Is it better?
Sitting around a fire on a recent trip to Southern Africa, with a huge extended family singing a capella seems a distant echo from the immense amount of technology and power around me required to assist with, and generate today’s music in the studio. But is it better – and what’s next for music? Let’s unpack it a little.
Firstly, what’s the difference between Generative and Assistive AI?
Generative AI uses pre-existing sources and learned patterns to create at your prompts. For example Splice Stacks or CassetteAI.
Assistive AI on the other hand speeds up and enhances tasks that a trained human would traditionally undertake. An established name in the audio world is Landr with new apps arriving all the time to assist with each part of the process.
Using technology to assist with human tasks has always been a part of music. The stops (control levers) of a pipe organ are all labelled to represent different instruments, to enable one person to be able to play all of them (including the use of their feet). So assistive AI, (bringing together technology, with the digital capacity to store and access patterns in data) is a very natural, and useful next step in the revolution.
Generative AI is something altogether different. The creation of new music from learnt patterns, data and prompts has never been possible before.
How can I use AI in my music making?
Firstly, decide whether you are looking to use AI for assistive or generative purposes.
Izotope for a while now, includes some useful AI functionality within their suite of plug-ins. Within Nectar, for example, it can ‘listen’ to your isolated vocal track, and, once given prompts, combine what it heard with what you ask for, and it’ll help to select and modify the necessary elements of the plugin to create the desired sound. It’s a great starting point when working quickly. The initial processing and learning time is completed in a matter of seconds, and then all the usual variables can still be altered to suit creative goals.
Are there down-sides to assistive AI?
Each occurrence of assistive AI in a mix (for example, creating the perfect vocal sound, or the amount of compression on a kick drum) works in isolation. As yet, each component cannot speak to another, to know how they hold together. I describe it as being like asking different artists to combine ideas on one canvas to create a single work of art. Each element will have its own character and desired outcome – but it will take a human to understand and pull these together into a complete piece.
My biggest concern is that I love nature. I love the organic world around us – and how ‘un-uniform’ everything is.
Hearing a choir for example, in one region of a country will sound different to a choir in the next village – for many tiny nuanced reasons. As the global music world moves towards a reliance on the same suite of Assistive AI plug-ins, there’s a danger – like with car-design – that regional characteristics – and creative differences – the things that make us human, will decrease, meaning a steady disconnect with perfected mixes, without the listener or engineer really knowing why.
Is Generative AI really music?
Generative AI can absolutely take the place of music. For example, when used by a student film-maker needing music under a particular scene. Budgets don’t allow for an orchestra, and copyright restricts the use of existing music. So to replace music, Generative AI can certainly be used as a place-holder for music.
However, look at many of the descriptions of Generative AI music tools. Here are some examples: “Never suffer from lack of inspiration again…” “The answer to writer’s block…”.
The worrying promise is that these tools will allow progress, where there was none. Which, to a degree, can be helpful – if there’s little care for human-connection with the end result.
True music creation however often comes through blood, sweat and tears. Patiently waiting out those periods of writer’s block and frustratingly starting a mix again from scratch. True creativity comes from seeking inspiration – and knowing when it’s not the right time. To stop – wait and reflect.
And sometimes the ideas that we first had (which would have shaped the ‘prompts’ that we were giving Generative AI) turn out not to be the best natural and organic direction after all.
If I had to look to the future, I imagine a trend towards un-plugging. A move towards a world where perfection isn’t celebrated (because it can be artificially created anyway – and turns out to be not all we thought it would be) but a wholesome move towards celebrating human creativity in all its imperfection and vulnerability.
Is Generative AI really music?
AI brings so much speed and efficiency which gives more time for creatives to do what they do best – create. It won’t be long until we’ll be able to see the true effects of AI on music creation. Whether it brings better music, enriching our lives and financially supporting the creators – or whether it brings mundanity and takes from the creators more than it gives. We will see.
As a side-note – AI was not used at any point in the creation of this blog post. Which, after running pretty much everything I put out through ChatGPT (to ensure I’m speaking to the right audience), feels a little scary.
