I will add to it. You can sing to an metronome and hum a melody line segment from an arrangement on paper and put together a title from just a vocal track, if the metronome is also recorded on it’s own track. But of course you need the headphones to hear it, or you must later replace the vocal track with leaking metronome sound on it. Though this is often done as a “scratch vocal” a reference for arranging the song, when the song is arranged the singer returns and sings full force and emotion to the entire track like a perfect band. Nowadays most songs are from workstations that can give you a basic beat at a set tempo, and/or from metronome, a tape loop, or a drum sequencer patch. In all cases you need to hear the time reference to sing or play to it. Hence headphones Or if a band is playing together, but instruments are physically issolated for control of their sounds, the ‘mix’ they all play to is created in the several headphones by the engineer producer, with band input.