ANC Headset

Thread view:

Enable EMail Alerts

Start New Thread

Thread rating:

MCBSP - 23 Mar 2006 23:20 GMT

Hi everyone,

I am currently looking for some information about digital ANC Headsets.
I have bought an ANC book from KUO Active noise control. But in
conjunction with ANC headsets there is only one chapter. Therefore I
have done some more inquiries and found an really interesting Paper.
This paper describes some investigations with a feed-forward System.
The whole System consists of a headphone, a referencmicro which is
attached outside the earcups, an errormicro which is attached inside
the earcups below the speaker and a DSP.
The errormicro has to record the result of the acoustical superposition
between the noise and the generated "antinoise". The results which are
shown in that paper are pretty good. Up to 20dB attenuation between
40Hz and 750Hz. But a second measurement result shows that there are
big diffrenez between the attenuation next to the errormicro and next
to a micro which is placed into the ear canal. Up to 150Hz the result
are the same but above 150Hz the attenuation at the ear canal micro
decrease rapidly. There is now attenuation above 500Hz.
Is there anybody who could explain me this fact in simple words.
Because I am not a specialist in physics. I am a student of
electronical engineering.

Thanks MCBSP

Reply to this Message

The Ghost - 24 Mar 2006 00:58 GMT

> Hi everyone,
>
[quoted text clipped - 19 lines]
>
> Thanks MCBSP

The cancellation is the result of taking the difference between two
signals. One signal is that portion of the outside noise that leaks
through the cup. The other signal is the cancellation signal that is
produced by the ANC circuitry. For a variety of reasons, neither of these
two signals are the same everywhere inside the earcup or in the ear canal.
Because the ANC circuitry works to minimize the pressure ONLY at the
location of the error microphone, as one moves away from the error
microphone and into the ear canal, cancellation is degraded because of
spatially dependent amplitude and phase differences between the two
signals. The cancellation process is VERY sensitive to these differences.
For example (ingoring phase), 1-0.99 yields 0.01 or 40dB of cancellation,
wherease 1-0.95 yields 0.05 or 26dB of cancellation.

Reply to this Message

MCBSP - 24 Mar 2006 19:42 GMT

> > Hi everyone,
> >
[quoted text clipped - 32 lines]
> For example (ingoring phase), 1-0.99 yields 0.01 or 40dB of cancellation,
> wherease 1-0.95 yields 0.05 or 26dB of cancellation.

Thanks for your reply. I thought the wavelength of a 500 Hz signal is
so big compared to the distance between the errormicro and the mirco
inside the ear canal so the phase differences shouldn't be so enormous.
Could you advise me some papers or books where this problems are
discussed ? What do you thing about the idea to include the transfer
function between the errormicro and the ear canal into the caculation
of the cancellation signal? Is it a really complex transfer function or
could you model them by a simple delay or something else?

Reply to this Message

The Ghost - 29 Mar 2006 02:06 GMT

>> > Hi everyone,
>> >
[quoted text clipped - 38 lines]
> inside the ear canal so the phase differences shouldn't be so
> enormous.

That is only approximately true for sealed volumes which are very small
relative to the wavelength of sound. The ANC situation does not involve a
sealed volume because of acoustic leaks between the headset and the head
and because of energy flow into the ear canal.

> Could you advise me some papers or books where this problems
> are discussed ?

Sorry, but I am not familiar with the literature in this area.

> What do you thing about the idea to include the
> transfer function between the errormicro and the ear canal into the
> caculation of the cancellation signal?

It could work, but only under sinusoidal steady-state conditions.
Additionally, the transfer function would be unique to every
headset/individual combination.

> Is it a really complex transfer function

More likely than not.

> or could you model them by a simple delay or something else?

If you are talking about cancelling a real-time dynamic signal at a point
within the ear canal, you would need a physically-unrealizable negative
delay because of the finite distance and associated finite time delay
between the cancelation point and both the error microphone and the
transducer that is generating the cancellation signal.

Reply to this Message

Natural Science Forum / Physics / Acoustics / March 2006

ANC Headset