I will continue to do my own Metal vocals. It is more distinct and unique that way. It is amazing what AI can do but we might find everyone's metal vocals sounding the same when people start using AI.
I agree with you hundred percent. Plus I like doing extreme vocals in Person around ppl and getting the wow factor. You won't get that with ai. Ai is more about 💩 and giggles.
2:34 "so if you hate the sound of your voice, this is what you need!" So you need Audimee, not practise? It saddens me that anyone who follows this route will miss out on the best part of becoming a great singer. They will miss out on the journey and the deep satisfaction that comes from growing as an artist through practise, passion and dedication to ones craft.
No amount of training can make your voice sound like someone that you're not. If I don't want to sound like myself, then I'm gonna use technology to get the sound that I want. If I want to sound like a Japanese girl, for example, this is currently my best bet -- and I'm not missing out on anything whatsoever by taking advantage of it.
@@HyperNova137 Sure mate, I can understand you using it for that reason. I have on occasion manipulated a guitar to the point where it no longer sounds like a guitar, so why not completely change a voice? I guess this is one of many paradigm shifts I will have to accept as this technology becomes commonplace. I also believe there is no right or wrong in music and I have no right to tell anyone how they should make their art. So I apologise if my comment came across that way. My comment was a reaction to some of the messages in this video which I do feel are different in essence to your idea of changing your nationality and possibly gender(?) through Audimee. Particularly when Bobby recommends this for those who hate the sound of their own voice I feel that is maybe not the best advice. Being an obsessive reader about my favourite musicians I’ve discovered it’s surprisingly very common for even the most incredible singers to hate the sound of their own voice. A prime example is Rob Halford, one of the most iconic, original and influential metal singers of all time. Thankfully his voice was never changed beyond recognition through AI because millions love and connect with it. Who knows, if someone hates their own voice maybe they are onto something very unique. It doesn't necessarily mean they are a bad singer. I think we are usually the worst judge of our own voice. And this product is being sold as a solution for those who “suck at harsh vocals”. This is a skill that can be improved through learning technique and practising. A singer still may never sound like Audimee or their favourite extreme vocalist but that could be the best thing they have going for them. I absolutely love the process of practising and growing as a musician and reaching my goals through hard work and dedication. Audimee gives the destination without the journey, so that naturally holds no appeal for me. But I do see it’s not right for me to suggest others are missing out if they don’t feel the same desire I do to take that journey. I think Bobby may well be right in saying the use of AI will become standard and if you don’t get on board you’ll be left behind. But then again, what a great opportunity to stand out from the crowd and be the one who doesn’t use it. All food for thought anyway. One thing’s for sure, we are about to see some massive changes in the music scene. An interesting time to be alive.
I agree in case you are a vocalist. If you're a songwriter and you wish to pitch your ideas, this technology well be indispensable. However, it's not without limitations. I heard people use Ariana Grande's model and it sucked! Because they fed melodies to it that didn't sound like her choices, neither did the rhythm fit her voice nor the emotional impact. But there's a teaching moment here: if you lack emotions, flow and interest in your voice AI can't produce it for you.
@@ErikAndersdon’t worry, there will be bands that are nothing more than models who perform live to backing tracks that AI wrote, played and sang in their entirety. People already pay big bucks to see pop stars dance to tracks without actually performing music. The bridge to human beings who pretend to perform to AI tracks isn’t that big. The future is bleak
To me the cool point about music is….making music. Surely there is room for AI generates tracks just like drum libraries have been around since basically forever by now - but I‘d never replace my singer by an AI.
@@botengjoe8715not everyone can sing. Not everyone can play every instrument. Not everyone can afford studio musicians. That is why we have sample libraries. That is why we have physical modeling synthesis. We have been using AI in music for over fifty years. Get over it.
They're not that great. You can see the real vibration in the growls when he's done. I sound like the original, but I cannot do the end product. It's a good tool for those of us who can scream mediocrely and just can't quite get there but still enjoy making music.
Pretty creepy actually, but the lid has been removed and there’s no way to get it back in the box. It will make you wonder if anything is being done by a human anymore.
If you don't have access to a metal vocalist, this seems like a really powerful tool. Or, if you are a vocalist, this seems like a great way to double a vocal performance for added chorusing/texturing.
Old school early 90's production death metal trick . Pitch your full session up a couple have steps and sing your death metal vocal then when you pitch it down to its original key you have a nice natural warm and deep sounding vocal
If the choice actually does come down to getting on board with AI or getting left behind then I think I’d rather just be left behind then. Thats just me though.
I don't think singers are going to be left behind any time soon. I can't sing, I have tried the trials of four midi+lyrics to sing already, and they all fall flat specially for rock music. EDM or pop might be, but not for rock. Even smiling while singing changes your voice, the tension ion your mouth, neck and throat... I think that, first, there is a lot of things missing there, second that music to dance to is not the same than music to listen to, if it makes sense. I think that connecting with a real singer is still important. The nuances of emotion that a singer can do go beyond what we can pretend with programs. Then again, I would like to, at least, have an approximate idea of how a song I write would sound without finding a vocalist, the hours of recording in a studio I don't have access to and so on. On the other hand, producing music is a hobby to me, so why would my opinion matter, anyway?
5:00 your recording is clipped. Can't you hear that high end clipping going on? The AI version is NOT clipped. Actually, the AI version I like better. But godz, you love to hear yourself talk, man. Just play the damn files - THEN explain. Everyone who can't, thinks AI will save them. hah
As an engineer/producer who works primarily with acoustic and electro-acoustic instruments and mostly with artists who want to maintain a significant level of realism in their recordings, I could see using this tech to layer in under the actual performed vocals, in the same way that drum samples and impulse responses have long since been commonly used. Also useful for live reinforcement of performances as a backing track. From a broader creative standpoint, it would be fun to hear how far some of the parameters can be pushed for fantastic effects. To note, these synthesized voices have the same tale tale sound that other AI vocals do: a particular ringing echo-like sound. It’s very subtle, but once I started noticing it, I can hear it on all the AI stuff. Especially higher voices. I know most people don’t hear it, so from an audience acceptance standpoint, it probably doesn’t matter. And it will probably be refined away as the tech improves. Very interesting…and challenging us to question what our role will be moving forward. Thanks for the look at this tech 🙏🏽
So you have to be able to sing to begin with to use this AI? I’ve been using Suno to get some vocals on my songs that have been unfinished for years but the results are really unpredictable.
I've heard the scream type of vocals done by AI a few months ago for the first time. And it did surprise me... so this is not new to me. But also what is the application of this really? How do we benefit? how do we apply this in a live setting? have a backing vocals for a live band? It's fun to mess about with this but ultimately I find it useless
What I’m trying understand is after creating the voice model, how “off” can the voice that you feed into the model be and have the return still sound usable?
The terms "copyright free" and "cover" used in the same sentence says a lot about AI. Maybe someone should hack Audimee and throw it on Napster or Pirate Bay.
The Audimee voice sounds like it's false fold, while the original vocalist sounds more like fry or fry dominant hybrid screams, especially in the "pitched scream singing" section
What i think for now is for producers to use this as a form to make layers. I could see a producer mixing after a session and add these to the mix because metal vocalists dont really ever use one layer anymore. Would have loved to hear a track with all the layers and mixed ones to see how it could fill out tones
Meh~ it has to start harsh to start with presumably. Lots of artifacts in some of it that sound like FFT phase weirdness. I suppose in some situations. Thing is as a music fan and a musician I end up battling people constantly who don't understand tools like this over things like the accusations about tracks live and "autotune" and this just makes the problem worse. Just like with autotune "hey do you suck at singing? You too can make a hit record!". Getting all excited about this just makes it worse for real harsh vocalists who can't get away from accusations that they aren't really doing it. Especially the women in metal. "No way does a woman sound like that". A few can and do. These techniques take years to learn and perfect so no it's not a good thing. If you suck at harsh vocals maybe you shouldn't do them because faking it is fake period and by using this you crap all over people you probably admire by promoting something that is fake just to push yourself into the business IMO. No you aren't the same as someone who has been doing it for real on stage for 15 years. People who don't know better may think so, but what are you going to do when you can't do it live? Pop singers have a really bad reputation at this point for faking it, lip syncing, doing everything with tracks, not being able to sing with out autotune. Great you made a metal record with AI now what? Metal fans aren't like pop fans. If they find out you're faking it they won't let it go so you shouldn't be glad that you started out in your home studio making fake voices because if it launches you into something else, you can't live up to what you created.
Sounds surreal! It seems like a perfect songwriting tool for conveying ideas to vocalists, saying 'My vision is to capture that performance and grit in this song.' On the other hand, it might discourage many vocalists who feel their vocals aren't at that level, leading songwriters and producers to keep searching and ultimately hiring a pro, which isn't cheap. Personally, I'm looking for AI singing and voice cloning plugins that effectively mimic characteristics, emotions, and human factors, and work seamlessly in a native DAW environment as third-party plugin.Any recommendations?
I still think the original voice sounds better in this example but I do have to say that the AI did a pretty good job. I used one of their AI female vocals as a backing voice in one of my songs. I am not a singer so I just talked and I have to say it not only done a good job with the voice but it done well with Polish language as well. There were some frazes that didn't sound quite on point but overall it was acceptable. I'm yet to try it as a main vocal but like I said, I am not a vocalist so getting a good performance is a bit tricky. 😆
that's the correct use IMHO. get something to work with, then find a vocalist that can do a decent job (it's too hard to find perfect). wish I could still sing and do vocals similar alright Myself but after years of chain smoking...well...it sucks but I can't pull it off anymore 😕
For me you can use this as a hack tips to put your vocals much bigger into the mix. For those who doesn't want layering vocals with every take and to save your time. But the results will be different. I already used this technique quite often but I still recommended you to take every possible vox you want to take, because there is a element that you need from the samples. Good luck! And good content! :D
I made an RVC Model trained with a couple of vocalist and it sounds something like this (but you can run it for free!), and it took like a couple of hours to train
For me it's a game changer . I have a lots of songs I have the vocal melodies in my mind . But without a "pro" singer they are just plain music, not too interesting to show or share anywhere. But with this I can make them like a "finished" products. And I think this is all about. Especially if you don't have a "good" vocal voice . Like me . Very hard to find someone who can translate your ideas the way as you imagined. It's not gonna replace a real singer in any way. But you can showcase your ides and songs and that's a huge step forward . Or a big help to create your own music
Everyone worried about this needs to remember that the AI is "replicating" a style of vocal, and not creating a new voice. Voices like Ozzys, Chris Barnes, and Tom Arayas are unique metal voices that will always be relevant, so if you're afraid of getting "left behind", then BE outstanding. In other words, the cream rises to the top as usual. And as good as the AI is, it'll never replace the real artist. IMO
I think this could be good for me since I don’t know how to scream and don’t want learn to just for a hobby. Hopefully it can take my “whisper screams” and turn them into decent vocals
So, you have to have a file of good growling already to be able to convert it to good growling? Why can't you just talk into it or use Synthesizer V or something?
This is awesome for the solo producer or hobbyist. You can create background vocals done by five, six, or 6000 people. Unless you’re recording something with one microphone like Robert Johnson, you’re already partway down this path anyway. You might as well lean into AI it’s not going away.
i've used this before, its good for layering , and i've also used another program and trained my voice with a different voice model which is prettycool and other programs give you the option to blend two voice models together
Dude... voice modulators for screaming have been huge since 2008 and probably a few years before that. Still fly by TDWP (Big tymers cover) used voice modulation.
I don’t agree that you’re making “your” music sound pro, unless you’re using it as a tool for writing, NOT a tool for doing vocals that you can’t pull off in real life (good luck performing it live which is where the money is) this only helps the people who want to make music but don’t have the skills. They still need to learn it to perform it. This was okay in the days of people composing things alone it’s okay to be a good composer and not be able to play anything. Maybe you just have golden ideas. But here, the ideas are not your own, so I personally don’t understand why you’d even want to use it. That being said, I’m not a drummer, and I’ve been using Logics AI drummer since it came out like 6-7 years ago I think, maybe longer. And suggesting settings in plug-ins and what not? We already have presets for that. Presets styled by professional humans. All the AI is going to do is tweet things to make sure that your inputs and outputs are set correctly. That way you don’t have too much gain reduction happening on a setting that is only supposed to be hitting at like four or 5 dB of gain reduction. I will speed up the process on mundane things and make it easier for single musicians to produce demos, but people will always like real music, no matter what I will always be looking for something that’s made by humans that comes with an aspect of human experience. People don’t latch on to bands & make them superstars because their music is good necessarily they do it because they relate to the story and relate to the image and message of the band. I cannot re-create that, nobody’s going to “worship” AI like they do pop stars. No one is going to aspire to want to be like AI or want to dress like AI in case you’re not aware for last 50 or so years or longer music has influenced every single cultural and fashion trend, and is a huge part of how people identify in their lives, what music they listen to and what artist they like.those, will never replace that. You want a good answer on how not to get replaced by AI as an artist, just ask AI ha ha it’ll give you a beautiful answer on exactly what it cannot re-create. And in short, human story and human experience popular artists have a way of tapping into the current subconscious of the public, I can’t do that because it first needs to study what’s going on and create models. And that’s an answer from ChatGPT itself.
🙋 hi! I already do all of that. I write my own lyrics and all that. Anyways when it comes to EDM and dubstep I have a little more freedom to remastering each song but it's all good. If AI music isn't your jam that's cool but it's what I do. It's like 1/10th of the total video creation. I don't do cheap prompting and uploads like lots of others so. 😂 I put a lot of work into my music.
I can’t do vocals in any style that someone would enjoying hearing. I have a couple people that will do stuff for me. They are often busy with their own lives. This could be useful to get the ideas I hear down and send them something basically done and save tons of time.
I can see that as a voice enhancer tool that is not as cheating as autotune stuff. It picks your actual singing performance and just change the voice to something more palatable to the ears. I bet that before AI there was tools to do some of those tricks with analog or maybe digital means. AI is just more powerful
I tried Audimee and it was useless. I was looking for a powerful rock vocal, and it couldn't do it. It didn't sound very different to my own singing voice, which is crap. Even when I set it to convert at 100%. I'm surprised I couldn't get a better result because I've got some excellent results with Suno ai. The biggest problem with suno was, it didn't follow the tags I set. The tech shows great promise, but it's got a long way to go yet.
I would like an AI that creates demonstration videos of plugins and new audio technologies without talking so much about them and that makes the results heard without having to be hostage to a suffocating and useless narrative
TH-cam recommended me this video after I liked a video showcasing what is actually a separate offline tool where you can make vocals using the approach described below (now hold on really tight because it's really CREEPY compared to this). The tool I'm talking about is ACE Studio, where you just import a MIDI track of a melody you want the lyrics to follow, then you... WRITE THE LYRICS inside the markers showing what is what is the pitch of the sound in the imported track and it converts it into singing vocals using the voice selected from the library of vocalists, ZERO SINGING SKILL REQUIRED! 🤯
I liked the original high vocal. From this video it seems like it's just eq, and not really A.I. I thought you were going to take a singing type vocals and turn it into metal. Additionally, most DAW's these days have ways to add distortion and bring vocals up and down octaves.
sometimes I like to add in my mix like example female voice AI generated from my voice. I already try in different website and the result was very satisfied for me. BUT the language on vocals was not English. Most disappointment a least for me was change on octave voice in original vocals generated to AI , The AI is just keep same octave on voice.
The real problem is where our money will come from. Pressure needs to be put on governments to adapt; technological movement is inevitable and potentially good. Elon musk rightly Reframes the money issue in terms of Universal High (vs basic) Income. The real question for an artist is why do you do what you do. AI may, in regard to music, take away (or actually amplify, at least in the short term, if you know how to use it right) your: money, glory, attention, feeling needed by others. If you make music for your own personal sense/experience of achievement, growth, beauty or therapy, then relax and continue on!
Mhh I‘m not convinced because the usecase should be bad amateur vocals as input and good screaming ai vocals as output. The input vocals are already good so I don’t see much sense here in using AI
The original vocals sound more primal. I prefer the original. You could use ai for training your vocals so that you try screaming at different octaves to see what works. I wouldn't want the ai to do all the work though. As a musician creating music is my passion.
But now can it replicate?Something as beautiful as the high and vocalist of symphonic heavy metal with female front singers, Like Sharon den Adele and Charlotte Wessels.And Simone Simons and on and on?
Not so free to check the ragnar voice)) I dont know will it works good for me. From your video I sow that you dont actually need to use AI couse your raw sound is nice enough to use it it self.
It doesn't matter at all if the song was composed by a real band or a kid in his computer. The only thing that matter is if i like the song or not. And btw, already exist AI that compose full songs really well. I tested myself by prompt. So...... We are entering in a new era hard to read yet.
I think it should be upto the artist if its used and always disclosed in some way if AI is used as its not entirely ethical to use it without the artists permission.
It already can replace some modern metal vocalist, cause they sound the same. But I highly doubt that it will replace someone like Corey Taylor, Winston McCall, Sully Erna and so on
I think there's likely to be a limitation in what the audience accepts. The audience has a long history of disliking artists who sound terrible live vs how they sounded in the recording. That took a couple of additional disliked steps when bands did lip sync, or used auto-correct for anything other than fx on a tune. I don't think that having an AI do the various artists' roles is going to be useful for much more than a brief minute of entertaining friends with a recording before they skip to the next tune.
I think thall should have different type of vocals than the ordeniry metal screaming (including animal, guttteral, tunel throath and so on). I mean new experimental technic.Maybe AI could be useful in that case. It's like got the metal sound of 21st century, but the vocals are stuck in the 20th century. Hope you get what i mean.
Yeah get it to work on the Grimm vocals also.(Immortal, Dimmu) Not just brutal and deathcore. And for a true test, see if that thing could pull off Behemoth.
I'm in the same boat as you are. I've made far more minor videos on IG and FB about this, and I had a lot of pushback. To me, I get it... but whether I feel that way or not, the tech is moving forward. Maybe this will kill music, maybe it will reinspire people in different ways. We have yet to tell. I figure I will learn a ton about it so that either a) I can be ahead of the curve if this is where music goes, or b) I'll already know AI, so if this job goes away I have a foot in at the new one... 🤣
I see how this can be awesome for making demos. Does that mean you can record your screaming vocals in a normal speaking voice at say… 2:00 am and not freak out the neighbors and I can demo songs? And I did click on this because the title said “free”. Having something free for a limited time doesn’t make it free. I probably will try the app but won’t use your code for it.
Free accept that is probably uploading your performances without your permission... Then elements of your voice profile interacting your performance composition and recycling them including a compositions with Maybe enough changes to avoid copyright infringement.
this is so misleading. The sample used to create Ragnar's parts already sounded good and metalish so ofcourse that's what you get. As soon as you sing out of tune of terrible and use these as samples you'll get absolute shit for results.
We got a good decade or so before AI even approaches what a real musician does. It always lacks tonal dynamics in comparison to the real thing. Now that I think about it, that could be super useful in certain genres.
🤑 Use code "RAYTOWN" for 20% OFF Audimee for a LIMITED TIME (links in description)
(and grab my FREE vocal delay secrets book in the description 🤘)
I don't know man, I want AI to do the dishes and do my taxes while I make music and art, not the other way around...
There's a line from an old Flock of Seagullls song: "Man made machines make music for the man
Now machines make music while the man makes plans"
You can do that, indeed, people use AI for bussiness mostly, the artistic part is just for fun for now
Veeeerrrryyy nice idea👍👍👍
watch...there will be ai tax next🤦♂
💯% agree
I think it would be more impressive if the original track was not already a scream but a clean vocal and then Audime translated it to a scream vocal.
I would say thats pretty much impossible, but 😅 who tf knows at this point
for sure
Let me know when the Axl Rose or Sam Kinnison ai voice is ready.
@@NoiseFromTheAtticyou can make them yourself. Theres a free RVC software . Again it's not gonna be perfect but be good for fun cover versions.
@@armondtanzwhat is it called ?
I will continue to do my own Metal vocals. It is more distinct and unique that way. It is amazing what AI can do but we might find everyone's metal vocals sounding the same when people start using AI.
I agree with you hundred percent. Plus I like doing extreme vocals in Person around ppl and getting the wow factor. You won't get that with ai. Ai is more about 💩 and giggles.
To me, the joy of music is still in the actual playing and singing.
I agree with you.
Ok, but soon you'll easily be fooled thinking it's real while it's not.
@@ExiLeZH what does that have to do with what I wrote?
@@mrcoatsworth429 sorry, was not sure if you meant HEARING somebody actually play/sing, or doing it yourself. Nvm my comment if it’s the latter.
Totally agree, AI is cheating and has no soul
2:34 "so if you hate the sound of your voice, this is what you need!" So you need Audimee, not practise? It saddens me that anyone who follows this route will miss out on the best part of becoming a great singer. They will miss out on the journey and the deep satisfaction that comes from growing as an artist through practise, passion and dedication to ones craft.
No amount of training can make your voice sound like someone that you're not. If I don't want to sound like myself, then I'm gonna use technology to get the sound that I want. If I want to sound like a Japanese girl, for example, this is currently my best bet -- and I'm not missing out on anything whatsoever by taking advantage of it.
@@HyperNova137 Sure mate, I can understand you using it for that reason. I have on occasion manipulated a guitar to the point where it no longer sounds like a guitar, so why not completely change a voice? I guess this is one of many paradigm shifts I will have to accept as this technology becomes commonplace. I also believe there is no right or wrong in music and I have no right to tell anyone how they should make their art. So I apologise if my comment came across that way.
My comment was a reaction to some of the messages in this video which I do feel are different in essence to your idea of changing your nationality and possibly gender(?) through Audimee. Particularly when Bobby recommends this for those who hate the sound of their own voice I feel that is maybe not the best advice. Being an obsessive reader about my favourite musicians I’ve discovered it’s surprisingly very common for even the most incredible singers to hate the sound of their own voice. A prime example is Rob Halford, one of the most iconic, original and influential metal singers of all time. Thankfully his voice was never changed beyond recognition through AI because millions love and connect with it. Who knows, if someone hates their own voice maybe they are onto something very unique. It doesn't necessarily mean they are a bad singer. I think we are usually the worst judge of our own voice.
And this product is being sold as a solution for those who “suck at harsh vocals”. This is a skill that can be improved through learning technique and practising. A singer still may never sound like Audimee or their favourite extreme vocalist but that could be the best thing they have going for them. I absolutely love the process of practising and growing as a musician and reaching my goals through hard work and dedication. Audimee gives the destination without the journey, so that naturally holds no appeal for me. But I do see it’s not right for me to suggest others are missing out if they don’t feel the same desire I do to take that journey.
I think Bobby may well be right in saying the use of AI will become standard and if you don’t get on board you’ll be left behind. But then again, what a great opportunity to stand out from the crowd and be the one who doesn’t use it. All food for thought anyway. One thing’s for sure, we are about to see some massive changes in the music scene. An interesting time to be alive.
I agree in case you are a vocalist. If you're a songwriter and you wish to pitch your ideas, this technology well be indispensable. However, it's not without limitations. I heard people use Ariana Grande's model and it sucked! Because they fed melodies to it that didn't sound like her choices, neither did the rhythm fit her voice nor the emotional impact. But there's a teaching moment here: if you lack emotions, flow and interest in your voice AI can't produce it for you.
@@drrodopszin A good point. Emotions, flow and interest are vital to a good vocal take.
@@HyperNova137 you're not even responding to what they're saying, they didn't say anything about sounding like someone else
"oh boy... i don't know how i feel about that." pretty much sums it up for most of us. amazed and terrified all at once
Might as well generate the full track tbh
i wasnt scared until two days ago a punk track on udio i generated was.....shockingly good. the guitar had tasty bends. absolute facepalm. fuck.
I can’t wait to see AI live….. 😂
@@ErikAndersdon’t worry, there will be bands that are nothing more than models who perform live to backing tracks that AI wrote, played and sang in their entirety. People already pay big bucks to see pop stars dance to tracks without actually performing music. The bridge to human beings who pretend to perform to AI tracks isn’t that big. The future is bleak
Udio?
well some services does that
Thank you so much for using our song in this awesome video!
To me the cool point about music is….making music. Surely there is room for AI generates tracks just like drum libraries have been around since basically forever by now - but I‘d never replace my singer by an AI.
Good for you. I have no singer in the first place, and that need brought me exactly here to this video.
@@andrew_kay well then - awesome that it helps you making music! Same with drums/ drum libraries for me.
@@andrew_kay become your own singer
@@botengjoe8715not everyone can sing. Not everyone can play every instrument. Not everyone can afford studio musicians. That is why we have sample libraries. That is why we have physical modeling synthesis. We have been using AI in music for over fifty years. Get over it.
How do we know you're not AI Bobby ?
The original vocals are pretty good... probably not perfect, but I'm not sure I understand why someone would want to change them with A.I.?
They're not that great. You can see the real vibration in the growls when he's done. I sound like the original, but I cannot do the end product. It's a good tool for those of us who can scream mediocrely and just can't quite get there but still enjoy making music.
Just play it man holy hell all the yapping
Pretty creepy actually, but the lid has been removed and there’s no way to get it back in the box. It will make you wonder if anything is being done by a human anymore.
Literally anything. You could be a bot for all I know, or myself...
@@toddhatfield5329 ha! very true but not a bot here, just a dude that likes riffing and recording!
Humans still have to interact. I used this to help with my album I just released yesterday so check it out.
I see it as being great for backing vocals. No matter how many takes, and how many harmonies I sing, it still sounds like a choir of Daves.
If you don't have access to a metal vocalist, this seems like a really powerful tool. Or, if you are a vocalist, this seems like a great way to double a vocal performance for added chorusing/texturing.
Old school early 90's production death metal trick . Pitch your full session up a couple have steps and sing your death metal vocal then when you pitch it down to its original key you have a nice natural warm and deep sounding vocal
Yep. Especially on tape, it makes everything sound heavy af. It also emphasises the natural timing variations of real musicians.
Is there any tutorial showing that?
If the choice actually does come down to getting on board with AI or getting left behind then I think I’d rather just be left behind then. Thats just me though.
You’re not alone, friend.
I don't think singers are going to be left behind any time soon.
I can't sing, I have tried the trials of four midi+lyrics to sing already, and they all fall flat specially for rock music. EDM or pop might be, but not for rock. Even smiling while singing changes your voice, the tension ion your mouth, neck and throat... I think that, first, there is a lot of things missing there, second that music to dance to is not the same than music to listen to, if it makes sense. I think that connecting with a real singer is still important. The nuances of emotion that a singer can do go beyond what we can pretend with programs.
Then again, I would like to, at least, have an approximate idea of how a song I write would sound without finding a vocalist, the hours of recording in a studio I don't have access to and so on.
On the other hand, producing music is a hobby to me, so why would my opinion matter, anyway?
5:00 your recording is clipped. Can't you hear that high end clipping going on? The AI version is NOT clipped. Actually, the AI version I like better. But godz, you love to hear yourself talk, man. Just play the damn files - THEN explain. Everyone who can't, thinks AI will save them. hah
As an engineer/producer who works primarily with acoustic and electro-acoustic instruments and mostly with artists who want to maintain a significant level of realism in their recordings, I could see using this tech to layer in under the actual performed vocals, in the same way that drum samples and impulse responses have long since been commonly used. Also useful for live reinforcement of performances as a backing track.
From a broader creative standpoint, it would be fun to hear how far some of the parameters can be pushed for fantastic effects.
To note, these synthesized voices have the same tale tale sound that other AI vocals do: a particular ringing echo-like sound. It’s very subtle, but once I started noticing it, I can hear it on all the AI stuff. Especially higher voices. I know most people don’t hear it, so from an audience acceptance standpoint, it probably doesn’t matter. And it will probably be refined away as the tech improves.
Very interesting…and challenging us to question what our role will be moving forward. Thanks for the look at this tech 🙏🏽
Perhaps a DeReverb will do the trick.
So you have to be able to sing to begin with to use this AI? I’ve been using Suno to get some vocals on my songs that have been unfinished for years but the results are really unpredictable.
I've heard the scream type of vocals done by AI a few months ago for the first time. And it did surprise me... so this is not new to me. But also what is the application of this really? How do we benefit? how do we apply this in a live setting? have a backing vocals for a live band? It's fun to mess about with this but ultimately I find it useless
What I’m trying understand is after creating the voice model, how “off” can the voice that you feed into the model be and have the return still sound usable?
Their website says "create copyright free cover vocals" ... Someone still owns the song guys
The terms "copyright free" and "cover" used in the same sentence says a lot about AI. Maybe someone should hack Audimee and throw it on Napster or Pirate Bay.
thanks for the info 😍😍😍
The Audimee voice sounds like it's false fold, while the original vocalist sounds more like fry or fry dominant hybrid screams, especially in the "pitched scream singing" section
What i think for now is for producers to use this as a form to make layers. I could see a producer mixing after a session and add these to the mix because metal vocalists dont really ever use one layer anymore. Would have loved to hear a track with all the layers and mixed ones to see how it could fill out tones
Meh~ it has to start harsh to start with presumably. Lots of artifacts in some of it that sound like FFT phase weirdness. I suppose in some situations.
Thing is as a music fan and a musician I end up battling people constantly who don't understand tools like this over things like the accusations about tracks live and "autotune" and this just makes the problem worse. Just like with autotune "hey do you suck at singing? You too can make a hit record!". Getting all excited about this just makes it worse for real harsh vocalists who can't get away from accusations that they aren't really doing it. Especially the women in metal. "No way does a woman sound like that". A few can and do. These techniques take years to learn and perfect so no it's not a good thing.
If you suck at harsh vocals maybe you shouldn't do them because faking it is fake period and by using this you crap all over people you probably admire by promoting something that is fake just to push yourself into the business IMO. No you aren't the same as someone who has been doing it for real on stage for 15 years. People who don't know better may think so, but what are you going to do when you can't do it live? Pop singers have a really bad reputation at this point for faking it, lip syncing, doing everything with tracks, not being able to sing with out autotune. Great you made a metal record with AI now what? Metal fans aren't like pop fans. If they find out you're faking it they won't let it go so you shouldn't be glad that you started out in your home studio making fake voices because if it launches you into something else, you can't live up to what you created.
Sounds surreal! It seems like a perfect songwriting tool for conveying ideas to vocalists, saying 'My vision is to capture that performance and grit in this song.' On the other hand, it might discourage many vocalists who feel their vocals aren't at that level, leading songwriters and producers to keep searching and ultimately hiring a pro, which isn't cheap. Personally, I'm looking for AI singing and voice cloning plugins that effectively mimic characteristics, emotions, and human factors, and work seamlessly in a native DAW environment as third-party plugin.Any recommendations?
I still think the original voice sounds better in this example but I do have to say that the AI did a pretty good job.
I used one of their AI female vocals as a backing voice in one of my songs. I am not a singer so I just talked and I have to say it not only done a good job with the voice but it done well with Polish language as well.
There were some frazes that didn't sound quite on point but overall it was acceptable.
I'm yet to try it as a main vocal but like I said, I am not a vocalist so getting a good performance is a bit tricky. 😆
Try the AI tool ACE and you can check out DR MIX video about it and he do show how tou use it and what is capable of.
it feels wrong but would be cool to hear my potential vocal idea for demos only.
that's the correct use IMHO. get something to work with, then find a vocalist that can do a decent job (it's too hard to find perfect). wish I could still sing and do vocals similar alright Myself but after years of chain smoking...well...it sucks but I can't pull it off anymore 😕
For me you can use this as a hack tips to put your vocals much bigger into the mix.
For those who doesn't want layering vocals with every take and to save your time. But the results will be different. I already used this technique quite often but I still recommended you to take every possible vox you want to take, because there is a element that you need from the samples.
Good luck!
And good content! :D
Maybe you should try seeing converge or dillinger live. It’ll change your mind about screaming.
I made an RVC Model trained with a couple of vocalist and it sounds something like this (but you can run it for free!), and it took like a couple of hours to train
"WE HAVE THE TECHNOLOGY!" and "I'M JUST TRYING TO MAKE IT BRUTAL!" are my favourite quotes so far 😂
For me it's a game changer . I have a lots of songs I have the vocal melodies in my mind . But without a "pro" singer they are just plain music, not too interesting to show or share anywhere. But with this I can make them like a "finished" products. And I think this is all about. Especially if you don't have a "good" vocal voice . Like me . Very hard to find someone who can translate your ideas the way as you imagined. It's not gonna replace a real singer in any way. But you can showcase your ides and songs and that's a huge step forward . Or a big help to create your own music
Everyone worried about this needs to remember that the AI is "replicating" a style of vocal, and not creating a new voice. Voices like Ozzys, Chris Barnes, and Tom Arayas are unique metal voices that will always be relevant, so if you're afraid of getting "left behind", then BE outstanding.
In other words, the cream rises to the top as usual. And as good as the AI is, it'll never replace the real artist. IMO
I think this could be good for me since I don’t know how to scream and don’t want learn to just for a hobby. Hopefully it can take my “whisper screams” and turn them into decent vocals
I always forget there's like nine bands named "Suspiria" and it's never the one that I know.
So, you have to have a file of good growling already to be able to convert it to good growling? Why can't you just talk into it or use Synthesizer V or something?
"AI can't scream"
Neco-Arc model: Am I a joke to you?
Wow this is insane!!!!🤯🤯
Interesting video, cheers from Germany
This is awesome for the solo producer or hobbyist. You can create background vocals done by five, six, or 6000 people. Unless you’re recording something with one microphone like Robert Johnson, you’re already partway down this path anyway. You might as well lean into AI it’s not going away.
i've used this before, its good for layering , and i've also used another program and trained my voice with a different voice model which is prettycool and other programs give you the option to blend two voice models together
How bout people learning how to train their voice? Maybe too much work involved, but it’s rewarding.
Dude... voice modulators for screaming have been huge since 2008 and probably a few years before that. Still fly by TDWP (Big tymers cover) used voice modulation.
I don’t agree that you’re making “your” music sound pro, unless you’re using it as a tool for writing, NOT a tool for doing vocals that you can’t pull off in real life (good luck performing it live which is where the money is) this only helps the people who want to make music but don’t have the skills. They still need to learn it to perform it. This was okay in the days of people composing things alone it’s okay to be a good composer and not be able to play anything. Maybe you just have golden ideas. But here, the ideas are not your own, so I personally don’t understand why you’d even want to use it. That being said, I’m not a drummer, and I’ve been using Logics AI drummer since it came out like 6-7 years ago I think, maybe longer. And suggesting settings in plug-ins and what not? We already have presets for that. Presets styled by professional humans. All the AI is going to do is tweet things to make sure that your inputs and outputs are set correctly. That way you don’t have too much gain reduction happening on a setting that is only supposed to be hitting at like four or 5 dB of gain reduction. I will speed up the process on mundane things and make it easier for single musicians to produce demos, but people will always like real music, no matter what I will always be looking for something that’s made by humans that comes with an aspect of human experience. People don’t latch on to bands & make them superstars because their music is good necessarily they do it because they relate to the story and relate to the image and message of the band. I cannot re-create that, nobody’s going to “worship” AI like they do pop stars. No one is going to aspire to want to be like AI or want to dress like AI in case you’re not aware for last 50 or so years or longer music has influenced every single cultural and fashion trend, and is a huge part of how people identify in their lives, what music they listen to and what artist they like.those, will never replace that. You want a good answer on how not to get replaced by AI as an artist, just ask AI ha ha it’ll give you a beautiful answer on exactly what it cannot re-create. And in short, human story and human experience popular artists have a way of tapping into the current subconscious of the public, I can’t do that because it first needs to study what’s going on and create models. And that’s an answer from ChatGPT itself.
🙋 hi! I already do all of that. I write my own lyrics and all that. Anyways when it comes to EDM and dubstep I have a little more freedom to remastering each song but it's all good. If AI music isn't your jam that's cool but it's what I do. It's like 1/10th of the total video creation. I don't do cheap prompting and uploads like lots of others so. 😂 I put a lot of work into my music.
I can’t do vocals in any style that someone would enjoying hearing. I have a couple people that will do stuff for me. They are often busy with their own lives. This could be useful to get the ideas I hear down and send them something basically done and save tons of time.
The good thing is that we already know what it does for pop music...we have been hearing it for years! ;)
What about blending the real vocals with ai vocals?
I can see that as a voice enhancer tool that is not as cheating as autotune stuff. It picks your actual singing performance and just change the voice to something more palatable to the ears. I bet that before AI there was tools to do some of those tricks with analog or maybe digital means. AI is just more powerful
Any ideas afor generating classical heavy metal? (more like iron maiden, helloween)
It's not hard ,it takes alot of practice and testing what works for you and you'll get there.
I tried Audimee and it was useless. I was looking for a powerful rock vocal, and it couldn't do it. It didn't sound very different to my own singing voice, which is crap. Even when I set it to convert at 100%. I'm surprised I couldn't get a better result because I've got some excellent results with Suno ai. The biggest problem with suno was, it didn't follow the tags I set. The tech shows great promise, but it's got a long way to go yet.
It's not free to use Ragnar anymore.
Well said. I've been getting given shit from my mates for using AI, but I've got some absolutely crazy stuff in the mix
I would like an AI that creates demonstration videos of plugins and new audio technologies without talking so much about them and that makes the results heard without having to be hostage to a suffocating and useless narrative
TH-cam recommended me this video after I liked a video showcasing what is actually a separate offline tool where you can make vocals using the approach described below (now hold on really tight because it's really CREEPY compared to this).
The tool I'm talking about is ACE Studio, where you just import a MIDI track of a melody you want the lyrics to follow, then you... WRITE THE LYRICS inside the markers showing what is what is the pitch of the sound in the imported track and it converts it into singing vocals using the voice selected from the library of vocalists, ZERO SINGING SKILL REQUIRED! 🤯
I liked the original high vocal. From this video it seems like it's just eq, and not really A.I. I thought you were going to take a singing type vocals and turn it into metal. Additionally, most DAW's these days have ways to add distortion and bring vocals up and down octaves.
sometimes I like to add in my mix like example female voice AI generated from my voice. I already try in different website and the result was very satisfied for me. BUT the language on vocals was not English. Most disappointment a least for me was change on octave voice in original vocals generated to AI , The AI is just keep same octave on voice.
The real problem is where our money will come from. Pressure needs to be put on governments to adapt; technological movement is inevitable and potentially good. Elon musk rightly Reframes the money issue in terms of Universal High (vs basic) Income.
The real question for an artist is why do you do what you do.
AI may, in regard to music, take away (or actually amplify, at least in the short term, if you know how to use it right) your: money, glory, attention, feeling needed by others.
If you make music for your own personal sense/experience of achievement, growth, beauty or therapy, then relax and continue on!
Mhh I‘m not convinced because the usecase should be bad amateur vocals as input and good screaming ai vocals as output. The input vocals are already good so I don’t see much sense here in using AI
I wonder how the artists would feel with there voice being on a track they didn't make 😅
The original vocals sound more primal. I prefer the original. You could use ai for training your vocals so that you try screaming at different octaves to see what works. I wouldn't want the ai to do all the work though. As a musician creating music is my passion.
But now can it replicate?Something as beautiful as the high and vocalist of symphonic heavy metal with female front singers, Like Sharon den Adele and Charlotte Wessels.And Simone Simons and on and on?
How bad can the original performance be? 😆
Not so free to check the ragnar voice)) I dont know will it works good for me. From your video I sow that you dont actually need to use AI couse your raw sound is nice enough to use it it self.
Screams are technically pitchless octiving is hard
That vocal sounds like what Meatcanyon thinks Jojo Siwa sounds like.
The cold embrace of oblivion, my final breath
In the void of nothingness, I welcome death
A.I
I’d like this to work in real time while I’m singing in church like Voicelive touch 2 portable unit be cool
It doesn't matter at all if the song was composed by a real band or a kid in his computer. The only thing that matter is if i like the song or not. And btw, already exist AI that compose full songs really well. I tested myself by prompt. So...... We are entering in a new era hard to read yet.
I think it should be upto the artist if its used and always disclosed in some way if AI is used as its not entirely ethical to use it without the artists permission.
It already can replace some modern metal vocalist, cause they sound the same. But I highly doubt that it will replace someone like Corey Taylor, Winston McCall, Sully Erna and so on
Check out a power metal track I generated entirely using a prompt 😅 I was just blown away
That tool is so interesting
Nice Tom Cochran cover.
I had no idea that they was doing metal
I think there's likely to be a limitation in what the audience accepts. The audience has a long history of disliking artists who sound terrible live vs how they sounded in the recording. That took a couple of additional disliked steps when bands did lip sync, or used auto-correct for anything other than fx on a tune. I don't think that having an AI do the various artists' roles is going to be useful for much more than a brief minute of entertaining friends with a recording before they skip to the next tune.
You turn your voice into a scream, release the song, and then squeak normally on live stage.
I think thall should have different type of vocals than the ordeniry metal screaming (including animal, guttteral, tunel throath and so on). I mean new experimental technic.Maybe AI could be useful in that case. It's like got the metal sound of 21st century, but the vocals are stuck in the 20th century. Hope you get what i mean.
AWESOME!!! thing for music. 👍♥️💋
Yeah get it to work on the Grimm vocals also.(Immortal, Dimmu)
Not just brutal and deathcore. And for a true test, see if that thing could pull off Behemoth.
07:26. Intro!
Awesome!! 😮😮
People ask me if I’d ever take audio engineering advice from a Senior Research Chemist. Yes, I would.
13:34 HELL YES!!!
I'm in the same boat as you are. I've made far more minor videos on IG and FB about this, and I had a lot of pushback. To me, I get it... but whether I feel that way or not, the tech is moving forward. Maybe this will kill music, maybe it will reinspire people in different ways. We have yet to tell. I figure I will learn a ton about it so that either a) I can be ahead of the curve if this is where music goes, or b) I'll already know AI, so if this job goes away I have a foot in at the new one... 🤣
I see how this can be awesome for making demos. Does that mean you can record your screaming vocals in a normal speaking voice at say… 2:00 am and not freak out the neighbors and I can demo songs?
And I did click on this because the title said “free”. Having something free for a limited time doesn’t make it free. I probably will try the app but won’t use your code for it.
Free accept that is probably uploading your performances without your permission... Then elements of your voice profile interacting your performance composition and recycling them including a compositions with Maybe enough changes to avoid copyright infringement.
this is so misleading. The sample used to create Ragnar's parts already sounded good and metalish so ofcourse that's what you get. As soon as you sing out of tune of terrible and use these as samples you'll get absolute shit for results.
Woah, this is sonic magic
No ai for me. Well be doing the dishes for the robots
I just saw an AI version of Mayhems DeMysteriis but with Dead on vox.
We got a good decade or so before AI even approaches what a real musician does. It always lacks tonal dynamics in comparison to the real thing.
Now that I think about it, that could be super useful in certain genres.
It would be good for people that can't sing or scream properly for sure
The only people who use AI in "music" production are people who can't play an instrument or sing. So nothing changes.
I want Jason Aldean to cover "Better Days" by Staind.