I tried again with different settings in Hedy (and tried to make sure input and output on macOS settings are correct) and I could get a little bit more of text, but not the whole transcript. So, I don't know if the issue is with capturing audio or transcribing the audio.