The 48 phoneme recordings are stored inside this file, so they work with no internet.

Longer pauses help a reader who is still learning to hold each sound in mind.

Tap any of these to hear it on its own. Orange = a consonant blend, spliced from this voice's own phonemes. Plain = a recorded word part, or a part that is a real word and so gets spoken as a word. If one sounds wrong, it is a single line in the WORD_PART table near the top of the script.