I'm teaching myself to sing and had a major breakthrough when I realized that the sound of singing a particular word might not be the same sound as when speaking it.

For example the song Yesterday (Paul, Beatles). The first word, Yesterday doesn't sound right when singing 'yes-ter-day' but after dozens of focused listening I know hear it as 'yiss-ta-day' (or 'yiss-tuh-day').

And now when I sing 'yiss-ta-day' it sounds right.

Another example from Yesterday is the the second line which starts ... "Now it looks as though" (they're here to stay) which I simply not able to pronounce similarly to Paul.

But when I analyzed it and phonetically sing it as 'na-ow ih Lukes za-zo' it sounds the same as Paul's pronunciation.

So my question is ... is there a methodology, a cheat sheet?, of 'phonetic singing' equivalents or is this something one has to do for each song one wants to cover?

  • 1
    One problem with how a word 'sounds' is how different it is from your own accent. To my ear, Paul's Yesterday sound just like yesterday; because we have a [slightly] similar accent. He has no hint of the American Yearsterrday, so if you were native US, it would sound 'wrong' to you. I've had this argument/discussion with IPA 'experts' in the past who claim to be able to copy my [or any] accent if they can get the correct phonetic spelling.. all have failed. Good voice mimics/impressionists don't think in phonetics.
    – Tetsujin
    Dec 12, 2021 at 8:19
  • You may be aware that the Philly accent is rather different from Scouse..!
    – Tim
    Dec 12, 2021 at 8:54
  • @Tetsujin You say good mimics don’t think in phonetics but now when I sing Yesterday it’s a far better … almost tolerable to my own ears … charting out the phonetics of each word as in my example. Dec 12, 2021 at 9:19
  • IOW when I sing his-tub-day my cognition accepts it as Yesterday. Dec 12, 2021 at 9:21
  • 1
    The way people hear themselves is completely different to how they sound. Most people have a voice recorder on there phones. If you record yourself speaking and listen to the recording the majority of people will not recognise what they hear.
    – Neil Meyer
    May 12, 2022 at 17:18

3 Answers 3


Two key phrases to look up in that context are "speech level singing" and "vowel modification". Like with many important concepts, they are sort of diametrically opposed. The former is an active method so you can find authoritive information from the respective proponents of the method. At its core is singing technique with a focus more on popular music and musical theatre, trying to approach singing from speaking techniques.

The latter is sort of a classical technique for modifying vowels in order to fit the formants with the notes you are singing in a manner that leaves a distinguishable vowel set even if they aren't matching the formants of vowels at speaking pitches. It is most relevant for high voices, particularly high soprano.

  • Vowel modification is also pretty relevant for group singing, or at least vowel-matching since often two or more singers may naturally differ in the vowels they use for certain words.
    – user45266
    May 7, 2023 at 7:15

Something for each song you want to cover? More likely it'll depend on which version of that song you'd like to cover.

A British singer will make a quite different job of a song compared with an American. Let's face it, there are large differences in accents, (even acreoss England) and even British kids will put on an American accent when trying to sing a song by their favourite American singer.

But, as previous, what if there are many diverse covers that you want to cover? You'll have to develop accents which are appropriate. And that'll keep you busy for the rest of your singing life, I guess.

  • The fact that some singing is expected to be done accentless further complicates matters.
    – Neil Meyer
    May 12, 2022 at 17:20
  • 2
    @NeilMeyer - I'm struggling to understand what accentless words might sound like.
    – Tim
    May 12, 2022 at 17:42
  • Take the celtic women the sound of them in interviews compared to them on stage my be the complete opposite.
    – Neil Meyer
    May 12, 2022 at 17:47
  • @NeilMeyer that just means that they're using a different accent when singing than when speaking, not that they lack an accent in either context. Anyone who sings or speaks with pronunciation different from yours has an accent; anyone whose pronunciation matches yours has your accent.
    – phoog
    May 6, 2023 at 13:55

The first step in learning to sing phonetically is to learn to sing "pure" vowels.

The second step, loosely speaking, is to learn to add consonants.

At some point, one would study with a language coach, to get the subtleties of singing in a particular language or accent.

And the list goes on.

Separately, there is the International Phonetic Alphabet (IPA) (Wikipedia), which is a system of representing the various sounds people make.

These are steps that would be taken in a formal approach to learning to accomplish the question's goals. This kind of training would make it much easier to emulate a particular singer/recording.

But formal training or no, the only way to cover a particular recording is to go step by step through that recording. The more one does it, the more attuned the ear becomes to the subtleties of pronunciation.

But the only cheat sheet would be to learn IPA and find IPA transcriptions of the songs to be learned (which is not to say they exist — just that such a thing would be the only cheat sheet).

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.