Many requests for this in the past so I suggest you add your vote to this thread:
If you can split the audio in software on your core then you can drive any on-screen software VU meter or visualisation.