37Hz axial mode between side walls of my contol room

[Jens Eklund]

Originally Posted by tINY

Link?

I've read some information on this. In large rooms, it typically doesn't make enough improvement to worry about. In smaller rooms, it can help, but you typically make your room smaller with the probelms common to an even smaller room.

If you are designing dividing walls between rooms in a modest sized complex, angling the walls can help with axial ringing. If you lower the Q, the ringing will not persist as long.



-tINY

I tried to find these articles on the web without luck but if you have a copy of Master Handbook of Acoustics, they are used as reference in chapter 13.

[tINY]

Can you explain the mechanism or the reasoning...

I understand why it's not worth it in big rooms or by making a room smaller. I don't see why it won't help if you are building dividing walls between a control room and live room that aren't large.



-tINY

[DrFrankencopter]

Originally Posted by tINY

Can you explain the mechanism or the reasoning...

I understand why it's not worth it in big rooms or by making a room smaller. I don't see why it won't help if you are building dividing walls between a control room and live room that aren't large.



-tINY

The advantage of splayed walls is the high frequency diffusion, and easier reflection free zone. A room with splayed walls will still have modes; but they will be harder to predict.

Imagine a square sheet of metal (mass + spring system) pinned at the edges not unlike a plate reverb, subject to an impact. The pattern of resonances are predictable and fairly easily calculated. Now cut some of that material away to make it non rectangular, and there's still resonances developed...but the mode patterns become more complex.

Not a control room, but here's a nice page showing the mode shapes of an acoustic guitar: http://paws.kettering.edu/~drussell/...mmingbird.html

Cheers

Kris

[A2D]

Kris has got it right
if i remember this correctly to make a trap that will solve your problem with insulation it will need to 1/4 wave length deep

[tINY]

Originally Posted by DrFrankencopter

The advantage of splayed walls is the high frequency diffusion, and easier reflection free zone. A room with splayed walls will still have modes; but they will be harder to predict.

Imagine a square sheet of metal (mass + spring system) pinned at the edges not unlike a plate reverb, subject to an impact. The pattern of resonances are predictable and fairly easily calculated. Now cut some of that material away to make it non rectangular, and there's still resonances developed...but the mode patterns become more complex.

Not a control room, but here's a nice page showing the mode shapes of an acoustic guitar: http://paws.kettering.edu/~drussell/...mmingbird.html

Cheers

Kris

So, it doesn't lower the Q on axial modes between the two biggest walls because it's harder to calculate?

OK.



-tINY

[DrFrankencopter]

Originally Posted by tINY

So, it doesn't lower the Q on axial modes between the two biggest walls because it's harder to calculate?

OK.



-tINY

Okay lets approach this from another perspective. You have a mass + spring system (a room filled with air), and you change the dimensions. How does this affect the Q of the modes?

A: It doesn't, you need damping to make that happen.

Non rectangular dimensions don't add damping to the room, so they don't change the Q of a mode.

But, they can change the mode distribution...so you might end up with more or less modes stacked on top of one another at a particular frequency. This might be (incorrectly) perceived as a changing of the Q of a mode based on a spectrum analysis. But you could get the same effect (though maybe not the exact same distribution) by changing dimension in a rectangular (i.e. easy to calculate), or non-rectangular (computer solution required) manner.

Personally, I think splayed walls make more sense in a large room where diffusion would prove to be expensive. Modes will be less of an issue in these rooms since there will be so many of them overlapping in the critical mid band that a bit of broadband trapping will take care of most issues.

In a small room, rectangular dimensions in the planning phase can allow you to design with good room ratios for mode spacing, reducing the need for tuned traps.

Cheers

Kris

[tINY]

So, the Q of acoustic (axial) modes is not affected by how parallel they are?

If the walls were 18 feet apart at one end and 24 feet at the other end (lets say 26 feet down the center of the room), you'd expect the axial node to have the same characteristics in the 27 and 53 Hz regions as parallel walls about 21 feet apart?

I'd think that the resonance change from one end of the wall to the other would spread out the energy and lower the Q of the mode. Sure, it's hard to calculate with Excel or a simple on-line calculator (being that you have to do some integration)...



-tINY

[DrFrankencopter]

Originally Posted by tINY

So, the Q of acoustic (axial) modes is not affected by how parallel they are?

If the walls were 18 feet apart at one end and 24 feet at the other end (lets say 26 feet down the center of the room), you'd expect the axial node to have the same characteristics in the 27 and 53 Hz regions as parallel walls about 21 feet apart?

I'd think that the resonance change from one end of the wall to the other would spread out the energy and lower the Q of the mode. Sure, it's hard to calculate with Excel or a simple on-line calculator (being that you have to do some integration)...



-tINY

Nope...a new pattern of resonances will be developed. Look at that acoustic guitar structural model on the URL I posted earlier. I'm just saying that the modes will still be present, and strong, but less easily predicted by formulas. There is no pure lateral axial mode in a splayed room (assuming its the left & right walls that are splayed)...instead, it's more like a tangential mode.

Q is damping, and the only thing affecting damping in a room is the rigidity of the walls, air temp/pressure/humidity, and the presence of absorption. Why should changing room dimensions alter the Q? There's still the same amount of energy going into the room, and that energy is not being dissipated any differently (assuming the walls are the same material). The room will resonate...and needs damping to compensate.

Cheers

Kris

[jacko]

I brought the equipment in (desk, racks, chairs, etc.) and the situation improved a lot.

[DanDan]

It seems to me that splayed walls would cause travelling modes. Say the speaker is at the narrow end. It will pour energy into the room. Some of this will excite the width mode, which will proceed to head towards the back of the room. Now does the mode lose support from the walls as it travels back, or does it drop in frequency?
DD

[DrFrankencopter]

Originally Posted by DanDan

It seems to me that splayed walls would cause travelling modes. Say the speaker is at the narrow end. It will pour energy into the room. Some of this will excite the width mode, which will proceed to head towards the back of the room. Now does the mode lose support from the walls as it travels back, or does it drop in frequency?
DD

This thread has pretty decent coverage of modes in non-rectangular rooms.

Acoustics Forum • View topic - Non-Rectangular Room Project

I don;t have time to go into it suffice to say that the mode will not drop in frequency, and the overall shape will change (i.e. it won't be purely axial anymore).

Cheers

Kris

[tINY]

Originally Posted by jacko

I brought the equipment in (desk, racks, chairs, etc.) and the situation improved a lot.

That doesn't look too bad... except for the 50Hz buzz. Is that in your measurement mic, or acoustic noise in the room?



-tINY

[tINY]

Originally Posted by DrFrankencopter

This thread has pretty decent coverage of modes in non-rectangular rooms.

Acoustics Forum • View topic - Non-Rectangular Room Project

I don;t have time to go into it suffice to say that the mode will not drop in frequency, and the overall shape will change (i.e. it won't be purely axial anymore).

...and it will ring just as long?



-tINY

[DrFrankencopter]

Originally Posted by tINY

...and it will ring just as long?



-tINY

Downloaded a paper on Eigenmodes in nonrectangualr reverberation rooms in the journal of Noise control engineering, which sheds some light on this.

In the absence of traps & absorption it is the wall losses that result in damping of the modes.

In rectangular rooms damping of the axial, tangential, and oblique modes are all different/distinct. In non-rectangular rooms this is not necessarily the case since the modes are established between more than 2 surfaces.

As a result, the modes in a nonrectangular (3D-no walls parallel) room will tend toward a similar level of damping (some modes will be more damped vs a rectangular room, but others will be less damped). This is great for a reverberant room as you get a rich and full reverb field. Is this the best bet for a control room?...I'm not sure how well it would image for stereo.

Cheers

Kris

[tINY]

Well, symetry is key for imaging in stereo. That would dictate bilateral symetry for control rooms.

And, if you can make the room big enough, rectalinear is fine. Modal density is the key.

It's just that 20+ foot high ceilings can be a real hassle in some cases...




-tINY

[JohnPM]

Originally Posted by tINY

So, the Q of acoustic (axial) modes is not affected by how parallel they are?

The Q (or equivalently the decay time) of modes measured in the room would be affected, though its not really correct to say it is the Q of the axial modes as if one of those walls is angled those axial length modes would no longer form as the reflection path is no longer closed along the length axis. The modes with the longest decay times in a rectangular room with uniformly damped walls are the axial modes along the longest dimension - each time the wave encounters a surface, some energy is lost, the further apart the surfaces are the fewer losses of energy there are per second and the longer the corresponding mode takes to decay. Paths that involve additional surface collisions with shorter distances between collisions (e.g. tangential or oblique modes) have correspondingly higher rates of decay as there are more energy losses happening per second along the path.

[johndykstra]

now I'm confused.

I took "Q" to mean, the broadness of frequencies effected. A room with a single (rectangular) depth will emote a sharp note specific peak, while a room with a splayed surface of exponential depth will have more of a "hill" where a number of neighboring frequencies are supported.

[DrFrankencopter]

Originally Posted by dykstraster@gmai

now I'm confused.

I took "Q" to mean, the broadness of frequencies effected. A room with a single (rectangular) depth will emote a sharp note specific peak, while a room with a splayed surface of exponential depth will have more of a "hill" where a number of neighboring frequencies are supported.

A single mode has a peak resonance at a single frequency. How long it rings for is a function of its damping. Damping, when looked at in the frequency domain is called Q-factor (Q factor - Wikipedia, the free encyclopedia and also see: Damping ratio - Wikipedia, the free encyclopedia). Increased damping will shorten the time for a mode to decay, and will reduce the peak resonance...

Overall averaged frequency response of a room will be more affected by the mode spacing than the mode damping. If you have modes that are widely spaced (such as in small rooms) then you'll have some frequencies where there is no modal reinforcement (troughs), and others where there are too many modes clustered together (peaks). Your listening position WRT these modes comes into play too (of course).

Cheers

Kris

[JohnPM]

To put some formulas around that, which I'm sure is what everyone has been longing for the time to decay by 60dB is related to the bandwidth (in Hz) by T60 = 2.2/BW. The BW and the Q are related by BW=F/Q where F is the resonant frequency of the mode, so we can also write T60 = 2.2*Q/F. For modal resonances it is easier to use absolute bandwidths in Hz than Q figures, as the T60 and BW relationship is independent of the mode's frequency (e.g. any mode with a 2.2Hz bandwidth takes 1 second to decay by 60dB).