Every text-to-speech and speech-recognition engine uses its own phonetic character set, because standardized character sets are usually not as complete as necessary for the vendor's purposes. The vendor defines the way engine-specific phonemes map to standardized character sets.
The International Phonetic Alphabet (IPA) is a standard phonetic character set that is already supported in Unicode. Although the IPA is not used by any specific engine, it is an international standard. At least a rudimentary conversion between the engine's character set and the IPA set is possible, although the conversion may lose some characters. (It is unlikely that any engine will use this set internally.)
The IPA character set is defined in the Unicode standard. It occupies 0x0250 through 0x02AF for the sounds and some Latin characters, and 0x02B0 through 0x02FF for the modifier characters. If an engine's flags indicate that it understands the IPA, an application can use IPA characters with that engine.
Because of the size of the IPA character set, the only way to send IPA characters to or receive them from an engine is through Unicode. There is no 8-bit method of using a standard phonetic set. You can use the Phoneme Conversion object in Speech Tools to display an 8-bit phoneme set that is language specific.