[3ll3d00d]

I came across this paper earlier - http://vbn.aau.dk/en/publications/l...rs(2961def0-3b73-11dc-b148-000ea68e967b).html - and, in particular, paper D inside there which is "Controlled Acoustically Bass System (CABS), A method to achieve uniform sound field distribution at low frequencies in rectangular rooms"

This is an approach that was implemented in a few places, physical devices I've come across are the the Klein & Hummel ARAM (www.klein-hummel.com/klein-hummel/globals.nsf/resources/o800_o800aram_studio2005_e.pdf/$File/o800_o800aram_studio2005_e.pdf) and the bagend etrap. In software, it has been implemented in Ultimate Equaliser (http://www.bodziosoftware.com.au/) & it also seems to be the same thing as a VBA (virtual bass array) in acourate.

I have a spare sub atm so thought I might try it out to attempt to dial out my 1st axial length mode (which is quite a big peak to say the least). I was just curious to know if anyone has tried this before, i.e. not a full DBA/SBA but just a tuned bass trap where the trap is a subwoofer. It seems like it will be prone to total failure but could be interesting to try.

 

[popalock]

I love your threads man... Will be following along for sure.

 

[Mfusick]

Very interesting. I've never done it but the theory seems kind. The application and execution would be the tricky part. How do you plan to adjust phase ?

 

[3ll3d00d]

Very interesting. I've never done it but the theory seems kind. The application and execution would be the tricky part. How do you plan to adjust phase ?

I'd probably need to create a filter to implement this then convolve that with the low pass of my crossover before applying the relevant delay/gain adjustment. This would then be loaded into the convolver as another set of paths through which each channel passes & set each path to output to a 2nd SW output (and connect that to another amp).

Thinking about it I have a spare channel on my main amp, & 2 spare on my audio device, which should work for testing it out at least.

 

[Mfusick]

It could work. I'd love to see your results of you do it.

 

[Bill Fitzmaurice]

An unused speaker working as a sonic sponge is a well known phenomena. When measuring speaker response there should never be an unused speaker in the room, as it will alter the result. Intentionally using one as a trap could be done, but the placement would affect the result, as would the tuning, making it impractical. That's why you don't see it done.

 

[Mfusick]

I think he meant actually driving the woofer.

 

[Bill Fitzmaurice]

I think he meant actually driving the woofer.

OK, I fully read the links this time, and yes, you can use additional polarity reversed subs to cancel out response peaks, using both EQ and phase adjustments to dial it all in. It's not a simple process, and flattening response at one room location can worsen it in another. But it can be done.

 

[popalock]

@Mark Seaton mentioned this awhile back. Basically a pitch and catch scenario.

I personally wouldn't go this route because to me it seems like a waste of firepower, but a very very cool concept none the less...

 

[3ll3d00d]

I'm hoping that it means I can use a weaker/smaller sub behind me just to tackle that length mode and hence it's not so much wasting firepower as using what I have more efficiently

I only need the output in the 30-50Hz range to do that (theoretically) which means a smaller driver/box might well be sufficient if I can get it dialled in properly.

On the downside I just measured up the existing box and it's too big to fit where it needs to go, seems I need to build a new box for it (which delay things a bit as a I need to build some speaker test boxes first)

 

[Mfusick]

I'm hoping that it means I can use a weaker/smaller sub behind me just to tackle that length mode and hence it's not so much wasting firepower as using what I have more efficiently

I only need the output in the 30-50Hz range to do that (theoretically) which means a smaller driver/box might well be sufficient if I can get it dialled in properly.

On the downside I just measured up the existing box and it's too big to fit where it needs to go, seems I need to build a new box for it (which delay things a bit as a I need to build some speaker test boxes first)

Yeah for sure. A $60 12" sub and a cheap $100 amp is probably enough. You don't need a thumping subwoofer. The issue you are talking about is not very low (judging by normal consumer sized rooms in homes) so you don't need a ton of extension, and you don't need to hit +++ over reference with the sub. You just need enough output at that frequency area you need to knock it down and give you a smooth response.

It's the same idea with smoothing subs. Often when you look at a lot of the professional theater designs like the ones by Erksine, they use a modest smoothing sub. 10" and a couple hundred watts. Nothing crazy.

 

[3ll3d00d]

I don't think we do cheap $60 subs in the UK, proper sub drivers don't come cheap over here

The driver I have on hand is an Fi SP4 15" so it can certainly handle a fair bit of power if that turns out to be required, bit of a waste for this exercise but I have it here anyway so will give it a blast.

 

[3ll3d00d]

I found some Infinity Reference 1260w for £60 a piece in the UK on amazon. 2 of them in a ~60L box models as follows, I've shown my current sub for comparison.

Lets assume, for the sake of argument, that I can calculate a filter that actually works. Any views on whether a sub like this be up to the task without colouring the overall sound in some way?

If it makes a difference, I'd be dropping the sub behind the sofa as shown by the red arrow

I have a spare channel on my amp with the power to feed that so the total cost would be pretty small for an experiment (£120 for drivers, £20 for MDF).

 

[Mfusick]

Yeah that will work I think. It's enougn to be a smooth sub or cancel an unwanted reflection.

 

[3ll3d00d]

Some discussion of the mechanics of creating the filter for the 2nd sub here - http://digitalroomcorrection.hk/http___www.digitalroomcorrection.hk_/Double_bass_array.html

this describes a VBA but I would think the same process would be used with an actual sub (i.e. use a low passed, inverted & delayed dirac pulse as the filter)

 

[3ll3d00d]

I knocked up a quick example following the instructions above

red = my room
light blue = the filter based on measured distance (looks like it misses slightly, need to play with the length)
green = the resulting, calculated, response

the impact is perhaps more obvious in the step response

If I can play with the position of the pulse so that it matches the shape of the existing IIR filter (that tackles that mode in the direct wave) then I guess there is some grounds for optimism that it will just work.

 

[3ll3d00d]

This is the VBA filter overlaid with my existing IIR, it clearly has a lower Q and hits the spot a bit lower so I guess the room is acoustically longer than my tape measure says it is (or my measuring skills are not exact)

 

[3ll3d00d]

It seems that 4.18 is the magic number, using that length (in metres) allows the peak of the VBA filter to line up perfectly with the room mode. It still has a different Q so need to investigate whether it is possible to alter that but certainly looks like it could do the job of squelching the effect of room modes on harmonic distortion.

In this graph

red = the room response
blue = the VBA filter
black = my existing IIR filter
green = room * IIR
cyan = room * VBA

 

[3ll3d00d]

It seems increasing the order of the crossover results in a higher Q filter, a 5th order BW is a v tight fit for the upper frequency slope for example (vba filter in blue, iir in black)

as an aside, anyone interested in this sort of DSP really should get acourate and learn how to use it (easier said than done but still). The number of things you can experiment with is just absurd, for example I was drilling into timing issues earlier and was pointed towards an approach that allows one to compare the alignment of two drivers not just in stready state but also during the transient phase of the response.

 

[LTD02]

what is the theoretical benefit to having the rear subwoofer produce the cancellation vs putting that same filter in the system before the front subs launch the wave? I could see a benefit if the wave were passing the listener kind of like a plane and then reflecting back--you want to hear the direct sound and then cancel the reflection from the rear. but at 40hz, the wave is large relative to the room, so it's not a plane wave moving past the listener, it's kind of enveloping, then not.