Can a machine mix music better than a human? Spike AI seems to think so

Elise Shumacher | staff writer

Mark “Spike” Stent is one of the music industry’s most critically acclaimed. He has produced hit records for some of the most iconic names, such as Beyoncé, Madonna, Ed Sheeran, Oasis and the Beatles.

At nearly 60 years old, he looks back on a career spanning decades and studded with numerous accolades. But he does not intend to retire. Last month, he announced a plugin that might just be the future of music production. Set to launch in 2025, Spike AI was developed to mix at the level of the world’s greatest producers.

Mixing is the crucial stage in sound production when individual tracks are processed and combined. This is achieved by adjusting volume levels, layering frequencies and adding effects like reverb, delay and panning. Musicians hire producers to make this process less complicated. Over time, experienced producers lean increasingly on “feel” with their mixing choices. Thus, producers rely both on their skills and intuition to create a record that accurately reflects the artist’s intent while sounding impeccable.

But what of the musician who cannot afford to hire a professional? Are they forever subject to sub-par mixes, simply due to monetary reasons?

Spike worked with his son Joshua Stent and longtime friend and colleague Henry Ramsey to develop the plugin. They claim Spike AI will become vital to the production industry, opening new doors for musicians who previously found mixing themselves to be a grueling and confusing process.

This skips the “middleman” — the producer with deep understanding of audio whose presence was once required during this crucial stage of a record’s development. Much like ChatGPT and other AIs, Spike AI exists in the form of a chatbot that supposedly promotes a “human-like” interaction.

According to the website (which is still in development), you can ask, for example, “can you make the drums more punchy?” and “can you reduce distorted noise?” You can even make it sound like another artist – a screengrab reveals a prompt example, “make it sound like hip hop, Dr. Dre.”

Written in bold letters are a series of self-aggrandizing statements: “Our AI analyzes your audio for perfect mixing;” “Spike uses pioneering technology and high level expertise without compromising the human touch that makes a mix truly special.”

Spike himself, in an interview about the potential of Spike AI, said “the mix is where the emotion happens … you have to mix so … the feel and soul really comes through.” He goes on: “We wanted to offer something unique that everyone can access. That’s what’s so brilliant about Spike AI.” Spike AI is so good at mixing, it will move you to tears, according to Spike’s son, who claims he has witnessed this response with his own clients.

Spike claims to have created a tool for the burgeoning producer to harness their creativity. In reality, it does the exact opposite. He shrinks a lifetime of learning to five minutes and a few cursor clicks. Spike AI strips the user of their independence, forcing them to confine themselves to set parameters. It imbues a generation with the notion that there is such a thing as the “perfect” mix.

This negates his own claims that a great producer possesses intuition that transcends technical abilities. According to Spike, asking AI a question (“make it sound like hip hop, Dr. Dre”) is akin to the intuition that producers spend their lives developing. The AI is apparently so sophisticated, it infuses “soul” in its mixes.

As with all generative AI, to satisfy the user’s request, Spike synthesizes millions of bits of data, then regurgitates a result. After scouring the legendary producer’s discography in the span of a second, its conception of Dr. Dre-style mixing can only be a pantomime.

Spike AI locks the “bedroom producer” or “home musician” in a space that entices them to put in zero effort. It assumes that those who cannot afford someone with experience also lack the capability to learn basic EQ techniques. They are thus relegated to the “conscious” mixing choices of an AI. By making it so easy, the path to understanding is taken for granted.

The concern should not be whether Spike AI and other models will steal producers’ jobs. It is the next generation of musicians convinced they are harnessing their creativity whey they are really stunting their growth. The inevitable conclusion: thousands of “new” records that are mere imitations of the original, gold-standard.

Leave a Comment