On Tuesday, December 9, 2014 12:22:51 AM UTC-6, dspjo...@gmail.com wrote:
> Hello,
> 
> I'm hoping the experienced folks in this newsgroup would be so kind so point me to some good tutorials (and books) on the subject of active noise cancellation.
> 
> Being new to ANC, this is my understanding of how ANC works:
> 
> Ambient noise (n) is entering a geographically small zone. To suppress the noise, an interfering (corrective) sound is injected acoustically into the zone by a speaker. The interfering sound from the speaker is adaptively generated so that it - in theory - ends up being a 180 degree phase-shifted version of the noise ....When this phase-shifted signal is acoustically mixed with the ambient noise, the noise is ideally cancelled out. 
> 
> So based on the above, I would need a microphone to record sound as it presents itself in the "zone" and then let that drive an adaptive process which calculates a digital speaker signal to send to the speaker. The energy of the recorded sound would be the indicator of how well the adaptive process is performing (in terms of cancelling out the noise).
> 
> My question is : what type of algorithm(s) are used for this kind of DSP scheme?
> 
> Is one microphone enough?
> 
> What if the speaker has to output more than just the noise cancelling signal? For example: A user is wearing headphones with ANC. The user is listening to music but wants the "outside" noise to be cancelled out. I can't imagine that it would work unless the headphone has an "inside" microphone and an "outside" microphone.
> 
> Does ANC work in dimensionally small setups? For example: 
> 
> The speaker is connected to a tube through which the corrective sound travels. The open end of the tube is placed in a small zone where a microphone is also located. An algorithm receives audio from the microphone, calculates the speaker signal and sends the speaker signal to the speaker. Will that work? Or are there some fundamental requirements/assumptions which have to be met for the ANC to work?
> 
> It would be nice if some of the folks in this group could share their experience and knowledge about ANC with me. Any advice is appreciated.
> 
> Thank you.

Wow, this is what I get for not proof reading. Sen's last name is spelled Kuo, not Quo. If I remember correctly, he wrote a paper on different noise control algorithms.

Maurice Givens

Maurice Givens

On Tuesday, December 9, 2014 12:22:51 AM UTC-6, dspjo...@gmail.com wrote:
> Hello,
> 
> I'm hoping the experienced folks in this newsgroup would be so kind so point me to some good tutorials (and books) on the subject of active noise cancellation.
> 
> Being new to ANC, this is my understanding of how ANC works:
> 
> Ambient noise (n) is entering a geographically small zone. To suppress the noise, an interfering (corrective) sound is injected acoustically into the zone by a speaker. The interfering sound from the speaker is adaptively generated so that it - in theory - ends up being a 180 degree phase-shifted version of the noise ....When this phase-shifted signal is acoustically mixed with the ambient noise, the noise is ideally cancelled out. 
> 
> So based on the above, I would need a microphone to record sound as it presents itself in the "zone" and then let that drive an adaptive process which calculates a digital speaker signal to send to the speaker. The energy of the recorded sound would be the indicator of how well the adaptive process is performing (in terms of cancelling out the noise).
> 
> My question is : what type of algorithm(s) are used for this kind of DSP scheme?
> 
> Is one microphone enough?
> 
> What if the speaker has to output more than just the noise cancelling signal? For example: A user is wearing headphones with ANC. The user is listening to music but wants the "outside" noise to be cancelled out. I can't imagine that it would work unless the headphone has an "inside" microphone and an "outside" microphone.
> 
> Does ANC work in dimensionally small setups? For example: 
> 
> The speaker is connected to a tube through which the corrective sound travels. The open end of the tube is placed in a small zone where a microphone is also located. An algorithm receives audio from the microphone, calculates the speaker signal and sends the speaker signal to the speaker. Will that work? Or are there some fundamental requirements/assumptions which have to be met for the ANC to work?
> 
> It would be nice if some of the folks in this group could share their experience and knowledge about ANC with me. Any advice is appreciated.
> 
> Thank you.

I forgot to add, go to ti.com and search on active noise control.

Maurice Givens

On Tuesday, December 9, 2014 12:22:51 AM UTC-6, dspjo...@gmail.com wrote:
> Hello,
> 
> I'm hoping the experienced folks in this newsgroup would be so kind so point me to some good tutorials (and books) on the subject of active noise cancellation.
> 
> Being new to ANC, this is my understanding of how ANC works:
> 
> Ambient noise (n) is entering a geographically small zone. To suppress the noise, an interfering (corrective) sound is injected acoustically into the zone by a speaker. The interfering sound from the speaker is adaptively generated so that it - in theory - ends up being a 180 degree phase-shifted version of the noise ....When this phase-shifted signal is acoustically mixed with the ambient noise, the noise is ideally cancelled out. 
> 
> So based on the above, I would need a microphone to record sound as it presents itself in the "zone" and then let that drive an adaptive process which calculates a digital speaker signal to send to the speaker. The energy of the recorded sound would be the indicator of how well the adaptive process is performing (in terms of cancelling out the noise).
> 
> My question is : what type of algorithm(s) are used for this kind of DSP scheme?
> 
> Is one microphone enough?
> 
> What if the speaker has to output more than just the noise cancelling signal? For example: A user is wearing headphones with ANC. The user is listening to music but wants the "outside" noise to be cancelled out. I can't imagine that it would work unless the headphone has an "inside" microphone and an "outside" microphone.
> 
> Does ANC work in dimensionally small setups? For example: 
> 
> The speaker is connected to a tube through which the corrective sound travels. The open end of the tube is placed in a small zone where a microphone is also located. An algorithm receives audio from the microphone, calculates the speaker signal and sends the speaker signal to the speaker. Will that work? Or are there some fundamental requirements/assumptions which have to be met for the ANC to work?
> 
> It would be nice if some of the folks in this group could share their experience and knowledge about ANC with me. Any advice is appreciated.
> 
> Thank you.

Active noise control is being used for machine noise cancellation in assembly plants, road noise reduction in automobiles, etc. The last I remember, a variant of the LMS algorithm was being used called the "filtered-X" variant. At the 2004 comp.dsp conference, a paper was presented on active noise control, and can be downloaded here: https://www.danvillesignal.com/compdsp-conference/compdsp-conference-presentations.

Prof. Sen Quo at Northern Illinois University has students doing research in active noise cancellation. You might see if he or his students have papers for download on their website. Finally, "google" active noise control or active noise cancellation and you will get a plethora of information.

Maurice Givens

On Thursday, December 11, 2014 6:51:00 AM UTC+13, dspjo...@gmail.com wrote:
> > There was a previous thread on this topic a year ago or so. Most active >headphone ones are analogue and use fixed filters. That is good enough for most >problems.
> 
> But in this case it's not a typical headphone. In this case a speaker (of small dimensions) is channeling sound through an open-ended tube. The end of that tube is placed at the entrance of the ear canal. A microphone is situated close to the end of the tube (so that it 'hears' what the human ear hears and so that the ANC algorithm is able to adjust the speaker signal until 'silence' is observed in the space where the mic is located). The goal is to produce a speaker signal which cancels out (or at least suppresses) the outside ambient noise which is entering the ear. I don't know if this is feasible/possible, so that's why I'm asking for some guidance, thoughts and recommendations from the more experienced DSP folks in the group.

From a year ago:

With digital (non acoustic) noise cancellation we normally use something
> like LMS to match the phase and amplitude so I assume here you mean that
> they make the assumption that that adaptive filter is replaced by a fixed
> analogue filter which has a common characteristic based on tests in real
> environments.

Yes to replaced.  One important point, though, is that the controller that the LMS adaptation is providing you and the controller of a typical ANC headset are fundamentally differently.

The fixed analog controller in the ANC headset is a robust controller.  The controller does not know anything about the noise problem in which it is immersed.  The controller is intended to provide some predetermined quieting based on speculation of the designers(s) as to the disturbance.  It is okay everywhere and excellent nowhere.

If the sound at a particular location is your error function, the adaptation mechanism that you are using as part of your ANC system is causing your control filter to converge on an (instantaneous) optimal controller.  It will provide you with excellent performance for the operating condition for which it converged - and likely do little else anywhere else.

Here is an interesting exercise that may help to illustrate...

About 14 years ago, I was flying in a small prop plane with an leaky LMS-based adaptive controller that was reducing the noise at the pilot's head area.

The pilot reaches cruise.  I turn the controller on and it converges.  On-off-on-off - something like 8dBA of reduction.

Now I set my learning rate to 0.0 and my leakage term to 1.0 - what did I do?  I froze the coefficients of my adaptive controller.  The coefficients will not leak down nor will they adapt.  I now have a fixed digital controller.

I turn it on.  What reduction do I get?  8dBA, of course.   The spectrum of the plane hasn't changed.  The controller had already converged on the optimal solution for that operating condition.  The fact that I had turned off the adaptation meant nothing to the control.  Until the operating condition changes, there is no need for further adaptation.

The pilot flew for 2 hours before he changed the operating condition (RPM).  The fixed digital controller sounded great the whole time.

When the pilot changed the RPM, though, the spectral peaks of the disturbance moved out of the deep wells in the sensitivity curve of the now-fixed digital controller.  Only then did I have to turn the controller adaptation back on.  My fixed controller wasn't effective any longer.

Could I have designed a fixed controller that gave me 8dB everywhere?  Nope.  We are back to the Bode Sensitivity Integral thing.

On Thursday, December 11, 2014 6:51:00 AM UTC+13, dspjo...@gmail.com wrote:
> > There was a previous thread on this topic a year ago or so. Most active >headphone ones are analogue and use fixed filters. That is good enough for most >problems.
> 
> But in this case it's not a typical headphone. In this case a speaker (of small dimensions) is channeling sound through an open-ended tube. The end of that tube is placed at the entrance of the ear canal. A microphone is situated close to the end of the tube (so that it 'hears' what the human ear hears and so that the ANC algorithm is able to adjust the speaker signal until 'silence' is observed in the space where the mic is located). The goal is to produce a speaker signal which cancels out (or at least suppresses) the outside ambient noise which is entering the ear. I don't know if this is feasible/possible, so that's why I'm asking for some guidance, thoughts and recommendations from the more experienced DSP folks in the group.

I think still digital maybe if you can find the right filter.

> There was a previous thread on this topic a year ago or so. Most active >headphone ones are analogue and use fixed filters. That is good enough for most >problems.

But in this case it's not a typical headphone. In this case a speaker (of small dimensions) is channeling sound through an open-ended tube. The end of that tube is placed at the entrance of the ear canal. A microphone is situated close to the end of the tube (so that it 'hears' what the human ear hears and so that the ANC algorithm is able to adjust the speaker signal until 'silence' is observed in the space where the mic is located). The goal is to produce a speaker signal which cancels out (or at least suppresses) the outside ambient noise which is entering the ear. I don't know if this is feasible/possible, so that's why I'm asking for some guidance, thoughts and recommendations from the more experienced DSP folks in the group.

On Tuesday, December 9, 2014 7:22:51 PM UTC+13, dspjo...@gmail.com wrote:
> Hello,
> 
> I'm hoping the experienced folks in this newsgroup would be so kind so point me to some good tutorials (and books) on the subject of active noise cancellation.
> 
> Being new to ANC, this is my understanding of how ANC works:
> 
> Ambient noise (n) is entering a geographically small zone. To suppress the noise, an interfering (corrective) sound is injected acoustically into the zone by a speaker. The interfering sound from the speaker is adaptively generated so that it - in theory - ends up being a 180 degree phase-shifted version of the noise ....When this phase-shifted signal is acoustically mixed with the ambient noise, the noise is ideally cancelled out. 
> 
> So based on the above, I would need a microphone to record sound as it presents itself in the "zone" and then let that drive an adaptive process which calculates a digital speaker signal to send to the speaker. The energy of the recorded sound would be the indicator of how well the adaptive process is performing (in terms of cancelling out the noise).
> 
> My question is : what type of algorithm(s) are used for this kind of DSP scheme?
> 
> Is one microphone enough?
> 
> What if the speaker has to output more than just the noise cancelling signal? For example: A user is wearing headphones with ANC. The user is listening to music but wants the "outside" noise to be cancelled out. I can't imagine that it would work unless the headphone has an "inside" microphone and an "outside" microphone.
> 
> Does ANC work in dimensionally small setups? For example: 
> 
> The speaker is connected to a tube through which the corrective sound travels. The open end of the tube is placed in a small zone where a microphone is also located. An algorithm receives audio from the microphone, calculates the speaker signal and sends the speaker signal to the speaker. Will that work? Or are there some fundamental requirements/assumptions which have to be met for the ANC to work?
> 
> It would be nice if some of the folks in this group could share their experience and knowledge about ANC with me. Any advice is appreciated.
> 
> Thank you.

There was a previous thread on this topic a year ago or so. Most active headphone ones are analogue and use fixed filters. That is good enough for most problems.

On 12/9/2014 4:41 PM, dspjohnson3@gmail.com wrote:
>> i *really* can't see how this can work for an open speaker setup, except
>> *maybe* for a single person sitting in the "sweet seat".  there is just
>> no way for a single speaker to output a wavefront that would, for
>> everyone in a broad geographical area, destructively interfere with the
>
> I guess I didn't describe the setup too well :-(
>
> When I wrote "geographically small zone" I mean a small cavity -- like this:
>
> http://postimg.org/image/r0ccfkug9/

You can consider small in the context of the wavelength of the sounds 
you wish to cancel.  Once the dimensions of the area becomes a 
significant proportion of the wavelength it won't be possible to 
compensate for all noise over the entire space.

If you are using it to play a sound through the speaker, I would 
consider this to be the same as any other distortion in the signal, best 
compensated for using negative feedback.

     +----------------------------------+
     |  |\                              |
     +--|-\      /|                     |
Audio  |  >---[] |---| Noise |---O==---+
   o----|+/      \|
Signal |/    Speaker          Microphone

This should work ok for systems where there is not much delay between 
the speaker and the mic.

-- 

Rick