Hi,
    Whilst looking into the effect of the first fraction of time upon 
perception of the whole timbre of a sound I have been playing with 
some very simple FM synthesis scores. I'm very puzzled now. The .ORC 
that follows is a standard single carrier FM orchestra, and two 
.SCO's follow. I thought that the difference in tone and timbre 
percieved in the two was a result of the initial time segments 
differing, but the difference is an actual characterisic of the whole 
sound. If the resulting sound files from each are edited to have the 
first part removed, they still sound very different. What am I 
missing here, how can they be physically different in the latter 
stages ??



;
;			     Single Carrier FM				      ;
; 
;    Specific values and basic design come from John Chowning's
article       ; ;  "The Synthesis of Complex Audio Spectra by Means of
Frequency Modulation"  ; ;; ;	   coded by Richard Boulanger - Berklee
College of Music	      ; ;

instr 1
   ; p4  = amplitude of output wave
   ; p5  = carrier frequency in Hz
   ; p6  = modulating frequency in Hz
   ; p7  = modulation index 1
   ; p8  = modulation index 2
   ; p9  = carrier envelope function
   ; p10 = modulator envelope function

   i1 = 1/p3			     ; one cycle per duration of note
   i2 = p7 * p6 		     ; calculates deviation for index 1
   i3 = (p8-p7) * p6		     ; calculates deviation for index 2

   ampcar  oscil   p4,i1,p9	     ; amplitude envelope for the carrier
   ampmod  oscil   i3,i1,p10	     ; amplitude envelope for the
   modulator

   amod    oscili  ampmod+i2,p6,1    ; modulating oscillator
   asig    oscili  ampcar,p5+amod,1  ; carrier oscillator
        out     asig
endin

;score one

f1 0 1024 9 1 1 0
f2 0 513 7 1.04 21.63 .56 149.43 .54 85.37 .54 85.53 .54 85.53 .54
85.53 .54
f3 0 513 7 .54 85.54 .54 85.53 .54 85.37 .54 85.53 .54 85.53
.54 85.53 .54 i1 0 0.5 10000 440 400 10 15 3 2 e

; score two - difference is in p5 of f2 (.55 instead of 1.04)
f1 0 1024 9 1 1 0
f2 0 513 7 .55 21.63 .56 149.43 .54 85.37 .54 85.53 .54 85.53 .54
85.53 .54
f3 0 513 7 .54 85.54 .54 85.53 .54 85.37 .54 85.53 .54 85.53
.54 85.53 .54 i1 0 0.5 10000 440 400 10 15 3 2  e


*************************************************
Peter Kearton,
University of Surrey Psychology Department,
Guildford,
United Kingdom.
GU2 5XH
Tel. UK (01483) 259437
e-mail - P.kearton@surrey.ac.uk
*************************************************

	In FM, the starting phase relationships are going to have an
effect throughout the length of the note.  So if you have two notes, one
which is modulated slightly faster just at the beginning, when they settle
down into the part where the should seem to be the same, you are almost
certainly going to have a phase difference between the two modulators.
The only way to avoid this is to do a calculation that ensures that your
changes at the beginning result in no phase shift in the later sections.

Mike Berry
mikeb@mills.edu