Autotune is a tool, just like any other. It’s neither good nor bad.

I like it when it’s used deliberately and thoughtfully — like in Charli XCX’s songs (yeah, that is the official video). I dislike it when it’s used to conceal an atrocious singer. I can hear autotune use when I’ve been told it’s not possible that I could do so, despite it being quite easy for anyone who has any sort of musical training.

This article I also linked to on my other blog is good, but this bit is a complete exaggeration.

When we asked him to provide a simple explanation of what happens, computationally, when a voice signal enters his software, he opened his desk and pulled out thick stacks of folders, each stuffed with hundreds of pages of mathematical equations.

“In my mind it’s not very complex,” he says, sheepishly, “but I haven’t yet found anyone I can explain it to who understands it. I usually just say, ‘It’s magic.’”

The math is actually not that hard. I understand most of it conceptually and I’m completely terrible at anything in the mathematical realm. Combination of journalistic non-understanding and reputation-burnishing, I suspect.

It’s pretty standard signal processing. The innovation was applying it to voice, and especially doing it in a computationally cheap way.

And, just as importantly, the product having a good GUI interface. If a Linux person had designed this (or a Mozilla developer), no one would’ve ever used it.