I have been using voice recorder to record some of my meeting since I've been quite excited to try out the transcribe and summarise feature, esp coz Ive seen how well Copilot for Teams works.
Speaker differentiation is something it does really well. Clearly calls out Speaker 1, 2, 3 etc and again goes back to Speaker 1 when the original person speaks. So identification of speakers and voices is really good.
However, the actual transcription is pretty poor. I've got people from US, UK, India, Africa on the call so every accent there is is covered. But the transcription is awful in every accent. Not a single line is usable, and thus, the summary is of no use either.
I actually find it better to use Live Transcribe feature of Google to transcribe in real-time, copy paste it into Samsung Notes and use the Summarize feature there. However, Live Transcribe does not have Speaker detection. It's all in one big paragraph. And note, Live transcribe works Live, unlike Voice Recorder transcription which happens on a recorded clip.
Do others also face such a problem? I would love to hear some solutions.
