Subwoofer Frequency Response question

I just started testing my frequency response with a Radio Shack SPL meter (digital). My subwoofer is a Paradigm PW-1100 (10 in) with Green Mountain Europa speakers. The room is small, 12 X 14 X 9. The frequency response is generally very good with bass response down to 25 Hz. My issue is I have a dropout of nearly 20 db starting at 45 hz through about 70 hz. I have not been able to change it by moving the sub or changing the settings (phase and crossover point). I also have the sub on an elevated platform of 1 1/4 inch MDF with spikes between sub and platform.

I am wondering if adding internal bracing and additional acoustic dampening material to the sub woofer would change the response characteristics in a positive way.

Any ideas would be appreciated.
You should check the phase setting or your wiring to be sure the speakers and the subs speaker cones travel in and out together ie. in phase. If they travel in opposite directions that would cause cancellation of the sound for those frequencies where the main speakers have enough output to cancel the subs output.
Many subs have irregular frequency response in the rooms they are used in.If the SUB doesn't have its' own amp,then you could try a "plate" amp with parametric EQ.Parts-Express has 500/1000W units with this feature.

Are you taking readings in hz on your meter? I have a Radio Shack digital SPL Meter (cat No. 33-2055) but it only reads out in dB's and not HZ. How do you know your sub is dropping out 20dB between 45 and 70 hz? I've been looking for a digital meter that can read out in hz. Can you explain? Sorry if I sound naive. Thanks.
Your going to need to get a better handle on whats going on, before you can attempt to fix the problems.

If your going to try the spl meter route then you will need a good cd, or a computer with test tones on it.

For free go here:

You can download the test tones and burn them, or use them direct from your computer (better, because you can loop them). These are better than most because they cover more range, one db at a time (you can read about them first). There are some other good tools at the site also...and plenty of good reading.

I usually start with the main speakers only, and then I move to the subwoofer only.

And then move on to some crossover settings, that I think will pull the two together, as one. Kind of a divide and conquer approach in the first two steps, or at least a better understanding of the problem.

The "Virtual Minirator" download is another good tool for your computer.

I suspect you need a lot of bass traps in that room, but hey...what room doesn't :-)

Thanks for all the suggestions. I am using the Rives CD with 1/3 octave tones calibrated for the RS Meter. I am going to check out the realtraps site. Using this, I was also able to find a wire vibrating inside the speaker at mid-range frequencies. A little Blu-tak and voila! No more vibration.

PDN - Measure test tones with the meter in dB then graph the test tones and you will determine what your response is in a given frequency range.

I took readings with and without the sub and with the subwoofer by itself. This really helped clarify what was going on (thanks Dave). I was clearly having phase cancellation. After tweaking the settings on the sub and taking readings at various settings, I was able to eliminate the dropout at 50 hz and get a fairly flat response. The exception is at 35 to 40 hz where I have about a 10 dB increase. This seems typical of many subs and may also be the room.

Would this be eliminated by stuffing more material into the sub or external bass traps.

A 10db peak is not uncommon at all, you could eq that down. Peaks are better than nulls, you can't eq a null up and out.

Some subwoofers (they call them hometheater subwoofers), do indeed have built-in bass peak, I don't know anything about your sub though.

At any could eq that out too.

What you described, a deep notch at the crossover frequency, is the classic symptom of drivers out of phase. In fact this is the best way to find out what the crossover frequency is. Invert one driver, and look for the notch.
Most bass traps I've seen are designed to reduce peaks at 80hz upward though some provide a little boost at 50ish hz (bass busters). You can eq nulls, but it's not easy for a 20 db problem, I'd guess you might get some improvement via PEq, but not a full fix. I would have seconded Eldartford's recommendation, but it seems from your post that you've already tried that with no success.

I use a Velodyne SMS-1 x-over, analyzer, and PEq unit that is fantastic, but in your case I'm not sure what will help.

It could simply be room modal - this will not go away no matter where you
place the sub. Your only hope would be to add plenty of bass traps and EQ
down the peaks to make the trough relatively less big.

It also may be caused by your speakers - as the bass is omnidirectional it will
reflect off the wall behind the speakers. You get quarter wave cancellation
across the entire room.

Are your speakers on when you make this test? Or is it just the sub playing?

Just in case - perhaps I should explain what you should do with speakers...

You should try to avoid middle ground - either place your speakers within 1 M of a wall or more than 2.2 M to help reduce this problem. there are formulas to work out at what frequency you will get cancellation - remember this applies to the ENTIRE room. I put my large full range speakers into a soffit mount - only this way you can completely eliminate the problem - that is what studios do. I leave my surround satellites within two feet from the wall (they roll off at 60 Hz).

I quote from Genelec
For large loudspeakers, with very
low cut-off frequency (Active 3-way and Large systems), placed away from
the wall the necessary distance is far too long for any practical rooms. The
conclusion is that large loudspeakers should not be placed free-standing but
flush-mounted to perform in the best possible way.

The cancellation frequency is Fc = C / 4 * D.

D= Distance from front baffle of speaker to rear wall behind speaker.
C= 344 Meters/sec
Fc is in Hz.

In your case a suckout at 45 Hz equates to a Distance of 1.9 meters. And 70
Hz equates to a distance of 1.2 Meters. If your Europa speakers are in
between these two limits then it could well explain your problem. Typically
this yields a 20 db suckout. Most people have this problem as they use large
full range free standing speakers. Since most people ignore it - even a 20 db
suckout - as bad as it seems - is it not likely to ruin your enjoyment of music.
Sometimes other room modes can compensate at the listening position.
Hopefully you can see that speaker position can be important too - not just
the sub.

That is one very handy bit of info. It implies that there will always be an Fc, which can be raised as we place the speaker closer to the rear wall and lowered as we move it away. However, it leads to 2 questions:

First question: If the cancellation frequency is -round numbers- 90 divided by Baffle to wall (in meters), does the size of the suckout change predictably as Fc changes. That is, if we move a speaker to 2m out from its original position 1m from the front wall and Fc drops from app 90hz to app 45hz, have we reduced the magnitude of the problem or merely shifted it downward? If the latter, then is the idea simply to move the speaker so far out (the recommended 5m equates to app 18hz)that the suckout is below the range of the speaker's output?

Conversely, is the idea behind <1m to keep the suckout in the same region as bass peaks in a typical room? Given the variation in peaks from room to room, this strikes me as quite a risky approach. (I see why you like the soffeted approach.)

Second question: Which baffle? In a Mains/Subs set up, you will likely have 2 different Baffle to front wall distances. Do you know how to determine which is relevant (or in what ratio relevance can be determined) to the calculation? Is the crossover point relevant in making this determination? Does staggering the Baffle to wall distance create 2 problems equal in magnitude or 2 problems smaller in magnitude than the original?

I guess that's more than 2 questions Thanks in advance.


does the size of the suckout change predictably as Fc changes.

The suckout will be less as you get further from the rear wall (fall off with distance - double the distance = 6 db SPL drop in the strength of the rear wall reflected wave).

Bear in mind that there is more than one suckout - it actually looks like a "comb filter"...each succesive dip as you go up in frequency is lower (this is because less and less energy goes backwards as you go up in frequency - and above 600 Hz most of the energy is going only forward - this is because the wavelengths at higher frequencies are becoming shorter than the speaker baffle)

I see why you like the soffeted approach

It is not me. I simply copy what they do in the highest end custom installations in pro studios. What I have done is far inferior to a proper setup designed by an acoustician, like reknowned Roger D'Arcy but it is still a lot better sounding than nothing at all. I share this info becuase I know from first hand how important it is - I have directly experienced the sonic improvement and it a lot more tangible than your common run-of-the-mill tweaks. (tiny cups, spikes, pads etc.)

In a Mains/Subs set up, you will likely have 2 different Baffle to front wall distances. Do you know how to determine which is relevant (or in what ratio relevance can be determined) to the calculation?

You need software to do this. It is much more complicated. However, the effect with a sub is much less critical because it is rolled off above 40 or 50 Hz anyway and room modes can really dominate at such low frequencies anyway. The best thing to do with a large sub is to get a downfiring one and place it up against a side wall. This means you eliminate most of the quarter wave problems by coupling to both floor and side wall. Since it is to one side then the distance from the sub to the rear wall behind your speaker will vary - say 2.5 meters to one corner but 8 meters to the other corner - this in itself will help reduce the coherence of any cancellation or "suckout" (it is unlike the case of two full range speakers lined up symmetrically to a rear wall). Two subs symmetrically placed close to freestanding speakers should be AVOIDED - this is contrary to the popular audio myth of stereo subs (audio is full of myths). Ideally you can also soffit mount the sub and place the woofer close to the floor but it is simply not as critical as the mains because it is normally not placed symmetrically, room modes dominate anyway, and it only handles the bottom octave (no nasty lower mid bass effects).
Conversely, is the idea behind <1m to keep the suckout in the same region as bass peaks in a typical room? Given the variation in peaks from room to room, this strikes me as quite a risky approach.

I see I missed this. The idea of going close to the back wall (less than 1 M) is to push the "suckouts" up higher above 100 Hz where they become less noticeable. As the wavelengths get shorter then the suckout will be progressively less dramatic...
considering good overall seating location plotting and speaker placement for good coupling with the room for flattest response (use your sound level meter and test discs, at very least, also your ears), you should start by issolating your sub from your listening seat(s). You can easily do this by placing your sub where your ears will be in the seating location, and moving around the room down where you are considering placing the subwoofer. Then, you can play some steady bass music through the sub(set crossover to full open), and listen for a solid sounding bass beat. Then, you will measure to confirm. If you then get a solid bass note from 25hz to whatever frequency your sub will be playing up to (I suggest an 80hz crossover as a foundation), you have found a good potential location for the sub! Then, if your mains are playing solid down to their crossover point (with no sub), now it's just a matter of dealing with blending the volume between the speakers and sub, and tweaking the phase on the sub! You should then have no problem getting a good solid response.
Don't sweat it, you can get if you have patience, and go through ONE VARIABLE at a time.
So again, start with issolating the speakers, then the sub, and then work on the level settings. Good luck