Can someone explain dB increase by doubling power?

I may be wrong, but I think that adding a second sub would only increase output by 3 db. The other poster may have been mistaken.

If two subwoofers are co-located then the gains increase. The maximum output increase after co-locating two identical subwoofers will be 6 dB's due to coupling.

If two subwoofers are spaced apart then coupling between them starts to weaken and output increases are not as significant. Usually 3-4 dB's across the band. Until, of course, frequencies match the 1/4 wavelength distance that the subwoofers are spaced apart, you should see full maximum coupling and so should expect a 6 dB output gain.

The reason for this is because you get positive reinforcement when the wavelengths interact at 1/4 wavelength distance between two sources. The levels, frequency and phase also needs to be set identically to reap the full benefit of coupling.

Regards,

no.

[quote]
Can someone explain this to me?
[quote]

Someone is misunderstanding something.

Maybe, maybe not.

One mega-sub might have more output that several smaller subs.

Just depend on the numbers each produces and whether or not coupling comes into play.

The idea that two speakers will pick up additional *gain* if co-located is based on the fact that two speakers co-located will become far more directional because they make up one larger speaker.

The counterpoint is that the law of conservation of energy always holds. Two speakers receiving the same amount of energy will always radiate the same total amount of energy or perhaps less if there is cancellation. Never more.

However, if two co-located speakers are more directional than one, the intensity of sound in the preferred direction will increase. The intensity of sound in the non-preferred directions will experience a corresponding decrease. The law of conservation of energy is thus obeyed.

Subwoofers are a special case because they are generally non-directional. The drivers are so much smaller than a wavelength at the highest frequency they handle that they remain essentially non-directional. So, if you put two of them into a room, their joint directionality will remain negligable. There will be no preferred direction. There will be no non-preferred direction.

An obvious exception would be a subwoofer that is directional such as a fully horn-loaded subwoofer. Cardioid subwoofers might be another exception.




As soon as you start getting wavelength-related effects, response of course becomes very non-flat.

---------------------------

I'm reminded that speakers are basically antennas, and this is becoming more apparent as we start working with arrays of them.

We think of antennas as having "gain" but of course they can't have any overall gain if they are passive. What antennas do commonly have is directionality, and they obtain that property we call gain by the same means as loudspeakers - by means of directionality. If you measure the power radiated by a so-called high gain antenna in all possible directions, there is no over-all gain. Instead we see the power being concentrated in one direction at the expense of all the rest.

That isn't necessarily true. In some cases response can be made flatter. If one achieves positive reinforcement across the band and the subwoofer in question is deficient in a particular frequency or across a given bandwidth then co-location (or acoustic coupling) can yield a flatter in-room result due to more overall output within those frequency bands and/or bandwidth. Of course the opposite can occur and there might be an overabundance of output in a particular frequency range but as I said, it is not necessarily true that the response becomes "very non-flat".

Regards,

Ugh...

The explainations didn't cut straight to the point.

Doubling surface area of driver(add a second) will add 3dB to the theoretical maximum output.

Doubling the power(adding second sub with it's own amp) will add 3dB to the theoretical maximum output.

These together give you a theretical 6dB increased possible power output.

So either by itself or together do not automatically increase output. They increase the maximum potential output.

Since a good deal of the benefit of adding a second subwoofer comes from putting it somewhere else than the first one, to address a different set of room modes, the co-location max of +6 dB would not usually happen. Dolby and THX each investigated this question wrt to home theaters, and found empricially that a second sub adds 4.5 dB SPL if driven the same as the first.


All else being equal, yes. Harman/Floyd Toole recommend 4 subs for best taming of room modes (when placed appropriately).

Multiple subs in a room (all subs at same SPL and phase) will tend behave similar to a single sub located midway between them. The modal response of the room will nearly approximate the response expected if a single sub were located in this midway location. Individual adjustments of gain can "move" this logical location. If you change phase on individual subs, you can now create constructive/deconstructive interference patterns to alter modal response in seating locations. By applying individual parametric EQ, one can go further in smoothing bass response. (This is not a simple process.) I have not found any "automagic" room calibration systems which do well with multiple subs.

The Welti multiple sub research is often misinterpreted. This approach does not create smooth bass response over a seating area (multiple seats in multiple row). It does however create consistent response over a wide area. This consistent response characteristic benefits calibration efforts (EQ and treatments).

Nice analogy Arnold. I think traducer, which they are, but thinking fields and phases, emitters make a lot of sense.

How many subs are recommended? The answer seems to depend on if it comes from a bass trap manufacturer or a speaker manufacturer.

When I put my 2 subs together and moved them to a corner, the apparent level went DOWN even with the change from 1 pi to 1/2 pi. As my DCX is still in my lab, I can't time/phase match. I preferred them as mains stands.

He might be talking about the power giving to the sub???

Meaning

1W = the sensitivity rating of the sub..lets say 87dB @ 1m
2W = 90dB
4W = 93dB
8W = 96dB
16W=99dB
32W=102dB
64W=105dB
128W=108dB
256W=111dB

Remember this is at 1 meter.

Add a second sub at the same location, at the same watts and you add maybe 6dB (its all over the map in measurements for some reason) to the chart above.

Depends on what research you are talking about. Todd optimized for average frequency response flatness in "How Many Subwoofers are Enough," Welti, AES Convention Paper 5602, May 2002*. He optimized for minimum seat-to-seat variation in "In-Room Low Frequency Optimization," Welti and Devantier, AES Convention Paper 5942, October 2003.

Regards,
Terry

* Correction -- Nope, I'm wrong here. Todd took the standard devations from the spacially averaged response, which is the same as assuming that there will be equalization to this average. Don't know how many times I have read this incorrectly before!

If the rather significant benefits of the Welti/Toole recommendation of placing 4 subs at the midpoint of the room's 4 walls can be largely attained by one sub in the center of the room, I wonder why this has not become more widely recommended (other than the Dick Van Dyke effect--tripping over the darned thing because it's in the way). You stated here that >>With only one sub, you will have problems regardless of what you do.<< Am I mixing different issues?

While one could place a single sub in the center of the room get a real close approximation of 4 subs (see document I emailed to you), the single sub isn't practical:
1. The presumption is all four subs are on the floor. A sub in the center of the room on the floor is impractical (and creates problems...like trip hazards). When I use multiple subs around the room, they are not all at floor level meaning a single sub would need to be suspended (between the projector and the screen).
2. A single sub, during the calibration process, doesn't provide the same capabilities as multiple subs to further calibrate to the extent available with multiple subs.
3. Many times it's easier to reach reference with multiple subs than with one megamonster sub when space is at a premium (not too many are tolerant of using the basement or attic as an 'infinite' baffle (ala Tom N).

That only happens if you apply an addtional amout of power to the second driver, equal to the power applied to the first. If you fail to do that, the second driver is unpowered, and adds no output at all.

The idea that a second driver adds 3 dB more power *presumes* that is is receiving the same amount of power as the first.


That was presumed by saying that adding the second driver adds 3 dB more power to the room.


You're double-counting the amplifier power.


The do if you get really crazy and actually hook the second speaker up to an amplfier, which of course everybody does, and which is presumed up front.


That is true.

Mathematically, a single sub placed in the center of a rectangular room (up in the air) also excites only 1/8 of all possible room modes. That means 7/8 of the room modes have been neutralized and therefore can't hurt you!

- Terry

I think the real question was somewhat glossed over or overly complicated.

Fact: Two identical acoustic sources which are equal in level at the point of observation will combine to a level 6dB louder than either individual source.

The "close coupling" others have referred makes for coherent (matching level and phase) summation over a wide frequency range; typically below some frequency. When further apart two sources can still coherently add to +6dB, but this will be frequency dependent. The talk of 3-5dB is simply an average over a bandwidth with varying constructive and destructive summation.

Let's add 3rd example to 1 & 2 above:

3. With two subs, reduce

total power by 1/2 (which is 1/4 to each subwoofer) to produce the same level (100w -> 50w) --> 0dB change. This represents an increase in efficiency that comes from 2x the cone area and 2x the box size (drivers & box need to be the same). In this case #3 you only need 50W total (25W per sub) to produce the same SPL as one. If 100W is now delivered to each, we have a total of 200W, which is 4x the power.

The calculation of dB change per power change is 10*log(P2/P1) = change in dB.
10x the power is exactly 10dB. 2x the power is approximately 3dB.

If you are adding multiple subwoofers of the same type, you have an efficiency increase (3dB) from driver size/box volume, as well as an increase (3dB) from total power increase (each sub is powered separately). This results in the 6dB increase in example 2 above.

The calculation of maximum dB change per identical subwoofer addition is 20*log(N2/N1) = change in dB.
10x the subwoofers is exactly 20dB. 2x the number of subwoofers is approximately 6dB.

To have a doubling of efficiency, or +3dB in sensitivity, at all frequencies you need to double the drivers and double the enclosure. Doubling just the cone area (similar to making a 12" driver 18") requires 2x the motor(magnet/VC/etc.) strength to move it in the same manner as 2 smaller drivers which each have their own magnet, VC, etc.

This does start to break down at higher frequencies, or more accurately, the gains become very location dependent when the size/span of the driver or drivers is ~1/2 wavelength (ie >34" above 200Hz).

This stuff is way above my pay grade as of yet. I will be having 4 subs in a 2200 cubic foot sealed room. Would my best bet as a starting point be to put them at the midway points on the four walls?

Would I want some/all of them elevated artificially beyond the feet they come with?

Where is a good place to get the basic structure and work flow of how I would then calibrate the four subs? I will have the SVS AS-EQ1 (maybe 2 of them as each one is capable of calibrating dual subs. So wouldn't I need two? I am assuming that you keep the unit connected between your pre amp and the subs at all times, not just for calibration...) I will also have a pre amp with audyssey, and my understaning is to just bypass the sub calibration in the pre amp via audyssey and calibrate my subs with the SVS AS-EQ1. I will aslo have a SPL meter. But I am not sure when and if I would use that. Maybe after the SVS AS-EQ1 does its thing? I am new at audio calibration as you can tell.

I am looking for a good workflow and/or info? Thanks

This is just to answer the "Multiple subs is better than one big sub" for output.

You have smaller subs that have a max output of 100db.
You have a large sub with a max output of 125 db

2 Small subs brings you to 106db
4 small subs brings you to 112db
8 small subs bring you to 118db
16 small subs bring you to 124db

The point is to gain the 6DB output you need to "double" the amount of subs. Adding one from 2-3 is not a 6db output gain.

placement is up to you...move them around the room to find out what works the best.

To further add:
Placing a sub very close to you (V near field) makes intensity very sensitive to distance. Hence if you move your head few centimeters, your bass response can change several DBs... (not practicle)Also, it will be more critical to place the sub at Equidistance between ears since at those short distances, the difference in sub intensity between our 2 ears is very high and Lows becomes SURPISINGLY localizable.... Trust me, tried that in my concrete room.

I don't think even harmonics gets cancelled.... Only odd ones.... That is, if the cube room is 4m in dimensions, and you place the sub in mid air, u ll cancel 43Hz/129Hz/215Hz axial modes...but 86Hz/172Hz/258Hz... will be @maximum (unless you place your head @1/4th or 3/4th distance)

Hence in harmon paper, he said optimal placement for 4 sub (neglecting height axial mod to be equilized) is on 1/4 - 3/4 wall distances(lenght and width). the 1/4 is used to cancel the odd axial modes harmonics , and the 3/4 opposite position is used so both speakers cancels each other even axial modes harmonics.

Here's where the number 1/8 comes from.

For one dimension, say, lengthwise, a subwoofer placed in the middle cannot excite any room modes with odd lengthwise modal components (1, 3, 5, etc.). All of these have pressure nulls in the middle of the room. Push all you want with pressure from the subwoofer. These modes will just lay there like loxes. Only the even modes have pressure maxima here, and will therefore be excited. And I am talking all three flavors of modes -- axial, tangential, and oblique. This eliminates 1/2 of the room modes.

Now add another dimension, widthwise. The mid-room placed subwoofer cannot excite any of the remaining modes with odd widthwise components. That means we are down to exciting 1/2 * 1/2 = 1/4 of the total possible room modes.

Add the third dimension, height, and ... you guessed it. Only 1/2 of the remaining modes can be excited. Which means 1/2 * 1/2 * 1/2 = 1/8.

- Terry

Agreed.


I had to read this several times, to understand it . And now, not sure I actually did. I mean even if you cancel a mode in one dimension, there will still be energy at this frequency going from the driver to walls in the other dimensions and waves will superimpose and results add up. Unless you assume walls are perfectly reflective and energy @ this particular frequency is all sucked out from the first axial mode. right?



nop, you can't multiply since you mentioned a rectangular room that is random room dimensions ratio. Let take a paritular example of a cube. With sub at center, you took care of half the total room modes.

Sound intensity decreases expotentially with distance. This is an often overlooked issue with small rooms and many seats which results in a perceptable (annoying) seat to seat SPL variation. Even at that, it is highly unlikely low frequency localization can be achieved due to very close proximity to a subwoofer. The 6" delta between the left/right ear would not create enough of an intensity difference. Where subwoofer localization has been reported, that localization can be attributed to the higher frequency artifacts from the sub as a result of its operation. Aiming the subs drivers away from the seating locations or placing a 1" piece of fiberglass (say OC703) in front of the driver eliminates these localization cues.