🚀 go-pugleaf

In article <3936@randvax.rand.org>, edhall@rand.org (Ed Hall) writes:
|> Are you talking about two-way or N-way teleconferencing?  If it's
|> something like an audio "Internet chat" program (that is, N-way), with
|> one talker per station, then you simply place each data stream at a
|> different place in the stereo field at the receiving end with something
|> like:
|> 
|> 	L = S*P
|> 	R = S*(1-P)
|> 
|> Where "P" is some arbitrary position between 0 and 1, "S" is the
|> data stream for a particular talker, and "L" and "R" are the left
|> and right channels, respectively.  Thus, the talker assigned 0.5
|> would be right in the middle, while 0, 0.25, 0.75, and 1.0 would
|> be hard right, mid right, mid left, and hard left, respectively.

You may be interested in proper constant power panning:

	 /* pan pot position is in linear range -1 .. 1
		-1 is hard left, 0 is center and 1 is hard right */

	 /* convert pan position to angle in range 0 .. pi/2 radians, where
	    0 radians is hard left, pi/4 is center and pi/2 is hard right */

	/* shift range:  [-1..1] + 1 = [0..2] */
	    dvalue += 1
	/* range [0..2] * pi/4 = [0..pi/2] */
	    dvalue *= pi/4

	 /* left pan value is cosine(dvalue), right pan value is sine(dvalue).
		Power of both channels is:
			 sin^2 + cos^2 = 1
		Thus, constant power
	    */	
	   channelScale[LEFT] = cos(dvalue);
	   channelScale[RIGHT] = sin(dvalue);

|> : That leads to another question of mine.  Is there code for band reject
|> : and band pass filters out there?  I realize that filters should be
|> : specially tailored to the application.  However, since I don't know
|> : much DSP, I'd love to just try filters with certain specified
|> : tolerances.  I've read the FAQ without much luck.
|> 
|> Look in "Musical Application of Microprocessors," 2nd Ed. (by Hal
|> Chamberlin, Hayden Books, 1985, ISBN 0-8104-5768-7) for some practical,
|> non-mathematical material on the subject (and on stereo, too).  Some of

The book is an excellent starter.  However, I remember the filter
coefficient equations contained a chock full o' errors.