This function has the highest level of flexibility with input phonemes, and can take any multi-graph input. This should also take non-unicode characters, so if you wanted to count, say, Hangul ...