speech/music - thanks

[Sue Johnson <sej28@xxxxxxxxxxxxx>]

Thanks to all those who replied to me about the issue of determining
speech/music in a soundtrack.
I have looked at Slaney's paper, and now have the following problem...

Suppose you have some characteristics which you can use to recognise
speech (rhythm, harmonics, continuity, energy etc). If there is a piece of
the soundtrack which contains a speaker talking, but with music in the
background, then the system will recognise it as music.
Whilst it is true that it contains music, I want it to recognise the fact
that there is speech present, and mark it as speech.

Of course it is helpful to me to be able to recognise things with music
in, but it would be even better if I could almost re-phrase the problem to
be looking out for when speech occurs.
Has anyone any ideas on how you might be able to recognise the presence of
speech when music is also present. Does it have any special
pyscho-acoustic properties, or does the music drown them out?

thanks in advance for any comments.
Sue Johnson

 ------------------------------------------------------------------|
| email  sej28@eng.cam.ac.uk                                       |
| snail  6 Dalegarth, Hurst Park Avenue, Cambridge, CB4 2AG        |
|        (01223) 574087  Lab: (01223) (3)32754                     |
| w/ends 28 Basset Way, Kidlington, OXON, OX5 2HJ (01865) 374 925  |
| WWW      http://www2.eng.cam.ac.uk/~sej28/home.html              |
|------------------------------------------------------------------|