Music and calm: what we can tell you, and what we cannot
By Dan Cohen · Published 2026-04-04
This page used to survey the research literature and tell you how to find a credentialled practitioner. It cited a meta-analysis it did not link to. That was a mistake, and not a small one: this site says on its own about page that its author is not a neuroscientist and not an audiologist, and every body page carries the line Measured, not medical. A page that quietly claims the opposite undoes all of them.
So here is the honest version.
What we can tell you
We measure songs. For every track in the library we record how far it swings between its quietest and loudest moments, whether anything arrives without warning, how fast it moves, how dense the texture is, how predictable the structure is, and what the voice is doing over the top. Those are properties of the recording. They are the same for everyone who presses play, and you can check every one of them against the song in front of you.
That is genuinely useful, and it is useful precisely because it is narrow. If a song is going to jump, we can tell you before it does. If a piece never resolves into a chorus, we can tell you that too. Someone who has been ambushed by a song they were told was calm can use that. So can someone building four hours of background sound for a room full of people. So can a parent at bedtime.
What we cannot tell you
What any of it will do to you. We do not measure people, we have not assessed you, and the honest distance between "this song has a dynamic range of 3 out of 10" and "this song will help you" is enormous. Two people with the same reaction to sudden noise can want completely opposite music, and neither of them is wrong.
We also cannot tell you what a trained professional would. If sound is genuinely difficult — if it is shaping what you can do in a day — that is a conversation with someone who can actually assess it, and no measurement on this site substitutes for it. We are not being modest. We simply do not have the standing, and pretending otherwise would make everything else here less trustworthy.
Where that leaves you
With a filter, not an answer. Start from a song you already know behaves well for you, look at what it measures, and use those numbers to find more like it. If the numbers disagree with your experience of a song, trust yourself — you are the one listening. The methodology page sets out exactly how each rating is produced, including what is model-assisted and what is not, so you can decide how much weight to give any of it.