UPDATE: Apologies, I had to cut an example from my video for a copyright claim even though the face and voice of the actor was changed....So there will be a weird gap after 11 seconds and a missing example. In my last video 🎥, I showed you how to Deepfake videos 📹 and images 🖼, and in this video 📺, we will look at some free voice cloning 🗣 that is so good it can make your clone sing! 🎶🎤
at 26:40 in the video you can just click "false" for the option in step 1 for speech. I also go over the difference between true and false at that point in the video as well and show a sample of what regular text to speech sounds like.
I hope we can soon get control over the whole text like a DAW directly. Would be great to create narrated content such as youtube dub or audiobook intonation with "audio controlnet" guidance.
I was loling with some of these results, I did whitney houston, backstreet boys, lil wayne the video was too long already to include any of that. Also, I tried to pronounce that word a couple of times I give up. Thanks though.
7zip is more reliable than Wnrar when it comes to compression and decompression. Sometimes Winrar approximates the files, which makes it appear like it worked fine, but then the signature hash of the file is corrupted and different checksum than the original.
I just remember WinRar opening anything and being the most popular tool for torrents, so I tired that first. Did 7z work for you? I assume it did, since it was compressed with 7z.
Trained a few models successfully, 10mins of good clear audio, took out any bass "rumble" etc and 250 Epochs later (for each one), they all sound like the person they're supposed to, but there is an annoying "gravel" sound not just in between speech but throughout. I've downloaded some PTH files and some are the same as mine (noisy) but some are crystal clear. Any idea what I can do to make it higher quality?
thx great tutorial! I´m having trouble finding the version you use, I´m on Intel but none of the versions on huggingface that I dl:ed match that. What´s the date on yours?
I keep forgetting I have copilot but I tried chat GPT and it lead me to some dead end forums with people that asked the question but there was no solution. So I just figured I'd try an WinRar or 7z. I didn't notice it was compressed with 7z till after I installed it though. WinRar works though.
@@ArtificallyIntelligentAi funny thing is I use copilot/bing chat to find out things about Linux, like how to code something etc. chatgpt 4 which is what it is based on is excellent at telling the exact thing to do. Set the conversation tone to the right most button called most precise I think as that option uses the latest gpt model. I even installed Edge on Linux mint (which is a sacrilege for Linux users) because of the in built gpt chat.
Okay, so maybe I missed something but I followed your video and separated the instrumental from the voice and successfully changed the voice. How do I re-combine for instrumental with the new voice to make a complete song?
UPDATE: Apologies, I had to cut an example from my video for a copyright claim even though the face and voice of the actor was changed....So there will be a weird gap after 11 seconds and a missing example.
In my last video 🎥, I showed you how to Deepfake videos 📹 and images 🖼, and in this video 📺, we will look at some free voice cloning 🗣 that is so good it can make your clone sing! 🎶🎤
Epic, thanks for sharing.
Oh man, this is promising. Would love to do something with cow and chicken (with mr Red)
How can I use the voice clones I made with this and use them for text to speech?
at 26:40 in the video you can just click "false" for the option in step 1 for speech. I also go over the difference between true and false at that point in the video as well and show a sample of what regular text to speech sounds like.
I hope we can soon get control over the whole text like a DAW directly. Would be great to create narrated content such as youtube dub or audiobook intonation with "audio controlnet" guidance.
Ha ha ha ha that Jasmine voice was hilarious 😅. Accompaniment is pronounced A COMPANY MINT
I was loling with some of these results, I did whitney houston, backstreet boys, lil wayne the video was too long already to include any of that. Also, I tried to pronounce that word a couple of times I give up. Thanks though.
Excellent tutorial. Thank you
7zip is more reliable than Wnrar when it comes to compression and decompression. Sometimes Winrar approximates the files, which makes it appear like it worked fine, but then the signature hash of the file is corrupted and different checksum than the original.
If that is true then that's a major failure of a zip unzipper. 7zip ftw
I just remember WinRar opening anything and being the most popular tool for torrents, so I tired that first. Did 7z work for you? I assume it did, since it was compressed with 7z.
@@ArtificallyIntelligentAi Yeah I only use 7zip nowadays, coming from winrar until it was a problem which is what made me switch
Trained a few models successfully, 10mins of good clear audio, took out any bass "rumble" etc and 250 Epochs later (for each one), they all sound like the person they're supposed to, but there is an annoying "gravel" sound not just in between speech but throughout.
I've downloaded some PTH files and some are the same as mine (noisy) but some are crystal clear. Any idea what I can do to make it higher quality?
thx great tutorial! I´m having trouble finding the version you use,
I´m on Intel but none of the versions on huggingface that I dl:ed match that.
What´s the date on yours?
You could have used windows copilot to query how to handle sh files in windows. I find the gpt is very useful for file info
I keep forgetting I have copilot but I tried chat GPT and it lead me to some dead end forums with people that asked the question but there was no solution. So I just figured I'd try an WinRar or 7z. I didn't notice it was compressed with 7z till after I installed it though. WinRar works though.
@@ArtificallyIntelligentAi funny thing is I use copilot/bing chat to find out things about Linux, like how to code something etc. chatgpt 4 which is what it is based on is excellent at telling the exact thing to do. Set the conversation tone to the right most button called most precise I think as that option uses the latest gpt model. I even installed Edge on Linux mint (which is a sacrilege for Linux users) because of the in built gpt chat.
Okay, so maybe I missed something but I followed your video and separated the instrumental from the voice and successfully changed the voice. How do I re-combine for instrumental with the new voice to make a complete song?
Hi you can use a free editing tool like clip champ that can get from the windows store or DaVinci is another option
I am not able to install it, please :((( [ Windows 8.1 , AMDIntel ]
It takes a lot longer on my system- likely cause I have an old system
I think result not good. What about mangio rvc..,
Can i test it in colab