Path: utzoo!mnetor!uunet!lll-winken!lll-lcc!ames!ll-xn!mit-eddie!uw-beaver!cornell!rochester!PT.CS.CMU.EDU!cadre!pitt!cisunx!ejkst
From: ejkst@cisunx.UUCP (Eric J. Kennedy)
Newsgroups: comp.sys.amiga
Subject: Re: Sonic Tomfoolery
Message-ID: <7376@cisunx.UUCP>
Date: 1 Mar 88 02:41:41 GMT
References: <8802251858.AA21577@cory.Berkeley.EDU> <719@ur-cvsvax.UUCP> <4292@xanth.cs.odu.edu>
Organization: Univ. of Pittsburgh, Comp & Info Sys
Lines: 58

In article <4292@xanth.cs.odu.edu>, kent@xanth.cs.odu.edu (Kent Paul Dolan) writes:
> In article <7233@cisunx.UUCP> ejkst@cisunx.UUCP (Eric J. Kennedy) writes:
> >In article <719@ur-cvsvax.UUCP>, jea@ur-cvsvax.UUCP (Joanne Albano) writes:
> >> > 	
> >> > 	Huh, where did this come from?  I've played around with sound
> >> > quite a bit, and if I generate two tones of slightly different frequencies,
> >> > I can hear the phase quite fine thank you.
> >
> >That's not 'detection of phase', that's detection of two tones of slightly
> >different frequencies.  It's not the same thing at all.  Two tones of
> >slightly different frequencies will create a 'beat' between them, which
> >will sound like the tones are quickly increasing and decreasing in
> >volume.  The 'phase' here is two tones of the _same_ frequency, but with 
> >one slightly leading or lagging the other.  Here I'd have to agree with
> >Matt, we can't detect that nearly as readily.  
> 
> I don't think so!  I remember reading many years ago that although the
> human ear can only hear pitches up to about 20,000 Hz, that a stereo
> system, to maintain fidelity enough to allow a listener to pick out
> the second violinist playing half a tone flat in a symphony recording,
> had to keep the left and right channel phase relationship correct to
> the equivalent of 200,000Hz, because the human brain, processing the
> audio signals, is that sensitive to phase relationships.  
Show me the article and I'll believe it.  This still sounds like the
above misconception to me. This is two _different_ frequencies creating
a beat between them, not two identical frequencies out of phase.
  
>                                                               I believe
> the math can be done to prove this with a pocket calculator and a back
> of the envelope diagram of a symphony hall and a schematic human head.

My pocket calcutalor says:
 
speed of sound = (wavelength) * (frequency)

1000 ft/sec = wavelength * 20000 1/sec

so wavelength = 0.05 ft = 0.6 inch.

One half of one wavelength = 0.3 inch.  This means that if you are in a
symphony hall and move your head 0.3 inches to the side, the phase of
that 20000 Hz sound will shift by 180 degrees.  You're telling me you can
hear the phase difference when you move your head from side to side?
  
Now, granted, at low frequencies, it's an entirely different ball of
wax.  At 60 Hz, the half-wavelength is 8 feet.  This is why you make
sure your woofers are wired with the same polarity.  So, the bottom line
is it depends on frequency.  

For (relatively) high frequency sounds, I still think you can't detect a
phase difference.
 
> Kent, the man from xanth.


-- 
------------
Eric Kennedy
ejkst@cisunx.UUCP