Beat phenomenon question

[Robert Scott]

Richard,

The beat phenomenon that you reported in your experiement with the Korg
and TuneLab is probably the result of the way in which the audio tone is
generated in TuneLab.  You are right that TuneLab produces a sine wave,
but it is not a perfect sine wave because the sine values come from a
small look-up table rather than from a calculation for each sample time.
The resulting time distortion is equivalent to what you might hear if 
you moved your head back and forth about one inch while listening to a
perfect tone.  Regardless of the phase jitter, the long term frequency
stability and accuracy of the generated tone is unaffected.

-Robert Scott

Richard Moody wrote:

While experimenting with a digital "pocket" tuner (Korg DT-1) and Tunelab,
I fed the
Korg's audio output into Tunelab to be "measured".  I discovered that when
Tunelab
was in the audio mode both tones could be heard. That got me to wondering
what a
beat of one cent sounded like.  It was faster than I imagined, so I tried
.1 cent. 
Veeery sloooooow.   While I was marveling that a beat of  .1 cent could be
heard I
noticed "pulsation" of sound seemed to be on one note while the volume of
the other
note sounded steady. 
        The Tunelab puts out what I think is a sine, while the Korg gives a
sawtooth. In
the beat pulsation the sine seems to change volume while the sawtooth stays
the
same.  When the volume of the sine sounds low I can release the sawtooth
and the
volume returns to the sine.  When I put back on the sawtooth the volume of
the sine
returns to where it would be if I had not turned off the sawtooth.  At some
points
in the beat cycle the volume of the sine doesn't change when I release the
saw, and
it never seems to get less when I release the saw. Thus in the beat cycle
it appears
the pulsation is caused by the changing of volume of only one of the notes.
 Is
there an explaination for this, and why one note and not the other?