Finally someone whos got the point. Ai adds a lot of value to your productions. Look at you, your eyes are sparking. This tech is precious, and not that different from the classic work in the studio as a director.
@@iamdjsyko ai will never be enabled to speak your own words from the heart. same for each individual note you play. you are unique and irreplacable. i am currently working on a lot of lyrical stuff, with ai assistance. at the end of the day, after collaborating in the songwriting process with my ai assistant, i am always replacing almost the entire song with my very own words. this can't be done by any other individual than me. i have been very sceptical when the gpt hype started. today i am having a smart assistant with me everytime i need some help. it can communicate in more languages than i will ever be able to learn by myself in my entire life. we are unique, and this will never change. keep this in mind. your art is unique and precious. and you decide whether or not you will play with ai, or your human collaborators. for me i prefer to play with human musicians, when possible. no one can replace your friend in a band. not ai, not even another human, which might even be able to play better that him, or her. we are all individuals, unique. thanks for sharing your personal thought on this.
I’ve been using Udio uploading excerpts of my material and I have to concede to the fact that it baffles me as to how on earth, it comes up with amazing options! It replicated my guitar sound as well as, vocals! Well, 95%! It is disquieting, however.
If you want to get more prompt coherence, switch to manual mode. In its current mode the AI will change your prompt description, often changing the genres completely. Hit the "x" button to clear the suggested genres, you will then see a "manual mode" button underneath the "create" button. Click just to the left of "manual mode" and you will see a toggle switch, turn it on. Now the exact genre prompts you entered will be used.
*AS A SONGWRITER MYSELF OF OVER 40 YEARS... I HAVE WRITTEN OVER 2500 SONGS IN MY LIFETIME... I HAVE BEEN USING SUNO'S NEW EXTEND FEATURE FOR 3 DAYS NOW... TO BE HONEST... I AM COMPLETELY BLOWN AWAY... I FINISHED OFF ONE OF MY MANY MANY UNFINISHED SONGS IN LESS THAN 6 HOURS... THIS IS REVOLUTIONARY... I AM NOW MAKING A MUSIC VIDEO FOR IT... AND THEN ONTO THE NEXT SONG... THIS IS SURELY A GREAT TIME TO BE ALIVE... I ABSOLUTELY LOVE WHAT IT HAS DONE WITH MY MUSIC... UNREAL... I HAVE YET TO TRY UDIO'S EXTEND ABILITY COMPARED TO SUNO'S... BUT I WILL ... SOON... FACT... GREAT VIDEO BRO*
I always had idea for songs but not a band to create with. I started using udio one month ago and I am uploading all on my channel, for now mostly im the one hearing my musics, but feels good to listen those creations. Suno i didnt like that much because the music comes ready...in Udio as you said we need to extend and experience the music 30s at a time to feel if it is the thing that you designed in your head. I enjoy more this process. I always write the lyrics because gpt feels "cold" in my opinion. But Udio will never " win musicians over" but it is a nice tool that fills a gap of new content that it is real and expected due to the ton of work that a band needs to compose. I wrote a lot of diferent content and genres. It is a lot of fun to try it.
Yeah I honestly think this is what this technology was originally meant for and all it should be used for. As long as the courts will hold up their end of the law with the whole undermining clause, this technology will be so beneficial for all songwriters!
Just know that there's is a bit off sounding instrument or vocals, but your own is a plus, which renders like it needs work. Though I'm not a professional producer, you do hear it and need to fix the guitar or redo any other instrument that has a lot of noise, which I can't describe, but it's there.
@@pgarza1218 *I will say that so far... Suno seems to hit the bullseye more times than not... I put in 60 seconds of an idea... get back 3 minutes of a different variant of my idea... even with the same tone and structure... it will generate a violin piece within the 3 minutes that I can now lift and add to my original recording... and not only does it match perfectly with my recording... even the tone matches to where it even fools myself... it sounds that authentic... freaking TERRIFYING... it makes it sound like it was always there... and if I do it 100 times with the same song... it will give 100 different ways to do the song... like it was telling me... as a producer saying "Here is what I think your song should sound like"... that just gives me more tools in my box to use certain elements to add to my song already... a piano line here... a flute part there...a drum passage which just fits unreal... bloody brilliant... and the results can NOT be denied... I will be posting my new songs soon... WHAT A TIME TO BE ALIVE... GOD BLESS FAM*
You can't convince the naysayers that it's at all possible for anything involving AI to possibly be a tool for creatives. Literally the only purpose it can possibly have in their eyes is the total outlaw and banning of human creativity on Earth.
I’ve been messing daily with all sorts of generative ai for years now and I think the key to pushing AI to generate great unique interesting stuff in a prompt isn’t so much about picking qualities you want but picking wild things that can’t possibly work together like “saint-saenz, speed metal, meat grinder, synthwave, analog warmth, gypsy jazz” because prompts are like sticking concepts into a boxing ring and letting them all battle to the death, and the result of this chaos may be genres of music you’ve never heard before but holy shit it has some best aspects of them all.
Yeah I mean sometimes the word salad of genres produces some really cool and unpredictable output... But sometimes the output just feels random and disjointed from the prompt itself. Hard to know which part of what you put in there actually DOES what you hear.
I don't actually think it's an update - it's basically their training model - that's how they made it work in the first place, they fed it other songs and it spits out something very similar but not quite the same.
Well, the update is the UPLOAD feature which Udio didn't have until recently. Everything else about the service, in my understanding, is basically the same. And of course, I'm sure their model will improve and refine over time, particularly as any music uploaded with this feature then becomes part of the training data (at least that's my interpretation of what I read in the terms of service)
I really really enjoy it! At my age, 57, I dont really care to gig anymore with back and hip issues, but the creative soul never left. What I enjoy is writing lyrics, and now I can put my lyrics to an AI band, lol. What I think will happen is AI is basically going to create bands that create and copy AI output into real live music. So the sky is the limit really. For myself, I’ll just create music and throw it up on iTunes and SoundCloud. For me it’s never going to be a money maker but it will be a hobby to enjoy! Oh and btw, you CAN download the stems, and either reproduce in a DAW or use. I’ve found for my particular genre, Roots Reggae, the stems are quite good!! Then I just plug I. My guitar/bass to the DAW, and it somewhat custom, so like you said, a partner!
Okay, so watched the whole thing. You want Ripxdaw. You need it if this is the way you're using Udio. Now, to the rest of it. Totally agree with you on lyrics. None of the ai music machines are good for lyrics, chatgpt can't write anything but bubble gum pop well, and even there, it doesn't follow the same kinds of writing conventions that modern songwriters do. You can always tell when something's written by ai. I like it for vocals, and toplines. I use it against my own lyrics (which are designed to be bendy and challenging for the ai). Think of ai like a creative gateway drug. When you get deep into it, you'll realize that ai by itself isn't enough. If you're an artist, you're going to hit this place where you just want more. For me that meant learning how to play an instrument, and getting deeper into the technical aspects of music production. Doing it, "the old fashioned way," lol, if you will. Another trick you can do with Udio is backfilling. After you render your clip, there's no reason to keep the original piece you put in. You can backfill over that, and essentially have a new song, once your work has been spread over and through the new comp. Just remember, Udio and Suno are musical instruments. They're instruments that default to passable, but doing it well is still something that takes time, that you have to learn, like any other musical instrument. Getting it right is not always easy
The Version 2 (instrumental) sounds almost like 90's weather channel local on the 8's music. I bet these guys would have loved to have this tool in their box back then.
Like you, I definitely do see Udio and AI in general as a tool and a collaborative partner. What I'd like to see from Udio, Suno, or other, as soon as possible, is to enable us as users to have more control; to be able to act more as a producer. Ideally, have the music transcribed so we can go in and change notes and chords as we need to. It may seem like a lot to ask, but if Finale and Sibileus have it, I'm sure these AI companies can implement a MuseScore-type tool. As a musician, I am not at all intimidated by this technology. I've found it to be very helpful. My only desire at this time is for it to enable the users to have more control and act as producers.
Agree with all of this! I do wonder if someone is going to create an AI-forward DAW that really leans into this stuff. If I had the technical capability, I'd do it myself!
Outputting the actual stems is a feature on audio’s radar. Also, if you want it to transition with some kind of beat as it progresses, just mention that in your prompt. Might require manual mode, but might not. Claude Opus creates much better lyrics and you add those in manual mode. It’s best to give Claude a detailed textual walk thru of your thoughts, verse by verse. This is after you have started your Claude/ChatGP prompt like this: Write me lyrics for a synth pop, ambient, space anthem with a instrumental break, and a synthesizer solo in positions that make sense. I want 3 verses. Give me a drop near the end. I want an 8 line double chorus. Enclose section tags in square brackets like [Chorus]. Use simple wording that is more conversational speech format. I want to start with the idea that I’m having a bad day. Progress to meeting someone who magically makes me feel like the sun is peaking thru. Then end with the idea that the two of us are so comfortable together and might be soulmates. Make the chorus magically catchy. The above helps Claude, or ChatGPT give you better lyrics to work with and possible edit to really include your creativity.
I've been using this and suno for a while. Udio had an edge about 2 months back but Suno's new 3.5 blows it out in quality, length and value. It's also free to extend. I'm also a music producer and this is a game changer. When they integrate this into stems or within DAWs, you can kiss spotify bye bye. There's no way both companies did this independently. I suspect both Udio and Suno are arms of Google or OpenAI and they're testing it out. Both release features at the same time. It's a huge undertaking so it must have a big boy behind it.
Thanks for your excellent review of Udio. 2 things I would love to see from Udio or other A.I. platforms: 1- clearcut stem splitting capabilities. 2- The ability to modify and control BPM of rendered takes.
“AI Scene! Bringing beats to the team!” Your cringe was beautiful! You should cold open with that and your reaction before your video. That’s a great hook!
I love your explanation of my phrase I use on my radio show when I play 'music that does it for me'. 😉 I have tried out using Udio for 'sandboxing' station imaging for a new community radio station I will probably be launching in the near future. It did a good job and I just had to tidy it up and remix it myself to perfect it.
That's great. Actually it reminds me of using Ideograms AI to brainstorm logo ideas for my company. Ended up sharing some of those drafts with my human designer. A great starting point!
As a musician, I love what Udio can do. I could see using this AI Tool as a collaborative tool for creating great music. I go through my moments of creative block. Udio can be used as a tool to break that writer's block and add new ideas to enhance the musical piece.
Take a little snippet of what like, then throw in a keyword token totally off the wall in another direction. Don't use the same prompt each time. Then crop, extend, rinse and repeat. It gives you a lot of control in introducing instruments, vocals, lyrics, and such. And that's not including the inpainting!
i expect that they are using all the input from the many people trying this out to add to their machine learning algorithm. so you may find that your song ideas will show up later in someone else's song..what do the terms of service say?
It does appear that uploading material can then be used to train their AI. I wish I had read that before the video was released. Not because I care so much about sharing my own music with the Udio system. But because I would like others to know that explicitly. I will add that to the description. thanks.
My guess is, the terms of service are very vague and in a nutshell, "We can do what we like". ... It's gonna take years for human minds to sort out the legality of all of this.
Suno will generate full 4min songs per generation btw and quite a bit higher in fidelity. It also seems to make better melodic chord progression choices. Sometimes it generates themes it apparently likes so much that it refuses to break from it or ever end. It has an upload feature now as well so sometimes you just have to trim out a clip and re-upload to break it out of its pattern.
After seeing this video, I started using Udio and here are some songs I produced with it my words: www.youtube.com/@magnifico372/videos ruclips.net/video/-nv3TM2zaVI/видео.html Udio's words: ruclips.net/video/XMLjVK03wOo/видео.html ruclips.net/video/zjxLDsKvZEY/видео.html with a poem by Baudelaire: ruclips.net/video/CF3sPTgMSmM/видео.html
Great job on this video! UDIO's new Upload Feature is crazy good. I did a step by step tutorial video on my channel showing how you can take an 8 year old demo of an actual song written by Nashville songwriters and create a newer modern day version of the same song using Udio's upload feature. Technology is just getting better and better quicker than expected which is scary.
You have no idea how important the opinion of a real musician like yourself is when judging these tools. Please keep making content from that angle, most of us who ‘play’ with these tools are not musicians. I am subscribing to following your perspective from now and on… 👍🏻👍🏻
Appears to be a good product. I will try and spread to My youngest Son. I am a hobbyist musician since the 1970s - Drums, Keys/Piano, Bass Guitar, Music Software. My late-eldest Son put Me on to FL Studio in 2008. Once I started using FL Studio/etc. and got good just after a few months of learning it. Given My acoustic/analog music background I felt _kinda_ guilty making music in a fraction of the time a person who plays one instrument for 20 years to master. It took Me awhile to warm up to A.I. of any caliber - visual/construction/design/audio. This is a TOOL I will utilize still maintaining NOT to fully depend on it like anything else in life. *_Like You said, You use this A.I. for inspiration._*
WOW! We are worlds apart regarding musical genre interest but totally share how AI can get us over writers block and gives us content that then inspires us and tests our music ability to produce ourself. Cheers.
@@Techsploder Sorry an assumption based on this vid, I'm sure you have many musical interests! Mine are mainly metal but using AI to convert my ideas over to drum and bass and EDM. It's helping for sure.
A lot of real musicians think that this Ai stuff is cheating. I've been playing guitar for 40 years and I see Ai as no different than collaborating with other "real" musicians. For example: If you write a song or a guitar riff or a hook. You would take that idea to your bandmates and after collaborating with them you might end up with something totally different sounding. This is a common everyday occurrence with musicians. So for those of us that aren't in bands and don't have a network of seasoned musicians, now we can hear those ideas that otherwise never would have seen the light of day. BTW, if we have access to this for free just imagine what recording artists have access to in a multi million dollar recording studio.
Right on! My guitar-oriented music is on Spotify (search "Yellowgold"). Also have an electronic/Industrial project (search "WALL - This Here Life"). Both links are in the description of this video. Thanks for listening!
UDIO isn't as user friendly as SUNO but much more powerful once you learn how to manipulate it. You can get the quality of a master recording with the stems using a DAW. Yes, very impressive.
Udio is far and away the better tool of the two for musicians. Control might make for more complexity... but at least you HAVE control. Suno feels like a bit of a toy by comparison.
@@Techsploder I have found that if you can write great well structured lyrics yourself and not just words that rhyme and are synonyms like the cheesy lyric generator apps do then roll the dice on the same song a hundred times or more you might actually get a hit record with it. The slot machine effect. Lol!
Prompt percussion comes in at bar 16 change key up the scale 1/3 human inspiration and increase the speed of the automation of the beat to drop as a bridge. Echo, automation increasing volume to drop. Creative. Try that prompt.
Thank you Jason . If I were to use this I might use 15 percent of music augmentation .Just to enhance some effects perhaps . I am a hobbyist and like to create music as a matter of adventure and the exploration of experimentation .Someone still has to be behind the scenes giving this stuff direction . I think that it will slop into a routine of predictability .Something that I want to avoid . It may help with some aspects of mixing but it can't represent someone's experiences and certainly can't express them .Will the generation that buys into this will become musical zombies ?
Oh totally. My desire is not necessarily to use what it outputs, wholesale. Maybe sometimes, but really I'm thinking of it more as a way to workshop new ideas that I can apply to my music. Ways to work myself out of a hole.
I’d love to make music that sounds like a popular song but not actually the song for content rights. Is that possible? Like the intro to the Who’s “Baba O’Riley?”
Of course we are. Are we artists if we hit a button in the DAW to generate a random sequence locked to a key that we then manipulate into a form that makes us happy? We aren't handing over the keys, if we choose not to. We have agency as artists. So it really depends on how we use the tools. That's my take anyway. Thanks for watching!
You told it to write lyrics about a terrible topic and then said the lyrics are terrible. Wouldn't you have terrible lyrics in a song that has a bad theme? FADR does good stem separation for cheap. I'm sure it's only a matter of time before they add that in as an option to music generators, but for now FADR is probably the best subscription for unlimited stem separation, and you could just take the remix/extensions you like and break them down into their stems through FADR for the times that you like 'this part' but not 'that part'.
I suppose that's fair, I didn't give it much to work on with the lyrical approach. With practice I'm sure I can get a better understanding of how do get lyrics and vocals that match a particular esthetic that I want. As a vocalist myself, though, not sure if that's a need I really have to fill. Might as well just do my own and leave the AI as an instrumental element collaborator.
@@Techsploder You will have much more success if you START with ai music and then reinterpret it or redo it with real instruments and vocalists using rewritten lyrics... This is the approach I am taking in creating my concept album "Acrylique" ... From AI spark to Human Art ... Check out my song "Bhopal" as an example...
I have uploaded excerpts from some (my own) finished pop rock productions (1990 - 1995) (mixed and mastered). The songs were chopped up, the orders were reversed and the melodies were sung improvised by a different voice. But in the end it got very worse. If you have the talent to write the music yourself and record it yourself, you can't even get inspiration from it. The AI currently only reaches a very limited level. I would be interested to know if Udio plans to build a platform for all productions. So also for your own uploaded selfmade-productions “without AI".
@@Techsploder Clear yes. A musician can now mix AI-generated parts with real recordings on his DAW. For example, I would record all instruments with real singer myself - but add the choir to my singer with AI voices. That sounds better than having the same singer sing with herself 10 times.
Read the terms of service. Your work will (or could be used) for anyone to use. If you're ok with that, (of which I'm also) then great! Just want to make sure you are aware of this. Nothing is for free. But I also am really impressed with Udio & Suno. Also blown away!!! within the next 3-6 months the sound quality will also improve.
@@Techsploder yea, I'm in the same boat. 25+ year musician, hobbyist, artist...ect..at this point, not anyway I'm ever making money on music...lol.. Best wishes!! ---Jesse Lee Smith-
Also, (Not to bombard you with replies) I just launched a dedicated AI Music channel- www.youtube.com/@The-AIUniverse - I used a lot of my own Lyrics on songs I have already distributed, it's in beta still, but I'm doing a mass upload as we speak. I want to see where it goes. Ok, take care brother... -Jesse Lee Smith
GOOD DAY TO ALL, HAVE YOU EVER GOTEN A REVIEW FROM SOME ONE THAT, YOUR UDIO EXPORTS, AS IN TERMS SPECIFICALLY RELATIING TO THE QUALITY OF THE INSTRUMENTS TO BE SUB PAR AS IN SOUNDING OF LOW BIT RATE? DEALING WITH A PARTICULAR ARTIST I THINK JUST WANTS HIS BAND TO COVER THE TRACK AND USE IT FOR INSPIRATION LOL... THOUGHTS??
Using Udio for inspiration is a great way to use the tools imo. The sound generation quality is never 100% when compared to say a sample library, but taking a sound generation and recreating it is whatever quality YOU create.
LALAL.ai is great though I haven't used it in quite a while. Curious to know how it would do with this task. I do fear that every "interpretation" might result in diminished fidelity. I would guess this all gets refined over time.
I bet this Ai is using Theory methods based on optional suggestions through what it offers. However the big question here is once you upload Where's the libraries of songs going to unknown?
an AI to collab with , a tool like that could work for me. But i would like to have more control over it, it should respond to give me inspiration but then i want to takr it apart again, further than only stem separation. So it is nice to experiment with, learn to work with a premature tool like this and it wil become better working within your DAW.
I'm super curious to see how this idea develops. I totally see a time when someday, likely not too far into the future, our modern DAWs will have this functionality baked in, running locally (hopefully), to act as a music collaborator on demand. AND give me stem control to while yer at it!
Lets say UDIO generates an idea for a chorus to our song idea, and we create the performance our self, should we credit UDIO? And if we don't, are we lieing to people for taking full credit for the song? What is proper writers ediquette? ( damn, no spellcheck )
Well that's a great question. I think personal ethics will vary between person to person so if you feel inclined to disclose, I suppose you should. Let me ask you a question though. If you open a piano roll on your DAW and hit a random sequence that lays itself out, and it sounds great and you use it... Do you then tell people "I used the random feature in my DAW to create this sequence." No. And I do believe there are similarities here. No not 1 to 1, but if not one, then why the other? Just something to ponder.
@@Techsploder You make a good point. Referring to what some professional song writers have spoken about concerning guilt feelings that they did not write their song(s), Simon and Garfunkel - 'Fakin' it'. some writers already have guilt even without technology haha. Now we have tech that can possibly write the hardest part of a popular song...the chorus. (for example) "Did you write that song? It's great!!!" "Well...ahhhh yes!" When an established artist such as Bowie hires a musician to contribute on a project, I believe Bowie expects him to contribute his talents and skills in total. This may be some writing contributions in some smaller ways. The musician might consider this contribution as work security that they will be considered for future projects. They sign off the contribution for just a salary. This is understood between the artist and hired musicians. Writers like Bowie and Jimmy Page based their careers on other peoples contributions, so why not Ai then? But where do we draw the line for when do we give credit to the Ai ? Say the Ai writes the lyrics and we change a few words here and there. Or the Ai writes the chorus. Maybe it will just be 'understood' that 'this artist or that artist' notes that "Ai may have been used in the creation of this music" . End of story.
I actually preferred your pre remix to all the AI parts.I prefer to listen to many different genres for inspiration and ideas.This is interesting but for me doesn’t sound good.That being said I would love to hear AI do a remix of Japan song Life without buildings.
Interesting. I think that currently it's only good for ideas. As I could clearly hear the quality diminish drastically in the sections it added. The snare for example went from sounding sharp and snappy, to losing it's attack completely.
Totally agree. I'm certain the actual sound quality will improve to an acceptable point, and we won't have to wait too long for that too happen. But as an idea and brainstorming machine, this is excellent!
Yeah, you can always hear when it kicks in. But then, I remember last year, watching Open AI Juke Box or whatever it was called. I thought, yeah this is interesting and impressive, but just too weird and graney to ever fool anyone, but I wonder what's gonna happen in the next few months. ... Well, here we are, and it's kind of scary. So, what's it gonna be like in a year from now? LOL. AAAAAA! LOL
Are stems possible now with Udio? Can it regenerate the same thing with tweaks (remove this, add that)? Is tweaking by re-directing it really possible yet?
Udio does support breaking it stems at this point, but it's done after the fact, not before. It's also purely generative, so it's not a true removal, more of an approximation if that makes sense. I'm not sure you can easily tell it to "remove this" quite yet.
@@stephendevoreYes Udio can do stems now (see my other udio videos for a demonstration). But they are generatively driven, so they are imperfect. No, I do not believe you can add or remove a single element in generated content. Clearer?
I think IDM is so difficult direction for an AI, that's why it try to change main theme as soon as possible. I agree, AI is very handful for musician then all ideas ended.
I honestly think EDM makes the most sense when it comes to AI generated music. It's music based on machine, based on mathematics, based on algorithms. In my mind that must be easier for a machine to replicate or imitate than something organic.
@@Techsploder The thing is, the man only thinks he's making machine music. For a machine, these are completely incomprehensible, illogical combinations. IMHO
Yeah, AI music tools are getting incredible and the key is to beat AI into submission and by beating, I mean hitting that generate button a whole fuckload of times, using stem separation (like in FL Studio beta!) & z-plane decoda to pull all the most brilliant bits, variations, sounds out and make it your own.
I've been using Udio for a few weeks and a lot of it needs work but I have also gotten some really good results. I've uploaded a few to my channel if you want to give them a listen?
Ooooh that's a good question. I don't THINK you can get Udio or Suno to do just that gap at this stage. But I could totally see that coming sometime down the line. Sounds like a Photoshop AI feature applied to audio actually. Cool idea!
Actually yes. Waves is one company doing AI-assisted Mastering, there are others. And I'd imagine this aspect of the music industry will continue to see gains in the coming months.
This site can obviously create Female voice and a guitar sound? Wonder if it could be used to add a Rock Guitar & Female harmony vocal to a song I've been working on? I supposed I'd have to buy credits or something...
If I want it just to sample my voice, I am presuming that I would only have up upload a short vox file, which is fine, but how can I ask it to use my voice to create a sample so that their AI voice sounds like mine? What you did was a continuation of your actual music, but is there a specific way to sample something? That way, it's not exactly going to copy an entire song of mine, and therefore I'm not too worried about anything being used except for a few vocal lines because it will be unrecognisable. I just want to know how it samples my voice. Thanks :)
I've used Eleven Labs to clone my voice (not for music, for spoken word) and its pretty impressive. And improving. But still it wasn't TOTALLY there. Will be in due time, no question.
As a musician, I find this to be inspiring. Not because I want an AI to write my music for me, that's not it at all. I will never hand over the keys. But I'm always open to new perspectives and new ways to look at the music I create through a different lens. I think using the AI in this capacity is actually pretty empowering. Thanks for watching!
I will say, though... I'm really trying to find the balance between "overt RUclips thumbnail face" and "my actual, organic emotion". Trying to not go TOO overboard.
I agree, and given that Mr Beast spends $US10,000 for each thumbnail, putting work into them can really pay dividends. Having said that I think the whole production is getting more an more polished each episode, thanks for all your hard work Jason.
Of course it's just a tool and it still has to learn about meaningful and tasteful chord progressions, and listening to some of the examples I heard some pretty nonsensical and irritating melody lines that good musicians never would use. The lyrics Udio or other AI tools come up with mostly come across very uninspiring and soulless and totally lack ingenuity due to the fact that those tools don't have a lifetime of personal, emotional human experiences. Let's not forget: AI is just a copycat machine although in some respect a quite useful one and fun to play with.
Yeah, I mean I agree... and I disagree. Yes its fun to play with. Yes some of the output can be rather uninspired (its a machine after all, inspiration doesn't exist). But I would argue that some of the melody lines it generates might not be amazing as they are written but in the hands of the right producer, could be made to shine. It's almost like it can write the notes that might sound good if only the sheen was given to them by someone who really knows what they are doing. And beyond all of that, if it creates something that lights a fire in the mind of the producer to try something else because of it, what does it matter. In that case, IMO, it was the right tool for the job.
@@Techsploder What I am still missing is that you can't determine what chord progressions it should use or what key a particular song should be composed in. Because of the 30 seconds limitation it also only delivers a 32 bar structure which means that you still don't have the freedom to tell it how many bars of intro, pre chorus, Interlude or outro you envision and you also can't put in metric changes from 4/4 to 2/4 and/or other "odd" ones. Most often you have some specific lyrics but then you have to doctor them in a specific and quite restricted way so that AI can cope with them. If he still would be alive, I guess people like Leonard Cohen would be pretty frustrated working with AI as it is right now. All in all, AI still has to learn a lot from us to meet our demands. I don't expect it to be as ingenious as Jacob Collier is but I'd very much like to have more flexibility and variety to implement more than just average musical concepts and ideas.
@@Techsploder yes eventually, the daw and the ai will be completely integrated. I'm thinking any day now, someone is going to release an AI VST. Suno VST or Udio VST. I don't care which comes first! The power of AI in whatever DAW you like would be really nice.
Lot's of people point out the many AI sites that do stem removal. Personally, I'd love everything to be in one integrated tool. Kind of like a Perplexity for musicians.
Please define "register it." My analysis of the TOS is: I retain my original rights and ownership of my original audio doesn't transfer to them, I do grant them usage rights though. Udio can use my audio for its own promotional purposes as well as training data for its system. Now, I say this while fully admitting that this stuff is developing currently. I certainly don't recommend others do this without their own research and analysis.
Maybe I'm misunderstanding your point, but I believe I uploaded three totally different things of different variety. Even with a relatively simple drum and bass pattern (assuming you are referring to the first track I uploaded), my point was to see what kind of accompaniment ideas Udio would give me. Dare I say, it did a great job doing exactly that.
UDIO has always fallen short of harmonic structure and especially harmonic rhythm (on which beats or measures chords change), compared to Suno 3. But then, Suno tunes quickly become lame because they all follow the same overused chord progressions ala Taylor Swift.
I have done a couple (31+). I hope the emphasis will end up being on the songwriting rather than some of chance you happen to meet likeminded that contribute to the message and creativity to contribute to the intriguing loose interpretation that leaves the listener to interpret the text how it applies to their circumstances or emotional state, made by songwriters. Not worshipping the dull minded musician that can’t express the music with putting words to it. But I am biased. I find people unable to express themselves with words to convey a meaning that music uplifts, when music itself apparently is already replaceable by AI. My only issue is the compilation time that takes time. I would have been finished with most I made with better control over the end results. Lyrics take like, what? 2 minutes raw time to be written down? My patience is non existent.
hi ! just reacting to the title... impressed? I understand when RUclipsrs hat have nothing to do with music talks about AI music and are impressed... and while its pretty good at structure, generating ideas, and a few things, it actually sounds pretty bad and fake in the end, especially voices. Any sound engineer, artists/producer can hear it. I guess, we will have to wait a year or two. And not even talking about copyright issues... but sure, its funny.
I'm impressed with it's ability to create IDEAS that I can apply to my own productions, yes. As a system that creates new songs from scratch, wholesale, I'm less interested.
Careful man, anything you upload to the site you consent them ownership. Then their system will deconstruct your song and add it to their database for training the AI. You get no credit for making their system better. It's in their terms.
Your not worried about handing over ownership of your songs? You know your even giving then the right to distribute it. And snips of your song may appear in other people's generated stuff? Now you got me really worried lol @@Techsploder
So, Udio does not retain ownership. I still own the music I upload. They are granted a license to use, reproduce, and promote with that audio. I'm not a career musician. I honestly don't care nor do I see it as a huge threat to anything I'm doing. The license is non exclusive as well so if I really wanted to do something with that music, I still can. As for snippets of it ending up in other creations on the platform, I suppose I'm not that concerned if my music is part of the training base and output. Unless it goes totally rogue, it won't output my song wholesale, just snippets will inform the output it creates. I can be ok with that.
No worries. Thanks for your reply anyways man. Was just giving you a heads up on the matter. I can see how it can be beneficial, I just personally don't agree with their terms. I seen a post of a massive court case coming for udio and suno. Apparently they have used a massive amount of copyrighted music from the giant labels to train their software aswell as indie music. So A LOT of what it is creating are derivative works of copyrighted content in some way, shape or form. @@Techsploder
@@ZEKAVEO oh yeah, no worries! I appreciate it. And NO QUESTION it's derivative output. But, perhaps a controversial take, but I think that's what makes using the system in THIS way (as an inspiration machine) so appealing to me. Take the scope of modern music, now look at my song, and give me ideas of ways I can take it given what you know about the broad spectrum of modern music trends and techniques.
its an ego thing, why the person here gets sparkling eyes - its "his" song and it feels like admirition...like fake aplause ... bit pathetic, but if it makes him happy, who cares?
That's an interest take. I would say that admiration isnt what it feels like. More like partnership. I created something, then I fed it into a system and it gave me a different idea to try. I don't consider that admiration at all. But you know, "who cares".
I'm testing different approaches out. Believe me, I get it. But I don't know what works and what doesn't work if I don't try and fail and try again. Thanks for the feedback.
Alright, tweaked the title. Better? (And look. Click bait titles are NOT my instinct. But I've found that RUclips has a whole set of things that "work". And when I'm trying to build a business, I need to try all avenues. That's how I know what works best for ME. Thanks for caring enough to reply.)
Fair... But what are your thoughts on using AI to fill out these basic ideas and use them as a foundation for more? Part of my goal here is to feed the system ideas that aren't fully developed, and less than fully developed music is often, well, boring.
While this is interesting, a serious concern I have is that by feeding your songs into this AI, you and your music are in the AI's machine learning copy pool because you've trained it to sound like you. Yes, true that it's a miniscule drop in the AI's learned bucket, but it could start spitting out results for other people that erode your own originality because it's learned how to replicate you. That is not cool.
I realize that not everyone will feel the way I do about this... But I really do not care if I'm completely honest. I see no reason to feel like anything about my musical creativity is harmed by an AI that's influenced in a miniscule way by my own music. It's such a drop in the bucket. But I fully realize that my not making a living from my music puts me in a position of privilege in that perspective. I might feel differently if my livelihood was driven by my musical uniqueness and creativity.
@Techsploder All excellent points. And thanks for making this great and very informative video. The AI remix process is super interesting. It would be cool if you could use this AI in a closed environment where it didn't share back into the main pool. I wonder if that will become a feature in future models based on subscription choice? Some companies are doing that for AI systems for companies that don't want or cannot allow their proprietary information out in the public via AI, so maybe this will come to music to at some point?
![NOSFERATU - Official Trailer [HD] - Only In Theaters December 25](http://i.ytimg.com/vi/nulvWqYUM8k/mqdefault.jpg)
Finally someone whos got the point.
Ai adds a lot of value to your productions.
Look at you, your eyes are sparking.
This tech is precious, and not that different from the classic work in the studio as a director.
Value as collaborative partner, that's definitely where I've landed on it. Thanks for watching!
I totally agree with you! Ai adds to the creative process for serious Music Composers and Producers. That's how its suppose to be used!
for about 5 minutes you're right. Until it levels up to do all of our own part of the process itself too, making us utterly obsolete.
@@iamdjsyko ai will never be enabled to speak your own words from the heart. same for each individual note you play. you are unique and irreplacable. i am currently working on a lot of lyrical stuff, with ai assistance. at the end of the day, after collaborating in the songwriting process with my ai assistant, i am always replacing almost the entire song with my very own words. this can't be done by any other individual than me. i have been very sceptical when the gpt hype started. today i am having a smart assistant with me everytime i need some help. it can communicate in more languages than i will ever be able to learn by myself in my entire life. we are unique, and this will never change.
keep this in mind. your art is unique and precious. and you decide whether or not you will play with ai, or your human collaborators. for me i prefer to play with human musicians, when possible. no one can replace your friend in a band. not ai, not even another human, which might even be able to play better that him, or her. we are all individuals, unique.
thanks for sharing your personal thought on this.
If you need AI "to add value to your productions," you have a problem.
I’ve been using Udio uploading excerpts of my material and I have to concede to the fact that it baffles me as to how on earth, it comes up with amazing options! It replicated my guitar sound as well as, vocals! Well, 95%! It is disquieting, however.
Yeah it is quite something, eh? I haven't gotten any sort of useful replication of my voice however. How did you pull that off?
If you want to get more prompt coherence, switch to manual mode. In its current mode the AI will change your prompt description, often changing the genres completely. Hit the "x" button to clear the suggested genres, you will then see a "manual mode" button underneath the "create" button. Click just to the left of "manual mode" and you will see a toggle switch, turn it on. Now the exact genre prompts you entered will be used.
I started my experimentation with that feature on. But after some playing around, I understand it's purpose better. Thanks for sharing!
@@psychedelicafrica4817 Do you mean adherence? Prompt coherence is down to the person writing the prompt 😅
@@Techsploder Yeah but automatic use tags to guaranty the best result in term of audio quality. Not in term of audio coherence xD
*AS A SONGWRITER MYSELF OF OVER 40 YEARS... I HAVE WRITTEN OVER 2500 SONGS IN MY LIFETIME... I HAVE BEEN USING SUNO'S NEW EXTEND FEATURE FOR 3 DAYS NOW... TO BE HONEST... I AM COMPLETELY BLOWN AWAY... I FINISHED OFF ONE OF MY MANY MANY UNFINISHED SONGS IN LESS THAN 6 HOURS... THIS IS REVOLUTIONARY... I AM NOW MAKING A MUSIC VIDEO FOR IT... AND THEN ONTO THE NEXT SONG... THIS IS SURELY A GREAT TIME TO BE ALIVE... I ABSOLUTELY LOVE WHAT IT HAS DONE WITH MY MUSIC... UNREAL... I HAVE YET TO TRY UDIO'S EXTEND ABILITY COMPARED TO SUNO'S... BUT I WILL ... SOON... FACT... GREAT VIDEO BRO*
Nice, great work! Its cool to take dead ideas and bring them back to life!
I always had idea for songs but not a band to create with. I started using udio one month ago and I am uploading all on my channel, for now mostly im the one hearing my musics, but feels good to listen those creations. Suno i didnt like that much because the music comes ready...in Udio as you said we need to extend and experience the music 30s at a time to feel if it is the thing that you designed in your head. I enjoy more this process. I always write the lyrics because gpt feels "cold" in my opinion.
But Udio will never " win musicians over" but it is a nice tool that fills a gap of new content that it is real and expected due to the ton of work that a band needs to compose. I wrote a lot of diferent content and genres. It is a lot of fun to try it.
Yeah I honestly think this is what this technology was originally meant for and all it should be used for.
As long as the courts will hold up their end of the law with the whole undermining clause, this technology will be so beneficial for all songwriters!
Just know that there's is a bit off sounding instrument or vocals, but your own is a plus, which renders like it needs work. Though I'm not a professional producer, you do hear it and need to fix the guitar or redo any other instrument that has a lot of noise, which I can't describe, but it's there.
@@pgarza1218 *I will say that so far... Suno seems to hit the bullseye more times than not... I put in 60 seconds of an idea... get back 3 minutes of a different variant of my idea... even with the same tone and structure... it will generate a violin piece within the 3 minutes that I can now lift and add to my original recording... and not only does it match perfectly with my recording... even the tone matches to where it even fools myself... it sounds that authentic... freaking TERRIFYING... it makes it sound like it was always there... and if I do it 100 times with the same song... it will give 100 different ways to do the song... like it was telling me... as a producer saying "Here is what I think your song should sound like"... that just gives me more tools in my box to use certain elements to add to my song already... a piano line here... a flute part there...a drum passage which just fits unreal... bloody brilliant... and the results can NOT be denied... I will be posting my new songs soon... WHAT A TIME TO BE ALIVE... GOD BLESS FAM*
You completely nailed how AI is supposed to be used. It ADDS TO YOUR PRODUCTIONS!!!
Thanks! Totally agree, a tool to be used to enhance your workflow.
You can't convince the naysayers that it's at all possible for anything involving AI to possibly be a tool for creatives. Literally the only purpose it can possibly have in their eyes is the total outlaw and banning of human creativity on Earth.
You're so wrong. You really think in the future the human creation will be superior to AI?
I’ve been messing daily with all sorts of generative ai for years now and I think the key to pushing AI to generate great unique interesting stuff in a prompt isn’t so much about picking qualities you want but picking wild things that can’t possibly work together like “saint-saenz, speed metal, meat grinder, synthwave, analog warmth, gypsy jazz” because prompts are like sticking concepts into a boxing ring and letting them all battle to the death, and the result of this chaos may be genres of music you’ve never heard before but holy shit it has some best aspects of them all.
Yeah I mean sometimes the word salad of genres produces some really cool and unpredictable output... But sometimes the output just feels random and disjointed from the prompt itself. Hard to know which part of what you put in there actually DOES what you hear.
Logic has a stem splitter that's great.
Great idea! Future video!
Dope sound brother! I absolutely loved that first track you did. 🔥 Thanks for the video and explaining things.
Right on! Thanks for watching. I plan to dive into RipX DAW in the next week or two, keep your eyes posted!
I don't actually think it's an update - it's basically their training model - that's how they made it work in the first place, they fed it other songs and it spits out something very similar but not quite the same.
Well, the update is the UPLOAD feature which Udio didn't have until recently. Everything else about the service, in my understanding, is basically the same. And of course, I'm sure their model will improve and refine over time, particularly as any music uploaded with this feature then becomes part of the training data (at least that's my interpretation of what I read in the terms of service)
I really really enjoy it! At my age, 57, I dont really care to gig anymore with back and hip issues, but the creative soul never left. What I enjoy is writing lyrics, and now I can put my lyrics to an AI band, lol.
What I think will happen is AI is basically going to create bands that create and copy AI output into real live music. So the sky is the limit really.
For myself, I’ll just create music and throw it up on iTunes and SoundCloud.
For me it’s never going to be a money maker but it will be a hobby to enjoy!
Oh and btw, you CAN download the stems, and either reproduce in a DAW or use. I’ve found for my particular genre, Roots Reggae, the stems are quite good!!
Then I just plug I. My guitar/bass to the DAW, and it somewhat custom, so like you said, a partner!
Right on, very refreshing and I feel you on a lot of what you wrote here. Thanks for sharing!
Fascinating
Right on!
Okay, so watched the whole thing.
You want Ripxdaw. You need it if this is the way you're using Udio.
Now, to the rest of it. Totally agree with you on lyrics. None of the ai music machines are good for lyrics, chatgpt can't write anything but bubble gum pop well, and even there, it doesn't follow the same kinds of writing conventions that modern songwriters do. You can always tell when something's written by ai.
I like it for vocals, and toplines. I use it against my own lyrics (which are designed to be bendy and challenging for the ai).
Think of ai like a creative gateway drug.
When you get deep into it, you'll realize that ai by itself isn't enough.
If you're an artist, you're going to hit this place where you just want more.
For me that meant learning how to play an instrument, and getting deeper into the technical aspects of music production.
Doing it, "the old fashioned way," lol, if you will.
Another trick you can do with Udio is backfilling. After you render your clip, there's no reason to keep the original piece you put in. You can backfill over that, and essentially have a new song, once your work has been spread over and through the new comp.
Just remember, Udio and Suno are musical instruments.
They're instruments that default to passable, but doing it well is still something that takes time, that you have to learn, like any other musical instrument.
Getting it right is not always easy
Ripxdaw?
On the same page. I just heard from the folks at RipX so I plan to dive into it in the coming weeks! Thanks for sharing your thoughts!
Yep! I will be doing some vids on this in the coming weeks.
Rytr, it's an AI specifically designed for writing song lyrics in various emotions/context. It does a better job than Taylor Swift, that's for sure.
The Version 2 (instrumental) sounds almost like 90's weather channel local on the 8's music. I bet these guys would have loved to have this tool in their box back then.
You nailed my greatest inspiration. The weather channel.
Like you, I definitely do see Udio and AI in general as a tool and a collaborative partner. What I'd like to see from Udio, Suno, or other, as soon as possible, is to enable us as users to have more control; to be able to act more as a producer. Ideally, have the music transcribed so we can go in and change notes and chords as we need to. It may seem like a lot to ask, but if Finale and Sibileus have it, I'm sure these AI companies can implement a MuseScore-type tool.
As a musician, I am not at all intimidated by this technology. I've found it to be very helpful. My only desire at this time is for it to enable the users to have more control and act as producers.
Agree with all of this! I do wonder if someone is going to create an AI-forward DAW that really leans into this stuff. If I had the technical capability, I'd do it myself!
Wish the damn thing would spit out MIDI files.
Outputting the actual stems is a feature on audio’s radar. Also, if you want it to transition with some kind of beat as it progresses, just mention that in your prompt. Might require manual mode, but might not.
Claude Opus creates much better lyrics and you add those in manual mode. It’s best to give Claude a detailed textual walk thru of your thoughts, verse by verse.
This is after you have started your Claude/ChatGP prompt like this:
Write me lyrics for a synth pop, ambient, space anthem with a instrumental break, and a synthesizer solo in positions that make sense. I want 3 verses. Give me a drop near the end. I want an 8 line double chorus. Enclose section tags in square brackets like [Chorus]. Use simple wording that is more conversational speech format. I want to start with the idea that I’m having a bad day. Progress to meeting someone who magically makes me feel like the sun is peaking thru. Then end with the idea that the two of us are so comfortable together and might be soulmates. Make the chorus magically catchy.
The above helps Claude, or ChatGPT give you better lyrics to work with and possible edit to really include your creativity.
I've been using this and suno for a while. Udio had an edge about 2 months back but Suno's new 3.5 blows it out in quality, length and value. It's also free to extend. I'm also a music producer and this is a game changer. When they integrate this into stems or within DAWs, you can kiss spotify bye bye. There's no way both companies did this independently. I suspect both Udio and Suno are arms of Google or OpenAI and they're testing it out. Both release features at the same time. It's a huge undertaking so it must have a big boy behind it.
I plan to jump into Suno with these experiments too, I want to know which is good for what things, in THIS capacity!
Thanks for your excellent review of Udio.
2 things I would love to see from Udio or other A.I. platforms: 1- clearcut stem splitting capabilities. 2- The ability to modify and control BPM of rendered takes.
Yeah, both of those features would be great, totally agree.
“AI Scene! Bringing beats to the team!” Your cringe was beautiful! You should cold open with that and your reaction before your video. That’s a great hook!
Yeah that's a great idea... Too late now! Still learning the ropes of this stuff. Dang!
You can use software to split the stems. There are so many choices out there.
I know there are. I haven't used them enough to know the kind of quality I can get out of them. But I will!
Very good video. Thanks. So, Udio is a kind of partner to develop ideas and to grow-up the inspiration.
That's how I see it. I have another video coming this week along this line.
I love your explanation of my phrase I use on my radio show when I play 'music that does it for me'. 😉
I have tried out using Udio for 'sandboxing' station imaging for a new community radio station I will probably be launching in the near future. It did a good job and I just had to tidy it up and remix it myself to perfect it.
That's great. Actually it reminds me of using Ideograms AI to brainstorm logo ideas for my company. Ended up sharing some of those drafts with my human designer. A great starting point!
As a musician, I love what Udio can do. I could see using this AI Tool as a collaborative tool for creating great music. I go through my moments of creative block. Udio can be used as a tool to break that writer's block and add new ideas to enhance the musical piece.
Absolutely agree! I think that's its biggest strength (to me anyway)
Take a little snippet of what like, then throw in a keyword token totally off the wall in another direction. Don't use the same prompt each time. Then crop, extend, rinse and repeat. It gives you a lot of control in introducing instruments, vocals, lyrics, and such. And that's not including the inpainting!
I haven't even touched inpainting. Tell me what you use it for! Maybe I'll throw something together with that feature! I want to learn it.
i expect that they are using all the input from the many people trying this out to add to their machine learning algorithm. so you may find that your song ideas will show up later in someone else's song..what do the terms of service say?
It does appear that uploading material can then be used to train their AI. I wish I had read that before the video was released. Not because I care so much about sharing my own music with the Udio system. But because I would like others to know that explicitly. I will add that to the description. thanks.
My guess is, the terms of service are very vague and in a nutshell, "We can do what we like". ... It's gonna take years for human minds to sort out the legality of all of this.
15:05 - If Black Flag did synth-pop educational themes
That is PERFECT
Excelent review.
Thank you! More like this on the horizon, once I get over this COVID!
Suno will generate full 4min songs per generation btw and quite a bit higher in fidelity. It also seems to make better melodic chord progression choices. Sometimes it generates themes it apparently likes so much that it refuses to break from it or ever end. It has an upload feature now as well so sometimes you just have to trim out a clip and re-upload to break it out of its pattern.
After seeing this video, I started using Udio and here are some songs I produced with it
my words:
www.youtube.com/@magnifico372/videos
ruclips.net/video/-nv3TM2zaVI/видео.html
Udio's words:
ruclips.net/video/XMLjVK03wOo/видео.html
ruclips.net/video/zjxLDsKvZEY/видео.html
with a poem by Baudelaire:
ruclips.net/video/CF3sPTgMSmM/видео.html
It's like having Brian Eno in the room with his oblique strategies
Oooh i like that analogy!
Great job on this video! UDIO's new Upload Feature is crazy good. I did a step by step tutorial video on my channel showing how you can take an 8 year old demo of an actual song written by Nashville songwriters and create a newer modern day version of the same song using Udio's upload feature. Technology is just getting better and better quicker than expected which is scary.
Oh cool! I will have to check that out!
Loved it!
You have no idea how important the opinion of a real musician like yourself is when judging these tools. Please keep making content from that angle, most of us who ‘play’ with these tools are not musicians.
I am subscribing to following your perspective from now and on… 👍🏻👍🏻
Right on! Welcome to the channel. 😎
Appears to be a good product. I will try and spread to My youngest Son.
I am a hobbyist musician since the 1970s - Drums, Keys/Piano, Bass Guitar, Music Software.
My late-eldest Son put Me on to FL Studio in 2008.
Once I started using FL Studio/etc. and got good just after a few months of learning it. Given My acoustic/analog music background I felt _kinda_ guilty making music in a fraction of the time a person who plays one instrument for 20 years to master.
It took Me awhile to warm up to A.I. of any caliber - visual/construction/design/audio.
This is a TOOL I will utilize still maintaining NOT to fully depend on it like anything else in life. *_Like You said, You use this A.I. for inspiration._*
Inspiration, exactly. Not much different from listening to a song that catches your attention and suddenly having a new idea for your own music.
@@Techsploder - I agree with You
WOW! We are worlds apart regarding musical genre interest but totally share how AI can get us over writers block and gives us content that then inspires us and tests our music ability to produce ourself. Cheers.
Right on! What genres do you dabble in?
@@Techsploder Sorry an assumption based on this vid, I'm sure you have many musical interests! Mine are mainly metal but using AI to convert my ideas over to drum and bass and EDM. It's helping for sure.
A lot of real musicians think that this Ai stuff is cheating. I've been playing guitar for 40 years and I see Ai as no different than collaborating with other "real" musicians. For example: If you write a song or a guitar riff or a hook. You would take that idea to your bandmates and after collaborating with them you might end up with something totally different sounding. This is a common everyday occurrence with musicians. So for those of us that aren't in bands and don't have a network of seasoned musicians, now we can hear those ideas that otherwise never would have seen the light of day. BTW, if we have access to this for free just imagine what recording artists have access to in a multi million dollar recording studio.
I totally agree!
Where can we find your music? (I got hooked).
Right on! My guitar-oriented music is on Spotify (search "Yellowgold"). Also have an electronic/Industrial project (search "WALL - This Here Life"). Both links are in the description of this video. Thanks for listening!
UDIO isn't as user friendly as SUNO but much more powerful once you learn how to manipulate it. You can get the quality of a master recording with the stems using a DAW. Yes, very impressive.
Udio is far and away the better tool of the two for musicians. Control might make for more complexity... but at least you HAVE control. Suno feels like a bit of a toy by comparison.
@@Techsploder I have found that if you can write great well structured lyrics yourself and not just words that rhyme and are synonyms like the cheesy lyric generator apps do then roll the dice on the same song a hundred times or more you might actually get a hit record with it. The slot machine effect. Lol!
AI triggers people. Those it does have eliminated themselves from the future. They are so emotional about it
Sounds like Jamiroqaui!
...huh didn't see that one coming, but thanks? :)
Prompt percussion comes in at bar 16 change key up the scale 1/3 human inspiration and increase the speed of the automation of the beat to drop as a bridge. Echo, automation increasing volume to drop. Creative. Try that prompt.
If you want the stems use Moises it can seperate drums, vocal etc
Great suggestion, appreciated!
Thank you Jason . If I were to use this I might use 15 percent of music augmentation .Just to enhance some effects perhaps . I am a hobbyist and like to create music as a matter of adventure and the exploration of experimentation .Someone still has to be behind the scenes giving this stuff direction . I think that it will slop into a routine of predictability .Something that I want to avoid . It may help with some aspects of mixing but it can't represent someone's experiences and certainly can't express them .Will the generation that buys into this will become musical zombies ?
Oh totally. My desire is not necessarily to use what it outputs, wholesale. Maybe sometimes, but really I'm thinking of it more as a way to workshop new ideas that I can apply to my music. Ways to work myself out of a hole.
The find the upload button anymore, is it still there in the new Udio version?
I see it. It's there.
👏🏻
I’d love to make music that sounds like a popular song but not actually the song for content rights. Is that possible? Like the intro to the Who’s “Baba O’Riley?”
Btw, have you heard the song "Help, My Mom Is On Only Fans" on Udio? Freakin' Hilarious 😁 Actually it's here on RUclips as well...
Are we still an artist after all this AI-Remixes? To get idea impulses?
Of course we are. Are we artists if we hit a button in the DAW to generate a random sequence locked to a key that we then manipulate into a form that makes us happy? We aren't handing over the keys, if we choose not to. We have agency as artists. So it really depends on how we use the tools. That's my take anyway. Thanks for watching!
You told it to write lyrics about a terrible topic and then said the lyrics are terrible. Wouldn't you have terrible lyrics in a song that has a bad theme? FADR does good stem separation for cheap. I'm sure it's only a matter of time before they add that in as an option to music generators, but for now FADR is probably the best subscription for unlimited stem separation, and you could just take the remix/extensions you like and break them down into their stems through FADR for the times that you like 'this part' but not 'that part'.
I suppose that's fair, I didn't give it much to work on with the lyrical approach. With practice I'm sure I can get a better understanding of how do get lyrics and vocals that match a particular esthetic that I want. As a vocalist myself, though, not sure if that's a need I really have to fill. Might as well just do my own and leave the AI as an instrumental element collaborator.
BE WARNED: Read their Terms of Service carefully! Once you upload your ideas into their collective brain, they become more "Ours" than "Yours"
Yes you certainly need to read up on in TOS to be clear on rights going in.
any way add backing tracks to my vocals only
I would love to know how to do this: just add X ELEMENT to my existing music and structure. I wonder if this is possible.
@@Techsploder i will try
the source music is boring... i would use udio on its own ... lol
Fair... But that's not really the point of the video.
@@TechsploderI liked it, I was twerking to it
Respect
@@Techsploder You will have much more success if you START with ai music and then reinterpret it or redo it with real instruments and vocalists using rewritten lyrics... This is the approach I am taking in creating my concept album "Acrylique" ... From AI spark to Human Art ... Check out my song "Bhopal" as an example...
I have uploaded excerpts from some (my own) finished pop rock productions (1990 - 1995) (mixed and mastered). The songs were chopped up, the orders were reversed and the melodies were sung improvised by a different voice. But in the end it got very worse. If you have the talent to write the music yourself and record it yourself, you can't even get inspiration from it. The AI currently only reaches a very limited level. I would be interested to know if Udio plans to build a platform for all productions. So also for your own uploaded selfmade-productions “without AI".
Huh thats an interesting question. So are you saying the platform would display a person's AI generated as well as NON-AI generated music catalogue?
@@Techsploder Clear yes. A musician can now mix AI-generated parts with real recordings on his DAW. For example, I would record all instruments with real singer myself - but add the choir to my singer with AI voices. That sounds better than having the same singer sing with herself 10 times.
Read the terms of service. Your work will (or could be used) for anyone to use. If you're ok with that, (of which I'm also) then great! Just want to make sure you are aware of this. Nothing is for free. But I also am really impressed with Udio & Suno. Also blown away!!! within the next 3-6 months the sound quality will also improve.
Yes I am aware of the TOS. For this material, it doesn't bother me.
@@Techsploder yea, I'm in the same boat. 25+ year musician, hobbyist, artist...ect..at this point, not anyway I'm ever making money on music...lol.. Best wishes!! ---Jesse Lee Smith-
Also, (Not to bombard you with replies) I just launched a dedicated AI Music channel- www.youtube.com/@The-AIUniverse - I used a lot of my own Lyrics on songs I have already distributed, it's in beta still, but I'm doing a mass upload as we speak. I want to see where it goes. Ok, take care brother... -Jesse Lee Smith
Take your time next time I enjoyed it
Thanks!
GOOD DAY TO ALL, HAVE YOU EVER GOTEN A REVIEW FROM SOME ONE THAT, YOUR UDIO EXPORTS, AS IN TERMS SPECIFICALLY RELATIING TO THE QUALITY OF THE INSTRUMENTS TO BE SUB PAR AS IN SOUNDING OF LOW BIT RATE? DEALING WITH A PARTICULAR ARTIST I THINK JUST WANTS HIS BAND TO COVER THE TRACK AND USE IT FOR INSPIRATION LOL... THOUGHTS??
Using Udio for inspiration is a great way to use the tools imo. The sound generation quality is never 100% when compared to say a sample library, but taking a sound generation and recreating it is whatever quality YOU create.
How long did the generation of the first attempt take? Missed it because of the "magic of pausing and flashing forward into the future" :) 05:36
Not long at all. All generations took around 60-90 seconds or so.
I said the stems thing myself - there is LALALai which does it, so it could easily be incorporated....a basic sequencer might be cool '')
LALAL.ai is great though I haven't used it in quite a while. Curious to know how it would do with this task. I do fear that every "interpretation" might result in diminished fidelity. I would guess this all gets refined over time.
I bet this Ai is using Theory methods based on optional suggestions through what it offers. However the big question here is once you upload Where's the libraries of songs going to unknown?
Their TOS does state that music uploaded is used to train the AI going forward, so there's that!
an AI to collab with , a tool like that could work for me. But i would like to have more control over it, it should respond to give me inspiration but then i want to takr it apart again, further than only stem separation. So it is nice to experiment with, learn to work with a premature tool like this and it wil become better working within your DAW.
I'm super curious to see how this idea develops. I totally see a time when someday, likely not too far into the future, our modern DAWs will have this functionality baked in, running locally (hopefully), to act as a music collaborator on demand. AND give me stem control to while yer at it!
Lets say UDIO generates an idea for a chorus to our song idea, and we create the performance our self, should we credit UDIO? And if we don't, are we lieing to people for taking full credit for the song? What is proper writers ediquette? ( damn, no spellcheck )
Well that's a great question. I think personal ethics will vary between person to person so if you feel inclined to disclose, I suppose you should. Let me ask you a question though. If you open a piano roll on your DAW and hit a random sequence that lays itself out, and it sounds great and you use it... Do you then tell people "I used the random feature in my DAW to create this sequence." No. And I do believe there are similarities here. No not 1 to 1, but if not one, then why the other? Just something to ponder.
@@Techsploder You make a good point. Referring to what some professional song writers have spoken about concerning guilt feelings that they did not write their song(s), Simon and Garfunkel - 'Fakin' it'. some writers already have guilt even without technology haha. Now we have tech that can possibly write the hardest part of a popular song...the chorus. (for example) "Did you write that song? It's great!!!" "Well...ahhhh yes!"
When an established artist such as Bowie hires a musician to contribute on a project, I believe Bowie expects him to contribute his talents and skills in total. This may be some writing contributions in some smaller ways. The musician might consider this contribution as work security that they will be considered for future projects. They sign off the contribution for just a salary. This is understood between the artist and hired musicians. Writers like Bowie and Jimmy Page based their careers on other peoples contributions, so why not Ai then? But where do we draw the line for when do we give credit to the Ai ? Say the Ai writes the lyrics and we change a few words here and there. Or the Ai writes the chorus. Maybe it will just be 'understood' that 'this artist or that artist' notes that "Ai may have been used in the creation of this music" . End of story.
I actually preferred your pre remix to all the AI parts.I prefer to listen to many different genres for inspiration and ideas.This is interesting but for me doesn’t sound good.That being said I would love to hear AI do a remix of Japan song Life without buildings.
Well thank you , then!!
Interesting. I think that currently it's only good for ideas. As I could clearly hear the quality diminish drastically in the sections it added. The snare for example went from sounding sharp and snappy, to losing it's attack completely.
Totally agree. I'm certain the actual sound quality will improve to an acceptable point, and we won't have to wait too long for that too happen. But as an idea and brainstorming machine, this is excellent!
Yeah, you can always hear when it kicks in. But then, I remember last year, watching Open AI Juke Box or whatever it was called. I thought, yeah this is interesting and impressive, but just too weird and graney to ever fool anyone, but I wonder what's gonna happen in the next few months. ... Well, here we are, and it's kind of scary. So, what's it gonna be like in a year from now? LOL. AAAAAA! LOL
@@PeterCamberwick Always improving, eh? It really has come a long way in a short amount of time.
sounds like you could use ripx daw
I actually have a license and will be doing some content around this asap!
do you know Udio has a new version out -- 1.5
YES... I plan to do a similar walkthrough on that, probably tomorrow.
Are stems possible now with Udio?
Can it regenerate the same thing with tweaks (remove this, add that)?
Is tweaking by re-directing it really possible yet?
Yes, I watched the whole video.
Udio does support breaking it stems at this point, but it's done after the fact, not before. It's also purely generative, so it's not a true removal, more of an approximation if that makes sense. I'm not sure you can easily tell it to "remove this" quite yet.
@@Techsploder Huh?
@@stephendevoreYes Udio can do stems now (see my other udio videos for a demonstration). But they are generatively driven, so they are imperfect. No, I do not believe you can add or remove a single element in generated content. Clearer?
A little OMD from the album joan of arc album
That's what I'm talkin about!
I think IDM is so difficult direction for an AI, that's why it try to change main theme as soon as possible.
I agree, AI is very handful for musician then all ideas ended.
I honestly think EDM makes the most sense when it comes to AI generated music. It's music based on machine, based on mathematics, based on algorithms. In my mind that must be easier for a machine to replicate or imitate than something organic.
@@Techsploder The thing is, the man only thinks he's making machine music. For a machine, these are completely incomprehensible, illogical combinations.
IMHO
Yeah, AI music tools are getting incredible and the key is to beat AI into submission and by beating, I mean hitting that generate button a whole fuckload of times, using stem separation (like in FL Studio beta!) & z-plane decoda to pull all the most brilliant bits, variations, sounds out and make it your own.
This is the way!
Yeah, I think I'll just rite a song. LOL
@@PeterCamberwick yeah, “rite” a song lol
Don't touch 40%. And play with seeds.
RTFM.
Yeah I definitely want to spend more time tweaking the seeds to see how that influences things.
Get Moises, better than RipX for stems especially if you get the paid versions to get hifi wav and drums and other splits
Sweet, I will check out Moises. I have a license for RipX that I'll be testing as well.
Yeah, sometimes Udio has a 'make it stop' moment for sure😁
In some ways it is a bit like a slot machine, pull the lever and hope for the best. But when it gets it, it gets it!
@@Techsploder Exactly
The way I see it, do a bare bones track, and get the A.I. to do the rest, 😎
That's certainly one way to go!
Do we sue artists who say they get their inspiration from listening to various other artists?
Absolutely not!
Never!
I've been using Udio for a few weeks and a lot of it needs work but I have also gotten some really good results. I've uploaded a few to my channel if you want to give them a listen?
I have archival recordings that have dropouts of ten seconds or so. Is there a tool that can fill in these gaps?
Ooooh that's a good question. I don't THINK you can get Udio or Suno to do just that gap at this stage. But I could totally see that coming sometime down the line. Sounds like a Photoshop AI feature applied to audio actually. Cool idea!
Is there a way to give a file to an AI to improve sound quality.. or maybe even master it?
Actually yes. Waves is one company doing AI-assisted Mastering, there are others. And I'd imagine this aspect of the music industry will continue to see gains in the coming months.
Those lyrics, tho! 😂 Straight outta some really bad 80s movie! 🤭
Oh yeah, TOTALLY awful!
Oh yeah, TOTALLY awful!
This site can obviously create Female voice and a guitar sound? Wonder if it could be used to add a Rock Guitar & Female harmony vocal to a song I've been working on? I supposed I'd have to buy credits or something...
Buy some credits and test it out. I think it should be able to do something along those lines .
If I want it just to sample my voice, I am presuming that I would only have up upload a short vox file, which is fine, but how can I ask it to use my voice to create a sample so that their AI voice sounds like mine? What you did was a continuation of your actual music, but is there a specific way to sample something? That way, it's not exactly going to copy an entire song of mine, and therefore I'm not too worried about anything being used except for a few vocal lines because it will be unrecognisable. I just want to know how it samples my voice. Thanks :)
11 labs has the feature to clone your voice. This is infancy stage and it keep improving…
I've used Eleven Labs to clone my voice (not for music, for spoken word) and its pretty impressive. And improving. But still it wasn't TOTALLY there. Will be in due time, no question.
maybe add "fusion jazz" tag.....maybe sound better
Add that for a fusion jazz output? Or does it give something different and unexpected?
As a musician. This is depressing
As a musician, I find this to be inspiring. Not because I want an AI to write my music for me, that's not it at all. I will never hand over the keys. But I'm always open to new perspectives and new ways to look at the music I create through a different lens. I think using the AI in this capacity is actually pretty empowering. Thanks for watching!
Thumbnail game on target
I have to get over myself and just do the thing that "works"
@@Techsploder yeah you got to embrace that inner cringe !
I will say, though... I'm really trying to find the balance between "overt RUclips thumbnail face" and "my actual, organic emotion". Trying to not go TOO overboard.
I agree, and given that Mr Beast spends $US10,000 for each thumbnail, putting work into them can really pay dividends. Having said that I think the whole production is getting more an more polished each episode, thanks for all your hard work Jason.
Thank you so much!!@@webgenius937
I can't make music. But now I can.
Unfortunately.
@@bobbilly2397 Feel like a rockstar now💥
@@dj0ai It's not your music; it's only the rearrangement of other people's works, made by machine.
@@bobbilly2397 It's a new world.
Of course it's just a tool and it still has to learn about meaningful and tasteful chord progressions, and listening to some of the examples I heard some pretty nonsensical and irritating melody lines that good musicians never would use. The lyrics Udio or other AI tools come up with mostly come across very uninspiring and soulless and totally lack ingenuity due to the fact that those tools don't have a lifetime of personal, emotional human experiences. Let's not forget: AI is just a copycat machine although in some respect a quite useful one and fun to play with.
Yeah, I mean I agree... and I disagree. Yes its fun to play with. Yes some of the output can be rather uninspired (its a machine after all, inspiration doesn't exist). But I would argue that some of the melody lines it generates might not be amazing as they are written but in the hands of the right producer, could be made to shine. It's almost like it can write the notes that might sound good if only the sheen was given to them by someone who really knows what they are doing. And beyond all of that, if it creates something that lights a fire in the mind of the producer to try something else because of it, what does it matter. In that case, IMO, it was the right tool for the job.
@@Techsploder I agree. A tool in the right hands with the right mind can become a miracle. 😉
@@Techsploder What I am still missing is that you can't determine what chord progressions it should use or what key a particular song should be composed in.
Because of the 30 seconds limitation it also only delivers a 32 bar structure which means that you still don't have the freedom to tell it how many bars of intro, pre chorus, Interlude or outro you envision and you also can't put in metric changes from 4/4 to 2/4 and/or other "odd" ones.
Most often you have some specific lyrics but then you have to doctor them in a specific and quite restricted way so that AI can cope with them.
If he still would be alive, I guess people like Leonard Cohen would be pretty frustrated working with AI as it is right now.
All in all, AI still has to learn a lot from us to meet our demands.
I don't expect it to be as ingenious as Jacob Collier is but I'd very much like to have more flexibility and variety to implement more than just average musical concepts and ideas.
Try your own lyrics. Using Country as a tag is great for songwriters. Don’t select too many different styles.
Using country as a tag is great for songwriters even if they aren't trying to create country?
Well, use AI to split it up into stems! It's just as good at that!
Yes, although each of those steps can cause some degradation of the final sound. I'm sure that will all get better with time.
@@Techsploder yes eventually, the daw and the ai will be completely integrated. I'm thinking any day now, someone is going to release an AI VST. Suno VST or Udio VST. I don't care which comes first! The power of AI in whatever DAW you like would be really nice.
@@dspivey4284 Oh yeah, you know this is coming! Someone has to be working on this.
Yes, stems would be useful!
Lot's of people point out the many AI sites that do stem removal. Personally, I'd love everything to be in one integrated tool. Kind of like a Perplexity for musicians.
hurray!!! eye am kidding of coarse. we cannot eccept this .sometimes you need to stay in the corner...its a happy place.
???
as per the "it's useful when i am stuck in a corner" reference.
You uploaded your own music? Then, I guess it's not YOUR music anymore. Now you can't register it - read the terms of service.
Please define "register it." My analysis of the TOS is: I retain my original rights and ownership of my original audio doesn't transfer to them, I do grant them usage rights though. Udio can use my audio for its own promotional purposes as well as training data for its system. Now, I say this while fully admitting that this stuff is developing currently. I certainly don't recommend others do this without their own research and analysis.
Chatgpt 4 + Udio AI + Tensor art AI
ruclips.net/video/DXpihH0IHIA/видео.html
Sounds good!
You could upload something more interesting than a drum and bass pattern
Maybe I'm misunderstanding your point, but I believe I uploaded three totally different things of different variety. Even with a relatively simple drum and bass pattern (assuming you are referring to the first track I uploaded), my point was to see what kind of accompaniment ideas Udio would give me. Dare I say, it did a great job doing exactly that.
UDIO has always fallen short of harmonic structure and especially harmonic rhythm (on which beats or measures chords change), compared to Suno 3. But then, Suno tunes quickly become lame because they all follow the same overused chord progressions ala Taylor Swift.
You've picked up on a lot of subtleties that I haven't noticed yet.
Lalala ai if you want the stems
Yeah, might have to do a video that takes all of this further with LALAL.
I have done a couple (31+).
I hope the emphasis will end up being on the songwriting rather than some of chance you happen to meet likeminded that contribute to the message and creativity to contribute to the intriguing loose interpretation that leaves the listener to interpret the text how it applies to their circumstances or emotional state, made by songwriters. Not worshipping the dull minded musician that can’t express the music with putting words to it. But I am biased. I find people unable to express themselves with words to convey a meaning that music uplifts, when music itself apparently is already replaceable by AI.
My only issue is the compilation time that takes time. I would have been finished with most I made with better control over the end results. Lyrics take like, what? 2 minutes raw time to be written down? My patience is non existent.
hi ! just reacting to the title... impressed? I understand when RUclipsrs hat have nothing to do with music talks about AI music and are impressed... and while its pretty good at structure, generating ideas, and a few things, it actually sounds pretty bad and fake in the end, especially voices. Any sound engineer, artists/producer can hear it. I guess, we will have to wait a year or two. And not even talking about copyright issues... but sure, its funny.
I'm impressed with it's ability to create IDEAS that I can apply to my own productions, yes. As a system that creates new songs from scratch, wholesale, I'm less interested.
Careful man, anything you upload to the site you consent them ownership. Then their system will deconstruct your song and add it to their database for training the AI. You get no credit for making their system better. It's in their terms.
Thanks for the warning. I'm familiar with the TOS and not concerned personally. But I can totally understand why people might be.
Your not worried about handing over ownership of your songs? You know your even giving then the right to distribute it. And snips of your song may appear in other people's generated stuff? Now you got me really worried lol @@Techsploder
So, Udio does not retain ownership. I still own the music I upload. They are granted a license to use, reproduce, and promote with that audio. I'm not a career musician. I honestly don't care nor do I see it as a huge threat to anything I'm doing. The license is non exclusive as well so if I really wanted to do something with that music, I still can. As for snippets of it ending up in other creations on the platform, I suppose I'm not that concerned if my music is part of the training base and output. Unless it goes totally rogue, it won't output my song wholesale, just snippets will inform the output it creates. I can be ok with that.
No worries. Thanks for your reply anyways man. Was just giving you a heads up on the matter. I can see how it can be beneficial, I just personally don't agree with their terms. I seen a post of a massive court case coming for udio and suno. Apparently they have used a massive amount of copyrighted music from the giant labels to train their software aswell as indie music. So A LOT of what it is creating are derivative works of copyrighted content in some way, shape or form. @@Techsploder
@@ZEKAVEO oh yeah, no worries! I appreciate it. And NO QUESTION it's derivative output. But, perhaps a controversial take, but I think that's what makes using the system in THIS way (as an inspiration machine) so appealing to me. Take the scope of modern music, now look at my song, and give me ideas of ways I can take it given what you know about the broad spectrum of modern music trends and techniques.
If it something I didn't create you can keep it.
Fair
its an ego thing, why the person here gets sparkling eyes - its "his" song and it feels like admirition...like fake aplause ... bit pathetic, but if it makes him happy, who cares?
That's an interest take. I would say that admiration isnt what it feels like. More like partnership. I created something, then I fed it into a system and it gave me a different idea to try. I don't consider that admiration at all. But you know, "who cares".
Jason, Jason, Jason..... A clickbait title? You're better than this.
I'm testing different approaches out. Believe me, I get it. But I don't know what works and what doesn't work if I don't try and fail and try again. Thanks for the feedback.
Alright, tweaked the title. Better? (And look. Click bait titles are NOT my instinct. But I've found that RUclips has a whole set of things that "work". And when I'm trying to build a business, I need to try all avenues. That's how I know what works best for ME. Thanks for caring enough to reply.)
Man your music is so boring sorry…
Fair... But what are your thoughts on using AI to fill out these basic ideas and use them as a foundation for more? Part of my goal here is to feed the system ideas that aren't fully developed, and less than fully developed music is often, well, boring.
While this is interesting, a serious concern I have is that by feeding your songs into this AI, you and your music are in the AI's machine learning copy pool because you've trained it to sound like you. Yes, true that it's a miniscule drop in the AI's learned bucket, but it could start spitting out results for other people that erode your own originality because it's learned how to replicate you. That is not cool.
I realize that not everyone will feel the way I do about this... But I really do not care if I'm completely honest. I see no reason to feel like anything about my musical creativity is harmed by an AI that's influenced in a miniscule way by my own music. It's such a drop in the bucket. But I fully realize that my not making a living from my music puts me in a position of privilege in that perspective. I might feel differently if my livelihood was driven by my musical uniqueness and creativity.
@Techsploder All excellent points. And thanks for making this great and very informative video. The AI remix process is super interesting. It would be cool if you could use this AI in a closed environment where it didn't share back into the main pool. I wonder if that will become a feature in future models based on subscription choice? Some companies are doing that for AI systems for companies that don't want or cannot allow their proprietary information out in the public via AI, so maybe this will come to music to at some point?