[Dimitri_ISF]

Hi,

I've been using the BFD and REW for some years now with good results.

But with my latest setup I've been having difficulty equalizing my subwoofer(BK monolith) with the BFD 1124.

My speaker setup is: 2 KEF Q11 towers.
3 Kef eggs.

Crossover for all the speakers is 80 Hz. on my Onkyo tx-sr607.

I've posted some photo's of my setup and my first readings without EQ.

Main problem is boomyness and a sub which can be localized on many occasions.

My usual method for applying filters is to make a measurement, pull up the filter adjust tab on REW and just create manual filters to get a somewhat flat response, and this is usually enough. But I feel this situation calls for something more.

Can you guys help me to use the more advanced featured of REW like the spectral decay plot. I can recall JohnM talking about using this, but I wouldn't really know what to look for in this graph. I've posted 2 graphs to illustrate my situation.

If I left anything out please tell me.

Any help would be appriciated.

Dimitri.

 

[Wayne A. Pflughaupt]

Is this a full range, or sub-only graph?

Regards,
Wayne

 

[Dimitri_ISF]

Hi wayne,

This is the monolith and the Q11's in stereo.

Dimitri.

 

[Wayne A. Pflughaupt]

Thanks. Do you have a graph with the sub equalized, without the main speakers?

Regards,
Wayne

 

[Dimitri_ISF]

Here they are.

Just the sub with EQ and one without EQ.

Dimitri.

 

[Dimitri_ISF]

...

 

[Wayne A. Pflughaupt]

Okay, I see the problem. Notice the upper end of your unequalized graph, how response roughly traces the Target curve. On the equalized graph, everything above 80 Hz is elevated.

It looks like you applied filters that affected response all the way up to 150 Hz, or even higher. That’s why you’re having localization issues. There’s usually no reason to apply filters above the crossover frequency designated by the receiver, or even close to the crossover point.

I suggest eliminating any filters you have above about 80 Hz. If that doesn’t take care of the localization problem, try using a house curve. That would pull down the sub’s upper frequencies further.

Keep in mind that it might be impossible to achieve 100% non-localization. A program that has severe bass levels in the 90-100 Hz range could functionally “blow out” the crossover, much as your equalization has. There’s not much you can do about stuff like that, you’re at the mercy of the program’s sound mixers. Nevertheless, eliminating EQ filters above 80 Hz is the way to go.

Regards,
Wayne

 

[Dimitri_ISF]

Thanks wayne,

I never appllied filters above 85 Hz. orso. I was surprised to see the subs response at the higher frequencies though. I believe this is caused by the method I used to get the lowest points of my sub response at the 75 Db. level. I just raised the sub volume a few times to get the lowest points where I wanted them, and apply extra filtering to the peaks that have also been lifted. That's the way I use to get an even response along the whole range, And with the mains active in my measurements I didn't identify the peaks that were created as being the sub. I just figured it were the main speakers.

Is this a good method to get an even response?. Or should I just leave the dips where they are and just focus on the wild peaks?.

Another point is that I find the new KEF's to be a little too aggressive. To the point where the sound can become a bit annoying at times. The peak I have at 80 Hz. also affects my mains response I believe. I've added a graph of the response from the Q11.

Dimitri.

 

[Dimitri_ISF]

Here it is.

 

[Wayne A. Pflughaupt]

I just raised the sub volume a few times to get the lowest points where I wanted them, and apply extra filtering to the peaks that have also been lifted. That's the way I use to get an even response along the whole range, And with the mains active in my measurements I didn't identify the peaks that were created as being the sub. I just figured it were the main speakers.

Yeah, I’d say that’s a problematic approach. If you were cutting everything down to where the 78 Hz null is, and nothing above 85 Hz, that would result in the response you ended up with.

Cutting everything down to the lowest depression, or boosting everything up to the highest peak means you’re essentially using the EQ as a gain control. This is poor use of an equalizer. You might want to review my Minimal EQ article – there’s a link to it in my signature.

Also, leaving the mains running when equalizing is another no-no, because as you noted, you can’t readily tell what’s coming from the mains from what’s coming from the sub.

Take another look at your sub-only graph. Your best bet would be to raise the Target up to about 77-78 dB, and get response to track the Target, using both boosting and cutting filters.

Regards,
Wayne

 

[Wayne A. Pflughaupt]

Here it is.

Hmm, looks like the mains are running full range. There’s no evidence of the 80 Hz crossover reducing bass response even a little bit.

Regards,
Wayne

 

[Dimitri_ISF]

Hi wayne,

I'm not sure what you mean by raise the target?.

Dimitri.

 

[Dimitri_ISF]

Hmm, looks like the mains are running full range. There’s no evidence of the 80 Hz crossover reducing bass response even a little bit.

Regards,
Wayne

Wayne,

You're right. This measurement was taken with the mains at full range.

Dimitri.

 

[Wayne A. Pflughaupt]

Hi wayne,

I'm not sure what you mean by raise the target?

The Target Curve is the blue-green line that traces the response of the subwoofer crossover. Typically it sits at 75 dB, but it can be adjusted up or down. Look for the “Target Settings” icon on the left side of the screen.

Regards,
Wayne

 

[Dimitri_ISF]

Ok, will do.

It's back to the drawingboard then. I was wondering though, when creating filters, do you guys let REW calculate the filters, or do you just create them manually?. Is there a smarter way to get the right filters?, like using the spectral decay plot?.

Dimitri.

 

[glaufman]

You'll find that most people let REW create them, then tweak from there, sometimes eliminating a few entirely...

 

[JohnM]

Is there a smarter way to get the right filters?, like using the spectral decay plot?

REW V5 provides a predicted waterfall plot that updates live as you adjust filter settings, that can be useful if trying to target a specific resonance. The V5 filter assignment and optimisation algorithms are also more powerful than V4.

 

[Dimitri_ISF]

Ok,

I'll install the latest version tonight.

One more question before I begin filtering: When I want to introduce a house curve, should I start differently?. Does it take another approach?.

Dimitri

 

[Wayne A. Pflughaupt]

All you need to do is load a house curve file. Look for the “House Curve” tab in the “Settings” window.

Regards,
Wayne

 

[Dimitri_ISF]

Hi,

Do I need to create a house curve file myself?.

Dimitri.