Join us as we dive into the latest improvements to Udio’s AI Music Generation software, which is still in its beta phase. In this video, we explore the exciting new features that have just been introduced, including a longer context window for better music continuity, an extended maximum track length for more comprehensive tracks, and the innovative new trim feature that was added last a week. Discover how these updates set Udio apart in the competitive world of AI music generation. Don’t miss our hands-on demo of each feature to see how they can improve your music-making process!

By admin

26 thoughts on “Udio’s Latest AI Music Update!”
  1. That's a tricky question you ask, if people/we care. I'd say, for me personally, the answer would be both yes and no. As a songwriter, my #1 priority was always the melody, a strong melody, especially an emotional ballad with the impact to make people cry. Yes, music is so much more than just that, but to me, the real essence of music is to create songs that move you – ABBA said a lot of it in Thank you for the Music, really.

    So… with that in mind, can AI do this? With Udio, I've found the answer for the first time to be yes, although with some reservations. It depends how you "feed" it with combinations and orders of tags, styles, lyrics and instruments – and also your perseverance chasing THE generation that works. One of the biggest bumps, I've found to be ai-generated lyics, simply not sounding natural and re-using certain phrases over and over, stuff like "in a(n) … where…", using the word "embrace" in almost every song, as well as quite a few other words. But even a cliche-filled song CAN work, if everything else is right.

    So here is the thing: AI can't (at least not at the current stage) create music that really "works" on an emotional level, BUT it CAN be used as a tool to help create music that works on an emotional level. That's a difference, and the question is, who should be considered the "creator" of such a song, nurtured by the human, but actually suggested by the AI?

    But this still hasn't answered your question, has it? Do we (I) CARE if the music is AI generated or not? I certainly do care if people delivering songs in melodic genres no longer bother to invest enough into their songs to give them the best possible impact, if they become complacent and say "this'll do" just because the AI is able to quickly deliver something that resembles a melody. But if a songwriter has worked with dedication with an AI and the result, after countless re-rolls and fine tunings, is a deeply moving song, then I really don't care if he/she used AI to get there.

  2. On my good studio speakers I only hear distorted quality, instruments are not transparent, cruel mixing errors, compositional errors. Electric guitars sound terrible. In two years people will be so fed up with AI music that hand-made music will be traded like organic apples. In any case, there is an oversupply that no one wants to hear – except you.

  3. First time listener, loved the content. Feels like I’ve finally may have found my community ❤ IMHO on peoples opinions, I would agree that on instrumentals there is barely an issue. But on tracks with vocals, there is a lot of hostility albeit from what seems to be less people, but very strong nevertheless. I think the copyright conversation is likely to inflame things. But I agree that ultimately people will get over it and move on, use themselves.

  4. From what I've seen, Udio is simply plagiarizing existing music for their profit. I find that a deeply problematic basis for their technology and I'm not that interested in what it does w/o that being changed.

  5. It's only a matter of time before streaming services are buying the rights to bands so they can produce "new" tracks. Ringo and Paul need to get some of that action before it's too late.
    I recently made three tracks on Udio – all supposedly 1970s space rock/jazz fusion. Not only were the tracks very "authentic" sounding, when I sent them to three knowledgeable friends none of them could tell they were AI generated. Admittedly, I did cheat by giving the tracks a completely plausible origin story, but not only did my friends think the tracks were made by real musicians in the early 1970s, they all thought they were great tracks and wanted to hear more.
    The only way I could tell the difference was because the Udio audio output sounds very "mushy", which was why I chose the early 70s. That way I could pretend the tracks were bootlegs of one sort or another.
    Will people care? Well, they might, but will they even know the difference?

  6. Anecdotally I know the average person won’t really give a rats left butt cheek where it came from. I make joke/novelty AI music for my friends in styles I know they like and we had a good laugh and I sort of forgot about it. A couple weeks later when we got together s later they put the music on and were dancing to it and sining it word for word and just have a blast.

  7. It’s a strange time for those of us who’ve been musicians and composers since before youtube even existed. I don’t like this new way of doing things. It removes so much from the process and joy of creation for me. For example, it’s like being a guitarist vs playing “Guitar Hero” — not a perfect analogy, but captures how music creation is more of a “push-button” process in which you simply keep or discard outputs. That trial/error process exists in traditional composing, but it’s far more imaginative and has infinite breadth and variation potential. The AI platform takes so much away from the human creator, to the extent that it doesn’t even seem like “creation” anymore to me.

  8. Especially if you provide the lyrics and song structure so the AI is just providing the instruments and vocals it's becoming harder to tell and people won't really care. After all, look at modern singers and modern so-called artists. They're so heavily processed and auto-tuned already their singing is practically AI anyhow.

  9. Here's how I hope things play out. I've been a musician for 40 years and I do it for fun. I guess you could call me a serious hobbyist. I envision one day someone will hire us musicians to perform these songs that they created with AI but can't play an instrument. If not, oh well, I will still have fun playing instruments.

  10. Ai is just another tool for composers to write. Like will.i.am said: "Now i just have more time to spend on writing the song in different ways."

    As for the audience, many wont care, but we'll probably have a label. It'll be like organic fruit. A few will pay the extra buck for the "organic" version.

  11. I think consumers of music who don’t know anyone or anything in music will not care. People in the industry and gigging musicians hate it. They feel like it will take their job. I like it to help me create ideas because I am in so so much debt spending money working with others. I WOULD pay musicians and not use AI or VSTs if I have tons of money.

  12. Two days ago, I would have answered your question yes. Except, I tried Udio for the first time yesterday, so I fully understand the answer will be no in the future. People won't know the difference, and they won't care. I thought it was just a novelty, but it's become shockingly good in a very short time. In a few years, I imagine this technology will be indistinguishable from the real thing, and users will be able to customize any aspect of the production from instruments to final mix. Then it will be less about the machine, and more about anyone being able to pour their imagination directly into a WAV file.

  13. Personally I really do not care as long as the music is good and talks to me !
    I used Udio as well for a few days and it's just fantastic for a song writer without a band ! The AI is practically taking your lyrics and works with you to find the fitting music,it's like jamming ! I think many people really do not care what kind of music they listen to or who made it,as long as it's kinda good ! And keep in mind,it will just get better and better from here with this tool !

  14. I am in the music biz and i love all of this. In the 80s we were drawing in pixels and used to say "Won't it be amazing one day if we could make these photo-realistic?" And photoshop came along, and photographers all got angry and feared for their careers, and now there are more photogs than ever and they all use Photoshop. Audio-realistic music and sound generation is WAY OVERDUE! Go with it. Its here. Play and have fun. It takes me as long to assemble an acceptable song in udio or suno as it would take me to record teack by track. Its the same exact production process to decide whether a take is good. The computer creates a LOT of suck on the way to making solid gold, but the gold it occasionally makes is EPIC gold 😊 It still takes a lot of work and a lot of editing to make greatness. I wish I could copyright the stuff i produce with ai. My input and decisions are critical to wualiy production.

    I must say that with suno and with udio ive made a few of my favorite songs of all-time. Not kidding.

    Thanks for the amazing video. I had no idea about the updates. As of now i dont see the rollouts in my account. Im excited!

  15. Oh my god that trimming feature is a GAME CHANGER. Udio produces so much good stuff but very often there were parts I wasn't happy with and there was no way to trim them out. This is incredible

Leave a Reply

Your email address will not be published. Required fields are marked *