Why 46 sounds?
In English speech, words are distinguished by the sounds used to represent them. This is evidenced by what are called “minimal pairs”, e.g. “ship” and “sheep”.
Research shows that, if dark characters are counted but diphthongs are not, 39 sounds are needed to represent standard American and British speech. Professor John Wells former head of the phonetic department of University College London, the vestal virgins of English phonetics, has published the full list of them here: http://www.phon.ucl.ac.uk/home/wells/phoneticsymbolsforenglish.htm
Taking the number as 46 is in fact a compromise, reflecting the IPA origins of S4.
It could be argued that the T Esh and D Ezh sounds are just combinations of the letters that make them up, and should therefore not be counted. This would reduce the number to 44.
It could also be argued that the five diphthong sounds could be broken down into their component parts as follows:
ei to e·ii
ou to ə·uu
ai to a·ii
au to æ·uu
oi to o·ii
This would bring the number down to 39 but would complicate written text and would be a further step away from the IPA notation with which so many are familiar.