"While the chance of this occurring is not 0%, the world will probably not be destroyed by a vocal processor that allows you to reharmonize vocals and change their timbre." But imagine if it did. That is one heck of a dystopian movie idea. Move over, Idiocracy!

You DO know everyone uses melodyne (aka "the pitch editor) before using voice changers, right?

@@GizzyDillespee Right, but the whole point is that this is NOT THAT. This is not pitching a vocal by a 5th, or a 9th, getting weird tuning artifacts and then running it through a voice changer with those artifacts, where you will get an output with artifacts, based on an input with tuning artifacts. Instead, the tuner interacts with the voice changer. Instead of pitching the vocal and running the artifacted version through the voice changer, you effectively have a tool that is *telling the voice changer to sing a different note*. VERY different, in both premise and result. I get that this shift is subtle enough to miss the concept the first time you hear about it, but it is significant. It's what seems like a small change that has BIG implications, I think. I hope that helps! -Justin

Wait... We're expected to watch the video before forming an opinion? This reeks of common sense and reason. Ha. Great video.

🤣 BEST. COMMENT. AWARD.

You just hit a lot of people, pow right in the kisser, of their soul… 😂

Thats painful 🤣

For real my crappy songs can be AI upgrades on Spotify finally:)

For sure!

Yeah, its super useful, at least for sketching. I think they recommend not tuning before upload though.

So you’re not a real artist but want to be ?

@@kojoefante What are you on about, what's your purpose?

Exactly!

YES. That said, if there are excessive effects or roominess there could be some artifacts in how occasional consonants come across. But within reason, absolutely. Even with occasional subtle artifacts around consonants from very poor recordings, I think it’s still more than sufficient for background vocal harmonies. But if the original recording is really bad, it may not be suitable for leads. -Justin

Wait I made it to the part of the video where you discuss this lol that makes perfect sense.

No, this is incorrect. Applying tuning to the output of a voice model is nowhere near the same as telling the voice model to sing a different note. It’s also not the same as simply tuning a vocal, and then inputting it into a voice model. That’s not what this is. That’s why it is new and different and interesting. Please watch the video before commenting 😁

I’m working on it. I think we are close to having him locked in.

@@SonicScoop Well Justin, I know you're going to get it!!

It is. There is also a totally free version that you can use. There’s just a limitation on how much you can use it in a month.

No, you would not run into phase issues, because it’s a totally different note. More than that, it’s a voice model of a singer singing a totally different note. So the whole timbre changes too. The only potential problem is that the timing could be too accurate to sound natural. So you may want to shift the timing. I guess that the inflection might still sound too much the same... So if you want it to be even more natural still, do more than one take, and then reharmonize your second take. To hear the most unnatural version of this, just listen to the first harmony example. It sounds good. No phase issue. It just doesn’t sound realistic. It sounds kind of like a “vocal organ”. To hear a more natural version, listen to the second harmonized example. -Justin

the free version oly gives mp3 downloads--sorry if that's not what you were talking about

The tuning window runs low res FYI

Alvin getting strangled underwater etc - I think you've hit upon it!

I too, am not worried about AI artists replacing real artists. AI generated art generally makes me nauseous. This is not that though. At all. You might be more concerned about text and MIDI to speech synthesizers. But this is not that either. This technology takes a real actual vocal, recorded by a real actual living human artist…. …Then it allows you to re-harmonize that real human performance, and them morph the original voice into other types of voice sounds, or even instrument sounds. This is AI reshaping of real human performances. It is not an AI performer. -Justin

I would classify this as advanced DSP not AI….

No, it would not, and should not In the first harmony example, you can hear exactly the effect that would have. It sounds a little bit like a voice organ. No phasing, but not incredibly “realistic”. A cool sound, just not natural. Using multiple voice models for harmony would be my recommended used case here. And if you want things to sound even more natural human, you might want to shift the take slightly in time. Or do multiple takes of the same line. I hope that helps! -Justin

@@SonicScoop when I tried it phasing was a huge issue…?

Try saying it about 15 cents higher

“98% of people work on farms! Embracing tractors is shortsighted of you! It will put all those farmers out of work! Only extremely innovative and super creative farmers will survive!” We literally went from a world where 98% of people work in agriculture, to a world where 2% of people work in agriculture. And there are is more food and more jobs than ever. Also: AI mastering has been around for like a decade now, and yet I book more and more mastering clients each and every year without really even trying to. This is in part because more people are making recorded music than ever before, thanks to technology that “put people in studios out of work”. -Justin

Yeah thats a deal breaker imo

It could become a plug in. But at the current time, it would have to be a subscription. Unless you have a supercomputer in your house, that could run the AI! Years in the future, it could probably run natively. But home computers just aren’t that powerful yet. -Justin

@SonicScoop even if it's a website, it could have a perpetual license and not be a subscription. It could also be a standalone software that links to the cloud, but isnt in your web browser

In the video, Justin explained very clearly why it is currently almost impossible to use this AI functionality on your own home computer with your DAW 😅. Maybe it will be possible in two, three or four years, provided you have the necessary cash to buy such a powerful computer...😉

A good set of questions on the legal issues. That’s definitely something to look into. I’m not sure what the terms of use are. I don’t think it’s like a social media company, where they claim partial ownership to your creations. But definitely go ahead and read the fine print! The last part about power consumption I can’t help you with. There are all sorts of ways to get power, some of which are cleaner than others. There are extremely clean ways to get power, like nuclear, that people are afraid of-often somewhat irrationally. And, there are dirtier ways to get power, like fossil fuels, that allow energy to be inexpensive so that human standards of living can rise-especially in the most impoverished places. There are clean forms of energy that are intermittent, like solar and wind… Which also have an untold carbon footprint and environmental impact per kilowatt hour that come just from building them. There are clean forms of energy like hydroelectric that are pretty cool, but aren’t available everywhere, and sometimes wreak havoc on natural habitats around them. Then there is the question of whether or not fossil fuels are one of the only things preventing us from entering into another glacial period in our current Ice Age! …for which we are a bit overdue already. (We are currently in an interglacial period of an ice age. And glacial periods of ice ages are way scarier for humanity than moderate warming! The earth has recently been warming-and greening-which is arguably better than freezing.) All of that is probably above both our pay grades! But it is an interesting conversation, and one in which there is rarely sufficient nuance in context, as there are so many vested interests at play. Yes, data centers use power. But the cost of that power usage are distributed among tens of thousands or even millions of users. If they weren’t power efficient relative to the value that they put out, they wouldn’t be economical to use. There are cleaner and dirtier types of power people can use. But people tend to use whatever amount of power is economical for them to use. What they use it on, and how it is produced, is another set of questions. I don’t have much control over that part myself.

This isn’t exactly that. This is taking the performance of a real musician, and re-harmonizing that performance, and/or changing the texture of the voice to a different voice, or even a different instrument. -Justin

https:/:audimee.com but it’s browser based because it has to connect to cloud computing. It’s not a download.

I’m not a fan of auto tune effects or pitch correction, in general. That’s why my recommended use case IS NOT THAT. You’ll notice that in all of the examples I did in real time myself, I did NOT pitch correct the vocal. Rather, I used this kind of pitch correction to REHARMONIZE the vocal. Without touching the pitch as far as “correction” goes. I prefer this use case to light Melodyne. I actually have MORE of a problem with “light” pitch correction using Melodyne or auto tune. In that use case, you are sucking the humanity out of the vocalists performance. In my recommended use case, you are PRESERVING the humanity of the original performance and running it through interesting textural effects and spinning it out into new and interesting reharmonizations. I’m actually more hardcore about musical purity on this than you are, as I kind of hate light Melodyne :-) No! I want to hear real vocal takes. My recommended use case preserves the performative essence of the original vocal take from a pitch perspective. I hope that makes sense! -Justin

Nope, wrong, sorry :-) Melodyne is an more powerful tool for fine vocal tuning, duh. But that’s not the point. The point is that this allows you to radically reharmonize parts with extreme vocal retuning, with NO TUNING ARTIFACTS. Melodyne isn’t going to allow you to change your pitch by a 5th or a 9th or a 13th and have it sound just as natural as if you were tuning by a half step. THAT is the breakthrough here. Pay attention! :-D

@@SonicScoop Now I've read your pinned comment. At last haha))) Sorry for disturbing . Now I understand. Thank you.

Tractors put close to 100% of farmers out of business. The US literally went from having 98% of the population work in agriculture to 2%. The only thing wrong with this analog is that AI is unlikely to have the same effect in music. -Justin

My recommendation with this kind of tool is to preserve the honest inflection of a vocal performance, and instead of using tuners to “correct” performances, use them to REHARMONIZE performances without making them unnatural. My recommended use case is NOT pitch correction. It’s pitch Re-composition. THAT is the new part. Pitch correction is old tech at this point. How would that be new? John Lennon and Kurt Cobain both did vocal double tracking-a relatively new technology in the Beatles’ day, which makes the voice sound more “synthetic” in the literal sense of the word. If Lennon or Cobain could take their doubled vocal take and reharmonize it, using a totally different vocal texture they could actually make it sound less “synthetic”-by having the double be a different voice timbre. Or they could have it sound even more effected by instead using or creating a totally new and interesting instrument or voice model of some kind. Would the Beatles or Nirvana have used this technology is it was around in their day? The Beatles, absolutely!! Are you kidding me?? They were pushing the boundaries of the most cutting edge music production technology of their day, and using it to make new sounds that they weren’t even designed to achieve. Nirvana? Maybe not. But if you don’t think George Martin and the Beatles would be using this stuff, you’re either not thinking it through, or haven’t studied the Beatles very much. I hope that helps add sone perspective to it. -Justin

@@SonicScoop Thanks for sharing your thoughts

I just don’t see a processor that allows you to reharmonize vocals and change their timbre as the death knell for jobs in music. Not all AIs are the same. -Justin

I’m looking at synthesizer V right now, and it appears that it is a text and MIDI to speech synthesizer, exactly as I described! Are you saying it has a audio to midi function that allows you to use monophonic audio to extract MIDI from? If so, that’s neat, but it can’t preserve the exact phrasing, pitch and diction of the original voice that way. That’s a very different kind of technology. What you’re describing is taking MIDI (possibly MIDI notation extracted from audio?) with text and creating a newly synthesized voice from scratch that way. While in the same general ballpark, kind of, this is also VERY different. Services like Audimee preserve the original inflection, pitch and diction of the original performance, and remap it to a completely different tone. It’s a different process and end result. Both have their uses, but they are NOT the same. Hope that helps! -Justin

I’d imagine English is the biggest, most lucrative market, but once these tools become commercially viable there, I’d imagine they could quickly expand to other languages. -Justin

I think the earlier versions from other companies you are referring to are a little different. They are going to extract midi notation from a monophonic performance, and then use that midi notation to trigger some other type of instrument. This is different. It preserves the exact original inflection, timing and pitch of the original performance. It’s not converting to MIDI at any point. It’s kind of adjacent, but not the same thing at all in process or result. It definitely allows things that couldn’t be done before. I don’t think that Neutone is doing quite this at all. I hope that helps make sense of it! -Justin

@@SonicScoop Video describing what it does: kzfaq.info/get/bejne/o7WEiLF9mL6zoac.htmlsi=E9YcxonjFrdeMqdf It’s def reacting to your voice not midi What your describing sounds like dubler 2 More examples: kzfaq.info/get/bejne/ea9nbNSf2NqqfmQ.htmlsi=8kdAFdLxXdl7iUAy

The recommendations I’m making here ARE for people who can sing and play though! My recommendation is NOT simple pitch correction. I didn’t even really show pitch correction. The use case I recommend is vocal reharmonization off of a single singer or a single performance. Nowhere did I encourage simple pitch correction for people who can’t sing well. That’s not even what’s new here. That’s old tech already at this point. My recommendation is for using tools that could be used badly in creative and interesting ways instead. Watch the video! :-) -Justin

Why not?

@@SonicScoop I think AI is not like a tractor, but more la the combustion engine, and will revolutionize every single industry with comunded and very unpredictable effects. In all seriousness, AI (AGI, then ASI) might just be the last human invention, and that is the good scenario.

​@@SonicScoop It more comparable to the discovery of oil and the industrial revolution. ChatGPT was launched on November 30, 2022, which means we are not even 3 years in. Imagine the AI we have now is like the first steam engines in Scotland...

He's alluding to Luddites in England. It's apt.

​@@ThisGuyDude So is The Luddite Fallacy.

This can do that. You just have to use different voices or different notes. Phasing is not possible in that context. -Justin

Yeah, this isn’t that kind of AI. Please watch the video before commenting. 😆

A vocal processor that allows you to reharmonize vocals and change their timbre is going to obliterate mankind?

@@SonicScoop It takes part in destroying the craft and world of music making that you and I have been working in. One plug-in will let you skip hiring a singer, another will let you skip hiring a mixer. And some plugins will let you skip the whole chain of artists from the composer to the mastering engineer. I don’t see a way to stop it, I just find it strange to see you celebrate one of these destructive programs. Of course it’s possible to use it in a creative way, but we both know it will mainly be used to cut out yet another human element of music. It is so sad that I don’t have words for it. I have enjoyed many of your earlier videos about mixing, though, so thanks for that.

If you don’t think hat is sufficient voice change quality from female to make to use background vocal harmonies, I don’t know what to tell you, other than that you’re nuts! :-) Listen to the harmony example at 6:07 to hear what I mean. Absolutely suitable for this purpose, which is the primary use case I recommend. The other use cases I recommend are replacing/blending vocals with intentionally effected instrument sounds, and reimagining and retuning/recomposing vocal samples. Straight vocal replacement outside of sing demos is not one of my recommended use cases. That said, if you listen to the demo for each voice, it’s obvious that this can be done effectively with a little effort. It’s just not what I would ever use it for. -Justin

@@SonicScoop okay your comment is a bait-and-switch. What you said is that's a convincing change from a female to a male vocal. It's not. It's not a change from a female to a male vocal by any stretch of the imagination. If you want to change the goal posts after the fact to indicate that you meant a thicker female vocal that could be used as a background or a harmony then that's a completely different equation. But not what you stated in the video.

You don’t like the idea of being able to reharmonize a vocal take and change its timbre? How is that bad?

@@SonicScoop no, I don’t. I’m old school. I prefer to sing the parts properly. I think it’s all very deceptive.

​@@officialWWMthen go back been old man stop been nuts

*being

@@dgunzbeatz this difference between you and me is that I can actually sing! And spell.

There's more to music than singing. I use this to sketch vocals, which can then be replaced by a real performance later on. If you're a song writer but not a singer, then this is an excellent tool. What's interesting about Audimee is that anyone can create a voice profile, which may create a new venue for collaboration between writers and singers.

@@svensvensson8102 Yes, I'm (probably not obviously enough) talking in broad stroke generalities. Tech "gizmos" have actually RUINED the music biz, meaning, providing the most actual living wage jobs for the most number of - or let's suppose - deserving people. I think MUSICIANS (inclusive of singers, of course) deserve to be on the receiving end of the bulk of the work, $$ and accolades, etc. Speaking of non singer songwriters ( whom I greatly respect -- if they're actually good) My grandmother turned down Sammy Cahn when in the Great Depression she was renting a room and he presented himself as a would be renter, and she did not "believe" he was who he said he was... and I'm a big fan of this NON SINGING mostly lyricist (Top Five, Hall of Fame type) Anyway, yes, for certain "placeholder" applications, have at it. I'm not going to STOP, actually - any of this, of course, I'm just pointing out that music - the non hobby true business of music, is and should be the purview and realm of musicians and not people who can figure out how to push the "talent button." EGADS. Thanks for the response and Cheers!!

But, are you watching the video? Because this AI requires singing skills in order to use. You don’t input text and MIDI, and get a vocal performance. Rather, you take an existing human vocal performance, and can reharmonize it, or change its timbre. It’s like 33% of people aren’t watching the video before commenting 😆

@@SonicScoop Yes, I GET that. Chaka Khan was the first to really, or most prominently, do her own backgrounds, in the Rufus days in the seventies,. Her's is the sort of voice you might want to hear in triplicate. Really a pass was extended, IOW, public accepted it, didn't fully love it. It was, at the end of the day seen as a bit gimmicky, but - best case - perhaps the exception that proves the rule. WITHOUT EXCEPTION backgrounds benefit from the richer blend of timbres of different voices. I've done my own backgrounds, it's a compromise, no matter how good the performance/post production. Aesthetically NOT recommended. The beauty of the choir or ensemble, is that it is NOT the seam sound in triplicate. This is the problem with "the toys." It happens, all the time, in the video realm, as well. Notice now, how ALL commercials look like the same production house...gimmick rich, (cue in the drone shot, now give me that fast surreal zoom) but bland as hell..just lack of basic variety......Don't stand up as a bona fide engineer (and you are, I watched your videos prior) and take basically an anti art position in advancing "the state of the art."

By letting people know there’s an interesting tool that can let them change the tone of their vocal performances and/or cleanly reharmonize them?

I don’t make videos for short attention span people. I try to attract an audience who likes to understand things deeply. If that’s not for you, that’s ok.