by Julian Schmauch | 2,0 / 5,0 | Approximate reading time: 2 Minutes
Is Adobe's AI music generator music production's Photoshop?

Is Adobe's AI music generator music production's Photoshop?  ·  Source: Adobe

ADVERTISEMENT

We’ve seen plenty of AI music generators in the last few years. From Google to Meta to a plethora of startups, many have tried to achieve success and quality similar to image and video generation. But so far, none has come even close. Is the new AI music generator from Adobe the missing piece?

ADVERTISEMENT

Adobe Project Music GenAI Control: Photoshop for music production?

One of the biggest issues with AI music generators is that outside of the prompt like “lazy lounge music with saxophone and marimba” or “beach house with drop” you don’t have any control over the outcome. Which is desperately needed for something so detailed and granular as music production. It seems like Adobe’s new Ai music generator “Music GenAI Control” may solve that problem.

RipX Deep Audio
RipX Deep Audio

To be fair, “Photoshop in music” = Audioshop has been claimed by AI/VR-DAW makers RipX DAW for a while now. But here we’re talking about AI music generation. In a video demo Adobe just released, their new and still very experimental Music GenAI prototype is shown doing more than simple generation through prompts.

AI music generator from Adobe: How good is it?

Based on the video, it seems like the new Adobe AI music generator is still prompt-based. It also seems like a piece of software that goes pretty deep in terms of editing capabilities. Some of the spectral images remind of iZotope’s flagship post-production suite RX 10, which is also heavily AI-based. Adobe’s researcher Nick Bryan speaks of “music co-creation” in the demo video.

ADVERTISEMENT

You are currently viewing a placeholder content from Youtube. To access the actual content, click the button below. Please note that doing so will share data with third-party providers.

More Information

He starts with a piano melody and then builds arrangements around it with prompts “inspiring film” or “intense hip-hop”. Nick also shows a way to increase the intensity of an orchestra arrangement, not in terms of volume, but in terms of expression. They show an obviously AI-generated little RnB snippet to illustrate how Adobe’s music generator can build a song structure from a small piece of a song.

In addition, the prototype appears to generate fully loopable little compositions, which might be helpful for film and game music composition. Something we have seen from other AI music generators, extending existing musical pieces, is also part of GenAI.

When can we try the Adobe AI music generator?

Adobe’s AI music generator prototype was announced at the Ann Hot Pod Summit in Brooklyn. It appears to be in the very early stages and, in many ways, is seemingly still somewhat experimental. That’s why we probably won’t see any version of it in any Adobe software soon.

But given how quick the company was to integrate image generation through its Firefly platform, it be here sooner then we think, most likely in Adobe Audition or Adobe Premiere, if I was to guess. Adobe did mention what type of songs its AI was trained on and if any musician’s have been involved in the training.

More on AI and Adobe

Image Sources:
  • RipX Deep Audio: Hit'n'Mix
Is Adobe's AI music generator music production's Photoshop?

How do you like this post?

Rating: Yours: | ø:
ADVERTISEMENT

11 responses to “Is Adobe’s AI music generator music production’s Photoshop?”

    Diki Ross says:
    2

    Now is the time that composers need to assert their copyright, before this gets as out of hand as image AI has become.

    AI isn’t creative. It’s basically a bot that goes out and samples EVERYTHING and then does an elaborate mashup. But what it’s mashing up is YOUR music. Copyrighted.

    Plagiarism lawsuits can cost unwary composers a fortune, but we are giving AI a free pass, or at least we do until it’s too late. Legitimate protected content creators like newspapers and visual artists are already battling AI infringement, and most of them lament that they started so late.

    If we don’t jump all over this NOW, we’ll be in the same boat in a year or two’s time.

      Ab says:
      3

      People don’t realise how bad AI is at sooooooo many levels. I’m glad copyright makes a useful tool to fight back but that thing is way worse than that. Not because it will become sentient or take over the world (it won’t, it’s just advanced automation at the end of the day) but precisely because it works really bad and we’re putting it everywhere. Also it burns a developed country’s worth of electricity. Also AI evangelists are borderline cultists (look for the acronym TESCREAL, you’ll find a lot of serious scholars and journalists explaining you the ideology behind AI)

    Bc says:
    1

    In a week full of sad and horrifying by news stories this is the one that hits me the hardest. They are developing a robot to replace me. All the work I’ve put into developing this skill over my life and soon it will be impossible to compete with the AI. Yes the music I make is 100 times better than the AI. But the ai is cheap and fast. When it comes down to it many clients will prefer cheap and fast even if it means the music has no character.

      An. says:
      0

      Yeah this is incredibly sad. AI won’t kill concerts or bands because people want humans (they freak out for backing tracks!)

      However for musicians making a living doing corporate stuffs or TV or whatever, it will make hurt them a lot… I suppose it’s your case and we shouldn’t accept that.

        die beastie boys 😁😁 says:
        0

        Even if AI were to take over, they would still need people who know how to use the AI to create music. None of what I’ve seen indicates that we will be pressing a magic button that spews out perfect jingles. AI is a tool that will be used to assist some musicians in writing music, just as our current technology does.

        If you ask me, we already have technology that achieves many of the things that AI does. It might take a couple extra steps, but a good track ultimately comes down to the creativity of the user. Using AI to make anything good will still take creativity, which AI does not have.

        And if you’re worried that the results of what people/companies will start creating/using will be utterly terrible, then we can hardly blame AI for that. People are already resigned and content with, well, “content”.

        Content has been prized over substance for decades now, and AI isn’t going to make it any worse. I personally think it’s going to be more of the same. The same people “making beats” and drag-and-dropping MIDI chord packs will be making full tracks with AI, and they’ll have the same fanbase as they do now.

        As I’ve said many times before, people who appreciate “real” music aren’t going to suddenly stop appreciating authenticity and creativity just because of some new AI gimmick. I really think things will remain pretty much the same. AI is all hype for the wrong reasons.

    Mark Dolan 🇬🇧 says:
    3

    No artist need be concerned by this nonsense. Even when the algorithm is ‘perfected’ it will produce utter crap, as that is what it has ‘learned’ and is certainly what the market user will require. The only people who will be remotely affected by such an algorithm are ‘hirey’ producers who churn out total garbage in industrial quantities to the purchasers who are Adobe’s target audience. This is exactly what is beginning to happen with the image generation algorithms. This may be an economic concern for the Hirey, but for artists this is utterly irrelevant

      die beastie boys 😁😁 says:
      0

      I 100% agree, and I’m so glad someone here is considering this topic from a place of reason. Every time there is a post about AI, I feel like I’m the only commenter trying to think about this in a sensible manner.

      The type of music that may be “replaced” by AI is already generic, formulaic, and whatever other disparaging adjective you want to call it. The people (or companies) currently listening to (or using) this type of music will continue doing so.

      The type of people who want to listen to “real” music will also continue to do so. No one is going to suddenly stop supporting their favorite artists, even if AI-generated music was really good. People want to support people. Even people who are fans of pop artists who might be considered “fake”, they still expect a certain level of authenticity. It may be manufactured authenticity, or it may be a small amount of true authenticity, but it’s still something that people want. The reason it has to be faked is because people want it.

      The other point that people are missing is that AI is a tool. Just like any other tool or process, it can be used creatively, or it can be used to create garbage. There is no magic button that will materialize perfect audio from nothing.

      The other thing is that people are already using AI to make music. I forgot what song/band it was, but I once heard a guitarist mention that a fan had learned how to play this super-complex guitar riff (or is it a lick? lol). The guitarist was so amazed because he hadn’t even played the original himself. The original recordnig was created with a comparably basic lick that was modulated by some kind of wonky delay pedal to create a much more complex tapestry (I hate to use that word haha).

      I know that example is not exactly AI, but it goes to show you two things. 1) We are alerady using tools to make not just music, but everything. AI is ultimately just another tool. 2) Not only can people use AI to create better things, but we can also create things better than AI.

      You’re not going to get replaced by AI unless you’re making “boring”, generic, and formulaic music. But even so, that’s exactly what some people want. Honestly, more power to them. Maybe that’s what gets them through a relaxing, chill study session.

      But there’s a reason people gravitate towards something like “lo-fi girl”. It’s a personification or representation of something real. It’s a cultural landmark. There’s a reason why the lo-fi-girl stream lists what artists are playing, because no one wants to listen to an endless, nameless, faceless, AI-generated audio-hellscape that just rushes at them from the void. They want music created by people.

    SNKY says:
    1

    This comment was generated by AI:
    The abysmal music generative AI produced mediocre, nonsensical compositions that defied any semblance of musical coherence. Its algorithmic choices seemed oblivious to nuanced rhythm, melody, or harmony, resulting in the auditory equivalent of the color greige. This misguided attempt at artificial creativity demonstrated a fundamental misunderstanding of musical principles, yielding an unbearably bland result.

    bert says:
    -1

    i hate the rise of AI in all its forms so much. It is such a demise of the world.

    bob says:
    0

    I hope ai is a hype.
    When you want to make music, you want to MAKE music. In the end there is no fun in telling a bot to make something, Creative art is creative, i saw someone publish a childrens book, with ai images and ai text. How utterly ‘crap’ must this person feel in the end. There is no personal investment.
    However, companies going for ai generated designs and music, that is a whole different bad situation

    Alex Anders says:
    0

    ai doesn’t remove humanities ability to create, it will however greatly reduce commercial prospects for multitudes.

    Anyway, for better or worse, it won’t be stopped.

Leave a Reply

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