Victor):
OK so update. Kits is getting super shitty. Apparently you can only do 3 minutes of audio a month of youtube conversions, and audio files from what I'm being told

--

Wait wtf? That's so dumb.

I'm filling out their feedback form and this is what I'm gonna submit. This is bullshit. It used to be a good tool up until the premium was released. Why is it bullshit? It says, quote on quote, free voice generators on kits.ai and now everything is being put under a paywall? what? ? Just doesn't make sense.

If anyone wants to submit feedback, here's the link to the form.
https://docs.google.com/forms/d/e/1FAIpQLSddVzYmbIR6JZfliT4VehTlBx_hpgshOcxju2Ixpmxvz8cWYg/viewform?usp=sf_link

Alright, I'm actually testing kits and it looks alright? I think? This may depend on whitch country you live in now... Idk

I don'ts see the bit about the 3 youtube conversions a day anymore

Yeah me neither but they still come out in mono

Hi!
Here's what exactly i could find on subscriptions faq on kits.ai:
"When you convert audio with an AI model on Kits, your usage is tracked in minutes. Free accounts can convert 15 minutes of audio (including audio files and YouTube videos) each month. Once you reach your limit, you can upgrade to a paid plan to increase your limit".
In order to avoid speculations and wrong info, go to https://app.kits.ai/subscriptions and expand the button called "What are conversion minutes?" to make sure that the result is consistent for your country etc. I don't think that there are different conditions for different countries but who knows.

bad

OK, so esentially that proved it's getting worse

Of course it does. But did anyone expect that it will be as good as it was? Calculation resources are too expensive for lasting this party forever. They collected enough data from us to improve their models and now the doors are closing. Bad but inevitable and predictable.