MUST SEE 🔥 New AI Music Generator is INSANELY GOOD 🔥 udio Review - Has AI Replaced Musicians?
Vložit
- čas přidán 4. 06. 2024
- AI Music generator UDIO has just launched and it is amazingly good at creating professional level songs from a simple prompt.
Like Suno, Udio can create lyrics, and music - including all instruments like guitar, bass, drums, piano, strings, etc... and even vocals in just about 1 minute all from a short description of the sound you are looking for.
UDIO is still in beta, but it already sounds crazy good and generates professional level tracks. We'll check it out in this video, and I'll generate tracks from several genres, including: Metal, Rap, Hip-Hop, Country, Pop, Instrumental, and more.
This is one more step in AI taking over the music production process, and it's only getting better.
Leave a comment for your thoughts about AI music and how it may impact you as a musician or audio engineer.
Check out Udio here
www.udio.com/
➡️CHAPTERS:
0:00 - Intro
0:23 - First Look at Udio AI Music Platform
1:05 - Generating our first AI song
2:58 - Metal AI Songs
7:36 - AI Removes the Talent and Sacrifice Barrier
8:18 - Pop AI Songs
11:32 - Rap and Hip-Hop AI Songs
19:16 - Instrumental AI Songs
24:03 - Country AI Songs
28:58 - Funny and Video Game AI Songs
36:48 - AI Song about Fortnite & Metal AI Song
41:02 - MUST HEAR - Metal AI Song
45:24 - Outro
➡️Our Links
Website
benonistudio.com
EPIC Plugin Deals
benonistudio.com/deals
Link Tree
linktr.ee/benonistudio
➡️Buy your audio plugins and hardware with our affiliate links
Amazon - www.amazon.com/shop/benonistudio
Sweetwater - benoni.click/SWEETWATER-DEALS
Best Service - www.bestservice.com/index.htm...
Plugin Boutique - pluginboutique.com/?a_aid=5dc...
Waves - waves.alzt.net/b3jrVM
Native Instruments - benoni.click/NativeInstruments
SUBSCRIBE ::: LIKE ::: COMMENT - Hudba
What do you think about AI music? Are you a musician? Are you a fan of music? Would you be upset if you found out that a song you love was created with AI? Comment ⬇⬇⬇⬇
I think it's possible we're gonna see more AI generated songs used in more places, but I truly hope there's not this flood of clever lyrics writers who put out these meandering, unfocused and clearly generated tracks. Basically, anyone serious about the process of creating music will still do things traditionally. However there's no denying that if someone needs a fast bed track for a commercial spot etc can get real use out of it.
I also suppose that someone might find real joy in messing with these engines and doing mashups or remixes based purely on generated content. So in a way it's all up to how end users will utilize the tech.
Of course, the AI engine is only as good as the training data used on it, so things will invariably start sounding like, "oh, you made this with Udio" or something. The amount of time, data and energy used for training datasets is the work now, vs the work of physical people with instruments. Also I would rather write and perform the song myself. I don't like using drum samples either, but don't have a space for a kit. Ampsims are different; they impact a performances sound, but not the performance itself (eg, pitch and dynamic level).
It's a fun toy, but I don't like how AI is shoved into everything now. It's beyond a buzzword and I sigh sometimes when I read more about it.
cheers!′!
Ai not gonna make her wet
unlike instrument in hand
its a good short cut for ideas
It still takes a human to create. It is just easier.
@@cmd_f5Come check out my music.
@@ZeginMakesMusic I think the issue is how we define "create"
(most) Millennials and older will never think of a prompt as "creating a song"
Younger Gen Z and Alpha will grow up with this tech and consider it normal and will think of prompting as how you "create" a song.
Go to your great grandma and ask her how to make a hamburger. She'll tell you how you need to take one of your cows to the butcher, have them grind the meat, go pick it up three days later, come home, start making dough for the buns and put them in the oven, form the burger patty, put it in a pan, cook it, put it on the bun - and FINALLY eat it.
Now go ask your Gen Z girlfriend how to make a hamburger and she'll pull out her phone and open the Sonic app. 🤣
One of the best use case scenarios for Udio is for ambient background tracks for audiobooks. I'm working on a book that spans the late 60's and well into the 90's. I need Metallica-style thrash. I need Motown style. I need hip-hop tracks. It took less than a few minutes to generate a wide variety of music. Very impressed.
I can see music like this making a comeback in video games at least: indie rock, metal, punk, etc.
No way. A lot of money is put into music production for video games. Why would you leave music production on an expensive game project up to chance. Udio does not follow your instruction.
When I extend a song, it chance language. Anyone know Why?
Here is the thing I'm not buying into yet:
They are saying AI trains off of millions of samples/files/documents etc. but what we don't know is what 'training' actually means. What I am getting at is some of these examples seem like they could have been blatantly ripped off of some obscure local musician channel and we wouldn't know. Sure the final output is pretty amazing but I still think there is something shady going on under the surface...
Thanks to the CZcams algo for putting this vid in my feed! Rock on man!
Google is using CZcams videos to train their AI systems. Many others use Stock Audio and Video sites to train their models. I have stock footage on Pond5 and just got an awesome check for $0.67 because some billion dollar company (probably Meta) used my content to train their AI 🤦♂️ Thanks Zuck.
Sure, but how is any different than humans being influenced by existing music?
@@BinuralOdyssey It's not. It's just faster and better at it than humans.
@@BenoniStudioWhat Im getting at is it consumes like humans, and creates like humans, thus "blatantly ripped off of some obscure local musician channel" is not exclusive to AI.
Now that I think about it a little more, I see the biggest difference is what the human side knows about the origin of the music. Great video.
Its hard to complain when we all listen to existing music to learn an instrument or song. I personally don't care whether a song is played by humans or not if its good but do respect musicianship and know how hard it is to craft. I'm hoping humans can use this tech to improve what is being played on the radio, although do we really need the radio now?
You called it right. This tech removes the ‘talent barrier’ millions of people have while trying to learn / play an instrument. I am one of them. I tried to play drums: failed. I tried to play the piano: failed. I tried to play guitar: failed.
Now I have three kids running around the house all day. This tech is my dream come true!
God bless music and AI! 🙏🏻🎸
Im genuinely interested for what purpose you would use this? I rly mean no offense but its something different to show your kids a song on guitar you made and playing it for them vs "hey kids look what promt ive typed in"
@@ishimurabeats6108 I simply want to create music with my prompts. Like when someone creates something with MidJourney or ChatGPT. Famous musicians have been using songwriters and ghost writers for decades. Rhianna for example has a troop of people writing songs for her. That is basically like using IA but for rich people. Why couldn’t we use these tools for the same purpose? Even if we like it or not, AI music will dominate the charts sooner or later and famous musicians using AI to create music doesn’t really make it ‘real’ music just because their are famous. 😉👍🏻👍🏻
Awesome review 😎 I SUB'd
The Base of AI is bigger than one could experienced in lifetime
Needed more than 10 tries for a song i like with extensions and intro, and the lyrics are amazing
Downloaded the app on iphone but it doesnt generate anything..i click “create” and nothing happens
I've had a bit of a test out with the software, and as a musician and composer myself, this is scary. There's some generic-ness to it (is that even a word? Probably not, but you get what I mean), but it's also very good. Damn. If this is just the beta, it's only going to improve from here... how does this end? I dunno. Still trying to process this!
Being that most music and audio is created for video (movies, tv, commercials, youtube, product videos, etc...) I'd say the first musicians to be replaced are the ones nobody ever sees, so they don't have any attachment to them.
There's alot of people in the Stock Music/SFX world that won't be needed when AI matures. The ironic part is that most of the AI models are trained on royalty-free Stock Music libraries 🤦♂️
Then you'll have the commercial music guy who makes jingles be replaced by the video editor who can just prompt a short song to the exact specs the producer wants and they can save money and cut costs.
And then it just snowballs from there under the idea of cutting costs, but never realizing that cutting costs has its own cost.
And then the Terminator. 🤣
@@BenoniStudio I think the only way music will stay "human" (for want of a better term), is to know how to play an instrument live and perform live. And yet, technology is already impacting that too!
@@shouldvebeenonthealbum Like at the Super Bowl when you see the "musician" on stage, but their mic is turned off and they are blasting the album version through the speakers 🤣 I always love seeing these "musicians" take the mic away from their face, but yet you still hear them singing 🤣🤦♂
@@shouldvebeenonthealbum That is where artists and bands will still shine.
People, especially Metal Heads like myself appreciate and love live performances. I don't think I've ever heard of a metal band lip-syncing.
Anyway, metal head or not, people value the human connection in a live performance. I think it will be the same for their albums.
It'll be similar to those who walk through the art gallery. If you show someone who appreciates art two different pieces of art and they say they love both, until you tell them the one on the left was made by a.i., and then suddenly that piece is, "meh" to them.
Unfortunately for artists, most jobs for them are cut completely. I do not think it will be the same for music, especially for live performances.
I wonder if someone enters the exact same prompt would the output be similar
its possible. but I've entered the same prompt 5 times in a row and each time got something totally different. the AI model knows what it generated, so it could be programmed to never repeat the same exact generation.
Stunning, insane and scary all at once. Wow 😮. The lyric writing is deep, but how is that going gel with the 'short attention span' generation 🤷🏿♂️
When you release music today, one of the platforms it gets distributed to is TikTok (and CZcams) so then people can use your music in their short form videos. I was thinking 30 seconds is short for AI songs, but the more I think about it, I think the 3-4 minute song is too long for people today. I think songs of 30 seconds to 2 minutes is enough. There's already a trend among gen Z and alpha where they listen to music at 2x speed so everything sounds like the Chipmunks. GET OFF MY LAWN Whipper snappers! 👴🤜
Taco Bell Reggae song is just gold!
It honestly almost feels like science fiction
and it's already gotten better since this video 🤣 the new in painting feature is crazy!
AI music processing has come a long way in the past few years. It was just 5 years ago I did a thesis on making a neural network to identify music by decade and that algorithm alone took several days to run on a dataset of 250 songs. Now there's millions of songs categorized by many parameters that can be used as the building blocks for making new content, it's wild. We'll soon have AI generation be indistinguishable from human performed music; we're not there yet but soon.
It's a wonderful tool for creating generic soulless music. Quicker and more efficiently than when humans create generic, soulless music! Let's face it the minute we included computers in the mix this was inevitable. Humans will still be needed to perform the AI generated music until they perfect the replicants.
I guess next is making the music video with AI...
Adobe is including generative AI in Premiere Pro, others will follow soon. So you wont even have to leave your NLE (video editing program) to generate videos. There are some platforms that already generate Video from Audio like Plazmapunk and Neural Frames.
Heres a video from Matt Wolfe who basically only makes video about AI. He made a music video with AI last year in this video czcams.com/video/CaTqxkZTSV8/video.html
Alot of what he does isnt necessary today since things have gotten much better.
I had the same reaction, it is over. Music industry as it is now... soon we wil have more parameters to influence this. The near future will be singing something and you will be orchestrated immediately
or hum a melody and it will be orchestrated or played in any insturment you want.
yeah. Izotope has a plugin called Morph One that is still in beta that lets you take your voice or other audio source and morph it into a Saxophone, or guitar, or bass, or whatever. And Facebook (Meta) has a music generating AI that you can hum a melody into and it will generate the music based on that.
The second hip hop example sounded just like Wiz Khalifa....they might be using actual artists likeness lmaooo
Udio doesnt state what media they trained their model on. Many of them are trained on Royalty-Free stock media sites like Pond5, and they have permission to do so. I have stuff on Pond5 and they sent me a check for like $0.65 cents because my media was used to train a LLM by some billion dollar company 🤣
There are also more and artists who own their music who are licensing it to these AI sites to train the model.
So I can't say for sure what Udio is doing, but most of them (that I have seen) are actually legit.
In an ironic twist, the Royalty-Free music used to train these models will end up putting the Stock Music platforms out of business.
It's a fun thing to mess with, but I don't see it replacing real performances. Although for background music under a voice track I get someone hammering out a few of these tracks as filler.
Edit: I just had a mess about with it, and the drums in heavier stuff sound papery and fluffy, so it's no surprise there. Still, I could have a place for something like this. Not necessarily as a serious writing tool, but perhaps a rif idea or pattern could come up that I' not normally think about. Generating stuff with random prompts can be a good time.
bro at some point in the not so distant future, you will be able to go on stage with a mic and AI will autotransform your entire performance, similar to waves autotune live. you will have a personal artist setting that you can fully control and it will always sound exactly as lit as you want and can imagine/prompt
In this stage AI only able producing what they want - a.k.a hallucinations .
Useable in some case, but not really if you've already have a half baked song idea.
I bet the next stages of AI music would be finishing your original 8 bar loops . Thats would be thrilling and horror in same time
you can already load music into some AI generators and it will create a song based on melody 🤯 you can even hum a melody and a whole song will be created.
It`s like dating some chick in the club and realising it's a dude after the beejay-and afterwards eating a 3d printed burger: Humanity quo vadis ? Make sure u got ur boosters folks
I immediately used it to make songs about poop
thats what most people seem to do 💩
I imagine a future in which DAWs have their own plugins where you tell them to do such things and the daw creates the entire project for you... or at least the audio tracks separately, thus avoiding artifacts from any other instrument
believe me I don't see it very far away
It will be the only way the audio is truly high quality
I do hardstyle and it does surprise me that it recognizes the patterns, it even differentiates what a hardstyle kick is, obviously the audio quality is not good because there are a lot of trebles and it still suffers from inconsistencies, this definitely makes me conclude that they did use copyrighted songs author so that this AI learns since they are unique kicks, but today we ourselves use samples without knowing that they could belong to 1000 songs
These AIs are undoubtedly a great victory for video game developers or background things
The algorithm that it follows in purely its own. It does not follow basic directions, such as instrumentation preferences. IMHO, when you let it do what it wants to do, it sounds great, but when you try to design your own music, it's a hot mess that bears no resemblance to your instruction. AI is no where close to replacing musicians.
Reckon phil Anselmo would be pretty pissed hearing this.
again..WILD🤯
This is stunning, Fake musicians will be taking over the world. :) The gig scene will disappear or end up just as fake :( Or maybe musicians will duplicate what AI generates and call it their own.
most of the music industry is already fake. A mediocre singer that fits the look some agency is looking for is crafted into a superstar by audio engineers who take their garbage vocals and turn them into something listenable.
Holy sh💙t!!!!!
I think it wont change that much. There will be just more same old generic bs going around as it is right now. 90% of music today is already soulless. Background music for tv shows ad etc already go with a certain scheme its all the same so why not speed up the process. There will always be shows and talented musicians and those who buy their shirts etc otherwise no one would go compete in for example running sports since a ford Fiesta will always be faster than any human could be.
Copyright infringement claims will be coming soon!
they would need some sort of evidence first. most of these models are trained on royalty free music libraries with a license from the site to do so.
@@BenoniStudio I’m just not comfortable presenting something as mine that really isn’t mine. Great for businesses not for artists.
@@capitalcrud9062 do you use MIDI or virtual instruments? you didnt create those sounds either. do you use presets in plugins? you didn't create that either. its just a new technology that will take awhile to be accepted, the same way it was for MIDI, which is now a standard.
ummmm, thousands if hours of work to write some songs and program them, now theres this smfh
In the future it will be like: "Hi, what do you do?" the person responds, "I'm a musician." You reply, "Oh, really, that's so cool, me too! - What instrument do you play"
Looking confused and slightly disgusted like you're 150 years old rambling about putting onions on your belt (which was the style at the time), they dismissively respond, "Uhh...Yeah, instruments are so 2000, I prompt" 🤦♂️
I'd say within 5 years we'll have our first superstar "musician" prompter who will be having sold out live shows. It's not that different from DJ'ing.
Question is, would anybody listen to this shit ?
Its technically music, but where is the novelty, creativity, progress, inspiration?
Is AI going to ever move anyone on a deep level like a real musician can?
I dont know. Not yet.
have your heard music today? this is a real song, made by a real musician. tell me how it's "deep" or "creative" czcams.com/video/Hmee2zK6lpM/video.html
Or maybe this czcams.com/video/tsp7IOr7Q9A/video.html
If you want to see what music is out there today and what people actually listen to and you also hate yourself - then go on tiktok for a couple hours and just keep flipping to new videos. Your ears will be treated to a buffet of awful, extremely popular music.
I would go as for to say that the AI music in this video is better and deeper than 80% of music produced in the past 5 years.
Here's how I feel about AI music - unsubscribe.
Aww he gonna miss you soooo much 🥴🤡
@@katielowenikr, buddy’s scared about AI
@@moopy6446 Facts! 😂💯
czcams.com/video/JDHvYdi6Bu4/video.html