Multiple Small Subs - Geddes Approach

[Wayne Parham]

The problem is floor bounce usually occurs at frequencies where wavelength is somewhere between 5 feet and 10 feet, which is why the self-interference problem happens when the speaker is raised off the floor a couple feet. It's a 1/2 wave cancellation notch. So the radiator has to be spread out over a couple feet in the vertical dimension, and even the largest direct radiators are a little small for that. Better to have a couple of them offset in the vertical.

[soongsc]

Originally posted by markus76
Send me a soundfile and I'll show you that your idea of low frequency perception is completely wrong.
The drum in your example shows a transient part when the stick hits the drumhead. The frequency of that sound is nowhere near the low frequency range. That makes the drum localizable. The low frequency part of the drum sound takes some time to build up after the first hit.

Best, Markus

First of all, I don't think it's legal to just copy files like that.
Second, the point is exactly that the impact makes the drum localizable, however, the following waves are also localizable in live situation. But when room modes are excited, the the following drum rumble cannot be localized. This is a listening experience totally dependent on preference. How can you use data to prove what another person hears or cannot hear?

[markus76]

Are you saying that multiple subs is wrong or is it right or what are you trying to say?

The "kick" of a typical bass drum is at 2000 - 5000 Hz, not in the low frequency region.

[soongsc]

I am not saying that anything is wrong or right, I'm just explaining what difference you will expect to hear. You decide what is right or wrong for yourself.

[markus76]

Difference of multiple subs and what else?

[soongsc]

Listening to the actual instruments.

[inertial]

Hi Wayne,

I understand your point but agreed only partial.
If you are familiar with room's simulators ( yes you have said this previously) you sure know that floor is not the only issue.
There are six walls, you can virtually add one at time and see the difference in your listening position. A caos!
When you have added all the six walls, the influence of the only floor
is not so rilevant .
I am referring at european rooms, not very large, very solid walls and
listen triangle about 2 metres.
In this scenario, i always see peaks and dips in the midbass zone and basic measurement semms comfirmate this. Problems in the 80-300Hz zone. ( of course there are big problem also below 80 Hz)
It is easy listenable with normally a lack of punch, no body, no weight , very bad sound .
This is a very typicall performance in our "small" european rooms ( to say 20-25 square meters), not sure how much different can be in yours americans rooms .
I know simulators makes big simplifications respect real conditions, but something is interesting to my eyes .
So the question remain IMHO, who takes care of the 80-300Hz?

Cheers,
Paolo

[jan.didden]

Originally posted by soongsc

First of all, I don't think it's legal to just copy files like that.
Second, the point is exactly that the impact makes the drum localizable, however, the following waves are also localizable in live situation. But when room modes are excited, the the following drum rumble cannot be localized. This is a listening experience totally dependent on preference. How can you use data to prove what another person hears or cannot hear?

The transient is a high-freq component and will not go through the sub anyway. The transient will go to the high freq driver. The transient will not excite the room modes.

Jan Didden

[markus76]

Originally posted by soongsc
Listening to the actual instruments.

Now I know where your misunderstanding comes from. A loudspeaker is not an instrument! It can't recreate the original soundfield at your ears. If you want that, then look for dummy head recordings or wave field synthesis. A speaker for stereo or multichannel sound reproduction in a home listening room just needs to reproduce the original, which is NOT the original venue but what was heard in the control room where the recording was created.

Best, Markus

[badman]

Originally posted by Wayne Parham
The problem is floor bounce usually occurs at frequencies where wavelength is somewhere between 5 feet and 10 feet, which is why the self-interference problem happens when the speaker is raised off the floor a couple feet. It's a 1/2 wave cancellation notch. So the radiator has to be spread out over a couple feet in the vertical dimension, and even the largest direct radiators are a little small for that. Better to have a couple of them offset in the vertical.

Oh certainly, for the primary notch that's true. Perhaps I should have been more clear: "Floor BOUNCES". While it's most apparent at one frequency, there are effects at multiples thereof. So if your primary bounce is at 150Hz, then at 300Hz, 450Hz on could expect an issue as well, and here, a half wavelenth can fall within the variance of size between a 15" and a 6.5".

Oh, and while the vertical dimension spread is valuable to combat floor bounce when the subs are close to the speakers, in the multisub setups we're discussing it's going to be spread via the lateral/depth variations fairly effectively.

One interesting technique I've played with is a "bed" of acoustic foam under a speaker, cutting out some of the "down and forward" energy that contributes to floor bounce. Obviously this is only effective when thick, but can really help clean up the mids, when done appropriately. It also typically looks quite horrible.