NOTE: Currently available neural TTS-engines are not trained to synthesize „pure“ (isolated) vowel or
consonant phonemes, or consonant clusters (like „br“, „bl“ etc. which are very frequent in German).
Initially all audio was synthesized with Amazon „Polly“ (female voice). For the purpose of this demo short fragments for initial graphemes 'ä' to 'eu', 'sch', and 'z' have been
replaced with the correct pronunciation (recorded human male voice).
Short fragments for initial letters 'f' to 'y' are still spoken by synthetic voice to demonstrate the (erratic)
behavior of the speech engine.