What Behringer DEQ2496 Says About My Room - (pic)

[meagain]

S - Yes, no more switching things. Hubby's grabbing test tones: 100, 400, 1000. Not sure if there's others I need.

[sfogg]

"When I do the phase test with woofs disconnected - I hear a "center channel" when it's supposed to be out of phase. When it's supposed to be in phase - I hear the sound more from the l/r speakers, the air, or more on the sides of the room. "

This is exactly like it would sound like with one of the drivers out of phase with the other. Swap the polarity of one. Then consider phase *between the squawkers* a done deal and lets more on.

The next question was phasing between the squawker and the woofer. I think you said earlier you couldn't hear or see a difference between it when testing it. If that is the case lets not stress out about it for now. The can't really be in exact phase anyway with passive crossovers.

(Question for K'Horn owners... what is the path length difference between the K55 and the K33?)

So... if we have the woofers in phase with each other and we have the squawkers in phase we can move on to dealing with the hot tweeters.

Set up a parametric filter that is joined for L/R.

Set the center frequency to 8500hz.

Set the depth/gain to -6dB.

Set the Q to 3.

See how that sounds and how it measures at the listening position.

Shawn

[meagain]

I believe that when the Squawker and Tweeter are in phase in the left speaker and the 315Hz null goes away is the correct hookup and if I followed all this completely with both old and new crossovers being used that the left speaker is were the problem is and only with one of Deans Crossovers. If I'm wrong I will apologize now because this has been a bit difficult to follow completely.[]

Yup - 315 dip came up. Crazy huge lone spike at 160.

[sfogg]

Mike,

Thanks. 1/6 might help a little in trying to determine phase. Though it is also going to make everything else look worse too which is going to add to the confusion.

"I believe that when the Squawker and Tweeter are in phase in the left speaker and the 315Hz null goes away is the correct hookup"

50% chance you are correct.

The problem with basing it on just that null though is it is with the mic relatively close to the speaker and with the mic elevated above the K400 horn. The path length difference between the bass horn and the K400 is likely at a somewhat different relationship to the mic then when she is listening on her couch.

Move the measurement position up and down and with the squawker in either polarity you will likely be able to find null points. He measurements from the sweet spot don't show that null for example.

Shawn

[meagain]

"When I do the phase test with woofs disconnected - I hear a "center channel" when it's supposed to be out of phase. When it's supposed to be in phase - I hear the sound more from the l/r speakers, the air, or more on the sides of the room. "

This is exactly like it would sound like with one of the drivers out of phase with the other. Swap the polarity of one. Then consider phase *between the squawkers* a done deal and lets more on. Last time I did it, I got a totally different impression. Like it swings back & forth on me. I left it wired normally.

The next question was phasing between the squawker and the woofer. I think you said earlier you couldn't hear or see a difference between it when testing it. If that is the case lets not stress out about it for now. The can't really be in exact phase anyway with passive crossovers.

(Question for K'Horn owners... what is the path length difference between the K55 and the K33?)

So... if we have the woofers in phase with each other and we have the squawkers in phase we can move on to dealing with the hot tweeters.

Set up a parametric filter that is joined for L/R.

Set the center frequency to 8500hz.

Set the depth/gain to -6dB.

Set the Q to 3.

See how that sounds and how it measures at the listening position.

Shawn

OK - You mean target 8500 on the PEQ. (I can only get 8531). Take that point down -6db. Then you lost me on "set the Q to 3". Not sure what that means. I'm more used to using the GEQ

[PrestonTom]

"

The next question was phasing between the squawker and the woofer. I think you said earlier you couldn't hear or see a difference between it when testing it. If that is the case lets not stress out about it for now. The can't really be in exact phase anyway with passive crossovers.

(Question for K'Horn owners... what is the path length difference between the K55 and the K33?)

So... if we have the woofers in phase with each other and we have the squawkers in phase we can move on to dealing with the hot tweeters.

Shawn

Shawn, you hit it on the head.

The difference in propagation delay betwen the woofer and the mid-range is a few msec. If the crossover overlap is a few hundred Hz aound 400 -600 Hz then this difference in propagation delay (expressed as phase) is over a full cycle. These are first order filters, so the 90 deg phase shift from the network (and who knows what from the driver) is small. With the 1/3 octave analysis you are dealing with, it will be hard to tell whether the summed signal (within the band of overlap) is adding constructively or not.

In other words, I am not sure what one considers to be "proper phasing" between the woofer and mid range. Ahh heck, what ever sounds good.

Congrats on remotely identifying the out of phase (re: left vs right) mid range. I suspect that will clear up a number of problems. Doing this remotely is can be frustrating...

As far as the the hot tweeters, she has got an EQ waiting to be used. At this point, I can't imagine needing to spend money on the new tweeters or pads etc.

-Good luck,

-Tom

[sfogg]

"OK - You mean target 8500 on the PEQ. (I can only get 8531). Take

that point down -6db. Then you lost me on "set the Q to 3". Not sure

what that means. I'm more used to using the GEQ"

8531 is fine. I just wanted to get in the ballpark of the center of the hotter part of the tweeters range.

You should have three settings for a parametric EQ.

You have set two of them... frequency and the depth/gain. The last setting could be labeled Q or maybe Width. On other Behringers it is called 'Q.' Whatever they call it set it to 3 or 4 and take a listen to the system like that.

Shawn

[Erik Mandaville]

I don't recall the schematic for the network in question, but in a second order network there is a 180 degree phase shift when drivers are connected in parallel. That means that drivers wired in this fashion would have a null in response at the crossover frequency.

The Butterworth second-order thus calls for drivers to be wired in reverse with respect to polarity, and there will be an increase in FR response at the crossover point.

So, one is left with a choice: Confirm the correct polarity of the drivers in question, and then choose between a parallel (or not) connection between the drivers. As I guess must have already been mentioned, there may be an incorrectly wired connection associated with the driver itself, or possibly the squawker on one network is wired accidentally out-of-phase from the other.

IOWs, Meagin's plot may be showing the difference between the two types of connections available for this type of network, and the subsequent dip or peak in response.

If it's hard to tell if the woofer is wired out-of-phase, you can confirm this with the common penlight battery test (if you can get a visual of the driver cone). Doing this with La Scalas would really be a pain, but I can't remember what kind of speakers you're using, Meagin.

Also: Sorry for not responding sooner to your interest in that variable type 'A' network I mentioned. I can still do that with 3 different commonly used caps, but let me make sure I have what I need for another pair of simple networks for me to use in the meantime.

Erik

[mikebse2a3]

Mike,

Thanks. 1/6 might help a little in trying to determine phase. Though it is also going to make everything else look worse too which is going to add to the confusion.

Shawn,

Just to clarify what I was telling you is that The Behringer is a 1/6 Octave RTA and that you are already seeing the RTA Display in 1/6 Octave. If you look closely at the pictures (This Display is Page Three of the RTA MENU and is just an expanded Display of Page One and Page TWO of the RTA MENU) Lisa has been displaying. The Numbered Bands are at 1/3 Octave but there is a Band in between each Numbered Band. Actually in Page ONE and Page TWO of the RTA MENU DISPLAY you can scroll the pointer up or down the display and get a specific FREQUENCY Readout of the BAND (of the 61 Bands Displayed) you are pointing at along with a specific db readout for that BAND in a small window of that screen.

"I believe that when the Squawker and Tweeter are in phase in the left speaker and the 315Hz null goes away is the correct hookup"

50% chance you are correct.

The problem with basing it on just that null though is it is with the mic relatively close to the speaker and with the mic elevated above the K400 horn. The path length difference between the bass horn and the K400 is likely at a somewhat different relationship to the mic then when she is listening on her couch.

Move the measurement position up and down and with the squawker in either polarity you will likely be able to find null points. He measurements from the sweet spot don't show that null for example.

Yes I first thought that it very likely was a room or MIC position artifact and one good test would be to swap Deans other crossover into that speaker and if it disapears then the crossover warrents investigation. It will be interesting to see how this turns out! mike[]

Shawn

[sfogg]

" In other words, I am not sure what one considers to be "proper phasing" between the woofer and mid range. Ahh heck, what ever sounds good."

Right, that was my thinking too. For example 400hz has a wavelength of 2.8 feet. If the path length difference was 1.4 feet between the woofer/squawker if all else was even (which is isn't of course) either polarity on the squawker would be pretty much as in phase with the woofer as the other. It would be about 180 degrees out of phase either way.

Since she couldn't hear or see much of a difference either way I'm guessing it is something along these lines. That the difference is small enough to maybe not make a huge difference either way at the listening position. The rolloff of the k33 in the bass horn probably helps to make it less of an issue too. Acoustically that is going to act like a higher order crossover which will reduce the interference region.

"I suspect that will clear up a number of problems."

Agreed. Imaging/focus should be obviously improved, the mids will likely sound better balanced to the bass and I would think overall listener fatigue would be decreased. Conflicting imaging queues (in phase bass, out of phase midrange) I would think might drive the ear/brain a little batty over some longer term listening.

" As far as the the hot tweeters, she has got an EQ waiting to be used. At this point, I can't imagine needing to spend money on the new tweeters or pads etc."

Exactly right. The EQ can pad down the hotter part of the tweeter but leave the response above 12k or so alone. The end result will be a better FR for the tweeter itself. Can't do that with a L-pad or autoformer.

Shawn

[Erik Mandaville]

Meagin:

I've read through more of this, and I'm all but certain Shawn got this exactly right. If you have been switching networks in and out, it may be that there was in inadvertant misconnection made with respect to polarity. Or possibly the same for the actual connections on the networks, themselves. I found that I did the same thing myself once when I experienced a very similar situation to your own.

Erik

[sfogg]

Mike,

"Yes I first thought that it very likely was a room or MIC position artifact and one good test would be to swap Deans other crossover into that speaker and if it disapears then the crossover warrents investigation. It will be interesting to see how this turns out!"

NO SWAPPING OF COMPONENTS.... PLEASE I BEG YOU!!!

I think at some point she swapped the same type of crossover side to side (though I am hardly sure of that) and that null stayed in the one K'Horn. When she inverted the phase of the squawker I think the null went away. So if one of the two crossovers was miswired for the polarity of the squawker then that null would not have stayed there when she swapped crossovers left to right.

If both crossovers were inverted on the polarity of the squawker they would have passed the imaging test wired color to color. Since that didn't happen I think the squawker is what has its polarity inverted internally.

Shawn

[mikebse2a3]

"OK - You mean target 8500 on the PEQ. (I can only get 8531). Take that point down -6db. Then you lost me on "set the Q to 3". Not sure what that means. I'm more used to using the GEQ"

8531 is fine. I just wanted to get in the ballpark of the center of the hotter part of the tweeters range.

You should have three settings for a parametric EQ.

You have set two of them... frequency and the depth/gain. The last setting could be labeled Q or maybe Width. On other Behringers it is called 'Q.' Whatever they call it set it to 3 or 4 and take a listen to the system like that.

Shawn

Shawn and LISA on the PEQ of the Behringer the Bandwidth is set and displayed by a small window marked BW/OCT and the badwidth is selected by selecting from (1/10 octave to 10 octave range).

Also to select the center frequency two scroll options are avaible from 20 to 20khz.

The first scrolls in 1/6 octave steps and the second uses 1/60 octave steps by pressing the Large DATA WHEEL you can toggle between these two options.

mike

[meagain]

I didn't read the last few posts..... Still trying to figure out this "Q" thing or "width".

Shawn/Mike - does that mean the 'width' of the PEQ peak thing?

I took the 8531 area down -6db. Listened to some Sinatra as there are horns on here that kill me. It's actually sounding worse to me than before. This could be because I'm now on 5/2 and used to 4/x? IDK. But I can't find anything in the PEQ menus where I could change anything to "3".

EDIT: BW/Oct (a fraction)... this makes the points widen. It's currently on 1/6. Those are the 3 settings I have.

[sfogg]

Mike,

Is the 'bandwidth' setting the entire range (edge to edge) that the filter is active in or the distance from center to one side that the filter covers?

If you have a display on it... make up a filter that drops 6dB total and starts at around 6k and ends by around 10 or 12k.

Shawn

[meagain]

Mike,

"Yes I first thought that it very likely was a room or MIC position artifact and one good test would be to swap Deans other crossover into that speaker and if it disapears then the crossover warrents investigation. It will be interesting to see how this turns out!"

NO SWAPPING OF COMPONENTS.... PLEASE I BEG YOU!!! At this point, right now? I'd rather take a bullet to the head. Plate full.

I think at some point she swapped the same type of crossover side to side (though I am hardly sure of that) and that null stayed in the one K'Horn. When she inverted the phase of the squawker I think the null went away. So if one of the two crossovers was miswired for the polarity of the squawker then that null would not have stayed there when she swapped crossovers left to right. Are you talking about the "315 mega-dip"? It is now basically gone or severely fixed with proper wireing to taps.

If both crossovers were inverted on the polarity of the squawker they would have passed the imaging test wired color to color. Since that didn't happen I think the squawker is what has its polarity inverted internally.

Shawn

[mikebse2a3]

Mike,

Is the 'bandwidth' setting the entire range (edge to edge) that the filter is active in or the distance from center to one side that the filter covers?

If you have a display on it... make up a filter that drops 6dB total and starts at around 6k and ends by around 10 or 12k.

Shawn

If I remember correctly it is entire range (edge to edge)

[mikebse2a3]

Shawn/Lisa

The Bandwidth Options are:

1/10, 1/9, 1/8, 1/7, 1/6, 1/5, 1/4, 1/3, 1/2, 3/4, 1, 2, 3, 4, 5, 6, 7, 8, 10 in Octaves

[sfogg]

If it is edge to edge try 1 octave and see how that sounds. Also maybe take a peak on the RTA to see what that looks like on pink noise.

If that setting is distance of the edge from the center frequency try 1/2 octave.

Shawn

[meagain]

I don't think I'm capable of taking that whole section down -6db with the peq. I can dip it -6db in the center of the tweet mountain - but I can't suck down everything to -6db. PEQ always infuriated me and the few times I played with this thing back in Jan., used the GEQ. Reading the manual now.