Are three inexpensive subs better than one expensive sub-Dr. Earl Geddes Approach - Page 61

i remember that thread

headroom will be limited to the sub with the highest level and wasting headroom on the subs that are using lower levels

Yes, that is level-matching. But you could also gain-match your subs. Gain-matching would be running the subs at identical gain (volume) settings without regard for each sub's individual contribution to their combined measured SPL at the listening position. Mark Seaton presented this as an option to level-matching and is (apparently) what he prefers. It assures that both subs are running identically, with the same headroom constraints (among other things). He (or someone else) pointed out that truly gain-matching a pair of identical subs requires measuring their output in an identical and as neutral a place as one has available and adjusting their output to be identical. This, as opposed to simply relying on their volume pots to be a completely accurate way to adjust their gains identically.



Well, goneten, for one, wasn't too fond of the gain-matching approach. He argued that level-matching was the proper way to incorporate 2 identical subs.


When you think about it, both methods, level-matching and gain-matching, make some sense.

So does setting the level for the smoothest response, which is almost never going to be the same gain or the same level, especially if they are not in the same place.

fair enough.

can i ask about integrating multiple subwoofer sources where the drivers are of very different construction?

as you know, the typical home audio subwoofer is comprised of relatively heavy cones while pro audio drivers such as those employed in the summas employ relatively light cones. as a result, many home sub drivers have a qes of around .4 or higher, while the mains are .3 or even lower. lower qes drivers are often subjectively described (for better or worse) as being "tighter" or "faster" than their high qes counterparts. since the multiple subwoofer strategy may be comprised of drivers with very different designs, how should we be thinking about integration in this sense?

...if you use "level-matching".

However, the "gain-matching" approach recommended by Mark Seaton avoids this. I essentially use the gain-matching technique with my JL's, by using the Master/Slave connection. The Master sets the gain for the Slave.

Mark recommends gain-matching because level matching at the listening position does not factor out the room's influence relative to the two sub's different locations. If one sub has a large peak not offset by a null, and the other sub has a large null not offset by a peak, the two sub's levels will be set quite differently, (obviously that is an oversimplified explanation, but it portrays the "concept").*

Gain-matching ensures that both subs are using their amps and their excursion to maximum benefit.

Dr. Geddes, how do *you* set the gains in the systems you set up?

Craig

Edit: *The better way to express this is... if one's sub's "average" level using a noise signal is higher than the other sub's average level, their gains will be set differently.

Right, I suppose. That would be THE third option. I realize that your method could be made to work with identical subs, but most people with 2 (or even more) identical subs and EQ (and maybe some room treatment) are going to level-match or gain-match their subs. I think that accommodating or approaching your method with, for example, 3 identical subs, would require some effort (not that it is impossible) to high-pass 2 of the subs properly.

This brings to mind the main question I think I have regarding "your" method. Setting different low-passes on 3 different subs is a piece of cake, as most subs are equipped with variable low-passes. But the nature and implementation of the high-passes on the 2 subs covering the upper frequencies was a bit vague to me. Is this very carefully and specifically designed into, for example, a pair of DIY subs, so that they have very specific roll-off characteristics? Or is their exact low-end roll-off and slope not so important and instead what is important is only that they have relatively high roll-offs, frequency-wise, preferably different, and that the method is molded or fitted around their characteristics no matter what they may happen to be.

To me, the high-passes on these 2 subs are a very important feature, or part, of "your" method. But it wasn't clear to me how these are approached, exactly. With over-the-counter-subs would you recommend trying to incorporate some sort of fixed high-passes such as what can be provided by in-line filters such as Fmods? Or would a more flexible method that allowed variable high-passes, and even slopes, be preferable. I understand that a feature of your approach is that you take what you have and you "force" it to work, but could you elaborate on the nature of the high-passes or roll-offs on those 2 subs, what characteristics would be preferable, there, and how best (and conveniently) to set about achieving those characteristics?

interesting cj, is there a post or link to a page where mark or somebody explains the "gain-matching" approach?

Thanks for the explanation. The only thing I don't get is, if you gain match instead of level match, then aren't you going to lose some bottom end headroom due to only one subwoofer reaching lower and louder than the other? For instance, one of my subwoofers is pretty close to my right speaker, and in the corner. The other subwoofer is a little further away from the left speaker than the other sub is from the right speaker, but there is no wall directly next to that sub, so it's gain needs to be a few dB louder in order to get the sound level the same at the listening position. I would think this is not a good thing, though maybe it is because the subwoofers will have identical constraints.

I guess I will have to try both and see which one sounds better, then compare it to which measures better. I have a BFD, but I will not boost the low end, as that's not recommended by the "gurus" at home theater shack. It would be a lot easier if my receiver had the option to adjust the output level for more than one subwoofer (it only has one subwoofer out connection).

All that you end up with for the most part when you use a single filter (LP or HP) are phase mismatches between various speakers. HP and LP filters cause a phase change as they roll off, so nothing is a "piece of cake" unless you are putting together a standard crossover (AKA LP and HP).

If you gain match two identical subwoofers, both subwoofers will max out at the exact same time.

If you level match two identical subwoofers, you are taking into account room acoustics. The subwoofer that uses the higher volume setting will then max out before the subwoofer that uses a lower volume setting.

If you EQ the two identical subwoofers in a different way, they are no longer identical!

If you have a pair of large main speakers setup properly, you would gain match the speakers.

Thanks.

I actually experienced what you said about one subwoofer maxing out before the other, or at least one amp maxing out before the other. During the "Pod Emerge" scene on War Of The Worlds, the clip lights lit red on the EP2500 driving the subwoofer farther from the listening position. This was also the amp with the gain turned higher up, so what you guys said about gain matching vs. level matching does make sense. I'll re-run REW with them both gained matched rather than level matched and see if I loose the extra oomph down low.

My main's are the Salk SongTower's, so they aren't "large" by any standard, and they only reach down to 38Hz with authority. In my opinion it wouldn't be a good idea to run them full range with my receiver set to double bass. Now if they were sealed I'd give it a shot.

So when running a sub (s) and an equalizer such as the BFD I know it's recommended to set the distance of the subwoofer in the preamp/receiver 1 foot further than it really is. Now what are you suppose to do if you run more than one subwoofer, assuming you are NOT using the Geddes method? Set it according to the furthest away subwoofer, then plus 1 foot?

Thanks all.

Mark Seaton explained it in one (or both) of the two links posted in post #1114. There's not much to explain, though, that hasn't been said in the last several posts, is there?

i believe this was the thread:

http://www.avsforum.com/avs-vb/showthread.php?t=1068195

you might want to skip to post #67

I'm simply referring to the fact that most subs have a variable low-pass which very easily lends itself to use in the approach. The knob is there. Not that the actual implementation is necessarily easy.

Or start at #59.


Better yet, you could start reading at Post#7 in THIS thread.

Ricci's Post #23 in that thread is useful.

that's a good one too, kind of like a "spin off" thread of the one i posted

or maybe the sequel; Calibrating Multiple Subs 2: Electric Boogaloo

sorry nuance, but i must question the claim that a couple 5" drivers can reach to 38hz with "authority".

i'm not saying your system doesn't sound great, but there are physical limitations at work here. around here "authority" is represented by reference spl. those little cones just aren't going to get there.

everytime that i suggest this point, the stones come flying...that's not really what i want. i just want to shed light on the fact that a couple 5" drivers can't produce miracles. there is a reason why the old school designs employed 15" drivers, or even multiples.

Assuming identical subs, if they are running at the exact same setting, then this can't possibly occur. Can it? That's (sort of) the whole point.



It is unclear what you are referring to in that last sentence, but when your subs are set to different gains in order to level-match them, they are operating dissimilarly, especially with respect to their headroom. Craig John also explained how level-matching the traditional way does not really provide level-matched results because the room can affect each subs output differently.

Yes, if you gain-match 2 (or more) subs that are not located equidistantly and/or are very dissimilarly affected by the room, one sub is going to be "louder" than the other. Proponents of gain-matching would say "so what?".


THIS post of Ricci's should provide some clarification. He's describing gain-matching by measuring the subs' actual output instead of simply relying upon setting the volume pots identically.



Once properly re-calibrated, the low-end oomph shouldn't change at all.



I don't know anything about that "1ft. further" bit. Someone may have measured the delay imposed (if any) by the BFD and converted that to being equivalent to 1 ft. in distance (I think it is approximately 10ms/ft). What receiver are you using? If your subs have continuously variable phase knobs, I would use one of the sub's knob to adjust the subs to be in phase with one another. Then use the receiver's distance setting to adjust both subs to then be in phase with your speakers.

He actually posted, first, in the thread that was started second.

It was in the threads sivadselim linked to. Basically to "gain-match" you set the gains to equal levels. If you're not sure that setting both subs to, say, 50% gain on their amps, will result in "equal" gains, the way to test it is as follows:

Set sub #1 in the middle of the room, as far from the influence boundary walls as possible. Set the mic about 6" in front of the driver. Play a a noise signal in the bandwidth of interest and record the level. Now set sub #2 in the exact same position and measure the same noise signal. Set sub #2 for the exact same level as sub #1. You have now essentially "gain-matched" the two subs. You can now set them in dissimilar locations and know that, no matter how they interact with the room, they will reach maximum output at the same point. (In Mark Seaton's setups, maximum output is a very high priority.)

This may, (or may *not*), result in the smoothest possible FR. However, it ensures that one sub is not "crippled" by the gain setting of the other. If one sub is bottoming, compressing or distorting while the other sub still has gas left in the tank, that is an ineffective utilization of the subwoofer resources. Still, if one's priorities are smoothest FR, it may be preferable to sacrifice that last iota of output for smoother response.

As I stated in one of my first posts in this thread, (Post #86):

...this is all about differing priorities: those being; maximum output, maximum extension or smoothest response. Some of us are willing to sacrifice one for the other. Dr. Geddes's highest priority is obviously smoothest FR. He may well be sacrificing some max-output, or some VLF extension to get his smoothest output in his BW of interest. That's fine for him. For me, I am willing to sacrifice a few +/-dB for some more output, especially at <20 Hz. My gain-matched (Master/Slave) setting allows this.

(PS. I also first asked the question about Audyssey in post #86):

Yes... that's Audyssey.. which you have now declared as "flawed"! In fact, I think you referred to it as a "disaster".

IME, Audyssey has been a godsend. I have used "equalizers" for years. My first exposure to a graphic EQ goes back to the late 70's with the Soundcraftsmen PE2217 preamp/EQ. I have had many EQ's in the interim, the most recent being the digital, parametric, OSD Velodyne SMS-1.

IME, Audyssey trumps them all... and by no small margin.

Your complaint that Audyssy EQ's above the modal region is moot. Audyssey's primary effect above the modal region/Schroeder frequency/transition zone/"whatever-you-want-to-call-it" is to implement their "target curve", which is a shallow rolloff of the high frequencies, (-3 dB at 20 kHz). In my system, you can implement Audyssey EQ, and then turn it off. The only significant difference you'll hear is a difference in the articulation of the bass. The mids and highs are virtually unchanged. However, the improvement in articulation of the bass is quite impressive. Bass lines are much easier to follow. Distinction of the bass instrument, (guitar or standup, electric or acoustic), from the bass drum is much easier to delineate. OTOH, bass drums are still percussive and tactile, just as they are in real life, (of course this depends on the recording).

Bottom line, there are *many* ways to attempt to optimize the bass performance of subwoofer systems. There are also differing priorities about the importance of various performance criteria. One needs to decide their own priorities and then select the setup routine that optimizes those priorities.

Craig

It's cool, I am not insulted bud. I understand where you are coming from. Well, the SongTower's are a transmission line design, so that attributes to the deeper bass extension that, normally, a 5" driver couldn't handle.

By authority, I mean if I play a 38Hz sine wave, it rattles the walls. That may not mean your requirements of authority, but it does for me.

Hope that clears that up.

Edit:

As an update, I moved the gain in the second subwoofer down to match the corner loaded sub. Then I used a pink noise signal from my receiver and raised the output a bit to match my main speakers at 75dB again. I changed nothing on the BFD and left the filters as they were, and the results are pretty much the exact same as when I EQ'd the dual subs last week. Sweet!



Now my only question is, will the difference in output levels result in any negativity in the sound field? In other worlds, will I be able to tell that the left subwoofer is quieter than the corner loaded subwoofer, thus being able to localize the LF's (69dB by itself as opposed to the corner loaded subs 74dB by itself)? Combined they equal about 76-77dB.

Oh, and what about the distance setting in the receiver? Do I use the furthest away subwoofer, then add a foot, or the closer subwoofer?

You're the guinea pig. This is going to be reflective of how well your room is treated and/or behaves. In a "perfect world" it shouldn't matter. You tell us. Inquiring minds want to know.



I sort of addressed this in the final edit in Post#1818. Do you really simply rely on measuring your subs' distance(s), and adding the purported 1ft., for the receiver's subwoofer distance setting?

With that FR you posted, I wouldn't change anything, btw.

Yes, the delay of the BFD was figured and equated to 1 ft. distance adjustment.

My subs do not have variable phase control yet, as they are DIY 20" ^3 enclosures w/ about 3.3 cu. ft. internal volume when taking into account internal bracing. I'll add variable phase adjustment switches down the road, but until then I just have my Onkyo 805 receiver and soon to arrive Parasound Classic 2100 preamp.



Well, I watched a few movie scenes and have been listening to music since. It could be psycho acoustics, but I could swear that localization is impossible now, while before I could swear the bass seemed lean toward the left side sometimes. Now its all over and enveloping. I can't believe lowering the gain and a drop in dB from the left subwoofer (now 69dB in comparison to the others sub's 74dB at the listening position) alone would do this, but it actually seems to have balanced things more, so it seems to be true. Again, this could all be mind games, but I like what I am hearing so far, and it measures just as well as before, as you saw. I also had no clipping of either Behringer EP2500 amp during the "pod emerge" scene.

Wow, who would have thought? At least in my current situation and setup, it seems "level matching" isn't as important as gain matching, even if it does vary in dB's on paper at the listening position. Weird... Perhaps someone like Mark Seaton can explain why it doesn't seem to matter, and in fact seems to sound better?

Thanks for everyone's help, by the way. This is the best audio/video resource on the planet! I've learned so much about proper bass implementation!

One can attain equal gain levels and achieve less distortion with a disproportionate amount of bass produced per sub with respect to distance (output varies with distance) or one can achieve equal levels and slightly different "drive levels" but better coupling power (read my explanation for further details...). So in a way, it almost balances itself out.

Regards,

I have not seen a strong connection between "Q" and effectiveness in a real room. In a real room things like Q and resonance will all get swamped by things like Modes and damping. In other words in the room response it will be impossible to tell what the speaker Q is and often even what the resonance is because the subs couple into the room making one system that is no longer that well dictated by the specifics of the Sub, but more the specifics of the room. Short answer - it doesn't matter much.

I will state my preference for bandpass subs as I have just found that these work better. I can't say for sure, but I believe that the acoustic LP filter helps to minimize localization of THD harmonics, but that is a guess.

I'm very disappointed that my question wasn't answered. I really wanted to understand the difference between a null and an "interferometric null".

Very disappointed.

Regards,

I think that I 've been clear on this, I use multi point measurements to set all the variable parameters in all the subs. There isn't one solution nor one way to get to that solution.


I don't HP any of the subs.


Except that I don't sacrifice any of the thing that you mentioned. Just because my highest priority is on FR DOES NOT MEAN that I don't achieve all of the other things that you mention. My room can do < 20 Hz at ear splitting levels - and yes its fairly flat too!

What seems to be true?

Right. You rely upon the subs' own natural roll-offs. But how did you arrive at the subs you use for the upper two. What characteristics did you look for? Did you design and build them to have specific characteristics in this regard? If so, how would you recommend someone set about using your approach with store-bought subs?