Thanks for your comment - I hear you and share a similar issue from a past job. I figure I better learn the tools and adapt - these will be standard in a short time. Do you think they will replace singers or offer a way for people who can't sing well to create what they otherwise couldn't? My thought is that violin is probably the most similar to the human voice and you can still tell the difference between a phenomenally good orchestral VST and a real violin.

Hi - I tried it with an UDIO sharing link. The direct link method didn't work but could to download the udio file and upload to controlla. Thank you so much for asking!

That would be a cool idea. I can show harmony creation in Controlla and Kits, but I don't know how to generate it in Suno. If I find a way, I'm happy to demo it. Thanks for watching!

kits ai

Yeah - I just checked the harmony generator out this morning - pretty cool stuff!

Thanks for pointing it out now - the workflow video was recorded 2 weeks ago with a free account. I just tried it after seeing your note and it is now prompting to pay to download. That is new. I adjusted the subtitles to reflect that. Thank you for watching and thank you for pointing that issue out.

Barry thank you for pointing that error out! I just trimmed it off in the KZbin editor, so it should show correctly very soon. The In and Out points in the editor must have been incorrect and I did not catch that. Thanks for watching and commenting!

@@CreativelyMakeAIMusic lol, i thought u sneaked it in 8min + vids gets u more ad revenue ;)

lol :) That would be sneaky LOL - nope, just my learning curve!!! I fixed it with the editor and it should be about 3:20 any time now. Thank you so much for watching!

@@CreativelyMakeAIMusic, No problem! 🕊️🫶

After trying several workflows on suno I found it was faster and more reliable to take a track and regenerate the vocal and instrumental by putting the link into controlla than to keep remastering it in suno. Thank you for watching and for asking this!

@@CreativelyMakeAIMusic Yeah. i tested it but it doesnt remove the shimmer.

It is now fixed in KZbin editor and should appear correctly by 1/23. That was an editing mistake where the scrap footage was output by accident. I pinned a note to the top of the comments to let people know after I saw yours and Barry's notes. Sorry about that! Thank you for letting me know and thanks for watching.

Great question - Yes, you can do that but you will need a paid account with Controlla to create your voice clone.

@@CreativelyMakeAIMusic Can you do a video on this? Tq

Our pleasure - thank you for watching!

I really appreciate that you took the time to come by; thank you so much!

I don't see why not; I will talk about that in an upcoming video. Thank you for asking and for watching!

@@CreativelyMakeAIMusic Thank you

I am learning more about the difference of how AI separates stems vs what we humans think of as well. It will be interesting to see how it develops. Thank you for watching!

Thank you Joe! I'm so glad you could watch the video.

Thank you!

Thank you so much for your kind words. I'm glad you enjoyed the video!

Thank you for watching!

Hi Kerry - I'll check those out! Are you talking about using those on the ai output or on my narration for the video (asking honestly)... Let me know! Thank you so much for the suggestion!

This is to say you generally apply the noise reduction to your dry vocal stems. Rnnoise will work with Reaper/Audacity/Goldwave/Insert digital audio workstation software here that can accept a VST. OBS-Studio also has it built in. You sometimes need to tune it a little but it does a really good job without destroying too much of the source audio. Bad things happen if you use it for things that are not voice. SuperTone Clear is more flexible and can remove echos and other things however it is not cheap. Cheaper than Isotope RX etc.

Aha - yes, reducing background noise and echo make the conversions so much better. And I understand about costs; that is why I like to show free or tools that are an excellent value. Melda, Kilohertz, and Bluecat all make great free tools for this purpose as well; sometimes a noise gate may be a better choice. Thanks for the info on Rnnoise!

Glad it was helpful! I love sharing this stuff and am thrilled to read your comment. Thank you!

Great question! This is a HUGE concern. Controlla, ACE, SoundID and others etc all train their own voice models that they contract with, so after reading their terms of service I feel okay using those. I would suggest that anybody do that before deciding which program to use. Thanks for watching and commenting - I hope you enjoy the channel!

@CreativelyMakeAIMusic Thanks for the reply and insight. This is reminiscent of the DNA testing that was popular for a while, although that is even more concerning. As you indicated, it is an individual decision.

I would recommend using ACE vocal studio. You can type in the words and enter your melody using midi right on their screen - it's really cool. There is a free trial if you want to check it out - you don't have to sing at all. I have an overview at kzbin.info/www/bejne/gWKlXoSChpV_fKssi=uHCKVDntGnJ1DlAH Thank you for your kind words! My cat used to howl, charge at me, and then jump up and scratch at me whenever I'd try singing, so I really DO understand. That is why I feel so strongly that these tools can be helpful! I hope you keep at creating and thank you for your comment and for watching the channel :)

Also, you can generate a vocal in Suno without singing and, if you want to use a Controlla voice, just bring the url to controlla - here is a video kzbin.info/www/bejne/b6HEaZWqq5epqdUsi=TyJXzUhAVT8wGdcx. Thanks for watching!

Thank you for your comments!

@@CreativelyMakeAIMusic You are welcome.

The first part is literally so wrong. If it was right we would’ve already had a singularity meaning an AI passed the Turing test thus meaning it can think the way a human can and do things a human can with 0 human input at all. If that actually happened we wouldn’t be still trying to achieve AGI or ASI or any of those things cause we would already have it. So to say it’s just learning like a human while it is literally performing like a computer and unlike a human still with capabilities humans do not currently have especially to that extent. To say that while it’s doing things that under copyright even humans can’t do as it would be infringement is ok is just wow. Imagine if I as a human indefinitely listened to and remembered all of your songs even after deleting them all from my device and just used pieces of them to make something new, a derivative under copyright which means I used your work not as inspiration which is totally different but as the actual source material to pull from, doesn’t matter if it’s sampling, interpolating, covering or remixing or otherwise. If I can’t do that as a human without attributing and licensing why is it ok for a machine under human control to do so is something that should be thought about.

It's definitely got a looooong way to go :) Thank you for watching; I enjoyed your comment!

I agree - it sounds great and I love that you can run it locally. I'd like to use it for a video but need to buy it in order to do so as I've already used up my free trial.

I agree.There is a TON of trial and error in this. I've tried prompting the vocals as Crystal Clear (as an example) and that helps. I also check "remove reverb" in the advanced settings and that can help as well. The results are sometimes better using different splitters, or just having Controlla pull the vocal from the Suno link (where it picks out the vocal from the full song). If you find any good ideas to improve the splitting results feel free to share them. Thank you so much for watching!

I neat thing you can do with the RipX DAW is if there is any bleed over from the instruments into the vocals or vice versa in the stems you can simply drag the notes into the correct stem, it's pretty cool.

couldn't disagree more, I think these are tools that will revolutionize the way us as musicians make music.

@staysickmusic5971 the dumbing down of society and culture continues. It's not you as the musician making this. Prompting a large language model that has been trained on other people's work is not making music, and it sure as heck isn't you making the music. It's a plagiarism machine plagiarizing because you told it to. There is no artistic merit in this, and can be no pride in this. It's pathetic.

I hear what you are saying and respect your opinion. I'm a trained musician, and have worked as a professionally as such or in the music industry for years ("many many many" of them LOL). Not everyone has the ability or opportunity to pursue music in traditional ways and these tools provide access to learn and create in a different way than you or I had the opportunity to, similar to DAWs, VSTs, samplers etc created new opportunities for new and experienced musicians.

Honestly, your take comes off as pretty ignorant about the history of music and creativity. Let’s get real. artists have always borrowed, sampled, and built on the ideas of others. That’s how art evolves. From hip-hop producers flipping soul records, to rock legends inspired by blues, to EDM using presets and samples-great music comes from reinterpreting what’s already out there. Beethoven, Picasso, The Beatles-none of them created in a vacuum. They all took inspiration and elements from others, made them their own, and pushed things forward. So what’s the difference if an artist today uses AI as a tool? AI is just another instrument-like a sampler, a guitar, or a DAW. It’s not ‘stealing’; it’s transforming and innovating. Frankly, accusing someone of being ‘pathetic’ for using a tool they didn’t create is ridiculous when you’re probably using drum samples, presets, VSTs, and other tools someone else designed without crediting them either. By that logic, are you ‘pathetic’ for not making your own drum kit or synthesizer from scratch? Instead of whining about tools, maybe focus on how people use them to create something meaningful. Dismissing innovation outright makes you sound stuck in the past, and history shows us that those who resist change in art usually get left behind ;)

@@staysickmusic5971I was about to write the same thing, but you explained everything I wanted to say, even better 👍 AI is just another tool and it’s up to us how to use it

Hi - yes - as an example I take a stem from Suno and put it in controlla in this video kzbin.info/www/bejne/b6HEaZWqq5epqdUsi=TyJXzUhAVT8wGdcx. Thanks for watching!

Barry thank you so much! I'm really glad - I am working on a step by step tutorial for this process, and writing some full songs as examples of using the different tunes together. I hope you can tune in for more. Thank you so much for watching and subscribing!

You made my day :) If you want to hear a few more (all me as a real human) kzbin.info/door/ZbSS5ZnCrya3HS8EJdyhOg

@@CreativelyMakeAIMusic you kidding? I'm going to listen to as much of your stuff as I can !

I'm so glad you watched - here is the a new video using Controlla - kzbin.info/www/bejne/o2a2doxofZKMna8si=BuAZyKG64iQtcgAv

Thanks for watching! Good question - -I don't know about that specifically but If someone is using the free version, they don't own the copyright - its all in this link from Suno. help.suno.com/en/articles/2746945#:~:text=If%20you%20make%20music%20with,license%20to%20monetize%20those%20songs..

In Suno 4, if you have a paid subscription, Suno states that you own the copyright. Just open Suno 4 and look at what you get with the different paid tiers.

I’ve had no problems at all uploading songs I did in SUNO onto KZbin. No copyright issues because they are new songs.

The companies that trained their own models from their own singers instead of scraping get a + in my book - better quality as well.; Controlla, Ace, a bunch of others do that. By the way, I like your kid's channel. I had a successful kids music project circa 2019 (all human created music; ai didn't exist), lots of fun bringing music to kids!

@@CreativelyMakeAIMusic Thank you . I didn't know you were an actual person, or an A.I. creation. My music was for the most part produced the old fashioned way, with musicians in a studio, like this song kzbin.info/www/bejne/gpXChIqid9h7rJo

@@BuddyBearCreator I'm definitely a real person LOL - I've been professionally working in the music business for most of my life including touring with major acts and releasing music as a real human artist etc. I didn't realize my voice sounded robotic enough to sound like AI! I'll have to work on that.

That's impressive - Thanks for sharing!

I wrote a long answer on the other video - I'm looking into this. Thanks for the suggestion!

@@CreativelyMakeAIMusic I think most of us want some ai to enhance what we have, but not make us into Adele so our songs become unusable and unrecognizable. Tks. I'm looking too, but the canned ai voices seem to dominate currently.

You'll have to use something like Melodyne to make your own recorded voice sound better.

Thanks for sharing that and for watching the video!

I tried a few where I can blend just enough of the AI voice behind mine to improve it but still sound a lot like me, but I haven't worked out a good enough vocal on my end for a video yet. Meanwhile, maybe try Kits or Ace (you may have to pay for a month to try blending your voice with theirs) and see if that helps out? Just a guess; Thank you for subscribing!

@@CreativelyMakeAIMusic kits ai looks like a good trial option; I uploaded a test vocal isolated in a song in Audacity and applied one of the vocal clone options...it sounded better, so that's good. I'll do a 3-day trial and get back to you. Thanks a lot for the tip...have a great holiday season.

Hi - I took your advice and made an overview of this! I'll go into more detail in about a week. Here's the link! Share it freely :) kzbin.info/www/bejne/o2a2doxofZKMna8feature=shared

@@CreativelyMakeAIMusic Excellent, I've followed you and look forward to seeing what you create.

Thanks for watching!

Thanks for watching!

The quality is getting very good. Thanks for watching!

Thanks for the feedback! You can improve the output in other languages by changing the accent strength to “use source accent”. We’re also adding a ton more voices in the next few months! Appreciate your thoughts 🙏🏾

Thanks for watching and pointing that out. @Controlla-voice left some instructions to address this in their comment here.

Thank you for telling us how to address that :)

Thank you for watching. I will showcase more!

Hi chillwalker - that's a great question! I entered foreboding, tense, building and other descriptors when I generated my original Suno vocal but, for this video, the goal was more to demonstrate how you can get a cleaner vocal stem rather than tags. I touch on Suno tags in other videos and I will do more on that. Thank you so much for watching!

@@CreativelyMakeAIMusic I am working on a virtual project. So for 6 month I went deep (deeeeeeeeep) into suno and audime, build my own voice for "her". First I tried to build it up with a blend of known german artists. Did not work at all, to little clean material. Then, when suno got better with voices in german, I generatet a ton of Songs, learned how to prompt for the voices I want and got the trick how to get the same voices, long befor suno announced this "persona" feature. And how to prombt, that suno makes a vocal focussed mix or even an acapella, whenever I need. I have now a basis Tranig model, out of about 15 different Songs. On this basis model I edit 1-2 Songs of the Style I want. For example, when I need her to sing more Soul, or more Rock, I let Suno create something (thats why I came up with the question) that fits at the beginning of my model. So its always "her". Then I went thru audimees catalog and found a voice that has a supernice akzent...she "rolls" the "r" like you do when you are from a specific region in germny. Thats a cute and personal touch. I blend my actual modell with 30% of the audimree voice. Funfact: Train an audimee voice based on an voice that contends more than 10-20% audimee voice does not work at all. I always have to make a "mixed" voice with my basis model. Feeding Audimee with the end result is...."strange" Its somehow like inbreeding in real life: The offspring is nothing you'll be proud off ;-)(***) Now my Artists voice is somewhat presentable in its Version 5.2. And, yess, I should do videos about the process cause I know no-one who had gone so deep. (***) (This effect also accures, when you wanna make an image generation of a face, blending generated images from one source...it really starts with strange moles ad leads to creepy transformations....So the philosophicl question is: Is Nature or "Gods Plan" in everything? ) kzbin.info/www/bejne/qoLUkoKGdqaAnZI

Dude...what do you expect? To get everything for free, but on a professional level, so that you can earn money at others costs? May I tell you what you have to pay for an hour of a professional Studio Musician or Singer? about 350 (low end) to 900,-

Thanks for watching! The free trial gets 1000 seconds of conversion with downloads, which is quite a bit of time. Keep your vocal sample really short until you find a voice you like, then put in the full vocal that you want - you can probably get 4 or more song length vocals out of your free account.

I'd check out to your voiceovers but I see you don't make videos, so there's that...🤣

I find AI as Video Voice over always annoying af. Even the good ones...

Thanks for watching! I haven't tried SoundID voice yet but its definitely on my list. I'll talk about mixing etc in future videos.

Thanks for watching! This demo uses the raw output from Suno & Control. Mixing will get addressed in future videos.

There is a little bit about audimee here - more to come. kzbin.info/www/bejne/mn60f3eZmpV2esksi=90ZoT_4EbmjVfANB