So how do you feel about AI vocals and what they can be used for?

Got the same problem: Composing since 1999 for living...but can't hold a single note ;-) And now, finally, I can write Songs without the barrier in my head "I need a Singer for the demo...ohnoooo...so much work..." Now I can do the writing.

This is the best of the AI vocal demo videos I've seen .... honest, useful, straightforward. I'm finding these ai services most useful for harmonies and backup vox.

Very interesting. Fast forward a few years, and I see this included in Cubase. Track your lead vocal, and it will profile your voice, make an AI version of you and then instantly generate harmonies or background vocals for your track, or even doubling takes, in seconds instead of having to make a bunch of takes, and then comping them.

Your comment about the vibe that each voice brings is very good. A good way to try out different vibes.

Thanks for sharing this! I'm not a well trained singer but I still want to use my own vocals because it's fun & I want to improve. But this is wonderful in combination!

What I find amazing is the AI generated voices have your accent qualities incorporated. Amazing what technology can do.

perfect breakdown and exact use cases I have in mind! pls keep us updated that stuff evolves so fast

Very interesting, thanks. And these are just the early days! Fascinating.

AI has achieved many milestones, including ChatGPT, and I believe it is starting to significantly impact music production. Every musician should now learn how to integrate AI into their daily work, which will provide huge value.

This was a great run through of this site and it's ability. For recording demo's this is an unbelievable tool. Fantastic.

Thanks for the video, AI vocals sounds great, what a technology , I’m shocked)

Excellent professional video thanks I’m trying to find out how AI software can improve my singing voice for personal use as a hobby. This is the best video on this I’ve seen so far by a long way and a great help. I am amazed at what can now be achieved and will definitely consider buying this software.

..and not only your voice. I barely work with vocal samples, cause The few that are really good, inspiring and, above all, actually usable due to their quality are so rare that other producers have already used them. In fact, me and a friend of mine used the same sample independently and when our (shared) record company asked us about it, we wondered who had "officially" remixed whom. But I recently found a vocal pack by chance that was so incredibly good that I immediately had two songs in my head with it. During the production (they were female RnB/Funk/Gospel songs) I thought to myself: "It's a shame. THIS SONG would be a great duet!... wait..." would be??" So I went to my trusted AI Vocal Site, uploaded one of the two verses, chose a nice male AI voice and turned the woman into a man (DON'T FORGET to transpose down an octave in the defaults!) THE ONLY PROBLEM IS NOW JUST: I bought the license to use these vocals. However, according to German copyright law, for example, changing a work of art(even if it has been licensed) is prohibited. I have to clarify with the producer of this sample series first... whether I'm allowed to do that ;-)

Grappig te horen dat de AI vocal ook het Nederlandse accent overneemt van de input van je eigen vocal. Tip: je kunt ook gewoon Nederlandstalige teksten als input gebruiken, dat levert best heel goede Nederlandse vocalen op.

This video is excellent - the comparisons and your examples are very helpful. It makes me think that in about 5-10 years we'll have outboard gear or pedal sized devices that will do this in real time. Probably it will be added as a feature in the next generation of workstations or arrangers like autotune and harmonizers, then separated as a single use piece of gear. I have a Roland VT-4, maybe the VT-12 will have a selection dial of 4-8 AI real time voice processors. Maybe Apple Logic 11 will do this - much like the named drummers. I can't wait for more of this technology.

This technology is changing so quickly ... I wonder if you might have the time and desire to create a "Lanewood Studio Updated AI Vocals Subchannel" that can be updated as the next new thing comes around. Today ... Audimee. Tomorrow ...? Either way - I love what you do, across your multiple channels, thank you!

Been fiddling around with Audimee for a week or so and its pretty amazing considering its a piece of technology that's still in its cradle. That said, the output is still very much dependent on your input, the better the vocals you run the algorithm on the better it will sound coming out. I got way better results generating variations on the voice of a trained singer than on my own voice since it captures the vibrato, timing and pronunciation of the original performance. Still sounds good on my voice though, just not quite as convincing. There are also plenty of more or less noticeable artifacts like strange swishing breath and ess sounds that vary a lot depending on the voice you use. And while it captures some aspects of the performance extremely well it cant emulate dynamics very well, giving a somewhat flat result. Those are just minor issues though. Its an extremely powerful tool for any home producer. AI in music and music production seems to be inevitable, for good and bad,

thanks for the thorough overview and comparing AI from a singer to AI generated from a non singer.

Could be great for doubling vocals if you only have one track. This Sheet blows my mind.

"stoooorm is rising, the..." this melody is playing in my brain (in loop) this morning... I have just put 3 "like" (my accounts) on the Awash video, nice song!

PERFECT and thank you so much. I am so happy tech has gotten to this point. What an amazing video that addresses all my concerns and needs in real life. So much money, time and frustration dealing with ungrateful artist is going to be safes ALL YEAR long and the foreseeable future. Thanks, and stated wit ZERO ethical issues whatsoever.

For whatever it's worth, I preferred your vocal for it's vulnerability, honesty, and authenticity over your singers vocal. I appreciate that taste is subjective. But I really felt like you were being a bit unkind to yourself. Everything has its place. Maybe someday you will find that place for your voice in the context of a personal expression. I wish you success in your future endeavors, sir.

I like your natural singing voice even if it’s rather sweet than perfect. I find these things fascinating but I promise in a few years most of the people will appreciate everything natural and „touchable“

James sounds like Vertical Horizon vocalist Matt Scannell. Aston sounds like Stevie Nicks

Wow fascinating.....really impressive...and this is only going to get better. Already absolutely useable for backing and harmony vocals to strengthen a track. Thanks for sharing.

Sounds perfect for harmonies und backing vocals you're not supposed to hear clearly.

Very interesting video 😀 i think i Will test this tool shortly

This is fantastic for me. I can sing fine, but my voice is not great for some genres of music. Now I won't feel so limited when writing songs

Thanks for this walkthrough, great when a real producer test things like this! I feel this isn't really different than EZ Drummer, EZ Keys and more by Toontrack, now just with vocals. Same with guitars, where you profile amps with Kemper and what not. The backside is that this won't add any value if you're an artist yourself, to have AI vocals on your music, so for artist branding like if you have a band or a solo artist, using AI could cause issues with building a fanbase - at least if you are not upfront about it (But we have Gorillaz ect). Cool tools!

Thank you. This was very helpful.

Maybe one day there will be a box that you connect to your microfone and that will convert your voice in (almost) realtime. So you can use this transformation on a live stage.

There is the whole big blob of uncertainty regarding how this and other Tools like are Trained. If you hear a voice that sounds like an established Vocalist, you may want to be careful trying to use it in a Commercial Product for Sales and Distributions. Of course if you own a LLC in the United States and your Record Label is Sued; you may be able to try to get all of your Company Assets Transferred somewhere the Legal Hawks cannot find them. For now, thank you for sharing the Link and the Platform and your take on how it can be used. Very Useful Information.

Really depends on the input though. I think that it would be interesting if you showed a video with an out of tune vocal as input for the processed vocal. It will surely help the composers like myself along the way as usually the worst part is singing. I can handle guitars, bass and even drums, but singing 😂

Thanks for the video, AI vocals sounds great, what a technology , I’m shocked

Personally, I think that the technology could be useful when working with sketch vocals. I would not let it out of the studio though. Aslo helpful if one does not possess adequate vocal chops.

Wow, sounds pretty good actually. No glitches or drop offs clicks or other imperfections....What I did notice is that obviously the singers are all different but the vocal performances are exactly as the original. What makes singers different are all the nuances style and vibe they inject on each song. So all of that is not there. It's just the same performance with a different voice...

I think one strong application is for BVs. If you have to generate BVs from a lead and pitch shifting you're restricted to the timing and lyrics on that lead. With AI you can input unique musically credible BVs with proper phrasing, words, gender mix.

Dreadful: I had to endure nearly six whole minutes of YT ads without being given the option to skip them before getting to watch this clip. 😞 There s been a seismic increase of ads time length on YT.

great.. very helpful. Thank you very much

Great video! I think this technology will be useful for artist who don't sing and want to demo a song idea. The AI voices are lacking unique character and feel of a real human voice, but I think they sound pretty good.

Thanks. Great video! 👏🏻👏🏻👏🏻 Yes the legality is the questionable bit to me.

Thanks for another excellent in-depth video! I have a question - is Audimee just using to the pitch and character of the original vocal (the vocal to be transformed)? To explain further, if I was to record my vocals with a bit of background noise (maybe very soft traffic, birds, the hum of an air-conditioner etc) would all that noise be in the AI version? Or, if I was to use a cheap microphone or even an iPhone to record vocals, would the AI version have a similar low sound quality?

You should try ACE Studio works inside your Daw :) And its a complete new ball game :) Compared to any other A.I The original singers have given there permission and get paid for there work :) You can also Mix up to 3 different singers and make your own and it lets your input midi and use your own vocal :) Adjust your pitch vibrato etc :) Looks cubase VarioAudio.

One day it will be able to meld a voice with the characteristics, formants and timbre of any singer you choose.So that maybe I could mix my voice with Stevie Wonder and John Lennon and play with the ratios. It doesn't sound that silly any more ! Thanks.

I use Synthesizer-V, which is more of a song-writing tool than a vocal replacer. It is however, capable of results that can fool even singers that it was sung by a real singer and could be used on commercial releases. As a non-singer myself, it has revolutionized how I write songs. If someone gives me lyrics, I can compose a simple piano backing and topline vocal in an hour and provide a demo. Of course it takes longer to do a more complex arrangement, and to finesse, but to get the idea across to a lyricist, its gold. And it has become invaluable in writing my own lyrics, something I always struggled with when I had to rely on session vocalists. I have a couple of links to Alonetone playlists in my profile if you want to hear how that sounds. I struggle to finish productions, but its no longer ever the fault of the vocal line.

The Isabella model gives a cool country vibe. Suits the song better than a male vocal in my opinion.

It could come in very handy for backing vocals

I hate the idea of AI, but, having said that, this program is very impressive. Is it any different than adding reverb to my vocals? Who knows? I'll give it a try through your link. Thanks.

Great Video! Like you song. I Don't want to use an AI vocalist that thousand are using, as well. I want be the singer but I am challenged with an unfortunate nasally and repelling sound that make my friends cringe. You were not cursed with the nasally sound. You sound good. Haha! I guess you can modify some of the AI singers so they sound unique.

that"s impressive

That's cool

Very informative video and well demonstrated. I am a non-singer that desperately needs a singer when writing a song. I'm not sure I could even sing as well as you...okay, I'm sure I can't...lol. But you have given a great demo with just the comparisons I've been looking for. I work for a publisher and I can't even send him any more than lyrics to my songs. So this was exactly what I was looking for. My concerns are more to do with copyright ownership and buying some version outright instead of a subscription. Are my creative works able to be used by the company to further train their AIs? Adobe is getting some serious push-back from their users due to their new End User License Agreement. But this product is doing exactly what I need. As to your question about not using a real singer, my answer is no, this is both legally and morally acceptable in my view. There are so many legal implications that still need to be sorted by the courts, so time will tell. But these AI singers know exactly what their voices are used for and are being paid at the front end...so singers are being paid. Great video. Thank you.

I think it´s really great

Do you happen to know if certain models will add vibrato to sustain notes sung. For example, if I'm not good at singing vibrato, will one of the AI models add it to the final conversion?

great and interesting video. I can't wait until someone releases an ai plug-in that I can just buy outright. I never like the idea of subscribing to something unless I absolutely have to. I usually start off like gang busters and after a bit I use subscriptions much less and feel like I am wasting money.

Are all the AI vocals excessively warbly? Also ... they put some reverb on there ... can we get a dry version delivered? PS: Thanks, Lanewood Studios! You are one of the very few sites I *always visit to catch up on things - always most excellent!!

Nice

Yo dude, thanks for the video. It sounds like there's something up with the vocal processing on your speaking voice... Sounds like it's high passed and like you have a cotton ball on your mouth? Sorry if there's a reason for this but I thought I'd mention it as it doesn't sound natural. :)

I love this video

Is the future everyone staring into their phone /device, watching the movies they made, based on the book they wrote, while listening to the music they produced?

I actually enjoyed your vocals without autotune the most, they were the only ones that didn't feel fake popmusic-ish. all the ai voices are so annoying, also the falsetto thing it's more musical without it imho

I am trying to find one that fixes MY voice I don't want my voice to sound like someone else. I want to hear my voice what it would be like if I could actually sing.

Excelente video! Question: Theres a way to "autotune" my own voice with AI to perfect pitch a song?

Very interesting

i'm more curious about the copyright complications of those singers who sound like already famous singers.

Are their any tools that can convert the vocal if you feed it the original track, ie without isolating the duff vocal first?

it would be necessary to imitate as closely as possible a favorite singer, the accent and intonations, then cut into several parts to train the A.I. on each of its parts separately, with parts of the original that really correspond. I'm going to test this.

Actors, musicians, producers, directors, animators, editors, studios, programmers, engineers and many many more are now obsolete.

Cantamus has been nice to sing choir songs. It is physical impossible for me as a metso soprano to sing a bass line. Listening to my song is then cool on thst app.

The problem is... counterfeit people. You're replacing a real performance by a human being with a counterfeit vocalist that never 'sang' the song. Besides: Everybody (who is willing to pay) has access to the same catalogue of voices, making every song made with this technology sound more alike and less special or individual. Flooding an already overcrowded market with counterfeit vocals is a very slippery slope and eventually a race to the bottom.

I think this will ultimately stifle individual creativity as well, limiting diversity and ultimately killing music

What I really, really would like to know is how well you can capture your own singer, and what effort it'd take. Got the problem that he removed (to another country, sadly) and I have stuff open that I'd like to get done with...

Ilove the female vocal

Great, thanks. How will you voice be with a female AI ?

I like your voice best

When will this be available for realtime singing? sure, some latency will always be there but one can learn to sing ahead of time

Aston sounds VERY close to Hayley Williams - their demo of her even does her band Paramore's Misery Business. Damn, I was hoping you had the $12 account, because they don't specify the bitrate of the MP3s and the free MP3 is 192kHz. I'm not paying them for less than 256 or better still 320/VBR🙃 But now I doubt it is, 16-bit 40kHz??? screw this company LOL I'd accept 44.1 even, 40 is just weird o.O They should have an FAQ with the audio specs, or better yet ON THE PLAN BUY PAGE like a company not being sketchy would do Thanks for sharing :) In my brief use (Free Account) I definitely noticed that warble on longer notes, better on some voices than others.

Thank you very much! Does it work with other languages than englisch?

Can you use any language on the conversion or is it just English??

It won't be long before all those Elvis impersonators will be out of work😂

Hey will you make a tutorial about how to fix ai voice with auto tune?

Long time ago, singing was equal to live singing. Every song was uniqe, personal and unrepeatable. Then came recording, and every song became repeatable, impersonal. Now, here is AI, and we will never know, what comes from the loudspeakers, is a human, or just a computer.

I am using Audimee. Just started. When I import into cubase the converted file seems to be at a different sample rate so does not line up. I am tracking at 44.1K 16 bit. What should I set cubast at? Thanks for any info

Can i use this stand alone. I don't want to upload any of my material to their server.

what DAW are you using?

I've been using Kits ai, to model vocals I'd isolated and edited from old disco records. I then used the model with Synth V vocals I'd programmed laboriously. Did it work? Did it fu..... It does have potential though, but I was expecting Chaka Khan😂😂.

I guess an idea would be to pay someone on fiver to give you a good demo then use this when that doesn’t work out.

But thanks for the video, possibly helpful to me.

Oh dear. It seems to me that what AI can't (yet) do is create either the emotion or a subtle and unique personality to a vocal. Even for a demo, these are things that could be crucial to selling a song. And where, for example, the video author has to drop down to low notes because he can't handle the falsetto high notes, the AI only copies that: the inference behind the words is completely changed by singing the lines low instead of high. And then, I reflect on part of what makes a pop song a hit, and I think it is because the artist as a person resonates with an audience (hence all the TC interviews that follow etc). AI makes a complete nonsense of all of this by de-humanising what ought to have some human artistry (faults and all) and personality in it. So I'm out, as they say on Dragon's Den (and yes I'm a singer, and so are/were Lou Reed, Mark Knopfler, Dylan, Leonard Cohen etc, who could not be called great singers in the technical sense but who can nevertheless deliver a spellbinding performance - why is that I wonder, and no prizes for guessing).

Crazy times indeed... Does it also able to work on other languages then English with those "AI singers" , say Russian, Hebrew, German, Arabic etc...

Now I don't have to deal with Axl Rose

How did you autotuned your voice?

In paid AUDIMEE, can the models sing in Spanish?

The original vocals of your bandmate was 100 times better than the synthesized vocal.

Are you from Holland or so?

I think it’s ok for demos…..but I wouldn’t want to see this being used for released material.

14:22 I can't recognize any "ethical" problem here. You can use any available tool to produce your music. Also the new technologies always make somebody jobless... It's just a natural law.

Need real-time conversion

Interesting. Thanks. The vocal styles provided by the service are quite limited actually. Very poppy and pretty bland. Personally, I would be much more interested if they offered jazz or classical/operatic voices. Although that degree of expressivity may be beyond the AI at the moment. Cheers.