This AI Music Update ROCKS! (Udio AI New Features)
Вставка
- Опубліковано 5 тра 2024
- Udio comes out with a new update that (IMO) solidifies it as the best AI Music generator!
▼ Link(s) From Today’s Video:
Udio: www.udio.com/
Song we made: www.udio.com/songs/7Kkfnx8zat...
► MattVidPro Discord: / discord
► Follow Me on Twitter: / mattvidpro
-------------------------------------------------
▼ Extra Links of Interest:
✩ AI LINKS MASTER LIST: www.futurepedia.io/
✩ General AI Playlist: • General MattVidPro AI ...
✩ AI I use to edit videos: www.descript.com/?lmref=nA4fDg
✩ Instagram: mattvidpro
✩ Tiktok: tiktok.com/@mattvidpro
✩ Second Channel: / @matt_pie
-------------------------------------------------
Thanks for watching Matt Video Productions! I make all sorts of videos here on UA-cam! Technology, Tutorials, and Reviews! Enjoy Your stay here, and subscribe!
All Suggestions, Thoughts And Comments Are Greatly Appreciated… Because I Actually Read Them.
-------------------------------------------------
► Business Contact: MattVidProSecond@gmail.com - Наука та технологія
Join me for a livestream on wednesday! ua-cam.com/video/L3IqqXUNKZ8/v-deo.html&ab_channel=MattVidProAI
yessir 🖖
👋
Udio is the ONLY AI music tool I use
whatever you make in udio is not your product.. read their legal terms dude.. they are not stupid.. otherwise Universal Music would sue their A$$es 100x over by now. how do ppl not know this.
you know you are getting famous when people clone your voice because it's nice and soothing :P lol ua-cam.com/users/shortsKVUbiuK6_hY?si=UIh9H8mz8ic0dfve
I wish UDIO would generate seperate vocals and instrumental parts, that would be perfect for us home music song producers.
They claim that's their plan according to a statement they made on the Discord. A UA-camr who is working with them also said that was brought up as a probable feature.
Can't you do that with AI?? I know I saw a number of AI tools on Replicate that can separate vocals from the instrumental. Not sure how well it works - may be worth checking out tho!
There's some pretty good tools for separating vocals and instruments (search for audio splitter). Of course it would be great if it existed inside Udio.
They should also work on a merge feature. This would be useful to merge generations or other existing sounds and music into tracks to combine elements that might work well together in one track and make something new or similar to the target generation.
@@timsell8751 Not ideal it looses allot of quality, it's better to just split the 2 tracks from it's generative root, then I can use my voice over the track using the other voice as a guide to follow.
little bit of info: when you manual prompt it, it doesn't replace the artist's names
Doesn’t it? Going to try that out
Thanks for this!
Would love to see
-Save voices to use for another song
-Save different character / voices
-Add my own voice / upgrade voice range
-Create similar sounding songs, to be able to make a full length album.
Linearity is key!
One day I would also like it to become like a Guitartabs, where it can use Ai to auto generate, to teach me how to play my own songs.
We are not even in a full year into this type of tech. Music producers need a studio version, with all the bells and whistles. Midi, chopping, STEM separation, mix and mastering tools etc. Also, with remix capabilities for original music compositions. The tech is there but not combined into one application.
It's a lot tougher to do all that because it's generative so until they add features like control net under the hood these things won't happen for a few months if they work fast
Yeah but where are you going to store all these pretrained models? I think some daw company in the future will create an algoritm that will enable you to create temp models on the fly that would be optimized enough to fit in RAM or a small enough temp file on HDD to quickly generate something from a combination of a promt and an aduo input like an random mp3 file. This will be a killer on the markred plus with all the features you mentioned above.
I mean, take udio output, put into daw Like fl studios or ableton, and you’ve got most of that, no? May not all be ai but the tech that’s already there is beyond insane as is
Awesome stuff, man! The last track I made for my channel was also with Udio, and it turned out really great!
wow thats realy good!
The ghost of Johnny Cash is trapped in a metal garbage can.
The fact that the lyrics are autogenerated gives a quick peek into the mindset of an artificial being it mentioned "this metal skin and wire veins" makes wonder if in the future Ais might get a Pinocchio complex?
it might.
The only sad thing about future AI media is that it will be so ubiquitous and abundant, literal masterpieces will be one time generated and then forever discarded.
It’s only a masterpiece when a sentient master creates it. But the sentient masters who made udio created a masterpiece!
Depends on how good it is. It might be a song that reminds you of a time in your life and then you keep it to remember that.
Been thinking about this myself….so much incredible media is going to be generated that never sees the light of day…who knows though, maybe some elves in a nearby dimension will dig it and it will not be for naught….
I gotta say Matt, I've been getting some pretty amazing stuff from Udio, working on the theme for Little Miss Adventures and I have a lot on my plate. I could write something, but seeing as it's all going to be mostly AI for the most part, I figured I'd just let Udio create a song for me, it took some rerolling but I'm quite please with my results and I can move onto putting this AI cartoon together. You know, it's because of you and Corridor that I started looking into all this. So thanks for very informative and entertaining videos!
That crop feature is so massive, brace yourself. It won't take long for someone to absolutely pop off with a banger
Always look forward to your updates! Been messing with Udio quite a bit these past few weeks. This selection tool looks like it will be an essential update. It's a roll of the dice if a song extension would have noticeable errors. Sometimes you'll have a really good generation with weird errors in a few spots. The thirty second context memory made a repeating chorus or refrain near impossible. I think it'll be much harder to tell some of these tracks are AI once they have three minute context memory. It would also be cool if they implemented the equivalent of a seed to produce similar tracks.
So much fun, after watching this vid I spent 2 hrs on Udio and had a blast. Thanks for the walkthrough.
I'm really having fun with the latest udio update. To me it's barely fathomable that this is actually possible. Right now I'm mostly waiting for udio to not sound like a very low bit mp3 :D
It's actually a 320 kbps mp3. Everything I've downloaded sounds like pretty high quality. Suno, on the other hand, sounds like a low bit mp3 and is only 192 kbps.
You should try Suno then, it's incredible.
@@aimusictv1 You can convert any audio file into a 320 kbps mp3 - it's not an indicator of quality. That said, I agree Udio outputs usually sound pretty good.
@@aimusictv1 It *sounds* like low bitrate. The actual bitrate makes no difference
@@Octamed Yeah, the downloaded mp3 is 320 kbs, but to my ears, it sounds more like 64-128 kbs. My GF does not hear it, so I guess it might be fine for most people xD
Udio needs to add a button that will request to Add a Chorus. That was just one long verse. It would have been more of a complete song if it could have broken up the lines of the verse with a chorus or two. Probably in the works.
If you add [Chorus] in the lyrics it will do that
It made a song for me with a 4 line verse and then a 2 line chorus and ended. I extended and it just started with another 4 line verse and a different 2 line chorus. It sounded good and the lyrics were pretty good but Suno is still better ay making a structured song
I believe when you extend that you can change it around, so put in a chorus on one, maybe none on the next extension. Not sure how well it works, gonna have to try it out
When you're doing auto lyrics just using the prompt, type "only create a [verse]" and then when extending you can do "create only the same [chorus]" doing this will make the AI create specified verses or choruses or whatever you want when using auto lyrics.
Problem is that this guy really does not know alot about this. You can do a ton of stuff with [] in the custom lyrics field. You can see how little he actually knows by how he explained the "select section" he explained it the wrong way round all the time. Don´t wanna hate, but imho you should only do Videos about stuff you are actually knowledable about.
Bro This tool brought a tear to my eye, thank you so much.
I've been wanting this! This alone changes a lot!
"AI will never replace human music because it doesn't have a soul and can't reproduce or stir up true emotion"
04:05: "Hold my beer"
yeah, I have to disagree. 99% of modern music is soulless trash about shaking asses
@@quaterman1270 I have to disagree. It's only 98%.
@@quaterman1270 that's pop, there are many genres however
@quarterman1270 Sure, if all you listen to is the radio, this is an easy mistake to make. But ass shaking garbage music, as you call it, only represents 10% or less of modern music being produced. Don't blame the music landscape for your inability/refusal to take the time to look below the surface for things you like.
It won't, because eventually we will all realize the best use of AI is assisting humans, not replacing them. Some will learn this the hard way, some companies already are.
Great video, I enjoyed it.
if you distribute udio songs to spotify you have to credit udio according to terms of service
How do they prove it was made by udio tho
watermark
Seed ID
@@sethrono6013I haven't seen or read any evidence of watermarking, it would be difficult to maintain now that you can extend and cut back the audio. It would be easier for them to train a ai copyright model like youtube uses with their database of user made songs.
I'd like to see them try to uphold this in court.
Love the videos
Wow. And with the update that's coming out for ChatGPT this week, this is going to be a memorable week for me. Good one Matt.
Your videos are awesome.
This is how BBL Drizzy sample was made
do you know the prompt for creating BBL Drizzy beat alike? it'd be funny to diss someone with AI generated song
Fantastic. I bit more editing in Udio. Perfect.
That very last generation was dope. I hope it shows up in your Spotify.
Question: can you remix other ppl’s songs on Udio?
be nice to have layers with a stem for vocals and instruments separated , maybe even a basic sequencer for arrangement '')
It’s just one long verse, no chorus or fluctuations
Thanks!
4:56 - I’m assuming it’s not clever enough for a multi-lingual pun, but a song a bout a dog where “she ain’t” sounds like “chien” (French for dog) is 👌
emergent french
that's a stretch of a stretch lmao
udio can do other languages very well.. i mix English and Ukrainian.. in fact i am working on a track right now
Good gravy. "wire veins"...that is shockingly inventive. Quite novel!
I'm not seeing any option to open a drop down menu for existing songs. It looks like the same old Udio I've been using. Am I supposed to do something to get an upgrade?
I've not seen the updates yet can't wait to see them
Guys, there's a video about what's coming with UDIO from a music lawyer. I posted the link in #general on the Discord.
I've been using Udio a lot, and my entire channel is music I've been making with it. I think I'm makin some real bangers. It's amazing how technology is enabling small artists to dive into the music world and make some high quality stuff with less barriers.
Yo Matt! Music? Let's go. Liking Udio so far. 6:47 First tune was definitely better 🤣🤘
Is there an AI MUSIC APP/PROGRAM where I can record a rough demo of myself singing and playing guitar and have the AI part fill in like the drums, bass, or other instruments I need to song write easier?
Magix Music Maker .. You would have to buy their soundpools for it to work with the AI but you can do what you want with it. However, the soundpool doesn't adapt to your singing ect .. You have to adapt to the way the AI arranges the stems. You can manually edit as required but that is as close as you can get to doing what you want. Their ultimate edition is on sale until May 12, 2024 so you can try out their trial and see if it suits you. It requires Internet to use the paid version because of the AI. Think they have a free version too but I'm not sure it offers AI. Look for perpetual license not subscription unless that's what you want. Can't beat it for under $ 90 while on sale.
When I read about the rabbit at CES, I was trying to figure out how it was significantly different from copilot or other apps (maybe more polished and optomistic about potential integration with other apps or api). Then, I saw the reviews. It's shocking how many people bought into the hype and refused to acknowledge they were duped.
I have yet to generate anything on Udio that exceeds what I've generated on Suno.
First radio play of a Udio song was last night on Oystermouth Radio at just after 9:15pm. The show should be repeated tonight at the same time.
I absolutely love Udio!❤❤❤
Rabbit R1 is the minidisc player of the AI era
Important points to know. You can't extend it from a specific time and add a specific length of the generated part to it like in suno. And more importantly the Chorus are not the same. Yo will have a song with different chorus. Quality is better but you can adjust the quality easily and for free with audacity.
All you have to do is put in [chorus repeat] for your second chorus and it will repeat same melody. Just learned that today.
That was fixed about a week or so ago. and you do NOT have to put "chorus repeat". They changed it from a remembering a section within a 30 second window to a 2 minute window. All sections match every time now... at least they do for me.
@@jeiwesthesoundman - Not any more. The issue was fixed about a week or so ago. In fact, the video I saw where someone put out saying use "repeat" was already old news when it was posted, which was after the fix.
In regards the SORA multiple render thing--the BG vibe and color are the same, but there are many differences in every render. Not usable if you need a match, but still cool for prototyping.
you have the best AI discord server
Udio sounds way better than Suno, in terms of audio quality, but I think Suno is closer to being radio catchy (still only a quarter of the way there). It is like how Midjourney generally has better aesthetics but was beat in terms of photorealism and prompt coherence at different points.
I have gotten nothing but garbage out of udio and 20 good songs out of suno so my experience is much different.
@@SamuraiG I agree, that is what I was saying but udio has better audio quality. As an analogy suno has better musicians/song writers, udio has better mics and audio engineers.
@@13lacle Your probably right but since I have never had the opportunity to extend a song due to not finding anything I like after a hundred + generationsI don't have your experience with that.
UDIO IS DEVOID OD CHARACTER OR SPIRIT UNLIKE SUNO
As a musician/songwritter, I would never use Udio as it is today. I do admire the music production capabilities and only wish I could input my own MIDI song with my own lyrics and then have Udio produce the instrumentals and the singing. Now, THAT would be fantastic fun. But I don't want to just listen to machines make up stuff.
I hate the trash music mainstream artists pump out today. I love AI music because it lets me make the kind of music I want to listen to.
I love how I can literally make a theme tune for anything. My wife goes to the toilet, a theme tune. The dog barks at the telly, a theme tune. You get the idea. Creativity in the hands of everyone without the years and years of practice or luck (in some cases).
I've been having fun with it, but THIS dude! I agree, I think though that in the future we will be able to input our own music and then Udio can assist with what we already made.
yeah or put short wav files of your own song and let the ai continue to generate ideas
then learn how to do that cuz we can actually do that
I have created some uncanny and amazing classic concertos with Udio. Mozart and Vivaldi! Seriously great!
Hi. Have you checked Udio now has pricing options but the page has no SSL certificate? My antivirus activated when I access the page. Any info will be much appreciated...
I did an entire album for family members, and my own creatives using several genres, including an epic "Led Zeppelin-Metallica-Moody Blues style" 14 minute epic! Outrageously awesome! I even made a forbidden topic track and had Udio create the lyrics. I thought for sure it would get flagged, but Udio made the track and the lyrics were on point! Incredible program!
The select section in red is the part they not will replace.
I like almost all AI UA-cam presenters, but MattVidPro is the best for me!
I think the reason they don't just inpaint is because of shadows and other optical features (reflections etc)
Had the R1 for 24 hours. Sold it on eBay. Instead I use Pi. The free AI that I just Lauch with the action button. Much better than the rabbit 🐇. Chose between 8 different voices. Knows the liocql weather and much more!
Needs a bridge, chorus and hook.
Udio is probably the most shocked I've been since MidJourney. There's really going to be nothing we can do that the AIs can't.
Udio it's just great, I love the new features. No doubt it's the king right now. Now let's wait for it to give us separate tracks, also some optional nodes for specifing chords, bpm, time signature, the primal key. That would make it even more powerful!
I was just listening and through you put a Lumineers song on randomly
Udio: definitely not trained on any Johnny Cash songs.
Udios website is obsessed with the word “echoes”
Mind blowing 🤯 In just one year Ai has gone to a top level player in the world of music 🎵
I think the R1 is an amazing idea. Would have been nice to just get the app and the R1 but money. Regardless I see no reason to use it as an adult with a phone. But, it would be an amazing first smart device for kids. It's like giving a kid a Hitchhikers Guide to the Galaxy. I asked lots of questions when I was young that no one around me knew. So a kid walking up to something and asking what is this and how does it work to me sounds amazing.
Matt showed how easy it is to make a boring 2 1/2 min song. 🎉
This function is actually the core of Suno, to extend the music while keeping context, so a chorus for example repeats again in the same structure, this is something that suno already does very good, much better when they added the ability to continue at any point, anyway I tried some less mainstream genres with udio and it failed miserably, combia, salsa, electro swing, dowop and mixes of that tend to create clearly not musical results, and I did my best to work with udio because the final tracks are more easily separated into stems, but I will take less quality in exchange for creativity anytime, udio is good of you want luck to be the one in charge, but suno works better to align with the ideas that I already have.
Pro tip: Change up the prompt in subsequent gens to tell your story and have a resolution.
Is it possible to regenerate the same singer for all music audios?
Man i cant wait until sora releases. Any news on the adobe release?
Honestly I hate the AI that replace artists. I love any other kind of AI but those make me sad 😢
Hey Matt, what's your Spotify's Channel?
can i download in wav or flac format if i have pro plan ?
Would love to here some hip hop and Reggae
In the UA-cam video "This AI Music Update ROCKS! (Udio AI New Features)," the host discusses various AI news, including Sora's ability to change one element of a video while keeping the background the same. The main focus is on Udio AI's beta, which generates original music. The creator shares his experience using the tool to create a country folk song about a robot and discusses new features like the "select section" tool that allows users to crop extensions and the software's ability to take into account the full context of a song. The speaker is impressed by the consistency of the singer's voice and the ability to generate a full-length song in minutes. Milky Way also shares a useful trick for generating songs using artist names and encourages viewers to stay tuned for updates on his social media channels
Looking at the first clip quartet I wondered how many years until they will be generated in Unreal Engine 😮
The video is wonderful, although when you want to select a part of the song, you select the part you want to save instead of the part you want to cut, in other words, it's the opposite of what you explained. 😉👍
All Suno needs is audio quality and it will remain the king. I usually get trash songs from Udio. Maybe 1/10 usable while with Suna 9/10 are usable
the disadvantage for work with Udio now is that it's free
My requests are to add bpm and key selection, and be able to hum the melody I want ai to follow for the lyrics.
The Rabbit R1 only makes sense as a device because it would never be allowed to exist as an app. I believe the play store and whatever apple has would ban it immediately. Because it aims to replace all other apps or use them for you off-screen, you wouldn't look at ads anymore, and that wouldn't fly AT ALL with every other developer, including the big ones.
The song you spent most of this video on has really nice realistic AI vocals that sounded great, but then most of the other tracks reverted to Autotuned Speak and Spell vocals. I guess the heavy use of Autotune in the music business is messing up the models causing them to mimick that ear grating effect for modern styles of music.
cool
I just graduated in digital music.. hoping to become a producer and now this comes out... 30k debt and pointless skills... I have however made some absolutely glorious anthems using Suno (on my channel)
Has an uncanny [even concerning?] similarity to Johnny Cash's "Hurt"....
I have to say I prefer Suno. Udio mostly ends up with weird phasing sounds on vocals and sometimes music. Unfortunately you can hear it on most of your generations.
I made a full album on Udio:
ua-cam.com/play/OLAK5uy_kVpymEyEQazuUAdpFMDH7ktVn_MqR7IQI.html&si=sFsVWqkfQobdhKz_
I think it's a case of "no contest" at this point, especially since Suno has started messing up lyrical cadence and v3 quality really isn't sounding any better than v2. Not sure what happened to v3 quality and the lyrics because Suno hasn't confessed. But comparisons between the current v3 and current v2 don't sound any different to me. Udio isn't just slightly ahead now, they're crossing the finish line while Suno is just coming off the starting line.
And thanks for not making us listen to that full song. I was starting to feel suicidal.
To me, Udio sounds okay but not contemporary. All the music sounds like it came from the 60s.
The only update I want is API access
Pro tip: With the next music review, choose a song that doesn't make viewers want to jump off a bridge- over and over and over ;)
😅
I see you are very young, I started in computers in 1977. I can tell you from loads of experience that this Music AI is in its very early stages. Only time will tell what happens , What is best today might not be best tomorrow. Ive seen this happen with the operating systems of the 80's and 90s, and so many other things. We can only wait an see what happens. Both of the AI music platforms still need serious work and both are pretty good.
:)
I agree that this goes years back. But Matt is pretty good 😅
@@gerardolopez9368 Sure he is but we cant predict what is going to happen next week. Udio came out of nowhere.. I know of 3 other unreleased music creators in development right now ;)
@@AllenParks1omg for real? Where did you hear about this?
Echoes of creation needed a bit of reworking but I liked it better
udio is awesome but it still has a lot of problems when it comes to following.. but is awesome.. would be great if they could let you change the language.. ie if i did a song in english but to be able to change it to another language using the same voice and back music.. and to be able to train are own voices and use them would be awesome.. but the 30 second sections are a pain would be better if it was 1 min or longer.. but yes one of the best. but i still use suno to create my lyrics as i like to see what it creates and then alter before use.. but i use far more extensions with the new setup.. i tried the tree system found it very confusing when you get lots or extensions and retries.. so went back to the old system.. but man i love Udio
and i tend to use about 30 to 60 credits to make one song.. as there are many extensions that just do not follow the flow.. but i am so glad we can now make tracks up to 15 mins long.. but its a lot of hard work..
UDIO has definitely much better sound quality than Suno, but the music itself is just not that good! So I will still be using Suno, becasue it somehow makes the music I want! Maybe it is because I have specific requirement like mostly instrumental, particular instruments or certain cultural vibes, but with SUNO I will not get Western Classical music when I ask for Chinese Epic Cinematic ;)
you cant have enough pockets :)
I almost got the Rabbit R1 but then I did not, I am glad I didn't lol
The problem of audio generation tools' review is that they make the pace of your videos so slow! :(
Would help he didn't choose the single most depressing song I've ever heard 🤷♂
Embodied ai
nice robot song
This sounds like a Peter Gabriel song
Bro made a whole new Johnny Cash song years after he died lol 😂
The robot song sounds way better at 1.5x speed. When I've tried Udio the lyrics are too fast like those near the end.