How a song actually gets made here

People imagine a studio as a room where music appears. It is mostly a room where somebody says "again" for four hours.

Here is the real order of it.

It starts as a voice note

Almost always. Somebody hums a line into their phone in a car, or at two in the morning, and sends it over without listening back because listening back is how a line dies.

That file is the whole song at this point. Everything after it is argument.

Then the words, which take the longest

A melody can arrive in a minute. The words take weeks, and they take weeks because the first version of them is always nearly right, which is worse than being wrong.

• The wrong word is obvious and you replace it.
• The nearly right word sits there sounding fine and quietly makes the whole verse mean nothing.

Then one take that ruins the others

The vocal session produces somewhere between twenty and forty takes. One of them has something in it the others do not — usually one where the singer was not concentrating properly — and once you have heard it you cannot use any of the tidier ones.

Then the part nobody photographs

Mixing. Somebody sitting still for a very long time moving one thing by a very small amount. It is the least interesting footage and the largest share of why a finished song sounds finished.

You cannot hear a good mix. You can only hear a bad one.

And then it goes up

No launch, no countdown. It goes on the channel, and then we find out whether it was any good from people we have never met.