Future of AI and DJ for ChatGpt

Options
2»

Comments

  • lord-carlos
    lord-carlos Member Posts: 2,472 Expert
    Options

    Here is a relativ new example of what AI can generate https://stability.ai/research/stable-audio-efficient-timing-latent-diffusion

    Far from perfect, but it shows how far we have come. And I can see that for some people that "epic trailer" music would be usable for a quick youtube video they are making. The death metal is .. horrible though.

    I would assume that in a few years AI is in some shape or form part of most DAWs. Expect maschine because NI always is behind :P

    I used ChatGPT to help me write a minimal ~soundcloud~ like website to selfhost my sets and make it easy to listen to it.

  • Milkman
    Milkman Member Posts: 219 Advisor
    Options

    Its ok! Before you realize the end result of this technology, it will be too late for you to restore your prior value at the labor and ideas marketplace and your new competitors will be software. You should ignore all cautionary advice from any senior technologists who bring it, and leap straight into "what you cannot stop", especially in brand-centric places like the NI forums.

  • Kayya
    Kayya Member Posts: 197 Advisor
    Options

    Yes, friends, I am sharing with you an example that I came across on the internet from the microphone item. I am in a daze.

    https://www.instagram.com/reel/Cwp0-PSNGHn/?igshid=MzRlODBiNWFlZA==

  • Stevan
    Stevan Traktor Mapping Mod Posts: 1,731 mod
    Options
    https://community.native-instruments.com/discussion/comment/96074#Comment_96074

    Beatboxers are going to love this.

  • Kubrak
    Kubrak Member Posts: 2,808 Expert
    edited September 2023
    Options
    https://community.native-instruments.com/discussion/comment/96074#Comment_96074

    Yes, it is where AI has its place. To transform something to something else. Multilayer neural networks do not make anything else than transform points from n-dimensional space to m-dimensional space. And projection is learnt by presenting many, many pairs - corresponding points - from those two spaces.

    NI uses this approach for ICM modelling in Guitar Rig.

  • Fritz
    Fritz Member Posts: 28 Member
    Options

    We already use AI in a lot of things like beatgridding and harmonics. AI, however complex it may be still behaves according to the inputs given. It is not creative at all. It just puts together other stuff that was made up by humans. Its a tool that we should use to help us. It will never take the human factor out. And when we look at the most succesfull DJs, we can seriously doubt that their individual technical skills are perfect, as it more their human outside and character that makes them likeable. AI isnt going to change that.

  • Kayya
    Kayya Member Posts: 197 Advisor
    Options
    https://community.native-instruments.com/discussion/comment/96117#Comment_96117
    https://community.native-instruments.com/discussion/comment/96124#Comment_96124

    I'm just learning that artificial intelligence is used in subjects such as Beatgrid and key. I was thinking that artificial intelligence was used in dj for the first time with stems. Also, while opening this topic, I didn't take the issue of making music with the microphone very seriously, it seemed to me that he had time. I need to remind you that the person in the video is a dj. An interesting level.

  • Kubrak
    Kubrak Member Posts: 2,808 Expert
    Options

    AI is pretty old thing, 40 years at least. I have already used neural networks more than 30 years ago.

    So, it has been used for many little tasks for years.... Just it did not have publicity.

  • Kayya
    Kayya Member Posts: 197 Advisor
    Options
    https://community.native-instruments.com/discussion/comment/96191#Comment_96191

    At this point, I can't comment because I don't have knowledge. It is necessary to be able to distinguish artificial intelligence well with algorithmic applications.

  • Kubrak
    Kubrak Member Posts: 2,808 Expert
    Options

    Neural networks like they are used now are known for about 40 years, I have used them more than 30 years ago. That time, in early nineties, it was big hit. But computers were rather weak to use it for other tasks than relatively simple ones. Mind that current computers are about 1000 times stronger than that time....

    The danger then and also now was/is that AI gives good answer when question hits region of space that is mapped well (is well covered by training data) and gives very wrong answer when asking sparsely trained region or not trained at all.

    And one never knows, how much reliable the answer is. If one asks human he may answer, I think it is this and this, but I do not know, I am guessing. But AI does not give any guidance to significance of its answer....

Back To Top

Categories