Further to the idea of micromanaging each syllable a singer utters, I want to point out that the natural way we pronounce words when we speak isn’t always the best way to pronounce them when we sing.
Why did Mick Jagger, and the great majority of all British rock’n'rollers from the get go, lose their limey accents and sound convincingly American when they opened their mouths to sing? It’s so common that a lot of us think it’s impossible to sing any other way. And yet Lily Allen retains every bit of her Britishness on record. It’s a choice.
And if a singer doesn’t make cool and pleasant choices–naturally or through trial and error–sometimes a producer helps them along.
In 2000 a young female R&B singer was signed to an American major label cutting her first album and a friend of mine was producing. One of the phrases in the chorus was ‘like a needle in a record groove’. In the girl’s natural speaking voice, she pronounced the word ‘recorrd’. Her sung pronunciation was more like the cooler and often-used ‘rekkid.’ But the producer asked her to sing ‘reckerd,’ because in the context of the song it the coolest choice: more retro, and therefore more current.
Don’t kid yourself…choices like this are made at the microphone every day. Whether the result is an honest representation of who the artist actually is, or more a persona the producer is designing, the identity of the artist on the final recording has to be likable! If the vocalist doesn’t make good choices naturally upon opening their mouth, the producer is left with no other choice but to instigate experimentation of this kind.