[sdiy] Harmonic bandwidth
Jim Palmer
jimp at pobox.com
Wed Jan 9 02:05:51 CET 2008
thanks for sending me down the rabit hole...
i was sceptical of this, but i did some tests in soundforge
and i can indeed hear a difference with only the phase
different per harmonic.
first i mixed 4 harmonics (1 second sine waves at 60,120,180,240)
at -6dB each to avoid clipping. then clipped a random portion of
each sine wave file from the end and pasted it to the front, playing it
looped to make sure no artifacts were introduced. then i made a
composite wave of the random-phase-start clips the same way
as the i made the in-phase composite clip.
they do sound quite different.
at first i was convinced the difference had to be psychoacoustic
and related to the difference in the "click" as it changed from one
wav to another, but that turns out not to be the case
next i mixed 7 harmonics at 120, 240, 360, 480, 600, 720, 840
again all the same level this time -18dB
i made the clip with both composite waves this time with a fade in
and fade out to avoid the "click" sound. again i can easily tell them apart
next i mixed the same 7 harmonics with -3dB*n to get a pseudo sawtooth.
here it was easier to see the difference looking at a single waveform.
in all these cases the in-phase clips are symmetrical, while the
random-start-phase clips are asymetrical. this was easier to see in these
clips.
the most interesting thing is i reversed the in phase pseudo sawtooth clip
and can
definitely hear a difference between the normal and reversed sound. i think
this is
a clue as to what is going on here. inverting the clip sounds identical to
reversing it (at least this makes sense)...
i'd really like to know why this is so, since i've been told and believed
for a long time that we couldn't hear phase differences directly.
maybe it's the speakers or audio card responding "wrongly" to the
assymetrical waves?
if so why does the inverted sound of the symmetrical psudo-sawtooth wav
sound different??
ok, anybody crazy enough to read all this ought to be crazy enough
to do the same test in their audio editor of choice
More information about the Synth-diy
mailing list