Little OT : Software for Active Noise Cancelling or Reduction
Mirco Piccin
i was considering to buy headphones with Active Noise Cancelling / Reduction.
But before spend money, i'd like to know if there's a software that could do the same job (for free).
Thanks in advance!
Regards
M
John Hasler
> i was considering to buy headphones with Active Noise Cancelling / Reduction.
> But before spend money, i'd like to know if there's a software that
> could do the same job (for free).
No. Not feasible.
--
John Hasler
--
To UNSUBSCRIBE, email to debian-us...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listm...@lists.debian.org
Mark Neidorff
> M writes:
> > i was considering to buy headphones with Active Noise Cancelling /
> > Reduction.
> >
> > But before spend money, i'd like to know if there's a software that
> > could do the same job (for free).
>
> No. Not feasible.
> --
> John Hasler
Is it technically not feasable, meaning that a room is too large to do noise
cancelling in, or not feasable from the linux software prespective?
Mark
Jeff Soules
> Is it technically not feasable, meaning that a room is too large to do noise
> cancelling in, or not feasable from the linux software prespective?
The former, I'd think. For that to work on the room level, you'd need
to have sensors surrounding the room and speakers in all the walls.
Sound waves inside a room aren't uniform enough for the a simple
speaker setup to cancel them out. You've probably noticed how the way
a room sounds can change substantially just from turning your head or
shifting it a few inches one way or the other? The same would be true
with a microphone assembly that you tried to use for noise
cancellation. And if the noise cancellation weren't calibrated
correctly, you'd wind up with spots that were noisier than they'd be
without the setup.
Really this is a job best suited to headphones.
Mirco Piccin
i find this thread:
http://forum.audacityteam.org/viewtopic.php?f=28&t=442
It's quite old, but it could be a start point.
The aim is to create an anti-phase in realtime.
This job is done by a don't-know small (and suppose poor) chip into some headphones that provide noise cancelling/reduction.
So i suppose it's not so difficult to do the same using a pc process and Debian (or other linux distro, of course).
Of course i don't want absolute noise cancelling (i think that it's also impossible).
A good reduction is enough.
Regards
M
John Haggerty
it's an interesting idea but doing dynamic sound calculations for an exact counter to all the acoustics of a room would be something that at best would take longer than the sound conditions of a "public" area would work. Maybe private.
Think about it if you had a private area in which conditions (if you theoretically ran a digital recorder would give you several hours of the same sound with no new ones) are going to remain static then a math intensive analysis of the direct diametric oposite of the wave would work because the work wouldn't "expire" (ie get new sounds).
Public you might get new sounds at least every 10-20 minutes. Let's cheat and say that the math can be crunched out in 15 minutes that would leave you with 'overlap' or 'delay' of 5 minutes where your program "wouldn't work".
That's at least what I would think. Then again I've been suprised about what gets posted to freshmeat.net these days anyway.
John Hasler
> Is it technically not feasable, meaning that a room is too large to do
> noise cancelling in, or not feasable from the linux software
> prespective?
Noise-canceling headphones only cancel noise right at the headphones so
the size of the room is not relevant. Even so, dedicated DSP hardware
is required. A process running under Linux cannot possibly be fast
enough. It would have to process at least two samples per cycle of the
highest frequency of interest. It doesn't matter how powerful your
processor is: the latency would kill you.
--
John Hasler
Micha Feigin
Mark Neidorff <ma...@neidorff.com> wrote:
> On Wednesday 05 August 2009 04:42 pm, John Hasler wrote:
> > M writes:
> > > i was considering to buy headphones with Active Noise Cancelling /
> > > Reduction.
> > >
> > > But before spend money, i'd like to know if there's a software that
> > > could do the same job (for free).
> >
> > No. Not feasible.
> > --
> > John Hasler
>
> Is it technically not feasable, meaning that a room is too large to do noise
> cancelling in, or not feasable from the linux software prespective?
>
Pretty much technically not feasable and from a scientific point of view, not even mathematical
Think of it this way, if you want to cancel a sound wave you and another sound
wave traveling in the same direction but inverted. Think of sea waves, you need
to invert the wave. If they travel at different directions they will cancel at
some points and add at others and may not even exist together at some places.
You would also get different effects at different wave lengths due to the
differing relative error.
This mean that you need to cancel the wave exactly at the source or on a
complete sphere around the source (with an accurate rendition on that sphere
which would mean nano speakers). The second problem is that you also need exact
measurements to create the cancellation wave, also on the entire sphere, and to
take account the delay between measuring and reacting (even assuming zero time
computation).
This is only partially feasible at the headphone level where the listener and
speaker are close together with a know orientation relative to the microphone,
minimizing the relative error, this also explains why you get noise reduction
and not noise cancellation and different effective with different noises
(depending on the uniformity, pitch and direction). A good algorithm also needs
to take into account where the speaker, microphone and listener are all
relatively located and take an assumption on the direction the noise is coming
from.
It may be feasible to improvise noise cancellation headphones though, by
sticking a microphone on the headphones and feed the input back inverted with
the correct delay and volume. At this level it would only take some
electronics, no processor at all.
I don't have the time though to create a simulation at the moment to see if you
need some processor based optimization of not. It does sound like a fun test
though.
Tzafrir Cohen
> Mark writes:
> > Is it technically not feasable, meaning that a room is too large to do
> > noise cancelling in, or not feasable from the linux software
> > prespective?
>
> Noise-canceling headphones only cancel noise right at the headphones so
> the size of the room is not relevant. Even so, dedicated DSP hardware
> is required. A process running under Linux cannot possibly be fast
> enough. It would have to process at least two samples per cycle of the
> highest frequency of interest. It doesn't matter how powerful your
> processor is: the latency would kill you.
Considering audio voice processing is done at the host anyway, why not
also cancel acoustic echo?
http://speex.org/docs/manual/speex-manual/node4.html#SECTION00450000000000000000
--
Tzafrir Cohen | tza...@jabber.org | VIM is
http://tzafrir.org.il | | a Mutt's
tza...@cohens.org.il | | best
ICQ# 16849754 | | friend