[music-dsp] measures of noisyness and transientness

Andy Farnell padawan12 at obiwannabe.co.uk
Thu Sep 30 15:56:58 EDT 2010


On Thu, 30 Sep 2010 11:34:25 -0700
Yan Li <phdlabs at shaw.ca> wrote:

> Hi,
> 
> What can be used as good measures of noisiness for singing voice? 

large RMS, small autocorrelation

> What about transientness? Spectral flux does not seem to be a good enough measure.

Transient-(ness??) isn't really a quality so much as confidence
in an event boundary. For untreated speech (no dynamic compression)
simple time domain ratio of peak amplitude in short windows. Speech 
often wants us to look for two or three specific types of transient:
plosive or fricative stops/glottals etc.  



> 
> Thanks
> --
> dupswapdrop -- the music-dsp mailing list and website:
> subscription info, FAQ, source code archive, list archive, book reviews, dsp links
> http://music.columbia.edu/cmc/music-dsp
> http://music.columbia.edu/mailman/listinfo/music-dsp


-- 
Andy Farnell <padawan12 at obiwannabe.co.uk>


More information about the music-dsp mailing list