We use Baidu's TTS (text-to-speech) engine to deliver high-quality, natural sounding audio for all of our words and example sentences.
Synthesized audio is not as good as native audio from human speakers. However, there are several advantages:
Having said that, TTS is not a panacea. Occasionally (in about 1/1000th of the words), the pronunciation isn't perfect. If we know about these, we can usually correct them, so let us know if you hear something strange!
Heads up! We expect to kick off a project to introduce native audio starting in summer of 2021!