Looper's Delight Archive Top (Search)
Date Index
Thread Index
Author Index
Looper's Delight Home
Mailing List Info

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Date Index][Thread Index][Author Index]

Re: OT: Personal Audio Space?

From: "Allan Hoeltje" <ahoeltje@best.com>

>Have you ever been in a loud restaurant and couldn't make out what the
>person sitting across from you was saying because the background noise was
>so loud?
>What I want is a "perception aid".  It seems to me that DSP's are good
>enough that if combined with a couple good microphones you could cancel 
>the noise and enhance the nearby sound.  Fit the electronics and mics 
>into a
>battery powered, walkperson sized box, plug into it with head phones (the
>ear plug kind, like what some performers use on stage) and have yourself a
>"personal audio space".
>Anyone know if this is possible, or available?
>For looping or recording, it could be used as a pre-amp (a de-amp?) in 

This is a great idea and it's certainly not implausible at first glance.

You all know about the sound cancellation earphones (that were invented
for the first round-the-world plane flight when they worked out that
the pilots would be deaf before the halfway point because of the

But a similar idea is used in professional sound reinforcement in large 
where speakers with signals out of phase with the original are placed
in order to reduce the effect of reflections and improve intelligibility.

The trouble is setting the correct phase and thus delay for each
frequency and direction.  The pros do it with a complex series of
measurements and experiments in the space itself...

Now suppose you just had a speaker and a mic and a little microprocessor 

the mic faces outward.
the speaker faces inward.
the uP is logically "in between".

         blah \                                 / -blah
  THEM   blah -  ->  mic -- uP -- speaker - -> -  -blah  YOU
         blah /                                 \ -blah

1. the mic gets a signal and send it to the uP.
2. the uP inverts the signal,
   (then adds a negative feedback/delay term in order to damp out
    the feedback through the mic -- this is where the tricky
    part comes in)
3. and sends it to the speaker, thus cancelling out the noise, at
    least in the direction of that speaker and YOU!

have five of these things and you are set!  (you'd want one
overhead because a lot of the sound would be coming from
ceiling reflections, particularly in a restaurant where
the ceiling is usually the only good reflector!)

even two might work pretty well.

now, with a bigger uP you could do stuff like this:

... have a lot of small mics

... compute the actual 3-d "sonograph" of the sound

... reduce it similarly to holography to a three channel version

... send the inverse to three speakers and "completely" cancel the
     sound in a small area.

the amount of computation involved is large but not inconceivable
and the mathematics is "well-known" (to people other than myself,
mind you...)

<http://herodes.feld.cvut.cz/akustika/akhol.htm> is a really
nice short though slightly technical summary of this technique.


                                  that was fast

.......all legal games of chess 
.....programmer's documentation