Udio & Suno AI just got STOMPED on.. New AI Music is SCARY GOOD

Elevenlabs just announced EASILY the most insanely good AI music to date.
▼ Link(s) From Today’s Video:
Announcement: / 1788628171044053386
More examples: / flavioschneide
Elevenlabs: elevenlabs.io/
► MattVidPro Discord: / discord
► Follow Me on Twitter: / mattvidpro
► Buy me a Coffee! buymeacoffee.com/mattvidpro
-------------------------------------------------
▼ Extra Links of Interest:
AI LINKS MASTER LIST: www.futurepedia.io/
General AI Playlist: • General MattVidPro AI ...
AI I use to edit videos: www.descript.com/?lmref=nA4fDg
Instagram: mattvidpro
Tiktok: tiktok.com/@mattvidpro
Second Channel: / @matt_pie
-------------------------------------------------
Thanks for watching Matt Video Productions! I make all sorts of videos here on KZbin! Technology, Tutorials, and Reviews! Enjoy Your stay here, and subscribe!
All Suggestions, Thoughts And Comments Are Greatly Appreciated… Because I Actually Read Them.
-------------------------------------------------
► Business Contact: [e-mail]

[@MattVidPro]

What's your thoughts on this? I mean it's mind bending...

[@phen-themoogle7651]

I watched a few minutes, it's awesome, but question: Since it's ElevenLabs, could you clone your own voice and they put YOUR voice into the song? or is that a feature already (sorry didn't watch whole video) xD

[@MattVidPro]

@@phen-themoogle7651 We don't know yet but great point!

[@leebebbington3413]

Voice quality 💪

[@MrNobodyX3]

Can you link to the page that your viewing these examples from?

[@MattVidPro]

@@MrNobodyX3 I did

[@afti03]

Im a life long music producer and i just wanted to say this is way better than i thought it would ever be.. i don't know how to feel about this..

[@JohnSmith762A11B]

For awhile there might be an exploitable quality gap, though pretty obviously that won't last. I would suggest a few things: since live performance isn't going away, look to repurpose your skills and experience there. Also, if you can't beat 'em join 'em so embrace these tools and pull them into your workflow. Finally, join the movement for UBI as we are all headed for the glue factory as slow, expensive, error-prone humans who need things like sleep and vacations and weekends in a looming world of tireless AGI humanoid robots. We cannot and should not be made to compete with them for money. That is a contest we will lose. Good luck!

[@Raphy_Afk]

You should be afraid imho. Or enthusiastic and start your own AI project

[@spol]

Then it came for me.. but seriously the future will be fantastic even if our jobs disappear. As people we’ll simply consume and enjoy each other.

[@EddieAdolf]

This is how I felt years ago realizing designers wouldn't be designers anymore... We have to use our skills in a different way. At least that's what I've realized it's more about composing the content and the strategy stuff not the actual content since it can be done instantly now. But there's still much for us to do. It's just a tool,good ideas will still matter more, but I was also very conflicted at first.

[@fpham8004]

Today music producers are already using too many AI-adjacent tools like autotune, ace studio, stem separation, melodyne and then infinite loop services like Splice - so today music is mostly produced artificially already. The step from that to AI generation is not really that much. Using Splice or generating your own stem is not much different for the end user except with generating they can go further. The people who sell stems and loops would be the one going away, not producers. Imagine uploading your unfinished song idea and AI generating properly produced stems for it with the correct song structure.

[@youri655]

THIS PASSES THE MUSIC TURING TEST

[@afti03]

Oh man, it passes and leaves it in the dust

[@WastedContender]

😂😂😂😭

[@user-nq1ze4jt9m]

And how

[@michaelmmanzini]

Finally you can now play some music online without the risk of getting Copyrighted

[@hypersonicmonkeybrains3418]

As a guitarist / musician i welcome this interesting tech, and i don't feel that it threatens live music performances in any way.

[@darksoulnj]

💯🎯

[@OldNations]

What it threatens is the big boys who just create more of the same of the same of the same of the same...

[@Pawel_Mrozek]

Until AI would start making live music ;) So this "never" would not stay too long with us.

[@JanPapiezGaming]

till you get robot bands

[@hypersonicmonkeybrains3418]

@@JanPapiezGaming That would defeat the purpose of going to see a band.

[@danielrodio9]

Imagine every time you asked a robot a question, it answered in music

[@JohnSmith762A11B]

And could literally stomp on you for real afterwards. "Was that frighteningly good!? Tell me human, tell me!"🤖

[@NextGenart99]

Like those Disney musicals, where the actors spontaneously break out into song.

[@OliNorwell]

Haha great idea, I can hear it now “as an AaaaaaIiiii language model laaah laah laaaah”.

[@Windswept7]

@@NextGenart99omg imagine a game where everyone e responds like an absurd musical production! 😂

[@nathanbanks2354]

I enjoy asking LLM's like ChatGPT to write poetry for me. They're pretty good at limericks & sonnets, but occasionally get the meter and rhymes wrong. (Eg. It could think "prove" and "love" rhyme.) The crazy part is that they generate poetry at the same speed as prose!

[@knightride9635]

That first song is incredible, I would never have guessed it was AI

[@immaculatetoilet]

I mean, it sounds like a rip off of Tom Petty and later Incubus. Sound quality is good but it's completely derivative. I like AI music for the meshing of multiple genres. These demos aren't that impressive to me when it comes to unique composition.

[@cesar4729]

Dunno. Voice dinamics are so flat.

[@domehouse79]

We must have been listening to different songs

[@boggers]

The voice sounds very processed to me, but that said, pop vocals have sounded processed for decades now.

[@saulg195]

@@boggers That's where artists messed up. Using autotune and voice modifiers. They always sound different live

[@rossibytes]

I was actually wondering where ElevenLabs was in all of this... now we know!

[@Edbrad]

The funny part is Udio has more realistic text to speech than Elevenlabs, so maybe it's the same with ElevenLabs' text to music model, in that if you prompt for regular spoken word stuff it sounds better :D

[@Orange-tk4hc]

@@Edbrad Udio definitely doesn't have more realistic text to speech than Eleven (even if it's good).

[@starblaiz1986]

Yeh clearly they were COOKING! 😮

[@melodyinwhisper]

The jazz lacks cohesive structure. I actually think Udio is quite a bit better at jazz. It's nice that eleven labs can create whole songs more easily, without that business of putting 30 second clips together like Udioo. But to me, Udio is way better at old school music sounds. When i ask Udio to make me something from the 1920s, it actually comes really close, even emulating the quality of the audio recordings of that era.

[@joelface]

I think the 30 second clips is good for having more personal control of what you create though. It gives you a ton of choices about how the song should go. By the song you've made a 4 minute song, it really feels like you had a major effect on the song.

[@addict8229]

If there was a Turing test for music, this would be the first AI music generator to pass. I can’t even wrap my head around how good this going to sound in the future.

[@NGU7754]

This almost sounds like real music, borderline scary.

[@MattVidPro]

I wouldn't be able to tell personally

[@schuey7]

Me too.

[@Shakalakahiki8]

As an average Joe I can't tell the difference.

[@jooptablet1727]

Almost?? Admittedly I'm more of a visual person but I listen to a huge variety of music every day and I can't even tell if Udio songs are real, let alone this!

[@JohnSmith762A11B]

@@jooptablet1727 Yes. I think a bunch of people are lying about being able to tell. Some songs are obvious AI gens but it's not hard to make something in Udio that sounds legit like just some indie band or folk musician.

[@FRareDom]

the fact that both udio and suno got taken over by 11labs like this in like the span of a month, is crazyyy

[@MattVidPro]

it's really just nuts. they know what they are doing

[@NicVandEmZ]

@@MattVidPro I don’t think you can use it yet

[@brainwithani5693]

Eleven labs has always been awesome, I'm not really surprised but it is exciting!

[@Walexo45]

It shows that what is available to the public is NOTHING compared to what AI companies have in their HQs. The world should take notice by now.

[@R123Productions]

UDIO JUST BEEN RAN OVER BY ELEVENLABS 💀

[@IceMetalPunk]

I *really* hope this is the consistent quality we'll get from the model. And knowing ElevenLabs already has an API for their TTS, if this *is* as good as it sounds, and they have an API for it? I'm canceling my Suno subscription and switching my Synthia Nova AI framework over to ElevenLabs, for sure. And maybe... since ElevenLabs does TTS where you can choose the voice model... just maybe we'll be able to choose consistent voice models and get multiple songs with the same voice? Maybe that's wishful thinking; but that'd honestly just be a bonus. A consistent quality like we hear in these examples, plus an API, is all I'd need to switch.

[@tedstriker2000]

an AI top 10 each week wud be cool...going through the best tracks trending out there '')

[@ZhBorg]

For sure

[@Liminal-Mystic]

Good Idea.

[@DorothyJeanThompson]

I love your enthusiasm soooo much. You would be so much fun to talk to ❤. I always look forward to your videos! Keep up the great work!

[@houseofhigherself]

As a musician & producer, I'm so stoked about this....forreal. I'm all about ease and effortless manifestation, so I can see this being used for cloning in ways to get projects done without having the artist even attempt to finish physically. This is useful if the vocalist isn't able to use their voice for some reason or otherwise engaged. I would do this for my own vox, if I'm being honest, and simply learn the song later for performing live lmaooooooooooooo

[@haileycollet4147]

Since the lyrics are the same both versions of the song, my assumption has been that "single text prompt" means the prompt they provided plus the lyrics. Only other interpretation is you can give a full song as input and it remixes it.

[@PseudoProphet]

Lyrics are not the same, just the prompt is the same + ( jazz ) Lyrics are obvy made with a LLM, does it matter if you donit or the eleven labs? 😂😂

[@haileycollet4147]

@@PseudoProphet the importance in identifying source of lyrics is just important in guessing whether it can do style remixes

[@thekerb9545]

It matters if you are using it as a tool and not a party favor.

[@haileycollet4147]

@@thekerb9545 also very true, of course it's important whether it can accept custom lyrics, but it'd be shocking if it couldn't

[@notnotandrew]

Or the underlying model produces embeddings in multiple latent spaces (e.g. one for lyrics, one for genre, one for chord progression, one for melody, etc.) and they have some internal way of tweaking each individually

[@Not-all-who-wander-are-lost]

Once AI wins an Emmy, a Grammy, an Oscar, the Pulitzer and the Nobel it will never lose again.

[@avsync-live]

Doubt it. For a long time we will use AI tools in conjunction with traditional production. The only way “AI” can win a Grammy etc is if it has real agency not just controlled by a human

[@boggers]

Yeah, this. Prettymuch the same as Chess or Go. It starts to get a bit scary once you have AI designing similarly award winning neural networks, chip hardware and training techniques for its more advanced descendants to inhabit, and you can be sure that is already happening.

[@mr.guzwee7695]

Those prices are for humans

[@zaczerlany-band]

yeah yeah, there is already Paint It, Black by The Rolling Stones

[@FrotLopOfficial]

Nah at one point humans will vote a genuine human song out of spite lol

[@k9bfriender672]

One thing that I have noticed about AI produced music over "traditional" is that I can actually understand the lyrics. It tends to enunciate better than humans. That's not a bad thing, IMO, but it is a notable difference.

[@Lar_me]

That's... kind of hilariously ironic, now that you mention it.

[@uranus8592]

Imagine what Jukebox 2 ( from Openai ) would sound like given that the last version was from 2020 💀

[@JohnSmith762A11B]

G-rated lyrics only. And you will need to be wearing a handmaid's bonnet while submitting any prompt. OpenAI will slowly become strictly a B2B business for the MIC.

[@uranus8592]

@@JohnSmith762A11B unfortunately, I hate what Openai is becoming

[@Felipe-zl1rj]

To be honest, Udio is pretty much the same level. Just need to expand the length of the generation.

[@RNDMTRNGR]

Not even close.

[@guythatmakessense2033]

I also think it is. I think it’s just about prompting and knowing what to want.

[@shaunralston]

I KNEW you'd be on this, Matt. Brilliant 🎉

[@MattVidPro]

Thanks! It's AMAZING right?

[@shaunralston]

​@@MattVidPro No doubt, brotha ' . . . rumor has it that work is underway to make our products' API compatible with ElevenLabs TTS + STS (speech-to-speech), voice cloning, dubbing and (now) music generation. ElevenLabs is WORLD CLASS. 🗺🎶🍋🤖

[@sanderhoogeland9161]

Meanwhile, Open AI is probably just sitting there laughing.

[@DangerManQ]

@@sanderhoogeland9161 why?

[@sanderhoogeland9161]

@@DangerManQ I just wouldn’t be surprised if they have something even better hidden in their lab somewhere.

[@DJ-Illuminate]

Its just a matter of time before a Beatles AI app comes out I can buy that spins new Beatles songs on a whim.

[@jooptablet1727]

And at first it will be OK, and then it will be way better than even original Beatles songs...

[@K.F-R]

Let it be.

[@STONJAUS_FILMS]

bro, you might sound like a dreamer, but I know you are not the only one

[@bosthebozo5273]

Here comes the sun

[@aexetan2769]

Hey Jude, just imagine the possibilities!

[@mohamedzewail8907]

The rap song made my whole body ask for a reality check.

[@saymydomain9504]

Me too

[@mohamedzewail8907]

@@saymydomain9504 This is truely amazing. Weird feeling when you get fooled by something that you know is fake.

[@TheFeedRocket]

Yeah, we had the edge on creativity....now, not so much, LOL. We thought AI and robots would replace plumbers first, odd now all the creatives and intellectuals will be looking to become plumbers and electricians.... But I can still beat AI at Basketball! Humans! I think it will be 10+ years before a robot can take a Human one on one in Basketball. So we still got something.

[@mohamedzewail8907]

@@TheFeedRocket I don't think that AI will replace humans, I think it'll make the world require less humans to run certain jobs. and it will make up new jobs and professions that people with less AI knowledge will occupy. It's gonne be a new line in our CV.

[@jhdesigner]

Broke My Heart to Fill you UP is SICK!!!!!!!!!!!!!!!!! 🤯

[@rebeccamiller8772]

Yeah the music doesn’t just sound good,it’s cleverly and artistically written.

[@verence333]

Absolutely! I'm amazed!

[@mrobotguy]

It’s giving lorde

[@handle__]

​@@mrobotguyexactly, it felt similar to Lorde for me too

[@AmandaFessler]

I was washing dishes and zoned out before it started playing. At first I thought it was one of those music videos that show up on ads once in a while. Then I turned around and saw it wasn't.

[@haileycollet4147]

What you meant by "sub-beat" is a "breakdown" (where it chills out/simplifies)... The "drop" is where it goes through a, er, more different section and transitions back to normal (a drop is really any significant/sudden change in rhythm/bass, but usually preceded by a buildup and then breakdown).

[@swancollective]

Think he meant the sub bass tho

[@LeviQFoster]

Not noticing a huge difference from Udio. Udio is shockingly good as well

[@amkire65]

Wow. Being Elevenlabs, I wonder if we'll be able to pick a voice to do the singing or even a cloned voice as the singer, you know they can make that happen.

[@jaredf6205]

I’m glad so many good models are competing. Looking forward to seeing how crazy the next versions of these models are. I wonder if OpenAI will join in, they had AI music before anyone else, Musenet in 2019 and Jukebox in 2020. Meta has been working on voice and sfx as well, wonder if they’ll do music.

[@RockyBergen]

I use AI for all my backing music now. Stock Audio is so done..

[@JohnSmith762A11B]

Stock video is next on the chopping block once Sora ships.

[@pklemming]

Same, all my background tracks are AI now. No uuntimg for something I want, just design it.

[@MrNuance]

It’s dynamic! That’s the word you’re looking for. This is wild. Will be using on first day.

[@Markalodon]

I should have read the comments. Came here to say this.

[@SignumEternis]

Wow, that really is amazing. I honestly don't think I'd be able to tell for any of of those songs that it was AI without already knowing. I didn't expect AI audio to advance this fast. Gonna be wild to see where it goes from here.

[@lutaayam]

I love your visceral reactions. So entertaining to watch😂

[@TheRemarkableN]

AI Music Wars, begun they have - Yoda

[@bombibjornene]

Bro I feel like Udio just came out and blew my mind what is this and I want it right now

[@phizc]

Udio came out 1 month ago. I'd say that qualifies as "just came out", even when talking about AI. 🤯

[@cowprez]

The jazz version is incredible! Thanks for your work in showing us this excellent software.

[@BrentStewart]

What I would like to see some of these implement is the ability to generate stems which could be a super useful tool for DJs and perhaps bridge the gap between "is AI replacing musicians or is it a helpful tool to be utilized by musicians".

[@brockhampton3078]

No offense to MattVidPro, but Suno makes music easily on par or better than this if you prompt with human written lyrics that have structure as opposed to letting the AI write everything..I feel like he needs to make more effort to test things out before making hyperbolic claims that overshadow the time and effort these companies are spending on their projects. Elevenlabs sounds great because a decent lyricist prompted these examples, the same results or better could be achieved with these same prompts in Suno. Claiming Suno or even Udio got "stomped on" based on these cherry picked examples is an extreme take IMO.

[@VisionTracks101]

Suno has worked well for me, and been able to produce some great melody.

[@dommyshan]

Wow gotta try that out. Thanks matt! :D

[@chanpasadopolska]

How long prompt does it process? Really long prompts are really something that I wish for. That's the reason why I use Dall-e for image generating, no other model can process so long prompts, I wish any music generator would be able to process half long prompt as Dall-e can

[@BallinLikeMike23]

You can tell the last song was trained on 21 Pilots. Human artists will have to step up their music to compete. It’s a win win for consumers

[@JohnSmith762A11B]

Humans competing against AI sooner or later has only one winner, and it ain't the skin-bags of mostly water.

[@Recuper8]

Human artists will be a novelty soon, like handmade furniture.

[@simonsanchezkumrich8489]

This is calling for an enormous lawsuit lol, very clear plagiarism

[@j00bert]

@@simonsanchezkumrich8489 Lol, no. Look up Damien Riehl

[@LandersWorkshop]

@@simonsanchezkumrich8489 Good luck with that I guess.

[@sixstanger00]

With Udio, you have to add your intro. When you click "Extend," it gives you the option for an "intro," "outro" or "new segment" either before/after the existing one. Dubstep & techno are probably going to be easiest for AI models, since the genres are very "calculated" and "artificial" to begin with.

[@bloodyiceberg6827]

For dance music they need to improve their mastering first, because you can't get away with a flat mix in EDM, even the so called "lo-fi" genres actually have really good mixes to begin with. What i've noticed in udio "oldies" sound a lot more believable because most of the old stuff if it's not the bilboard top had subpar mixes with various issues. Compared with modern stuff that tends to be more polished & overproduced soundwise and has a significally bigger processing chain wich gives this specific sound, and it is very noticable when it's missing.

[@liberty0758]

Electronic music in general still sounds way too compressed and "AI like" in any of the generators to be considered a threat to proper producing, with proper tools. Yet... I use both Suno and Udio and the results of what I'm trying to achieve are simply not there. Again, yet.

[@erics7004]

This sounds amazing!

[@djstraylight]

Scary good is an understatement.

[@bobtivnan]

Another great video matt. It made me just a little bit sad that you didn't recognize that the first song was inspired by Tom Petty (RIP). Or perhaps even trained on his voice?

[@stephenrodwell]

Can’t wait for this! Question is, do I release the album I made on Suno, or wait to put the songs back through this?

[@FredPauling]

The voices in Udio feel more soulful and intricate, but the sound here is much more crisp in the high end. I wonder if they are using an audio version of an upscale to do it.

[@MattGoldenberg]

Question for people who know music - is this really better than Suno and Udio - what makes it better? Between Suno and Udio, Suno still seems to generate more pleasing songs to my ear (and I don't see this as much better). But others have a clear preference ranking - what is it that makes Suno < Udio < ElevenLabs?

[@Athari-P]

Suno sounds clearly AI-ish, but it's consistent and predictable. Udio is wild and creative, but can produce messy results. ElevenLabs wins in consistency, judging by examples, but seems lacking in creativity.

[@ChristianIce]

Suno and Udio are, at this time full of artifacts. The sound gets worse on the high end, feels like an mp3 with very low bitrate. What I've heard here is flawless. Feels like CD quality out of a mastering session.

[@Athari-P]

@@ChristianIce Maybe, but I'd take an interesting song with a couple of misspelled words at 128kbps over a boring song with zero creativity and intonation variation.

[@eatmybutt42069]

@@ChristianIce It seems flawless because it was cherry picked

[@marilynlucas5128]

Suno is still king. It follows prompts perfectly. Udio is junk to me. It never does what I like. Suno blows my mind Everytime by clearly getting my intention

[@tfebster]

From a professionals perspective -- the jazz still is lacking. But it's pretty amazing it's at this point.

[@JankJank-om1op]

fomo/clickbait...its good for sure but Udio is still objectively better composition-wise. possible best-case scenario: mix elevenlabs TTS w Udio instrumentals but otherwise nope nope nope

[@MarcusBuer]

Absolutely, but it is progressing fast. There was a huge improvement on the last 6 months, and it will only get better.

[@KerryMarsh]

Seconded. The tech will eventually get a grasp of idiomatic vertical harmony and progressions (more complex than the pop material it can already just nail), and it may happen within 6 months, but it's not quite there yet. This stuff is my livelihood, so it's concerning, but hopefully there will remain value when artists capture their process along with producing the final work (a la ArtistShare)

[@karlwest437]

Give it another week 😮

[@dbt4869]

Why are you a professional in music? BS

[@MrTk3435]

well well well, great job Matt! Dude the rap song ! OMG . 🔥🔥🤟🔥🔥

[@subuntu]

I've been using udio a lot, I've got a number of tracks on my channel, but the sound quality on the elevenlabs music gen is amazing, this is like going from cassettes to CD :D I can't wait to play with this.

[@soulxavier]

We are gonna end up hearing these ai songs on the radio one day 😂

[@Happ1ness]

I knew that this day will come relatively soon, but I didn't think we'll be there in almost no time

[@FrankysAItime]

thank you for your great videos. I'm currently making a lot of music on my channel with Udio

[@Glowbox3D]

Ya, this is insane. Didn't expect to hear EL coming out with the top dog here, but I guess it makes sense. They clearly have some good heads on their shoulders, and some sick training data, especially with voices. I've been using EL for TTS and STS for a while now, it's super impressive.

[@marklondon9004]

"I got the algorithms, I flop them like I'm Turing" - extra marks for AI believing in the Creator.

[@Deckardb25364]

Saw this last night, shits insane.

[@MattVidPro]

it's a mindfuck

[@MisterWondrous]

Just checking that out. Far out! Looking forward to this.

[@DragonBallDetroit]

Thanks for guiding us through this dystopian future!

[@jpmallick367]

i clicked faster than the speed of light, especially when its about music and sounds, i didn't even fully start moving on from suno to udio 😭

[@harnageaa]

udio is inferior to suno when comes to creativity, but excel at quality. Hope elevenlabs to be suno+udio cuz that's what we need.

[@The1QwertySky]

finally a music generator i dont need lyrics to understand what they are saying

[@phoenixrisingmedia569]

Absolutely amazing!

[@shawnbibby2934]

The music is always so much more friendly and inspirational!

[@justtiredthings]

I mean, that's not jazz, man. That's like big band swing edit: I was wrong. I just don't think of swing as jazz. L taken

[@tigrisparvus2970]

Swing jazz. Big band, danceable, 1930s. You're probably thinking more like Bebop and the 1940s onwards. That being said once you get past the quality it's a terrible attempt for that era. Far to much crooning and the swing rhythm is........weird.....

[@Ricolaaaaaaaaaaaaaaaaa]

@@tigrisparvus2970 The rhythm was actually perfectly fine for what the song was going for lol. The issue you guys have is with prompt adherence not the output.

[@soulxavier]

​@@Ricolaaaaaaaaaaaaaaaaa You are right

[@AgustinCaniglia1992]

Not at all. It sounds amazing.​@@tigrisparvus2970

[@justtiredthings]

@@tigrisparvus2970 fair enough, you're right

[@DaDa-kf4vp]

implications are that only live performances will be real. Musicians will have AI create their music, then they will use their own voice to copy the AI song and claim they created it.

[@Remour]

It’s a terrible future….

[@Aldraz]

@@Remour Why? We will have infinite music, great music for every mood, will generate a hit you've never heard before based on your current mood, which would be impossible otherwise. Also, this is not the end of human performances.. humans can still do really cool dances into these musics and be superstars without actually being the original singers. This will only improve the quality of life for normal people. And also will mean that live performances will suddenly become more popular, because people will still value people.

[@phizc]

This is ElevenLabs.. They can clone their voice and let the AI sing in their voice from the start.

[@JohnSmith762A11B]

@@Remour Arguably this is an improvement on using autotune.

[@shadydragon22]

@@Remour You're absolutely right and you get it. Ignore the ai tech bros in the comments who have absolutely no understanding of the value of art, human expression and originality. They're comparing auto tune to not creating the whole song yourself ffs ugh

[@powerpackip112]

Impressed, can't wait to do my own prompts

[@Darrenithell]

Every week I find myself holding my head in disbelief and astonishment, my eyes begin to water. I don’t know if it’s joy, surprise or concern, it’s probably a combination of all those things.

[@ListenGRASSHOPPER]

Ai music coming to Billboard top 10 next year.

[@2CSST2]

Man first half of 2024 is coming to its end but we've already have so much crazy stuff. With this I'm gonna say AI music is what stole the show so far. Sora was pretty crazy and robotics has had some wow moments also, but AI music straight up went from barely listenable to practically undiscernible to real music. Just crazy I'm also just as mindblown, mad mad stuff 🤯🤯

[@aresaurelian]

Well done. I am pleased with the result. Now on to creating completely new genres, and inventing new styles of music.

[@peacefulman2196]

Create please. AI developers need fresh input to train their AI.

[@Thierryhavefun]

Just wow!

[@Tanaka.Minoru]

The narrator voice in the beginning is Kevan Brighting from the video game “The Stanley Parable”

[@thomassynths]

Their sound effects are actually quite jank. Not sure why you pinpointed that as something good. The rest of the stuff they do is genuinely impressive.

[@MattVidPro]

best AI sound effects I've heard... no one does them better

[@thomassynths]

@@MattVidPro That's fair. I'm comparing more toward free synths that can do this stuff easily.

[@Aldraz]

That's like when artists said the AI can't do fingers well.. but 6 months later.. it can do it perfectly.. so these are just details for AI.. like really small details

[@Aimerxd]

​@@AldrazSome artists stayed in the past and continue saying that the girl doesn't know how to do fingers xd

[@Grassland-ix7mu]

That "It started to sing" is amazing. I love it. Searched for it on youtube because i want to add i to a playlist, but found it only on X

[@MoreBusinessWithAI]

Absolutely thrilling to see Elevenlabs pushing the boundaries of AI in music! The way you detailed the versatility and authenticity in each genre is truly remarkable. It's exciting to think about how this technology will continue to evolve and enrich the music industry. Fantastic work!

[@TheArtOfCreedStonegate]

That song is the video killed the radio star moment for AI music. I'm going to remember hearing that is the first real AI music. I've added it to my tunes playlist! :]

[@mohamedzewail8907]

Imagine this with their top notch voice cloning, we're all gonna be good singers.

[@JohnSmith762A11B]

Or, we will all be wrinkly, blubbery hell-bellies with receding hairlines pretending to be good singers and sultry chanteuses. At least we won't have to put literal bags over our heads like The Knife lol On the Internet no one knows their favorite pop star is a dog.

[@waynelai354]

Getting better. It would be good if we had the same controls that people developed for picture generation. For instance to get the wave form and allow us to regenerate sections (like inpainting). Suno/Udio generate some nice tunes but awkward areas in the middle of the songs is difficult to correct. Also reference songs as input (like me humming something). I know it will get there esp with open source but we are still very limited in these ways.

[@waynelai354]

Ah inpaint was mentioned at the end, I just finished the whole video. Yes we need a control net to get the song toward our vision. Soon writers, musicians, media people will all have a fair shake at creating content end-to-end. Was hoping for this back in the 90s. Great to see it finally getting to be a reality.

[@LoveIsWhoWeAre]

Fascinating vid! thnx

[@jvlbme]

It's so well-trained it includes 'auto-tuned' voices. 🤣 It is crazy good.

[@jashall30]

This just means song witers can bring the own music to life them self i dnt think its neither good or bad i just think its another tool for creators i dnt think this means singer that cnt right music will be out of a job personal there still something more intimate for a person thats why we go to concerts to see the artists rather then just listen to it at home

[@Bebo18]

I’ve already made the songs i wrote into Suno. I can’t wait for this. I NEED THIS!

[@MarcoMugnatto]

The post on my AI blog (it's in portuguese. Translated by GPT): First, it was the illustrators, but I wasn't an illustrator so... Wait! Well... This post isn't really about that. But it's true that now with Suno, Udio, and this recent announcement from Eleven Labs, musicians are panicking on KZbin, calling for a "ban", "regulation", saying it's "piracy", a "plagiarism machine", etc. I have mixed feelings about the subject, which essentially break down into four points: 1- On one hand, I am indeed saddened... It's almost certain that many kinds of professions involving musicians will disappear, like jingle composers, film music composers, etc. It won't make any sense to hire humans for this kind of work. History has proven, however, that live human performances will not lose their value. Nevertheless, I can't help but feel a pang in my heart for the many music students and professionals who harbor so much love for this art that now loses the human exclusivity in its production, which is quite profound. 2- On the other hand, lamenting is one thing, and for that, I lend my shoulder. But wanting to prohibit technological advancement? Oh, that's not possible (PS: I meant "don't count on me for this" in portuguese). It's another thing proven in history: in practice, technological advancement has far more benefited than harmed humanity. So the musician who wants to "ban" the thing, to me, is already seen as selfish, and in this aspect does not have my support. 3- There's also in me the side of euphoria with technological advancements. Since childhood, I've liked technology and seeing how far it can go. It's always been my passion, and therefore this obviously weighs in my opinion in favor of these advancements. 4- Lastly, there's the issue of personal musical taste. Like many others, I can't stand contemporary music. I miss, for example, guitar solos. Therefore, having now a way to get new music, but the way I like it, in a world of so much bad music dominating the charts, is undeniably a very positive point. So that's it... I think these points well express my stance regarding this revolution in the world of music. In favor of and excited about it happening, opposed to attempts to stop it, but with a sad feeling for human musicians, while at the same time knowing they will continue to be valued in live performances, even if we no longer know if the music they play wasn't composed by artificial intelligence. Below is one of the demonstrations from the announcement of the new tool from Eleven Labs, not yet available to the public.

[@adrianmunevar654]

It's nice and sad at the same time, what about the real musicians? Is this the end of human art? From now on all the next human generations are going to listen is transhuman music? 🤔

[@MattVidPro]

I think that human art will always have it's place - my definition of art is that it fuels the individual's sole

[@justtiredthings]

how will human art always have a place if anyone and everyone can fake it, and we have no way of distinguishing? this pandora's box should never have been opened

[@Recuper8]

Creative works will be seen in the same way as handmade furniture is seen today.

[@Diogo85]

​@@justtiredthingsAI has existed for decades. It got better.

[@Alex-nu6ug]

This doesn't replace live music

[@cpaps8500]

Wait until openai gets involved with music generation...

[@JohnSmith762A11B]

Only G-rated lyrics allowed and the CIA will have to approve each song.

[@PixelsVerwisselaar]

😮 Amazing. Im testing Udio now

[@aiaudiosecrets]

RIP Beatstars - and R.I.P. any publishing music company on this planet. The only jobs left (beside live music) will be Audio Prompter. One or two for each company.

[@user-xedwsg]

Yup. Thats right.

[@lpanebr]

I'm speechless. I can't stop sending it to my friends

[@Sargamai]

Send them this song too, made with AI: kzbin.info/www/bejne/jJbVeqhqr7GAaJo

[@OverdriveDiecast]

It’s not that it can’t stop , it’s just that it won’t stop !

[@oufukubinta]

Love the Bold and Brash in the back

[@ScottSummerill]

I’m blown away. WTF.

[@ALFTHADRADDAD]

Wasn't convinced at first but that broke my heart was crazy

[@Aldraz]

Guys, just remember, we are living in the future and each day are closer to the picture of a real utopia. In 1000 years, it's likely that this will be noted as one of the very important milestones. Each sense being a huge milestone - text (thoughts, thinking), image (imagination) and now audio (hearing and speaking) and soon video and 3D understanding and generation. After that not much is left for full human like brain simulation.

[@EddieAdolf]

What's interesting it the spectrogram that's scrolling across. I had seen earlier attempts where they basically use image diffusion to generate an image of the audio and then use code to play it back like a needle on a record. If it's literally an image, inpaiting or outpaining or adding new tracks as layers would seem highly likely.

[@cesar4729]

I mean... thats how every AI audio works. 🤷‍♂️

[@AaronBoothe]

As a singer songwriter instrumentalist it’s sad that in the near future people just won’t believe that you created a song without this technology. Also, people who didn’t dedicate their lives to the craft and instead choose to crap out this soulless music will feel that they are now “an artist” and they somehow compete with actual musicians. People want to exist because they feel they have purpose. AI is stealing human purpose in every vocation. We’ve gone too far.

[@vxery]

I’m not gonna lie, I’ve been making music in Suno that sounds just as good as this. I’m confused on my people are so blown away by this. It’s all about the lyrics and prompts you use. Suno is the best AI music generator out there, you just have to use the custom section and actually have good lyrics.

[@bambino3624]

Are you kidding me? Suno has so much more artefacting this sounds miles better.

[@vxery]

@@bambino3624 again, it’s all about the prompt and lyrics you use. It’s not going to get it every time, but I’ve made songs that sound as good if not better then the songs in this video.

[@HCG]

That rap sample is trash lmfao. That’s the best example they had?

[@busyworksbeats]

FIRE 🔥

[@polskizwiazeknaczelnychhej1100]

You are fired (from work)

[@EdGreenTO]

The vocals are really impressive, but on every other song or so, the music feels like it still needs to come along (and these are their hand-picked best renders so far). As for making a whole song at once, I just tried Udio after Matt's vid on that and I really enjoyed the creative process of making parts at a time and then making choices. It might be because they want to restrict server use, but I really enjoyed it and made a compelling piece that I felt more involved in that just the prompt engineer.