The Effect of Parametric EQ (RABOS) and Audyssey MultEQ on Freq Resp w/ 2 subs

I'm not surprised you didn't get any feedback on your extremely thorough posting. I have a pair of Infinity SW-12 SWs and are using both RABOS and Audyssey. I'm very happy with the results for music. It seems for movies, there's so much noise it overwhelms the bass these subs are able to produce.

I haven't found anyone who is interested in RABOS or Infinity subs. They all seem to be busy with the other more popular brands. Do you have any updates on your CSW-10s?

There are two ways of approaching RABOS for two subwoofers.

1. Apply RABOS to each subwoofer separately
2. Apply the same RABOS settings to both subwoofers when playing simultaneously

I've seen arguments for both, but because multiple subwoofers can interact in different ways, depending on location, #2 may make more sense.

However, sticking with #1, like you did, it looks to me like you may wish to widen the Q slightly and shift the center frequency downward significantly, unless you REALLY like the peak in the 30s Hz and the dip around 56 Hz. I think you want the parametric EQ curve wide enough to cover both peaks in the curve for the left speaker, which will likely leave you with a dip around 40 Hz, which will probably be filled in by the right speaker to some extent. For the right speaker, shifting the center frequency downward should help with that big 30 Hz peak, too.

Looks like the RABOS calculator is back up. Could you let us know how things turn out using that?

Strange eq'ing by those systems. You could do much better manually.

Rabos can work really well. I used it with 4 subs with very good results.

Couple of suggestions. Make sure the mic is in EXACTLY the same spot
every time you do your measurements. Even a slight difference can make
a big difference.

I found that using rabos for both subs at the same time worked the best. You can even measure from a couple of different positions and average it out.

Next, use the Infinity website. They have a feature that you can plug in
your numbers and it will tell you where to set your dials. This was very helpful. I have done it both ways and the website was always better.

Because multiple bass sources can interact in different ways, it may not be that simple. For example, suppose you had a rectangular room and put one subwoofer in the middle of the left wall. This would drive a set of room modes (standing waves) corresponding to the width of the room. A listener in the middle of the room would experience peaks and nulls in the bass response because of this. Put a second subwoofer in the middle of the right wall, and this would cancel out some of the room modes but reinforce others. This example doesn't correlate well to your case, since your room is irregularly shaped and you wouldn't be equalizing up the nulls, anyway, but I hope you get the gist.


Like I said, I might consider the two peaks in the left subwoofer as one broad one (albeit one shaped like a camelback) and set the Q and center frequency accordingly. Once corrected, you might have a depression that the other subwoofer might fill in. However, again, you might consider using the same RABOS settings for both subwoofers after you get measurements when they're both playing simultaneously.


Yes, if you get their co-locate their drivers as much as possible, you should expect a significant increase in output, though less than the theoretical +6dB. The problem is that you may have more uneven bass than when they're located away from each other (for example, if they cancel out standing waves like in the example above, some of the energy goes into the mode cancellation), but then that's what the parametric equalization is supposed to help address, to a limited extent.

Unless you're optimizing for only one listener in only one listening position, a little spatial averaging is likely to be helpful.

The procedure you described in your response to swgiust sounds just about right.

Hi hellerbrewing,

I wonder if you might PM PL Hart and ask him to join this discussion to give us his opinion. I've seen him on AVS Forum in a couple of Infinity speaker forums. He designed the CSW-10 and I believe is Audysssey certified or some such thing as well. I'm just not sure if he's still around. Just a thought.

Attached are some jpgs of my room dimensions and layout (not to scale, but it should give you an idea); a picture of my front / center speaker layout; and a copy of my RABOS results.

For the RABOS one, the black line is the measurement I took from my main listening position starting with the -10dB setting. I left both Subs on at equal volume levels when I did the measurement. I used the manual method of adjusting the subs for the 40-43Hz spike. I tried using the online tool, but it had me adjust the spike at 60Hz for some reason, and afterwards, I still had boomy bass. I much prefer the sound I'm getting after using he manual method.

Anyway, after I applied the RABOS settings to both subs, adjusted them with a meter so that they were at the same volume, I did the Adyssey eq using 8 spots around my main listening (per the Audyssey guide on the Audyssey forum). The pink like is the measurement I made with both RABOS settings and Audyssey, both subs on, again from my main listening spot.

I really like the way my bass sounds now. Unfortunately, I still have a dip.

Another weird thing about my room, is from my main listening spot, the bass sounded better with only the right SW on with the left SW off that it did with both of them on.
I can only deduce that I'm sitting in a null. When I had both SWs on, I moved around the room, and the bass got better in the corners and in other places. Just not where I sit.

So now what do I do? I tried one sub in the back of the room, and didn't like that. Then, I decided the use the left sub as a 'mid-bass' sub. I connected the left and right front speaker outputs from the AVR to the High Level In connectors on the SW, and connected the High Level Out from the SW to the L and R front Beta 20s, and set the SW crossover to 150.

All my speakers are set to small, the AVR crossover is set to 80Hz, and the right SW is connected to the LFE AVR output. In theory, I think, the right SW should be playing <80Hz (<120Hz for the LFE), the left SW should be >80Hz and <150Hz, and the main left and right speakers should be >150Hz. After I re-wired the front speakers, I re-ran Audyssey, and well, it works. I did do a little volume tweaking with each sub for personal taste. All I need to do now is re-run RABOS and see what the measurements would be. Maybe I'll do that this weekend.

It's bizarre, but then so am I.

There's not much more I can do with my room. Unfortunately, there's no space for traps, plus my wife is all ready to sign my commitment papers for the guys with the net. She cringes every time I talk about getting a super huber sub woofer.

The good news is that we're thinking of getting a lot and building our next house. I would be fine with a 6 car garage, a living unit with all the basics needed for domiciling, and a dedicated home theater designed as it should be. The rest is all fluff and woo. Needless to say, I am quietly ignored.

Hellerbrewing et al-

The lengths and the half lengths of a subwoofer’s bass frequencies will always approximate the various height(s), widths and lengths of the room(s) into which they work. For instance, to calculate the frequencies at which your 20.5’ room length will come into play divide 1130 ft/sec (the speed of sound) /20.5’ = 55Hz. The half wavelength, 10.25’ therefore is 110 Hz which puts this frequency past the high frequency roll-off of the sub’s high pass filter and , for the purposes of subwoofer calibration, out of play.

In a sealed rectangular room it is the ceiling height, the smallest room dimension, which will usually cause the largest bass peak. Thus, for an 8’ ceiling > 1130/ 8 = 141.25Hz , half wavelength is 70.5 Hz. In this case the half-wavelength frequency peak at 70.5 Hz is the one which needs to be RABOS’ed.

The problem with a non-sealed (dedicated HT) room like yours is that subwoofer is now seeing all the attached cubic footage, so to speak, of all (the irregular sized) attached areas. So the problem of being able to actually pre-calculate or pre-estimate where a room peak might be becomes virtually impossible. This is where the manual measurement-and-plot system of the RABOS becomes invaluable.

A second at least as important issue is your theater area(s) total cubic volume versus the size and output capability of a subwoofer. You’re expecting a lot of a CSW-10, even with it’s 650 Watt amp driving its sealed 10”er to be able to keep up its end on woofing when you’re driving seven 6 ½ “ two-ways to 100 SPL (combined) in a 2500 cubic foot room. It ain’t gonna be able to keep up! Assuming your ceiling is 8’ and solid (drop ceilings completely screw up the RABOS measurement during calibration because the ceiling panels move from the bass wave pressure) just your listening room alone would be approximately 20.5’ x 12.5’ x 8’ = 2050 cubic feet. Add the area for your office and I’m guessing you’re working into an apparent ~2500 cubic foot room.

Look up the Robinson-Dadson equal loudness contours and it becomes easy to understand why a subwoofer must be positioned against a wall or better yet into a corner to achieve the extra low-end boost required to sound equally as loud as the normal program material. For instance, if playing your normal war movie at the Academy standard 85dB at 1000Hz the subwoofer must be capable of hitting 110dB at 25Hz to sound equally as loud. The CSW-10’s max free air SPL at 1 meter from 30Hz to 100Hz is 104.5dB.

Now, if you take the CSW-10 and place it on the floor you get an automatic (theoretical) 6dB of “free bass gain”. Move the sub into the junction of the floor and a side wall and you get another free 6dB. Finally, move the CSW-10 into your right front corner and the boundary gain is now theoretically 18dB. Now, theory aside, since it is impossible to actually put the sub completely into the corner, shoving it as close as possible to the two room corner walls while keeping it on the floor should yield around +14 dB free (actual) bass gain. Therefore, +14dB boost from a single CSW-10 in your right front corner will give you 85dB + 14dB = 99dB best case. Therefore, you’re under-woofed to play your main program material at a 85dB level while still producing undistorted clean bass from the subwoofer.

This is why I always recommend a minimum of two subs in a room of your size. I would recommend placing your second sub in your right rear corner to get the highest potential corner-placement-driven boundary gain.

Equalization: Since there is only one true .1 channel and since you have to use a Y connector going from two subs to your receiver anyway that is how I recommend you perform the final dual sub RABOS measurement. Initially, I do perform a preliminary RABOS run and calibration of each sub separately with the mike at the money seat just to see if either of the subs happens to be in some sort of a weird null with regards to the money seat. I have actually never seen this situation exist with corner mounted subs since by definition a corner mounted sub drives the whole room in the most efficient (+18dB theoretical) manner of which it is capable. The initial individual sub RABOS calibration and plotted measurements will give you good indications for the peak height, peak width and overall sub level a single sub will be capable of producing when measured from the money seat. (I always call the front sub A and the rear or side sub B and give preference to A’s calibration flatness firs).

Once you determine that each sub (individually) in your right front and right rear corners work without any kind of extreme dips (nulls) in responses hook them up via the Y connector and do the combined subs RABOS calibration with Audyssey switched out.
The last time I did this in my system it took me 11 trials and about two hours before I was satisfied with the relative combined bass response of my two subs. I was satisfied only after I had been able mess with each sub’s phase switch then finally by rotating my individual subs corner positions about their vertical axis’ relative to the money seat the measuring and plotting using the RABOS CD.

Audyssey’s calibration limits for the bass region are ±9dB. I have never done a single or two sub RABOS system wherein I had more than a ±3dB variation from 30Hz to 100Hz with RABOS before I let Audyssey loose to do its thing. (A ±3dB variation is total variation window is huge, 6dB.

Be aware of what a huge benefit two subs are versus a single. With two subs you get double the cubic air volume displacement (acoustic gain) potential for a +3dB gain. Plus by using the Y connector going into a receiver’s single sub input your halve the electrical input impedance for another +3dB of (electronic gain) headroom. The improvement of two subs over one always seems like way more than double, because it is.

Punchy bass versus tubby bass: The 60-80Hz frequency range is where you get punchy bass. 30-40Hz bass gives the “slower” bass-“fill” effect to the room. After all, a 40 cycle per second note IS only half as “fast” as an 80 cycle per second note. Lastly, the more you take advantage of side wall or corner subwoofer placement the greater extreme low frequency (~15-30Hz region) room gain effect you’re likely to hear at the money seat.

Have fun guys,

Patrick Hart

I would suggest taking out all drop ceiling panels except the ones which have a light fixture or other intrusions such as HVAC vents attached prior to running and fully calibrating your RABOS sub(s). When you are satisfied with the smoothness and consistency of bass response from your money seat replace all ceiling panels and run Audyssey.

Corner mounting will produce the highest peak, say 14dB as in my original example of any place in the room. Moving the CSW-10 diagonally toward the center of the room will attenuate the 14dB peak. It will attenuate pretty quickly too because you're losing the 3-fold boundary reinforcement provided by the intersecting walls and floor.

A physics course will refer to the area under a subwoofer's frequency response curve as the potential amount of work any given system is theoretically capable of producing. That amount of "work potential" is fixed. In simplified terms work potential is determined by the subwoofer's motor structure, the size of the cabinet, and the amount of power (in watts) in the amplifier. So when you move the sub out of the corner because it's too bassy that tells me you have not used the RABOS calibration system correctly in order to tame the single largest room peak.

Think of a Corvette with a 400HP engine but also with 600 lb-ft of torque that comes on just off idle at 1500 RPM. Now put skinney 4" wide tires on the
Vette and tromp on the throttle....

The 600lb-ft of torque at 1500 RPM will shred those tires as the Vette claws to get traction. So the fact that the Vette has 400HP @ 5000RPM is irrelevant because the only way you'll ever be able to get up to 5000 RPM is if you baby the Vette off the line. This is what you're doing when you move the CSW-10 away from the corner because the RABOS is not properly calibrated. You're wasting the entire system's potential because you are not locating it in the corner where its full potential can be realized.

Corner mounting is equivalent to putting 12" wide drag racing slicks on that high torque Corvette motor. With corner mounting you allow the whole CSW-10 system to hook up and blast down the drag strip with no wasteful wheel spin. A properly calibrated RABOS sub ( i.e. a flat response at the listening position where originally there was a peak) placed in a corner is the equivalent of giving your Vette the wide set of hook-me-up slicks.

No, you need to level match each one in the same position.

stepyourgameup is correct.

All levels for all speakers are always calibrated for level from the money seat. Place the mic at your ears' level, the subs into their respective corners, and adjust for level. Do this for each sub by itself initially, then together when you do the final calibration. You should be able to start with each CSW-10's level controls set at the 11-12 o'clock position.

The primary reason for a sub giving out its location are from an untamed room peak. Once you've got the room peak tamed you still need to use the low frequency steady state tones on the RABOS CD to rid the rattles from objects like glass vases on shelves etc. I use a clear, soft clay-like product called Quake Hold underneath my wife's tchotchkses in her display cabinet which is behind our couch in our living room/home theater. (Available at True Value Hardware stores.)

If after listening to a few movies you feel one sub overpowers the other in a non-realistic way feel free to mess with the levels until it seems correct.

Here's a hammer. I'll hold this nail. Now when I nod my head you hit it.

In order to truly level match your subs, you need to place one sub in one of final locations and check it's level. Then, move that sub out of the way and move the second sub in the same position as the first and level match that one. If you have the subs in their final position and then level match, they may not be the same because of a null or a peak caused by your room. You could potentially have one sub working harder than the other and that is bad.

Ah.....incorrect. Last sentence first "You could potentially have one sub working harder than the other and that is bad." "Working harder" is what a subwoofer is designed for. Working harder implies "within its design limits”. Working harder implies that the amplifier does not (or is not capable of) overdriving the bass transducer so hard that the voice coil goes beyond the driver’s ±15% Xmax excursion limit, at which time the amplifier is no longer able to control the transducer.

With a cheap, mismatched transducer/cabinet/port system or underpowered ported subs this limit can be reached before the amplifier starts to clip. This scenario is not a problem with the CSW-10. The California SubWoofer 10 inch (that’s what the three letters stand for) is a small as I could make it (using LinearX’s LEAP software). All Thiel-Small parameter’s which I specified for this 10 driver were geared to matching the transducer’s Xmax± limits to the same time the amp reached its 650-watt dynamic power limit. (Also, in play with this design was getting the overall package as small as possible while keeping the frequency response flat down to 30Hz. To hit that flat @ 30Hz performance figure we had to build +6dB of boost into the system at 27Hz.)

Using the clock-wise rotation of the CSW-10’s level control this Beta series system will maintain it’s controlled flat response up to 104.5dB in Harman’s large 4pi anechoic chamber at the 1:30-2PM rotation position of the level control. This is the position at which the CSW-10 amp will begin to clip and the THD will begin to rise.

Now back to the two subs in hellerbrewing’s room. Place the sub’s in their respective corners. Set the level controls of each sub at 12 o’clock and calibrate first one sub then the other from your listening position. If with the RABOS SPL level read at the listening position with only one sub operating can’t get up to the 75dB level called for in the RABOS manual (before hitting a 2 o'clock rotation) try to set the system up at a lower level, say, 70dB. Then do the full RABOS calibration.

Remember, adding a second sub Y-connected into your receiver will automatically give you a +6dB dual-sub in-room overall level increase once both subs are hooked together. Also remember that adding a second sub will tend to interact with both the first sub and the room, and, either add or subtract from the overall system’s response at the listening position. And that is all that matters >> AT THE LISTENING POSITION.

Now redo the RABOS calibration with both subs operational. You should be able to make minor overall tweaks in level, frequency and width to the sub nearest your listening position to get the flattest possible response. Once you’re happy with the frequency response, then and only then do you want to raise or lower both sub levels in unison.

Last note: It is only if you read a huge suck-out in frequency response at the listening position with both subs operating that you have to move either your sofa or one of the two subwoofers to rid your ears of sitting in a bass dip. Narrow band dips cannot (and should not) be equalized out under any circumstances.

I just wanted to thank Patrick for his highly informative posts here. I now realize I need to peform some futher calibrations on my two subs. One thing I've noticed is that I'm not satisfied with the RABOS adjustments recommended by the tool on the Infinity web site. Looking at the attached graph of my measurements, the black line (my RABOS frequency measurements. Again, the red line is after applying Audyssey.) indicates I have a peak to tame in the 40Hz - 43Hz range. However, when I plug the measurements into the Infinity tool, it tells me to use these adjustments:

The computed frequency: 66 Hz: 18 ``CLICKS`` from fully counter clockwise position.
The computed level for Level: -13.90 : 2 ``CLICKS`` from fully counter clockwise position.
The computed bandwidth: 39 : 14 ``CLICKS`` from fully counter clockwise position.

When I use the manual method, I obtain results around the 40Hz - 43Hz range instead of 66Hz. Unfortunately, I don't have the exact numbers with me at the moment.

The bottom line of this, is that the sub sounds better to me when I use the manual calculations. The sub isn't boomy, which is noticable with the Infinity settings.

I'm not sure if there is anything that you can comment about this. It's just an interesting observation I've made. And makes me wonder why Infinity calculated a 66 Hz frequency instead of one around the 40Hz - 43Hz range.

Why would anyone believe that relying on a calculation on a website would be more accurate, or even accurate at all in comparison to measurements which you yourself have taken of your own listening space? That's the double-click mentality at work. Trust yourself. Read and follow the RABOS procedure step by step. No skipping in between. Do that (skipping steps) and you're inviting your own unclear thinking to creep into the measurement process.

Alan Devantier, one of Harman's brilliant engineers who worked along with Dr. Floyd Toole at Harman's Northridge, CA facility was the inventor of RABOS. Alan proved out the correctness of his RABOS approach beforehand, through hundreds of RABOS runs in dozens of rooms. It was only after this test method was proven conclusively to be a viable approach that RABOS subwoofer were put into production.

If you're a member of AES or other recognized acoustic or electro-acoustic societies access to Alan's RABOS paper is easy. But it is not free. New technology by AES members is always put out among us members at our conventions. It is called peer review.

Thanks for your response Patrick. I'm glad to hear my common sense was right for a change. And as for peer reviews, I'm married to a PhD, so I know all about peer reviews. I get them all the time. They usually go something like this:

Wife: How much did you pay for those speakers? (Referring to the side and back channel speakers in a 7.1 system.)
Me: Why do you ask?
Wife: I can't hear them. All I can hear are the front speakers. I want to hear the other ones too.
Me: Well for music, you don't hear the sides and back like you do the fronts. They're only for presence and to surround you in sound. For movies, they help the sound track envelope you and they're for special effects.
Wife: If I'm not supposed to hear them, why did you need so many?

And on, and on it goes. We also have a similar conversation about sub woofers. Now, if I ask her about the protein channel she's been researching . . . well, I guess you get the picture.

Anyway, I've come to the conclusion there must be a surround sound gene. Some people have it, some people don't. I do, she doesn't.

Now, back to the topic of this thread.