Eliminating an active crossover

My speaker system was designed to be used with a manufacturer-supplied active crossover unit. The crossover offers several adjustments, notably:

- choice of xover at 63 or 80Hz
- independent level controls for the subs
- independent phase controls for the subs
- 0 or 180 degree phase adjustment
- 2 xover slopes (24db and 18db/octave (or maybe it's 12db - hard to tell, and I have no manual))

But when using REW with a FBQ2496, I'm wondering whether it might not make sense to try and subsume the functionality of this active crossover. It just seems better to me to eliminate components in the signal path if possible, and doing so certainly simplifies the room optimization problem (believe me - I've spent hours tweeking the knobs and switches on the front of this active xover and measuring the effects with REW, but the ability to control the frequency response of the system this way is crude at best).

To see if eliminating this xover is possible, I tried measuring the following:

a. prepro full range output -> active xover -> sub output (80hz xover)
b. prepro sub output (configured with 80hz xover)

Here's what I see (a=green, b=yellow, smoothed 1/24db/octave):

nf xover vs prepro sub output.jpg

From this graph, there are a few things that are clear. First, I've got some problem going on around 23 and 36Hz. I'm not going to worry about that here (since this is the measurement for one sub, and I actually have two... plus I probably need to reposition it to avoid this drop-out). Second, we can see that the active xover is operating with a different output level than the prepro sub output (~15db). If I adjust for this (so that we can compare similar curves) I get this graph:

nf xover vs prepro sub output2.jpg

From this graph it appears that the active xover is applying some equalization to the sub since the sub as measured directly from the prepro shows 18db or more attenuation at 20Hz. So the question is, what exactly is this equalization, and can I reproduce it with the FBQ2496...

This is where I'm venturing off into territory I'm unfamiliar with, so please bear with me, and let me know your suggestions. Taking a cue from some things I've read in this forum about designing a house curve and shelf filters, I came up with the following:

filter1.jpg

That's 3 parametric filters: -31db, 8 BW Oct @ 80Hz; -36db, 10 BW Oct @ 1400Hz; and -12db, 1.5 BW Oct @ 52Hz (seat of my pants); plus a 33db gain in the sub output level (I think my amp has the headroom, although I'm not sure what this means about noise I might be injecting into the system). To get a better idea of what these look like across the entire spectrum, here's an expanded picture:

filter2.jpg

Clearly I can apply some spot filters to smooth this out a little more, but my main question is whether this approach seems reasonable, and/or whether there's another route I should take. (I haven't yet tried this filter in my system, just with REW... so I can't tell you how I think it sounds.)

Thanks for your input...

---

Hi Wayne,

Sorry if I wasn't clear. I'm basically trying to go from the setup on the left, to the one on the right:

You don't need any filters above the crossover, your prepro with an 80Hz xover is doing the same as your active xover with the same setting. Just try some boost at around 25Hz, bandwidth around 1 octave, so that the "corrected" trace of your measurement without the active xover is more or less overlaying the trace of your measurement with it. Looks like it will need about 10 or 12dB of gain to do that. Then you can start putting in some filters to deal with some of the peaks at 45Hz and up.

I was under the impression that applying a boost was a bad idea because it amplifies noise in the system... although (a) at 20Hz, maybe that's not something I'm going to hear, and (b) maybe that's just what my active crossover is doing anyway. This is why I was going down the path of trying to attenuate the large rise between ~40-100Hz.

But if I have to boost my sub output level (in the prepro) to compensate, maybe the net effect is the same. Or maybe it depends on which component is noisier... or whether I even have that much gain available?

Looks like the usable output of your subwoofer is only down to about 23Hz from those graphs. Can you tell us the model.

brucek

---

Hi Bruce,

My speaker system is a Nearfield Acoustics Pipedreams system. It has 2 line array towers and 2 18" subs with an active crossover. You may recall my challenges to optimize them with REW some time ago in a previous post. I'm just now returning to REW after a hiatus, and I'm determined to overcome my room problems as well as drop either the active crossover or the FBQ2496 if possible. My prepro is now an Onkyo PR-SC885, 4x 350Wpc amplification on the pipedreams. I know this isn't your typical HT setup, and my room presents numerous challenges, but what can I say... my hobby has gotten out of hand in the last few years.

Also, about the limited low-end of this sub -- first of all, I only see it on the left sub (the right goes all the way to 20Hz), but my room is asymmetrical. I'll try swapping the 2 sub positions, but I suspect it's more about the room than something about the sub itself.

---

Hi Wayne - That's not a typo. I have 2 18" subs with dual drivers each. They're approx 4' tall and 20" in diameter. Nearfield Acoustics (now High-Emotion Audio) has downsized their design in recent years.

BTW, here's an attempt to fix my 23Hz problem by boosting the response at 20 and 23Hz by 15db (i.e. 2 15db boosts) + a few misc filters to fix some peeks:

boosting 20hz.jpg

My intuition is that applying this much gain with the FBQ2496 is a bad idea since it will introduce a lot of noise, but maybe my original trick of doing the 33db gain with my prepro and tilting the entire response curve to give a similar looking graph is an equally bad idea?

I know you guys are saying "why not just stick with the manufacturer's xover?", but honestly, tuning this system is a massive multi-variable equation, considering the gain and phase adjustments of the active xover, the fact that I have 2 subs, the room response in general, the Onkyo output level settings, Audyssey EQ, parametric EQ with the FBQ2496, etc, etc. Throwing something overboard seems pretty attractive right about now.

---

I Wayne can i as a noob and even if i'm a mile off base here but would like to learn more have a guess in that it maybe 1/ where the sub/s are positioned or 2/ the phase could be off some or the 3rd could be something maybe wrong with the xo used.

Steve.

---

You're right that the phase adjustment would be the one thing I'd lose. Since there are 2 subs, this could be important to correct for nulls at the listening position due to sub placement. My Onkyo 885 only has one sub output (unlike my previous Lexicon), so there's no way for the prepro to compensate for this... but to tell the truth, with all my playing around with the phase adjustment and measuring with REW, I can't say it helped much at all.

The active xover does no equalization to the main speakers, as best as I can tell.

So let me try to get this discussion back on track... I'm not really asking how to correct for my left sub's 40Hz problem (I can deal with that myself)... what I'm trying to get at is how to reproduce the active crossover's sub equalization curve using the FBQ2496 -- the difference that can be seen in the first graph in my first post.

---