Hi All,

I'm just wondering if anyone has used noise cancellation technology in / around their systems please?

If so, I'd appreciate your advice / thoughts / evaluations.

I was thinking an internal microphone and case speakers with some free software programme (hopefully with a small cpu footprint) to drive it.

From what I understand, the microphone and speaker environment should be quite limited to save on power and be more effective. (I was after a less restrictive option than using headphones)

Many thanks in advance.

Ive used regular noise cancelling headphones -- and yes they work.

but to use one around your PC would have some interesting issues:
1) quality of the mic to pick up sounds from the PC
2) direction of the mic - is it picking up JUST the pc sounds? or other stuff in the room?
3) power of the amplifier vs distance from the PC (if you get too far away the cancelling effect might stop -- speculation on my part)

But IMO - the most importand factor will be the quality of reproducing the noise cancelling sound. An inexpensive cheap speaker probably will not work.

your best bet is to try and quiet the existing components.

The level of design and engineering to make active noise cancellation work reasonably well in this situation is probably very high. It's done in headphones because they provide a nearly ideal environment; you know the exact response of the speakers and mic you are using, acoustics are almost a non-factor, you know the exact distance from the ear, you have complete stereo separation which virtually eliminates phase issues, you can have the mic at almost exactly the same place as the speaker, etc.

I'm sure it could be done, but it's not something you are going to be able to DIY and get the result you are looking for, it would take a team of specialized engineers. Spending your time and money on actually quieting the system would provide much better results.

There was a discussion on this forum some time ago about this idea...

Found it !!!
http://www.silentpcreview.com/forums/vi ... cancelling

^ Thanks for the links. I didn't manage to pull those out when I searched (I did seaerch first before posting, honest!)

If anyone has actually tried (and failed / succeeded) then feel free to post, else see the previous interesting discussions (although theoretical only, except for the commercially produced pc).

Thanks again!

Active Noise Cancellation (ANC) isn't too feasible for computers for a number of reasons:

1. Generally the noise in computers is of mid to high frequencies - active noise cancellation works with low frequencies better because of the longer wavelengths. It's easier to have the two waves (primary noise source, and the cancelling source) match up at lower frequencies if the waves are longer, there's some error tolerance. At higher frequencies, the wavelengths are on the order of centimeters, and it's tough to get them to match.

2. Cost - you'll need several microphones, and ideally several speakers. And you'll have to have enough processing power to compute the cancelling noise, which isn't easy.

3. Processing power - this may not be too much of an issue with the newer computers we have today, but it has been an issue in the past. The problem comes when the noise being measured by the (several) microphones changes. Say, when you talk, or someone walks by in the next room, or you're crunching away on those Doritos. The program has to figure out what noise to cancel, and if it's not a smart algorithm, it will try to cancel everything, which will likely be a bad thing.

4. You'll get higher noise levels on the other side of the speaker(s). With the speakers outputting sound that is opposite in phase to the noise towards the front, behind the speaker you'll get noise that is IN-phase with the noise, making that side louder. If you have the speaker mounted in the case, then you'll likely have more noise inside the case, which you'll have to deal with.

5. Where do you put the speaker(s)? And for that matter, where do you put the microphones? If you put speakers in the case, you risk increasing the noise levels inside the case. Put the speakers around you and you have more noise outside your "cone of silence". Then you have to contend with placing several microphones around you to measure the noise so that the ANC controller will know what to cancel...

6. Also at higher frequencies, speakers tend to be more directional, adding another level of difficulty to the mix - directivity. A smaller range of directivity at the higher frequencies leads to a smaller area in which the noise cancellation will be effective.


There are probably a few more factors that I haven't considered here, but overall it's a very technically challenging problem that a lot of people are still working on to get working...

It's far from that simple.

The setup will create an infinite number of sound minima and maxima in space around it, so by moving around you will find some quiet positions and some that are more noisy than without the cancellation.

It only works well with headphones, because the headphones only need to cancel noise at one point in space - the ear canal - which is not moving relative to the headphones and which is very near the cancellation device.

Let's do a very simplified example:

Assume that noise is not attenuated by distance from source. Assume that we have 2 noise sources in free space, producing the same sine wave at frequency f at the same intensity and in the same phase. Assume the distance of the listener from source 1 is d1 and source 2 is d2.

Now the other source is d=(d1-d2) further away from (or nearer to) the listener than the other.

This difference in distance creates a 2*pi*f*d/speed of sound difference in phase for the 2 signals in the listeners position. This creates interference. The intensity of the sum signal at the listeners position is 1+cos(phase angle), which is a value between 0 and 2. So depending on the position of the listener in space, these 2 noise sources either totally cancel each other or add up perfectly or something in between.

Now take away the simplifications and you notice that it becomes totally impossible to do. For example, the location of the minima and the maxima depend on the frequency of the sound, and it will be very hard to do it for all frequencies at the same time. But it also depends on the attenuation and echoes (by frequency) created by the listening space.

It's definitely far from simple, but I tried to explain it so that it's easy to understand.

The difficulty with phase and positioning of the sources (computer and noise-cancelling speaker) will definitely impact the volume of space where you will get 'good' noise cancellation. Within that volume (where hopefully your head/ears will be), you'll get good noise cancellation, but everywhere else in the room as well as in the computer, it will average out to be louder on the whole. (You'll probably get some areas in the room that are quieter too, but generally still louder overall)

The areas where the noise will be cancelled will largely depend on directivity of the speakers, where the speakers are positioned, and where the microphones are positioned - not to mention the directivity of the source, the computer fans/HDs/etc. What you're trying to do is make it so that the sound from the speaker is matching and 180 degrees out of phase AT YOUR HEAD position, which will likely be where the microphones are positioned.

It's difficult, but not impossible to reduce the overall average noise level in the room with ANC, and it's been shown even at the UBC labs that it's possible with a steady tonal source (not sure if it was tried with broadband noise) to optimize microphone positioning and speakers such that you'd get average 3-5 dB reduction over the entire room. Even with the minima and maxima present, as you explained, occurring because of the non-180 degree interference, they showed it was still possible. However, my previous points still stand, that you'd have to have quite a bit of processing power and a very good algorithm to handle the extraneous non-steady state noise that the ANC system would invariably also try to cancel out (i.e., music, talking, etc).

Dear All

Thanks so much for your detailed reply to my question. I can see that it is definitely a 'no' to a simple solution at least!

Now that I think about it, even when my noise cancelling headphones try to do it, there is a broadband hiss, which I'm guessing is going to be similar to the quiet fan noise I now have.

Thanks again! Great forum, great discussions here!!!