This is not really a problem, if one keeps in mind that only humans use language.
It follows that:
- if any thing makes noise, it's not speech sound.
- if any person speaks or writes those sounds, they're speech sounds.
Thus, if anyone reports sound in speech, it's all speech sounds.
That's how easy it is to tell the difference.
As to what Part of Speech the words are, you can call them anything you like, and it won't matter a bit. Knowing an official "Part of Speech" to label a word (or anything else) contributes precisely zero to one's knowledge of English grammar. As I point out in the post linked above, not everything they teach you in school is correct. Alas.
Real parts of speech (nowadays they're called Grammatical Categories) are simply those categories that are necessary to describe the grammar of a language, and they vary greatly -- like vowels or syllables -- from one language to another.