Every time I try to edit something in the text-to-speech text box (after first generation of the audio), I get corrupted audio, with something like a weird echo. The solution is to delete the whole audio and create a new TTS audio. Of, course, a huge waste of time.
Is this a known issue? Are you working on it?