AI Music Made Me Cry

When I was young, I wanted to be a recording engineer. Every month I’d rush to the newsagent and buy Home & Studio Recording magazine, reading it cover to cover. Those were the days when Trevor Horn was king and sampling was in its infancy. I should have pursued it as a career, but I didn’t have any musical skills and instead drifted onto the path of getting a useless degree.

When I moved to Japan at the end of the 80s, I built a home studio: Pro Tools II, a Mac Quadra 950, and a 25MB hard drive that barely worked. It all cost around $20,000 at the time. My friend Chris and I spent countless hours in the studio, honing guitar parts and fine-tuning his vocals. Looking back, I think the songs we made together still sound pretty good, given the equipment and our level of skill.

Over the years, as I moved around, built up a business, and started a family, music took a back seat. I sold the studio gear and barely played guitar – until about six years ago, when I committed to practicing acoustic guitar for an hour every day.

I didn't intend to write songs, but a couple of years ago, songs started to come to me, some fully formed, as if they’d been transmitted from somewhere else. It felt like a superpower.

So I began building a studio again. But now, decades later, everything was “in the box.” No need for hardware – just download plugins to access the best EQs, compressors, reverbs, everything at your fingertips.

Yet with all that equipment, my first demos were terrible. Even after upgrading my microphone and stripping things back – using fewer shiny plugins and focusing on the sound – they didn’t improve much. There was still a huge gap. So many questions: How on earth can I go from my acoustic guitar and a single vocal to the sound of a professional recording? How do I add other instruments when my keyboard skills are rudimentary? And how do I sing this better? I’m an okay singer live, but recording exposes every flaw. It’s so hard to hear your own voice objectively.

I sent my demos off to various places. The feedback was clear: the poor recording quality overshadowed the songwriting. I was embarrassed.

Last week, on a whim, I decided to run one of my demos through Suno’s cover feature, just to see what it might sound like.

I couldn’t believe it.

It didn’t sound like a demo anymore. It sounded like a record. Lush arrangements. Tasteful guitar lines. And real drums instead of loops. One of the most striking things about AI music is the lack of repetition. Everything moves, breathes, lives.

But the real revelation was the singing. Suno had taken my inadequate vocals and transformed them into something extraordinary. As the song progressed, I started to cry. There it was. This is what it would sound like if I’d managed to get my song to a major artist, with the best session musicians, producer, and arranger behind it. In seconds, something that had always been out of reach, and that most likely would never be in reach, became possible.

And I realised this feeling, this emotional release, built on years of hope and frustration, isn’t just mine. It’s happening to thousands of creators around the world. People who’ve carried lyrics in their heads for years but never had the means to turn them into songs. People like me, with melodies and ideas they could never fully realise before.

Surely, anything that can evoke that kind of emotion has real value. 

Now I find myself in the car, listening to Sing To Your Ghosts on repeat and drowning in that amazing vocal. And every time that singer, this impossible, soaring voice, hits those high notes in the final verses, I feel myself tearing up. 

Not just for me. Because this song reaches back across decades. It’s a prayer to Chris, who I made music with all those years ago, and who is no longer here with me to hear it.

Mark Devlin is the founder and CEO of UPCHART

UPCHART is the home for AI Music creators. Share your songs, get feedback and meet new friends. Join now for free credits...