From: ziert@beloit.edu (Tom Zier)
Subject: Re: SCI: sound cues for 'depth'?
Date: Tue, 27 Apr 93 15:36:30 CDT


> From: stevep@snowhite.cis.uoguelph.ca (Steve Portigal)
>   I'm interested in some suggestions, pointers to literature,
> expertise etc in some non-speech audio cues. I'm trying to use
> different dimensions of the same sound to convey various info:
> duration to indicate size, pitch to indicate distance moved, and
> something convey depth.

Literature, I don't know, there are some good references in the FAQ though.

You have two distinct elements to work with; the 'envelope' and its
'contents'.  The envelope, then, can be characterized by 'attack',
'sustain', 'duration', and 'amplitude'.  Attack (the leading edge of the
burst), sustain (the trailing edge of the burst), and duration (length of
the burst) can, due to atmospheric diffusion, yield information about
'distance' from the source of the sound because near to hand (ear) sounds
are always sharper relatively speaking, just as details in vision are
sharper near to hand.  A deterioration amplitude with distance from the
source (for the very same reason, atm. diff.) would be the most common cue,
but can be far more convincing when used in concert with changes in attack,
sustain, and  duration.

>  I'm not sure how to convey depth. I've thought about "reverb" or
> "echo" as indicating something very similar to physical depth, ie: the
> deeper underground you go, the more "cavernous" things sound.

Yep, sonar works pretty good.  And may be even more effective (in VR) if a
human can hit a key, or make a gesture, which coincides with the original
pulse.  That way you hook the sensory motor network into the task in a
fashion that randomly generated, or automatic, impulses would fail to do. 
Our understandings of 'location of residence' are processed, in every case,
based upon kinesthetic relationships.

The deeper you go in the cave, the more complete your feedback will be, but
how deep you are in the earth has very little to do with the reverberation's
timing or percent amplitude returned of the primary burst.  I think you are
confusing the depth in the earth of your space with its proximity to the
cave's mouth.  Near (1/4 mile or so) to the cave's mouth echoes can
actually, barring any more certain indicator like knowing what you're doing,
tell you in which direction to go for the nearest exit.  Perhaps we could
allow our perceiver to turn in each of the cardinal directions and clap;
they would then know in which direction there was a larger space (or how
close to the exit they are).

>  Perhaps more simply is (and I lack the technical expertise to specify
> this properly) something like "thickness" of the sound. I don't know
> what specific characteristics of sound convey thickness.

Now you're talking about the contents of the envelope specifically.

The number of harmonic components contained in a wave-form packet will
generally determine the 'thinness' or 'thickness' of a sound.  For instance,
the thin sound of a triangle could be compared to the sonorous, thick,
sounds of a cello.  The thin sound may have only the primary wave and a
partial secondary harmonic; whereas the cello is rich in vibrations from the
sound box, other strings, even the technique of the musician, enhancing the
harmonic blend.

>  If someone could enlighten me a bit here, that'd be great. Any
> suggestions as how to generate such samples for use on a Macintosh (I
> can vary pitch and duration myself) would be useful. I'm doing some
> experimental research and need to come up with some audio cues...

Only minimally inclined to computational entanglements, I will leave this to
more expert commentators.  Check out the sound recording utilities on your
Mac, though, you may be pleasantly surprised.

>  Thanks to anyone with any info...

*****************************
T. Zier                ziert@beloit.edu
*****************************
Pay no attention to the man behind the curtain, he speaks for no-one.