Seperate drivers for LF and ULF or one driver?

Different, not better. Pro drivers trade off response below 30Hz for sensitivity above 30 Hz because pro drivers need to go loud and they don't need to go low.

Fitz has it right. Different is the correct term.

2-way subs have been suggested in the past from Hsu's MBM on up to Velo's flagship 18+12 2-way sub.

A contemporary comparison is that it would take 10 of the JBL drivers to equal the output of a single Mach5 UXL-18 or LMS Ultra-18 at 20 Hz, but only 2 of the UXL/LMS to equal a single JBL (if the manny's claims are accurate) above 40 Hz.

The question is; is it worth the effort to overcome the crossover woes you'll experience at such a low crossover point vs just using a "full range" UXL-type driver. (search old articles regarding the crossover problems and see Slarti's test results of the Velo flagship at AV Talk)

After having tried it in several iterations over the years, my answer is no. You can easily test this for yourself, should you proceed with the experiment. Just get the system to its ideal in your room, then shut the JBLs off and run the system 'full range' and tweak to its ideal and compare the results.

It costs less, takes up less than 1/2 the floor space, is simpler to calibrate and gives more accurate results. Makes it a no-brainer, IMO.

Yeah, sensitivity is almost always a tradeoff with bass performance, per Hoffman's Iron Law. Heavier cones will actually move the response towards the bass region, but will also lower sensitivity. A stronger motor coupled with a cone that can stand up to the extra force (as compared to a weaker motor) will help with distortion but the stronger motor moves the response up and away from bass. And also there's a tradeoff with enclosure size (which is another part of Hoffman's Iron Law).

IMO, a 3-way tower with a high Xmax 7" (like my Adire Extremis or CSS SDX7 or whatever) or bigger (AE TD10/TD12 maybe) is ideal when used with a subwoofer. Well, the TD stuff might not even need the sub depending on your requirements.

I've been wanting to build a new 3-way tower with like 4 Peerless SLS per tower for the low-end... That would be a monster of a slim speaker, I think, and fairly more efficient than my Adires (which are something like 86dB/1w).

It's a bad idea. There is no other input necessary. If you prefer to attempt it any way, have at it.

The troubling part here is the "how it sounds at 60 Hz" thing. Bass is bass. There is no 'sounds better' or 'sounds bad' at 60 Hz.

What we hear is frequency response. If the response is non-linear (not flat), then your perception will be influenced accordingly. In the case of the PB Ultra vs the JBL, unless you're using a much larger power plant to drive the JBL and pushing the Ultra past what it can give, there is no audible difference at 60 Hz. The Ultra gives less than 5% THD at 60 Hz at maximum long sine sweep, which is as good as it gets.

Personally, IMO, JBL gets the biggest pass of any product ever discussed on these boards. I don't buy any of their published specs and would love for someone to send one of these particle board boxes to Josh for the skinny on the "extremely low distortion" blurb.

If you want to claim a difference in SQ at 60 Hz between the JBL and the PB Ultra, you're gonna hafta attach some data to that claim. If you match the FR (filter the low end out of the Ultra with equivalent order filter) and match the level, there will be no difference in SQ at 60 Hz. If there is, one of them is adding something that doesn't belong, and we know from test results from Ilkka, Slarti and Josh that it isn't the PB Ultra.

On to the question posed, which is that of a 2-way subwoofer system. Any time you use a crossover, you have phase issues (GD) that affect FR. Can you post a FR before EQ smoothing, focussed on the crossover region of the 2 subs? What exactly is the crossover scheme (point, orders, method)?

It's really pretty basic: The LMS Ultra is a perfect example. In order to get more displacement (throw) from a driver, you need a longer voice coil. The longer voice coil means heavier moving mass, which lowers Fs. Perfect, just make the VC longer, right? When you do that the sensitivity drops and the top end rolls off from high inductance. No problem, just add more motor and sleeve the pole with copper. When you add motor, Qes drops. When you sleeve the pole with enough copper to flatten the top end, you need even more motor.

There you go; the LMS Ultra. Huuuuuge motor, low Fs, low Qes and 90dB sensitive. Oh, and a $1k price tag. The JBL evades the need for a massive copper sleeve by offering no displacement. Instead they have a little copper ring centered in the gap. The JBL keeps its sensitivity up by requiring far less motor, it has a high Fs and low Qes. So, why does the JBL cost as much as the LMS? You got me. Why is your JBL said to be tuned to 25 Hz when its drivers Fs is 35 Hz and its box volume is 8 cubes? Because it really isn't tuned to 25 Hz. What is its THD at 60 Hz with maximum slow sine sweep? I would love to know, but I bet a beer it ain't <5%.

So, if you'd care to post the complete picture, I'd be glad to check back into this thread and add my 2 cents, FWIW. From the LP, post the JBL alone, then the PB Ultra alone, then the combination of the 2, all of the above with no smoothing and no post smoothing EQ. Then list the details of the crossover you used for the experiment.

Bosso, Josh did do a test on the OP's exact JBL system.

http://www.data-bass.com/data?page=system&id=68

Thank you Scott!

As I suspected, the little 13" Ultra turns in lower distortion performance at 60 Hz than the 18" uber driver from JBL. Although there is a slightly different harmonic by component profile, I seriously doubt that has anything to do with sonic signature at playback levels that don't evoke maximum distortion using program source that completely masks what distortion there may be. Scrolling to the FR of the 2 units it should be a no-brainer as to why they might sound different with actual program source in a typical home listening space.



Looking at long term sine sweeps vs THD vs compression vs the difference in FR provides overwhelming evidence as to what might explain a difference in sonic sig. As far as the "FR doesn't show everything" argument goes, I need more than that.

This, of course, is a separate issue from a 2-way sub. My opinion leans toward the 'to each his own' philosophy regarding perceptions and preferences. But, that doesn't mean I believe in magic dust-motors that 'sound better' inexplicably, because they don't. I've been around pro sound stuff as a musician since the 70s. As a bass player I'm aware that the typical bottom is open E, 40 Hz, but that bass amps and the E string itself produce 2H (80 Hz) at equal or louder than 40 Hz. There's no exception seen in the JBL.

Some here believe that if you pull down the JBL response 'where it lives', above 40 Hz, that you'll somehow have a better version of the PB ultra because of 'headroom' or some other baloney reason. Some others believe that you can have the JBLs performance, but with 20 mm of throw and that's gonna be somehow better yet... more baloney.

If you don't have the measurements or prefer not to repeat them, I understand and it's no problem, but you asked why there wasn't more interest in the thread and that's my simple answer. You have to have the data or it's just an opinion column. Most here don't have an opinion about 2-way subs, but might decide to join in if there's meat on the table.

Apparently, AV Talk doesn't display Slarti's graphs anymore and I'm not that interested to log in and find out why, but when he tested the Velo flagship 2-way sub, he had this to say:


Obviously, the cancellation at cross can be 'fixed' in-room with the on-board EQ, but that doesn't change the fact that the problem exists.

I'll bite, as right now I am in the process of doing something very similar to what you are wanting to do. I am trying to incorporate a pair of jbl 4648's with a pair of Resident Engineering XXX18's. I started to incorporate both together and then decided to stop as I only have one of my 4648's in my possession right now, and all the tweaking, eqing, and delay settings would be voided once I get the second one into the system. My plan as the 4648 have a tuning of 35hz was to run them down to either 40 or 50 and up to around 120hz where the mains would take over, while having the RE go from 40-50 on down. Sounds good in theory, I know, but getting them all to play right with each other is a BIA!!! I quit trying after a while and just put that project on the back burner until I get the other JBL cab in the theater. Im beginning to see along the lines of what bosso and fitz were telling you already, and I got early indications of that, but a pair of the 4648's was way cheaper than another pair of 18's, but now it is looking more and more like another pair of 18's will be just way easier to incorporate. My entire reason for doing this was to help the RE's out where they are very inefficient which is up past 40-50hz. while my response graph when I had it set up was at it's best didnt look TOO terribly bad, there was still some pretty serious integration issues.

I'm telling you guys.... just run them all full range. Don't try and 'cross them over'. If the cabs are vented, plug em. Measure. Enjoy.

I own a pair of 2242 and have them mounted in their own 12 cu ft enclosures tuned to 22 hertz. Here is my build thread:
http://www.avsforum.com/t/1363719/jbl-2242h-build-thread


I also own a pair of Avalanche 18s mounted in 23 cu ft sonotubes, tuned to 13 hertz (I don't have a build thread for them though).

I much prefer the sound of the 2242 when compared to the Avalanches. They sound better to me and I can hear more bass definition with them. There is also that mid-bass punch that the 2242 have that the Avas don't. In terms of low-end output, yes, the Avas have more but the 2242 are no slouch down low either. I was actually surprised at how much low-end the JBLs have.

Here is another point of reference, take a look at post #30 in this thread:
http://www.avsforum.com/t/1328971/need-help-jbl-2242h-with-a-slotted-port

There are no measurement graphs here, just opinion on what people are hearing when comparing pro-subs with long-throw drivers.

As everyone already mentioned, mixing a subwoofer and a woofer together can be troublesome due to phase cancellations. If you have the time, space, funds, and the patience to tweak the system then go for it. But I would look into drivers that are in the same league in terms of sound quality as the 2242 but with more low-end output. Try modeling the JBL w15gti and the Aurasound NS-18 in sealed enclosures and see if there is enough output down low for you. Both of these drivers are known for their sound quality.

"The troubling part here is the "how it sounds at 60 Hz" thing. Bass is bass. There is no 'sounds better' or 'sounds bad' at 60 Hz."

the problem is that is not what folks are hearing. there is more to it than frequency response and thd.

This is great stuff.

FR/THD is indeed what "folks are hearing". What the minority of folks may 'prefer' is irrelevant and most certainly does not give evidence for some mystery (non-existent) ingredient.

Take the LMS, filter it and add 2HD at 20 Hz and tilt the response upward and voila, you have a pro sound 'sound'.



Have another look at the FR difference and tell me it's not that, but really it's magic dust at 60 Hz that no one on earth can explain.

Sheeeeeeesh.

So how do you guys feel about the idea of a mid bass module, such as the one that HSU offers?

Did you try to HP your SVS at 25hz and see how it sounds above 40hz? If the SVS is playing normally then what you are hearing are the longer decay rate of the lower frequencies from the SVS. Once they are tuned the same you can compare. The only time I can really tell a difference between great subs are when one is compressing and or distorting and the other can play higher levels. This is what is usually happening at these GTG's, the more powerful sub from 20hz and up usually wins but once they are played within all subs limits things change. The cheaper subs now score much closer to the expensive ones. Of course the frequencies are never exact either.

I'll repeat here what I said in the Master List thread. I recently compared a full BW system to a 20 Hz tuned ported sub. Same location, same level and FR calibration. The FR at the LP is shown. A scene from WOTW was repeated using each system and the results are shown in the animation.

The sonic signature is, of course, very different. As MKT says, and as I've posted about many times, a single wavelength at 4 Hz is over 282 feet and exists for 250 ms. A single wavelength at 20 Hz is 56 feet and exists for 50 ms. The 4 Hz part of the effect has bounced back and forth off the walls of my room 10 times before it leaves the sub. So, of course, the higher tuned sub will sound as though it's tracking dynamics (start/stop) better because the higher freqs simply decay in 1/5 the time or less.

As far as auditory masking or any other such irrelevant phenomena, in the comparison, one is producing nothing while the other is, and even in the one that is there is nothing to mask the lower frequency decay because it lasts longer than any higher frequency component of the effect. Thus, if that part of the effect stands alone, there is nothing to mask it.

You've shown a FR at the LP that covers 10 Hz and up, assuming accurate hardware and proper setup for the measurement, this is a very commendable result, but as I mentioned before, the details of how that result was arrived at are everything to this discussion:

After the crossover filters but before the post smoothing EQ, we need to see the subs measured separately, as measurements for any 2-way system are always presented. I find it very difficult to believe the result is a flat line across a 30 Hz cross point given the results shown in Josh's site of the PB Ultra vs the JBL. There's a 10dB disparity at 30 Hz before any filters are applied. Of course, it would help to know what mode the PB Ultra is in? Once you've laid the before traces on the graph, a trace of both playing together on the same graph should be added, then the final response with what was done to the input signal to affect that final response

Dynamic tracking is an interesting point. My first question would be what source? If it's modern music, most of it is so compressed that there are no dynamics to track. Since all of the "what folks are hearing" and "what folks think" citings are devoid of any mention that this ot that source presented a bigger or less difference, I tend to dismiss dynamic tracking. I also tend to dismiss the dynamic tracking issue (assuming dynamic source is used for the comparison) because it is a product almost entirely dependent on available power and not the drivers ability to respond. Can you tell us what amplifier you use with the JBL?

I see absolutely no reason to believe that the PB Ultra can't tack dynamics at 60 Hz as well as a pro sound woofer can. Mms, sensitivity or any other physical property notwithstanding. Of course, as Josh's tests reveal, unless you're driving both subs to 130dB.

This is very basic stuff that anyone who has the ability to produce a FR at the LP can include with the subjective comments, but is curiously missing from these discussions.

My thoughts have been, if you truly are doing ULF properly (reference capable headroom 10Hz or lower), the power you have on hand combined with the number of UXL's/LMS's/etc... means that you are going to have reference capable headroom at 40-80Hz.

If you aren't doing ULF properly yet, add more drivers/amps and voila, your midbass performance increases.

Not needing to be said, if you have reference capable headroom, bass is going to sound good at both 20Hz and 80Hz.

Question: is there a documented experience of notnyt running his setup with his JBL's hitting midbass alone vs combined vs letting the LMS's do everything up to 80Hz?

[quote name="bossobass" url="/t/1422068/seperate-drivers-for-lf-and-ulf-or-one-driver/0_50#post_22265590"... a single wavelength at 4 Hz is over 282 feet and exists for 250 ms. A single wavelength at 20 Hz is 56 feet and exists for 50 ms. The 4 Hz part of the effect has bounced back and forth off the walls of my room 10 times before it leaves the sub.[/quote]

That strikes me as a peculiar way of thinking about it.

How can a 282-ft. wave bounce back and forth in a small room?

This is the pressure region; as the cone moves at low freq the pressure rises in the room in proportion to its displacement, as allowed by boundary flexibility and leaks.

The wave hits a barrier (wall, ceiling, floor) and is reflected off that barrier. Pretty simple. What's peculiar about it?

A sound wave is a pressure wave that moves through the air in a room. The rooms pressure doesn't change.

Someone please sticky this^^

It really is as simple as that.

Smooth bass above the fundamental resonance requires multiple bass sources with varying location, phase, and/or amplitude.

Logistically that's easier to accommodate with smaller boxes perhaps containing smaller drivers.

Of course, small boxes dictate less efficiency if they come with extension (you loose 9dB/octave) and will be severely excursion limited for home theater if allowed to reproduce ULF by the mechanical (a small box which forces early roll-off can do the trick) and electrical filters you've applied.

You're looking at different transfer functions (high-pass behavior and distortion). The pro-sound early roll-off may avoid gain below your room's fundamental resonance so it sounds faster.

No, although design constraints on driver and enclosure size which are sufficient for reference level second octave bass and beyond may be woefully inadequate for lower frequencies.
Edited by Drew Eckhardt - 7/31/12 at 3:18pm

I've always been under the impression that below a certain frequency, which is dependent on the room, sound stops behaving as a wave and starts behaving like a pressure change.
It sounds to me like you are saying that changing the volume of an enclosure, which is exactly what a subwoofer does to a room, doesn't change the pressure.

Don't these charts of Ricci's show that at 63 Hz, the JBL was over 10 dB higher with less distortion than the PB13-Ultra?

If that's always been your impression, someone misled you.

'Exactly' how does a subwoofer change the volume of a room?

Subwoofers cause a pressure wave to move through the air in a room. That pressure wave reflects off of hard surfaces (the law of reflection).

A pressure zone is near a reflective surface where the reflected wave reinforces the next wave (see; pressure zone microphone).

The barometric pressure in the room does not change because of those phenomena. I'm surprised at how this notion of a subwoofer 'pressurizing the room' has become accepted as a fact.

It seems pretty clear to me that the volume of the room could be said to change by (Sd * xmax) at a rate of f times/second. The absurd example being if one entire wall of the room was constructed as the cone of a driver, moving in and out and obviously changing the room's volume in the process.

While we're being absurd, let f approach zero and wave propagation stops mattering. Set f at one cycle per hour and put everything in an ideal perfectly sealed room, and indeed the salient result is that the barometric pressure rises and falls.

I don't know at what point (if ever) this model becomes a useful way of thinking about ULF in audio, but it's possibly worth discussing.