DiyMobileAudio.com Car Stereo Forum banner

1 - 9 of 9 Posts

·
Registered
Joined
·
77 Posts
Discussion Starter #1
I was wondering if anything exists or anyone is working on something like this. Basically a DSP that can filter and cancel external noises with phase cancellation. With todays DSP units having the power to pull this off it would be amazing. It could map out the freq response of the cabin, like autotune, then perform a constant on the fly RTA of externally generated noises and run a cancellation algorithm through the signal chain, restoring the audio to how its supposed to sound in a dead environment. basically have a mic feeding it constantly and readjusting the signal on the fly. or am i just crazy.
 

·
Premium Member
Joined
·
6,045 Posts
I was wondering if anything exists or anyone is working on something like this. Basically a DSP that can filter and cancel external noises with phase cancellation. With todays DSP units having the power to pull this off it would be amazing. It could map out the freq response of the cabin, like autotune, then perform a constant on the fly RTA of externally generated noises and run a cancellation algorithm through the signal chain, restoring the audio to how its supposed to sound in a dead environment. basically have a mic feeding it constantly and readjusting the signal on the fly. or am i just crazy.
If I'm not mistaken, the wavelengths are too short to do this effectively at midrange and high frequencies.

For instance, let's say the ambient noise level at 1khz is 80dB. To null that out, you would want to produce an inverse of that. But 1khz is 13.5" long. So if you moved your head even three or four inches, the null will no longer be out-of-phase.

Does that make sense?

Go to Gedlee.com and check out some of his patents. IIRC, when Geddes worked at Ford he made a bandpass loudspeaker which was positioned next to the muffler of the car. The bandpass sub would play out-of-phase with the muffler to cancel out the noise generated by it. Here is one of his noise cancellation patents (there are more listed on his site.)

http://www.google.com/patents/about?id=RLInAAAAEBAJ&dq=5119902

There are also cars which take the OPPOSITE approach, and use speakers to augment the sound of the muffler.

I am not a fan of sound deadening because it's all but useless, especially at low frequencies. What you are proposing might be effective because it works best at low frequencies, which is exactly where sound deadening is LEAST effective.

 

·
Registered
Joined
·
77 Posts
Discussion Starter #3
yeah thats exactly what im thinking. I know high to mids wont work. Like nc headphones. They cut the low freq rumble and drone but no good for human voices and chatter.

Basically, i was tuning the other night and worked so hard on the lows, but when the engines running, road noise, bumps in the road (the bass head next to me in traffic) constantly changes my low end response and i thought "man, i wish i could have a 'subroutine' constantly firing off phase inversions for these freqs and squash them on the fly. Let the system mix and feed it right into my signal.
 

·
Premium Member
Joined
·
6,045 Posts
yeah thats exactly what im thinking. I know high to mids wont work. Like nc headphones. They cut the low freq rumble and drone but no good for human voices and chatter.

Basically, i was tuning the other night and worked so hard on the lows, but when the engines running, road noise, bumps in the road (the bass head next to me in traffic) constantly changes my low end response and i thought "man, i wish i could have a 'subroutine' constantly firing off phase inversions for these freqs and squash them on the fly. Let the system mix and feed it right into my signal.
Instead of mixing it into the signal chain, you'll likely get better results by putting another speaker near the things that are making the noise (like Ford did with the muffler.)

The reason for this is that it preserves your dynamic range.

It's an interesting problem though. For instance, you learn to appreciate how much noise a tire can generate, because of it's large surface area.

(Instead of using sound deadening in my car I simply replace the tires - the idea is to kill noise at the source, not mask it with sound deadening.)

An ounce of prevention is better than a pound of cure, especially when it comes to noise problems.
 

·
Registered
Joined
·
5,234 Posts
The Acura TL-S comes with this. You can't hear a difference over the non ANC models and when adding aftermarket subs you have to unplug the mics or you get a horrible rumble. This particular application fails miserably.
 

·
Registered
Joined
·
77 Posts
Discussion Starter #6
Okay so i guess people are sort of utilizing the idea. But i cant help to think the Acura thing is more of a gimmick, i mean im sure it works to some extent but not what im thinking. And if you were to change ANYTHING out of the acura it wouldnt work because its tuned to that particular setup. youd have to rerun whatever the factory is doing again. know what i mean?

So Im going to do some tests and talk with a few engineers around here and see if im not chasing a ghost. Because right now my biggest problem is how to not cancel out the inverse frequencies that are being generated. I dont want to cause a time continuum loop problem here and end up in the old west.

The time delay between picking up a sound, processing it and having it sync with source is iffy too. a lot of variables in the time domain with tight tolerances that its going to take a lot of horsepower and patience to get it right. maybe pigs will fly by the workshop someday.

Thanks for the input guys.
 

·
Premium Member
Joined
·
6,045 Posts
Okay so i guess people are sort of utilizing the idea. But i cant help to think the Acura thing is more of a gimmick, i mean im sure it works to some extent but not what im thinking. And if you were to change ANYTHING out of the acura it wouldnt work because its tuned to that particular setup. youd have to rerun whatever the factory is doing again. know what i mean?

So Im going to do some tests and talk with a few engineers around here and see if im not chasing a ghost. Because right now my biggest problem is how to not cancel out the inverse frequencies that are being generated. I dont want to cause a time continuum loop problem here and end up in the old west.

The time delay between picking up a sound, processing it and having it sync with source is iffy too. a lot of variables in the time domain with tight tolerances that its going to take a lot of horsepower and patience to get it right. maybe pigs will fly by the workshop someday.

Thanks for the input guys.
Ah, good point. Hadn't considered the latency.

A quick review of the math:

Let's say you want to null out 100hz. To do it properly, you need to "inject" the reversed waveform within a quarter of a wavelength at the most. So that gives you 2.5ms to do it. With a 1ghz cpu that gives you, what, 2500 clock cycles?

That's not too bad. Of course when you move up to 1khz that drops to 250 clock cycles.

And all of this ignores the entire overhead of the operating system :D

 

·
Registered
Joined
·
204 Posts
The 5th Generation Honda Prelude had an Acoustic feedback system. The Front speakers were the main stage and the rear 6x9s were bandpassed midbass woofers at 2ohms each wired in series so it was mono. They has small microphones on the magnet and a feedback box that actively equalised the woofers to compensate for noise. Very good system.
 

·
Registered
Joined
·
1 Posts
You can try Road Noise Cancellation, it absorbs noise and can also prevent them from passing from road to residential areas. Specifically, the most commonly used acoustic solutions, as well as the cheapest residential soundproof barrier products, are noise barriers. These blocking road noise products are sold by standard size blankets or rolls, making the transformation of existing structures easy and cost effective.
 
1 - 9 of 9 Posts
Top