ChatGPT – the online artificial intelligence tool with seemingly limitless potential applications – is all the rage right now, so it’s no surprise that some intrepid music creators have sought to push the boundaries of what it’s really capable of.
After making light work of some straightforward music critiquing – Louder tasked it with reviewing Metallica’s Master of Puppets, with impressive results – ChatGPT recently came face to face with its greatest challenge yet: to create a VST plugin version of the Ibanez Tube Screamer overdrive pedal.
The futuristic experiment was conducted by YouTuber Burned Guitarist, who sought to find out whether ChatGPT could be used to code and construct a virtual DAW-compatible take on the iconic pedalboard mainstay.
And, owing to the outcome of the test, we’d wager those who have remained skeptical about ChatGPT’s application are about to rethink their standpoint on the impossibly versatile AI tool.
First up is a simple knowledge test, which ChatGPT passes with flying colors: it is a literal computer, after all, so it comes as no surprise when the app pulls up an eerily accurate definition of what the Tube Screamer actually is and what some of its components are.
What’s more surprising, though, is what happens when Burned Guitarist then asks ChatGPT to go one step further and create some code for an implementation of a Tube Screamer VST plugin.
In response to this simple request, ChatGPT wrote, “Here is a simple example of a Tube Screamer VST plugin written in C++,” before offering the caveat that the following code is “intended to give you an idea of how a Tube Screamer plugin might be implemented, but it is not intended to be a fully functional or production ready-plugin”.
Impressed but not entirely satisfied with the initial effort, Burned Guitarist asks for a more specific implementation – something ChatGPT is more than capable of, quickly serving up “the actual algorithm for the Tube Screamer effect” that “should be a fully functional plugin”.
Once that’s done, ChatGPT then reveals it is also familiar with a handful of frameworks for deploying VST plugins, including the Steinberg SDK, JUCE, and LV2.
After flexing its knowledge on such topics, ChatGPT then sets to work on implementing its Tube Screamer effect in the Steinberg SDK framework, with some helpful prompts from Burned Guitarist.
Though the first workable attempt is pretty harsh on the ears, some further back and forth between Burned Guitarist and ChatGPT results in a much smoother-sounding plugin, and though it may not technically sound like a Tube Screamer, it’s an impressive effort considering its provenance.
Is this a glimpse into a not so distant future where guitarists can easily conjure up their own plugins with the help of artificial intelligence? Quite possibly, and if Burned Guitarist’s experiment is anything to go by, such a future might be closer than you think.
It's not the first time guitarists have experimented with artificial intelligence. Google's Tone Transfer, for example, allowed players to transform their electric guitar into a saxophone, trumpet, flute or violin, while KnobsAI used artificial intelligence to help users tweak, save and share their pedal presets.