Subject: Re: Granular synthesis and auditory segmentation From: Peter Cariani <peter(at)epl.meei.harvard.edu> Date: Thu, 15 Oct 1998 19:11:29 +0100Richard J. Fabbri wrote: > ... Didier's example of listening to a Halfwave Sine vs a Sine is > not equivalent to the transduction function provided by the > neural HALFwave action caused by a neural refractory period > when the acoustic waveform is of the opposite polarity. I agree that Didier's example of listening to a half-wave rectified tone is different from half-wave rectification in transduction. Didier had suggested listening to the half-wave rectified waveform and hearing the difference (I think many of us misinterpreted what you had said initially; I think he was offering a counterexample constructed out of what he thought was your logic). Irrespective of the relevance or irrelevance of Didier's suggestion to what you believe, if you actually listen to the half-wave rectified waveform (mine was a 1 kHz pure tone), it (unexpectedly) doesn't sound all that different from the unrectified version, despite the addition of extra components in the power spectrum. There may be a reason for this -- when you look at the form of the autocorrelation functions of the two signals, rectified and unrectified, they are very similar. So it is also for peak-clipped speech. My general point here is that how various nonlinearities play themselves out perceptually depends critically on the nature of the neural representations that the auditory system uses to subserve those percepts. We need to be more self-conscious about our neural coding assumptions. So, I agree, we are all talking about different things here. Peter Cariani McGill is running a new version of LISTSERV (1.8d on Windows NT). Information is available on the WEB at http://www.mcgill.ca/cc/listserv