Since AI is a pervasive part of our lives now, I thought I would share with you what it has done for my music. Maybe you’ll find the same.
AI, in one form or another, has become a part of our daily lives whether we know it or not. The same goes for music, at least for me.
Listening
Spotify is my streaming service of choice and I’m really sure they use AI for predicting what music I may want to hear based on my past listening. What is surprising to me is how they have predicted music I would never have sought out myself, and hey, I really like it.
For example, I listen to a lot of traditional rock music and I grew up on Led Zeppelin and AC/DC. But for some reason, I’ve found I like artists like The Knocks, Alt-J, Remi Wolf, and Foster the People. I would never have thought it. But Spotify did. And, they’re right.
So AI has expanded my listening and given me music to explore. It’s even helped me create playlists that I like to listen to.
Creating
I like to make music and it is nothing like the music I like to listen to. Mainly I’m trying to concentrate on ambient music and I have a lot of artists in the ambient music genre that I like and want to somehow emulate. I’d like to take some of their ideas and make them my own. But out and out copying is not what I’m after at all, so I need ideas on how to go forward.
Ideas for Music and Tracks
AI to the rescue – it excels at ideas. In fact, it is probably the number one thing that I’ve found AI can do best, and that is generate ideas. I’ve been using YOU.com as my go-to source for all my AI needs and have made my own agents for generating music ideas. I mainly like the Claude series of AIs right now, and my agents take in a musical genre from me along with an “inspired by…” or a “using plugins of…” or even “the ocean, wind, or whatever,” and spit out in seconds some well-thought-out combination of ideas that are not only innovative but would have taken me days to think of. Armed with a list from the AI, I can expand on it or mutate it into different forms that suit me.
Saving the ideas from the AI I’ve done using Notion and put them all on a page that I can refer back to at any time. I now have pages and pages full of ideas for music I can easily make. Will it all sound good? Who knows? But it is a source for me when I don’t have inspiration and it keeps me rolling onward with my music-making.
I’ve also found that having fifty or so ideas that are all in a related theme is not something that I would typically come up with on my own. I tend to have one-off ideas that are not related to others and I end up with a lot of tracks that do not seem to fit together. By using the AI to generate ideas, I get a bunch of them that fit together and could turn into a themed album of material.
Learning Mixing and Mastering
Making music as a music producer requires a lot of technical know-how. If you’re going to send your creations off to someone to finish mixing and mastering for you, then this probably does not apply. But if you’re doing this on your own, then it can look like a mountain of material to learn and this can stop you from progressing in your music.
I’ve used AI-assisted tools in plug-in form like Sonible makes, for example, to get my music to an acceptable mixed and mastered level. But, rather than just letting an AI do this for me, I’ve learned from the AI-assisted plugins what they’re doing. For example, compression using the Sonible plugins has taught me where I need to set my other compressors to match what the Sonible plugin is doing. So, in a way, it has taught me about compression by doing it for me, and then I just replicated what it was doing with one of my other plugins and thereby learned how to do it. This has been great because now I can use any compressor and set it up for my needs. I found that I can get better results from my own compressor settings that better suit my material than just allowing the AI of the Sonible plugin to do it for me. I’ve since left all my AI plugins behind and don’t need to use them anymore. They taught me, and now I do it on my own. This saved me a lot of time trying to learn it all from books or watching videos, and it was done on my own time and with my own music, which is better for me as that is the music I will be working with on a daily basis.
Analyzing Music
I like to know what makes music tick, so to speak. I like to analyze it and understand why certain music appeals to me and then I want to put those same elements into my music. AI has helped me in analyzing music in many ways and has shown me what I need to do to put some of it into my own.
Frisson (chills from listening to music) has fascinated me recently and I want to know why some music causes chills and some does not. By taking the spectrograms of tracks that seem to cause frisson and ones that don’t, I’ve used AI to analyze these and come up with a comprehensive list of their similarities. I’ve found what causes frisson and what does not, and so now can incorporate those elements that cause it into my music.
Beyond that, I’ve had AI read hundreds of sources in minutes and compile the research with new insights that have allowed me to explore further what causes frisson and come up with my own formulas that I can test.
I’ll be writing a paper on frisson and how it is created in music and how music producers can cause it in their own music, and AI has allowed me to do that through extensive research that would have taken me a ton of time to do.
Now, I can do this same thing with any aspect of music that I find appealing and want to incorporate into my own.
References
Where did I learn this stuff? Here’s a list of what I would recommend you check out:
- YOU.com – my AI provider of choice right now
- Inversity – where I learned about prompting for AI agents
- Markdown – learn markdown language (it’s easy) for writing good prompts
- Co-Intelligence by Ethan Mollick – the book that opened my eyes on using AI
- Atlas of AI by Kate Crawford – the book that gave me AI insights on ethics
Who am I?