Filter Question

[aceinc]

Below you will find 3 graphs, the first is the sub unfiltered, the second is using the recommended filters, the third is a totally manual set of filters. All graphs have a 1/3rd octave smoothing.

I have two main questions;

1) Why do the filter recommendations from REW seem to be less optimal than what I did manually? Can I improve the algorithm, by changing parameters?

2) I can't seem to get the peaks above 140 hz to go away. I tried various settings. Any thoughts?

BTW, I might make an effort to get a midi interface operational, if an iterative process was introduced that would test, upload the settings, test again up to "N" times trying to get the flatest curve.

Paul

[brucek]

1) Why do the filter recommendations from REW seem to be less optimal than what I did manually?

It's normal to tweak the filters a bit. No big deal.

can't seem to get the peaks above 140 hz to go away

Because the sub has very little influence at 140hz. The crossover has reduced its signal at a rate of 24dB/octave after the crossover. The mains are the overwhelming influence by that frequency. The BFD is connected to the sub. We don't usually bother equalizing past ~100Hz.

I like the unfiltered graph the best. Why do you feel you need equalization with a great response like that?

brucek

[aceinc]

My testing was with the computer plugged directly into the BFD. No other speakers were powered during the sub test. The only sound should be from the subs through the BFD, so even though my receiver (when it is connected) will be crossed over at 80 hz, shouldn't I be able to control the freqeuncies above 120?

Why do you feel you need equalization with a great response like that?

I guess I was looking to flatten out the curve, and make sure it wasn't interfering with my mains. The graphs look much worse without the smoothing.

You may (or not) remember that I have three subs, 2 - 10" and 1 - 15". I spent an hour or two just moving subs, and adjusting volume, phase, xover and boost parameters on the sub system. I have all 3 subs side by side, as this was the best configuration based on empirical testing.

The sub system gets its signal from a Yamaha receiver's sub out (or from REW during testing) that is split into two input channels of the BFD, which drives a NAD amp. The main outputs of the NAD go into the two 10's and into the high level input of the 380 watt plate amp on the Rythmik Audio 15".

What other testing would you recommend? Can I test for persistence (echo, hangover, boom) of bass signal? Of course I would need room treatments to resolve that.


Paul

[terry j]

earlier you asked why the suggested graphs from REW were not as good as done manually, and in passing you mentioned that the graphs posted were smoothed. Did you ask REW to find the filters on a smoothed graph??? If you did, that may be part of the reason that the suggested filters weren't the best, (the find filters function should only be done on unsmoothed graphs).

If you already knew that then accept my apologies for a bum steer ha ha...

[brucek]

I guess I was looking to flatten out the curve, and make sure it wasn't interfering with my mains.

Always test your sub(s) through the receivers crossover to simulate the normal useage and equalize the sub(s) that way without the mains. Then add the mains and tweak the crossover area if necessary.

so even though my receiver (when it is connected) will be crossed over at 80 hz, shouldn't I be able to control the freqeuncies above 120?

No.

The graphs look much worse without the smoothing.

Turn it off so we can see it without.

I have all 3 subs side by side, as this was the best configuration based on empirical testing.

That's usually the way it goes. Multiple subs of different sizes and models is near impossible. Personally I'd retire the 10" subs and use the one 15". It will likely result in a better overall response.

terry says:

Did you ask REW to find the filters on a smoothed graph??? If you did, that may be part of the reason that the suggested filters weren't the best,

Exactly.

brucek

[aceinc]

Did you ask REW to find the filters on a smoothed graph???

Nope, did all the work on the raw data, it didn't even cross my mind to work on the smoothed data. I just did the smoothing for publication.

Turn it off so we can see it without

I am still working, will be for a few more hours, when I get home I will post the raw graphs. I believe I saved the .mdat file from the three graphs in question, hopefully that will allow me to rebuild the graphs without retesting.

This;

so even though my receiver (when it is connected) will be crossed over at 80 hz, shouldn't I be able to control the freqeuncies above 120?

Should have read;

so even though my receiver (when it is connected) will be crossed over at 80 hz, shouldn't I be able to control the freqeuncies above 120 with the computer plugged directly into the BFD?

Regarding the multisub configuration, you said yourself that the raw graph looked good, and it sounds rather impressive as well. The multisub, as is everything, is a tradeoff. Smaller subs used for the upper sub range might provide less doppler distortion, and be quicker. A larger sub for the lower sub range has the surface area to move a lot of air. When they are placed close together they behave as a single unit.

The 10" drivers are Dayton Titanics, and the 15" is a Rythmic Audio, so they're not el-cheapo, nor are they esoteric drivers. They are all good solid drivers in their own range.

Paul

[brucek]

When they are placed close together they behave as a single unit.

Well, that's true as far as they couple the same to the room. But if you take two subs and they both output 110dBSPL at 50Hz, but only one outputs 110dBSPL at 20Hz, do you see the problem that results?

You've turned them both down to take advantage of the increased headroom at 50Hz, but at 20Hz the output is lower and has less headroom. The expression used when a challenged output is combined with a not so challenged output is that the overall response dumbs down to the lesser sized sub....

so even though my receiver (when it is connected) will be crossed over at 80 hz, shouldn't I be able to control the freqeuncies above 120 with the computer plugged directly into the BFD?

Yep, make sure there isn't a dot in front of the frequency you choose for the BFD. That denotes kHz rather than Hz. A common error. that might have been it.

brucek

[aceinc]

Well I finally can post the raw graphs.

I suppose if I want to really see what I am working with I should take the subs into the front yard and close mike graph them indiviually, but that would be a real chore.

[brucek]

If the third graph is the filtered response, it's not a good idea to have the dip at 100Hz. That should be a level transition to the mains.

I still prefer graph number one. If that's your unfiltered, unsmoothed response, I would remove the BFD.

brucek

[aceinc]

You don't think the spikes between ~68 & 76hz and above ~ 83hz are problematic?

Well the good news is that I can AB test pretty easily.

My front speakers are floor standing KEF's which have pretty decent bass response by themselves. So my thought was to roll off the sub pretty sharply above the 80hz cutoff. I was going to use the BFD to help in that roll off, which is what started the thread.

Paul

[brucek]

You don't think the spikes between ~68 & 76hz and above ~ 83hz are problematic?

I doubt you'll notice. The main duty of a BFD is to remove resonant peaks caused by the room's modal response. You don't appear to have any of those peaks (or dips) and so I don't think you really require a BFD.

So my thought was to roll off the sub pretty sharply above the 80hz cutoff. I was going to use the BFD to help in that roll off, which is what started the thread.

But the bass management of the receiver already handles that for the sub out crossover with about 24dB/octave, (unless the receiver you mentioned in the quote below doesn't have any bass management, the rolloff provided should be suffice).

The sub system gets its signal from a Yamaha receiver's sub out (or from REW during testing) that is split into two input channels of the BFD, which drives a NAD amp. The main outputs of the NAD go into the two 10's and into the high level input of the 380 watt plate amp on the Rythmik Audio 15".

Normally, allowing (bass capable) mains to naturally rolloff isn't a good idea, and usually ends up to be a problem, since you can't equalize the low end bass, (normally handed to a eq'd sub). But in your case (given your rather nice room), it may work for you.

brucek