DPRG List  

Fw: DPRG: Re: Audio

Subject: Fw: DPRG: Re: Audio
From: Kevin Ross kevinro at nwlink.com
Date: Sat Jan 3 14:19:30 CST 1998

>>> >
>>> >1. The microphones should be of cardioid pattern (narrow beam) and
>>> >should be aimed up (on a short robot) and forwards.
>Why? Our ears arent. Neither are any animals'. I don't think anyway. A
>cardioid mic uses phased info to make it directional. It requres sound
>info to be introduced from behind the mic. Our ears have only one main
>entrance for sound. The others are for resonance (the tubes connecting
>the sinuses and ear, and eyes and sinuses, et al.)

Nonsense. Lots of animals have extremely directional hearing, such as Bats,
Horses, Bears, etc. When they detect a sound, they refocus their parabolic
aperatures (ie ear lobes) in the direction of the suspected sound. It allows
them to focus their processing attention on a particular direction for more
detailed processing.

The human ear is much more complex than a single entrance for sound. The ear
lobe functions as a signal attenuation device as well as a directional
finder. Our brains have been trained over time to detect the differences
between a noise from in front vs behind. A person who loses an ear lobe has
an extremely difficult time determining direction of sound with respect to
front/back. The lobe causes a predictable change to the sound pattern by
creating variations in the sound that are detected by the brain.  Extremely
cool, but way beyond this groups practical ability to duplicate.
>From a practical standpoint for creating a robot, a cardioid pattern
microphone is a great place to start. Your initial plan is seeking maximum
sound levels on both channels. To determine the direction, you want to have
maximum sound levels that are in phase with each other (ie start at the same
time). I would use the starting edge of a sound wave, and time the
difference between the two.

If you get good success with this, you might be able to try a more extensive
set of experiments.


More information about the DPRG mailing list