Can Subs also act as AVAAs?

[-Matt-]

Conventional room treatment often requires large volumes of absorption material which many of us would rather not have in our living rooms.

An alternative is active bass absorption, via AVAA devices like this one:
https://www.psiaudio.swiss/avaa-c20-active-bass-trap/
But these are currently rather niche and expensive.

If I understand correctly (and I haven't looked into it yet) the AVAA works in a similar manner to noise cancellation that is now common in headphones. I.e. Microphones in the headphones measure the ambient noise and an antiphase version of this noise is played back over the top of any music signal.

Many of us already have multiple subwoofers in our living rooms, so it makes me wonder, could those subs perform dual duty as subs and AVAAs?

The idea would be to have a small box of electronics (maybe raspberry pi based) and a microphone, for each sub. The device would have inputs for the microphone and for the signal sent to the subwoofer by the AVR. The output of the device would then be an AVAA modified version of the subwoofer signal from the AVR which would actually be sent to the sub.

I.e.:

[lfe] \
        [Processing Device] -> [Sub]
[mic] /

What do you think?

Could such a device have any hope of working?

 

[kipman725]

The mic has to be located at the sub that is doing the cancelation. PSI use an analog signal chain for low delay.

 

[-Matt-]

The mic has to be located at the sub that is doing the cancelation. PSI use an analog signal chain for low delay.

Yes, of course. I'm imagining one device and (at least) one mic, per sub. The mics would be attached to each sub, (measuring the ambient sound at each sub position).

You are right to draw attention to the latency, this is likely very important in this application.

 

[-Matt-]

Maybe a shortcut could be to cannibalize the electronics out of a pair of noise cancelling headphones?

 

[DVDdoug]

Noise canceling headphones work because the mic is on the outside where it picks-up noise to be canceled without picking-up sound from the headphone itself (on the inside) that you don't want canceled.

 

[voodooless]

You actually would not need a permanent microphone. Whatever you want to cancel can be derived from the original signal. Once you know the transfer function of that, you can just have a DSP apply that to the 2nd sub, and it will work just fine.

You can even do it without measuring: things like double bass arrays can be used largely with the room dimensions as the only input. Though also these should benefit from post measurement tweaking.

 

[-Matt-]

You actually would not need a permanent microphone. Whatever you want to cancel can be derived from the original signal. Once you know the transfer function of that, you can just have a DSP apply that to the 2nd sub, and it will work just fine.

You can even do it without measuring: things like double bass arrays can be used largely with the room dimensions as the only input. Though also these should benefit from post measurement tweaking.

This is not quite what I was suggesting.

The idea is that real time measurement of the ambient bass noise in the room (which isn't part of the original signal) is made at each sub, and the antiphase noise is superposed over the desired sub signal.

Noise canceling headphones work because the mic is on the outside where it picks-up noise to be canceled without picking-up sound from the headphone itself (on the inside) that you don't want canceled.

Yes, in this case the microphone would obviously also pick up what is played by the sub. So the intended sub output would have to be subtracted from the mic measurement in order to determine what the ambient noise is.

 

[voodooless]

This is not quite what I was suggesting.

The idea is that real time measurement of the ambient bass noise in the room (which isn't part of the original signal) is made at each sub, and the antiphase noise is superposed over the desired sub signal.

I understand. The point is just that you don’t need to do this in real-time.

 

[-Matt-]

I understand. The point is just that you don’t need to do this in real-time.

Well, "need" doesn't really come into this hobby much!

The point is, I already have multiple subs, optimised with MSO, using a MiniDSP 2x4HD. So what you are talking about (a fixed transfer function) is already done. Solved. Not what I'm talking about here!

The next step improvement could be either room treatment or AVAA to improve bass decay times. Hence the interest in potentially using the existing subs to perhaps also act as AVAAs just by adding some electronics inline between the MiniDSP and the subs.

Perhaps you are not familiar with the concept of AVAAs? (Or perhaps I have misunderstood them, I only heard about them recently myself). Since they are ACTIVE noise cancellation, I believe it is absolutely a mandatory part of their function that they are real-time and use real-time measurements.

 

[-Matt-]

With real-time noise cancellation...

When the dog lying on the floor next to the subwoofer manages to produce a 50Hz tone via its flatulence, this sound should be cancelled as it isn't part of the intended bass signal!

 

[voodooless]

Well, "need" doesn't really come into this hobby much!

The point is, I already have multiple subs, optimised with MSO, using a MiniDSP 2x4HD. So what you are talking about (a fixed transfer function) is already done. Solved. Not what I'm talking about here!

Not quite. What you do with MSO isn’t really the same as measuring the response at the sub positions and based on that create a transfer function for the sub at that spot.

Perhaps you are not familiar with the concept of AVAAs? (I only heard about them recently myself). Since they are ACTIVE noise cancellation, I believe it is absolutely a mandatory part of their function that they are real-time and use real-time measurements.

No, it doesn’t need to be real-time. You know upfront what you want to cancel, so there is no need to measure it constantly.

There are obviously advantages for the real-time processing: you don’t need an audio signal to your sub. A power socket is all you need

There is some more info here:

Active basstraps

Do people here experiment with active basstraps? An active basstrap is a subwoofer placed in a corner. A microphone a bit before the woofer records incoming bass frequencies, these get send to the woofer with a time delay equal to the distance mic - woofer and played back in anti phase. Result...

www.diyaudio.com

DIY the device of US Patent 4,899,387

This expired patent is entitled "Active Low Frequency Acoustic Resonance Suppressor". It was granted on Feb. 6 1990 to Threshold Corporation; Nelson S. Pass inventor. His inventive device eliminates standing waves [SW] in rectangular music listenning rooms. Thus it enhances the accuracy...

www.diyaudio.com

 

[voodooless]

With real-time noise cancellation...

That isn’t really a perfect analogy. What is noise, what is not? How does the system know with only a microphone? A noice canceling headphone has a reference that it needs to approximate, these AVAA’s done have this. That makes me very suspect about their performance already.

 

[-Matt-]

In this case I'd define noise as anything measured on the microphone that isn't part of the intended bass signal (for that sub). Therefore the noise that the dog produced (which couldn't be determined from the original signal) should be cancelled.

 

[voodooless]

In this case I'd define noise as anything measured on the microphone that isn't part of the intended bass signal (for that sub). Therefore the noise that the dog produced (which couldn't be determined from the original signal) should be cancelled.

So, that’s not how the PSI device works. It doesn’t have any audio inputs other than the microphone.

 

[-Matt-]

Thanks for the links above...

It seems that there are two uses of the term active in this context.

Active as in powered subs (but with a fixed transfer function). Like in a double bass array.

Active as in, based on real-time measurement of ambient noise. Like in an AVAA.

 

[-Matt-]

So, that’s not how the PSI device works. It doesn’t have any audio inputs other than the microphone.

Yes, but that is because it isn't trying to be a sub at the same time. That is the new idea here.

 

[voodooless]

Yes, but that is because it isn't trying to be a sub at the same time. That is the new idea here.

Without an idea of what to cancel, I don’t see these things working properly. Maybe someone else can elaborate on this?

 

[-Matt-]

Without an idea of what to cancel, I don’t see these things working properly. Maybe someone else can elaborate on this?

My idea is that you calculate this based on the intended bass signal and the microphone measured signal.

Logically, subtracting the intended bass signal from the microphone measurement leaves the ambient noise.

You then calculate the antiphase version of the ambient noise.

You then add this antiphase ambient noise signal to the intended bass signal and send this to the sub.

The calculations would be very simple in the digital domain. The main question is if this can be done with low enough latency.

 

[voodooless]

My idea is that you calculate this based on the intended bass signal and the microphone measured signal.

Logically, subtracting the intended bass signal from the microphone measurement leaves the ambient noise.

You then calculate the antiphase version of the ambient noise.

You then add this antiphase ambient noise signal to the intended bass signal and send this to the sub.

Agreed, this would work for very long waves in your room to also work at the listening position. I wonder what the effective bandwidth would be for this.

The calculations would be very simple in the digital domain.

This should also be easy in the a analog domain.

The main question is if this can be done with low enough latency.

I’d say to effectively work, I’d guesstimate you’d need the latency to be no bigger than about 1/5 of the maximum frequency. So at 100 Hz, it would be 2 ms. Doing this digitally would be hard. Even 10ms would be hard.

 

[-Matt-]

FPGA maybe?

Yes, I think you'd certainly put in place filters so that it only responds in the low frequency range.

I bet there are now integrated chips that do real-time noise cancellation processing for cheap noise cancelling headphones.