COMMENT ⊗   VALID 00003 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	RECORDED EXAMPLE 1:
C00005 00003	RECORDED EXAMPLE 3:
C00006 ENDMK
C⊗;
RECORDED EXAMPLE 1:
This example will consist of three individual violin tones,
and then two sequences of violin tones.  The first individual
tone will be the originally recorded and digitized tone directly
played back through the computer.  The second tone was
generated by the additive synthesis technique (described in
Section IIA1) and was based upon the complete data from the
analysis using the heterodyne filter.  The third tone was
generated by additive synthesis based, instead, on a vastly
reduced version of the data from that analysis.  Compare these
three tones to measure the ability of the analysis-synthesis
strategy to match the original tone, and to measure the 
adequacy of the data reduction.  Then listen to the following
two sequences, both generated by additive synthesis, where the
first was based on the complete data from analysis and the second
was based on the vastly reduced version of this data.

a) Tone from a violin which has been recorded on tape and then
   digitized at 25000 samples per second.
b) Tone generated by additive synthesis based on the heterodyne 
   analysis of the preceeding violin tone.  The results of this
   analysis are shown in Figures 1 through 4.
c) Tone generated by additive synthesis based on the 3-line segment
   approximations to the analyzed functions for the preceeding 
   violin tone.  The results of this data reduction are shown in
   Figures 5 and 6.
d) Sequence of tones generated by additive synthesis based on the
   complete functions from analysis, as shown in Figures 1 through 4.
e) Sequence of tones generated by additive synthesis based on the
   data-reduced functions from analysis, as shown in Figures 5 and 6.
RECORDED EXAMPLE 3:
This example will consist of a sequence of tones which gradually
changes in timbre from a violin to a saxophone, pauses, and then
gradually changes back to the violin timbre.  It was generated
by additive synthesis which was based on vastly reduced
versions of the data from the analysis of a violin and saxophone tone.
A simple algorithm produced the intermediate tones by performing
interpolations on the 3-line segment time-varying amplitude and
frequency control functions.  The amplitude functions for six
tones in this sequence are shown in Figure 13.