I'm curious how the audio was generated. It seems to be somewhat dependent on the character being displayed, as evidenced by jumping between different types of

I'm curious how the audio was generated. It seems to be somewhat dependent on the character being displayed, as evidenced by jumping between different types of characters (basic letters vs japanese characters)