In
linguistics, an individual
speech sound considered in its own right, without reference to how it functions in a language. This is in contrast to a
phoneme, which is "one sound" for a particular language, but might cover several phonetically similar variants. Phones that belong to the same phoneme are
allophones.
For example, in (most varieties of) English the P-sounds in pit and spit are different. The first one is aspirated: phonetically [phIt]; whereas the second would sound somewhat more like a B if you heard it in isolation. But these two phones function as the one "P-sound" or phoneme of English, and English-speakers generally don't notice the difference. The choice between them is automatically governed by the presence of the S.
Phonemes are tied to specific languages: the same two phones are different phonemes in Hindi, and the choice between them can change the meaning of a word.
The study of phones is phonetics. The relationship of phones to phonemes is phonology. From the adjectives 'phonetic' and 'phonemic' have been generalized the terms 'etic' and 'emic', for any similar contrast between low-level observational units and high-level structures.
Phonetic sequences are written in square brackets, while phonemic sequences are written between slashes. So [phIt] and [spIt] show the allophonic differences, but /pit/ and /spit/ show the phonological structure.
The following was a reply to a later write-up disputing the exact details; now deleted. You can skip this if you just want to know what a phone is: we weren't disputing that, just the correct choice of examples and analysis of English sounds.
The choice of spit is perhaps unfortunate, because it could have some onomatopoeic quality, but should be irrelevant. My discussion is equally true of pot, spot or car, scar, but I wanted as dialect-neutral a word as possible because I didn't want to explain the vowels in [phQt] or [ska:] (see SAMPA for phonetic symbols).
In my variety of British English, /p/ is unaspirated in /spit/. Perhaps not fully unaspirated, compared to some other languages, but certainly not equally aspirated. In Korean there are three contrasting stops, strongly and weakly aspirated and wholly unaspirated. In most varieties of English the contrast is between (middling) aspirated and (close to) unaspirated.
English /b/ (and /d/, /g/) is neither fully voiced nor voiceless, it is partly devoiced, a different phonation. That is, it is not exactly the fully voiced [b] of French. But it is not aspirated either. Voicing, devoicing, and voicelessness are properties of the stop segment, the [b] itself. Aspiration is a property of the following vowel: the voicelessness of [p] continues into the early part of the vowel segment, before voicing switches on. (Vowels are voiced.)
Although I find it hard to be sure I'm getting the phonation right when I pronounce [spQt] with unaspirated /p/ and [sbQt] with devoiced /b/, I think I can tell that I actually use [p] normally. Certainly, the breath test shows that neither is followed by aspiration. Compare to a language like Danish, where stops such as /b/ are (I believe) fully devoiced. Here a word like spa could use the same devoiced phone to realize the p, so its phonological representation would be /sba/ in Danish.
It is true that elementary linguistics texts do brush details under the carpet: they say /p/ is voiceless and /b/ is voiced, since that's the most convenient description for the difference between them, but imply by omission that there are no degrees, and that it's equally true for any language. Beginners can get by with that, but it becomes confusing if the true situation isn't eventually explained.
Others have tried to finesse the issue by saying the /p/ vs /b/ distinction is one of tense vs lax, but these are pretty meaningless terms. What, exactly, is tense or lax? To answer that would be to describe the details of vibration of the vocal chords and the tension in the arytenoid cartilages, viz the phonation, and there are a lot more than two kinds of phonation.
And of course not all English accents are the same. If you hear a recording of T.S. Eliot, one of the striking things about his voice is that voiceless stops are unaspirated.