More fun with crossovers

[mikebse2a3]

Mike,

"Dean has this ability to make adjustments in 1 to 2Hz increments using the PEQ section of the Behringer at low frequencies as well as small as 1/10 octave bandwith"

That isn't sharp enough. 1/10 octave bandwidth is 2hz in the first octave, 4hz in the second octave,8hz in the third....etc...etc. I mean the Q of the filter needs to be able to be set sharp enough to literally effect all of a single hz in width. That way when you target the tight room resonance issues you aren't effecting more then just the resonance.

Not much offers filters this sharp.

Shawn

Interesting Shawn.

Have you actually seen resonances that sharp in normal/average rooms in the home enviroment? What was the room construction/material like? I'm sure the more substantial construction of all room walls/floor/ceiling (Concrete for example)would setup sharper resonances than say wood studs and sheetrock type construction. I would be very interested to know what the situation was that exhibited this and how it was dealt with. I haven't really seen anything this sharp in any enviroment I've dealt with. I believe average construction methods and materials seem to have enough inherit damping to broaden most room mode resonances beyond 1 or 2 hz. I also read in F. Alton Everest (The Master Hanbook of acoustics that a typical studio room modes exhibited a bandwith of 5hz. A 1Hz wide resonance that you mention would be very difficult to deal with electronically or acoustically/which from my understanding would take extrodinary means to make a bass absorber/resonator with this sharp a Q. I have read of resonators being made of ceramic or concrete to acheive high Q values.

mike[]

[sfogg]

Mike,

"A 1Hz wide resonance that you mention would be very difficult to deal

with electronically or acoustically/which from my understanding would

take extrodinary means to make a bass absorber/resonator with this

sharp a Q."

I think the filters in my Lexicn can get down to 0.7hz in width. The

designer of the room correction system has posted quite a bit of info

about it. Meridian and Lexicon both offer automatic room correction

systems both of which set their filters based on resonance/decay time.

I believe both have said resolution/treatment down to 1hz is needed to

really effectively deal with resonances without causing other unwanted

effects.

Shawn

[mikebse2a3]

Thanks Shawn

Who is the designer of the Lexicon(AND Model?) you mention and If not to much trouble can you direct me too a few sites? Can I assume these are pretty expensive units also?

mike

Edit: I did some Googling and found this site on the Lexicon

http://www.lexicon.com/products/details.asp?ID=15

The Designer is Dr James Muller

To quote part of the article:

The too-slow decay of a resonance mode can be corrected only by applying the proper filter. The severity of a resonance mode is defined by its "Q" value. The higher the Q of a resonance, the bigger its amplitude and the slower its decay. The presence of a resonance with high Q is not a desirable condition in listening rooms. Proper treatment of a resonance requires a filter with not only the correct frequency and depth, but also the correct Q. Graphic equalizers offer neither a variety of Q values nor a fine enough frequency selection. In addition, the typical bandwidth of a graphic EQ is 1/3 octave, which is not precise enough to be effective. A good parametric equalizer might work, but you would first need to have the time, expertise and equipment to measure the room's resonances. Our analysis is capable of identifying problem frequencies to within 0.732Hz ñ 10 times more precise than a graphic EQ at the low frequencies under consideration.

------------------------------------

From what I've read so far it can get you within .732Hz of the problem frequency center but nothing is mentioned about how high the Q(Narrow The Bandwidth) capability is. Also from what I'm reading here they are using the decay rate to locate the most severe room modes for treatment but it appears to me by this information that it is basically acting as a very good auto analyzing Parametric EQ. What was most interesting to me in some of what I have read on the unit is how they mention by cleaning up these modes that clarity is improved at frequencies above the 250Hz limit of the EQ adjustments and that many listening positions will hear improvements when controlling these modes which corelates with the perception I have had with using the Behringer in Three very different Rooms/Systems that I've used it in.

This has been very interesting Shawn and any more information/articles describing the design features of the Lexicon would be appreciated.

mike[]

[sfogg]

Mike,

The Lexicon's room correction was done by David Griesinger and Jim

Muller. Along with designing the logic behind the measuring/determing

where to place the filters Jim actually did the coding of it in the

MC-12 too. The room EQ is an option for the MC-12/MC-12B processor. It

is in v4 of the software in the Lexicon and in the V5 EQ version of the

latest software.

There is some info on the Lexicon's room EQ in the MC-12 v4 users guide at:

http://www.lexicon.com/products/download-details.asp?ID=1&FileID=117

David Grisinger's home page is at:

http://world.std.com/~griesngr/

And it has *loads* of information on it. Check out:

http://world.std.com/~griesngr/asa05.pdf

For example as it covers a bunch of good topics including room EQ.

Jim Muller was online at the Lexicon forum answering a lot of questions about the room EQ when it first came out.

The Lexicon forum is forum three at:

http://forums.smr-forums.com:8080/login

Enter as a guest if you don't want to register for the site. If you

search forum three for posts with Muller in them you will find the

posts from Jim. Go back a years worth to when v4 was first released and

work your way forward. He gives lots of details on the system and even

posted a room response program he wrote in the early development of the

room eq software. He wanted to verify if the calculated theory matched

the actual measured response.

In a nutshell the Lexicon system used four microphones to measure the

decay times of every speaker in the room. In some manor it priorities

the resonances it finds by some type of psychoacoustical weighting

system and also with some sort of logic having to do with how the

resonances appear in each of the mics. It then applies filters between

19hz and 250hz to knock down the resonances it finds in the room. The

goal of the system is *not* simply to flatten amplitude response, Jim

has flat out said that is the wrong way to approach room EQ. Its goal

is to attack the long resonances that end up giving 'tubby' bass in the

room which also masks/smears details higher up in frequency in the rest

of the music. It can do 7 filters per speaker and it also applies a

sort of shelving filter when the EQ is on to make up for a perceived

loss of bass when the resonances are removed from the room. A resonance

effects the entire room (excepting null points) so reducing their

influence improves playback basically throughout the room.

Shawn

[mikebse2a3]

Thanks Shawn

I was editing my last post (figuring you might respond tomorrow) while you where replying to me.

I will definitly check this other information out and thanks for taking the time to list it all for me.

mike

[sfogg]

Mike,

I'll see if I can find the posts from Jim on the 0.7hz thing. They

built the measurement and the filters to have the same level of

precision. But Jim has said (and others agree) that

measurement/treatment to 1hz is 'good enough.'

"What was most interesting to me in some of what I have read on the

unit is how they mention by cleaning up these modes that clarity is

improved at frequencies above the 250Hz limit of the EQ adjustments and

that many listening positions will hear improvements when controlling

these modes which corelates with the perception I have had with using

the Behringer in Three very different Rooms/Systems that I've used it

in."

And there is a very good reason for that.

Contrary to somewhat popular opinion performing EQ in the room to treat

bass does *NOT* automatically limit the correction to being proper

where the microphone was placed. If you place the filters to treat the

areas with very long decays (resonances) it really does improve

performance throughout the room.

Everybody knows rooms have 'room boom' which causes huge amplitude

peaks/nulls at specific locations in the room. Those are symptoms of

the underlying problem in the room... not the problem itself. The

actual problem is the room is resonating at those specific frequencies.

And because of the resonance (sticking around longer then it should)

the room is getting standing waves which gives the amplitude problems

at a few specific places.

However if you think about it... what is a standing wave? It is

basically a wave that keeps bouncing back and forth between your walls

in a certain pattern so that it causes large amplitude shifts from

constructive/destructive interference.

When you realize that the wave is of course bouncing back and forth

within the room for longer then it should you also realize that this

problem is effecting the entire room. And you can see that in action if

you measure the room in the time domain... not just the amplitude

domain. Something like ETF will show you this in its waterfall plots

pretty easily. And if you measure from multiple points in the room you

will see the same resonances throughout the room excepting the null

points. The amplitude of the resonance will vary but the decay time of

it will be more consistent.

Therefor if you place your filters by analyzing the decay time

(resonances) improvements you make will improve the situation

throughout the room.

And the reason it makes improvements in clarity well above the range of

where you are treating is easy to understand when you think of the

correction as treating the room in time. Literally a resonance in the

bass is a note that 'keeps playing' in your room for longer then it

should. It literally is material that is playing that shouldn't be

there at all... that is pretty much the definition of distortion.

Though it isn't harmonic distortion... more like a distortion in time.

That material that is playing that shouldn't be there at all is masking

details and music the follows the original bass material which started

the resonance going. When you cut down the resonance you reduce/remove

that long decay which was masking details in your music throughout the

range.

So..... IOW.... EQing a room can work throughout the room. The key is how the filters are measured and placed though.

Shawn

[Deang]

You guys are amazing - my brain is completely fried.

Shawn - I built some 300/5500 networks and I was using the RTA to evaluate the result. It was especially helpful in determing squawker polarity. If you compare the results to the ones I posted in the 'Trachorn 400' thread (now on page 3) - you can see the improvement I garnered by making just a few changes. However, tweeter output comparisons go out the window because of the Beyma.

JC - I forgot to bring the camera home, so I can't take a digital pic of what I saw Friday while playing with the 100uF capacitor across the woofer terminals - I can tell you though: It does exactly what one would expect to happen - the response rolls off sightly faster, and there was a 3db reduction in output between 300 and 375Hz. I'll probably do this again tomorrow with the mic placed lower.

Dee - I'm very happy with the Beymas. I talked to Bob a few days who suprised me by saying he couldn't hear much if any difference between the Beyma and K-77 if the attenuation levels were matched. I'm running them with a slightly modified AA tweeter filter sitting behind a separate autotransformer swamped with a 20 ohm resistor -- which runs them at 8 ohms and 3dB of attenuation. As for 'integration', it sounds great, and as you can see from the pics on the previous page -- the transition and attenuation with the squawker is spot on. The whole top sounds much more dynamic and open with good sparkle.

[BEC]

Dean,

Why not give this tweeter a try?

Bob

[Al Klappenberger]

Bob,

The k79 tweeter is 3 dB less efficient than the K77. I don't think it will balance up in a Khorn.

Al K.

[BEC]

Al,

I haven't measured sensitivity myself, but specs claim it is 101 db in the Cornwall II. In that crossover, they use two 1.5 uF caps in series feeding the K-79. A single 2 uF cap might bring it close to what is needed for a Khorn. Haven't tried it. Just a thought.

The K-79 can also use a titanium diaphragm which may also add a bit.

Bob

[sfogg]

Al,

"The k79 tweeter is 3 dB less efficient than the K77. I don't think it will balance up in a Khorn."

What about using an autoformer in reverse? I think I read Klipsch did that in some model at some point.

Shawn

[sfogg]

Dean,

" I built some 300/5500 networks and I was using the RTA to evaluate the result."

You could likely measure their response directly without including the

drivers response in the output. Connect the Behringer across the

drivers terminals with maybe 10k resistor inline and you should be able

to see what the network itself is doing while loaded up with the

drivers..

Shawn

[Al Klappenberger]

Shawn,

Yes, you can use a transformer in reverse to set up the level to a tweeter, but the impedance seen by the network goes down rather than up. You can no longer use the swamping resistor to compensate. It really would require a complete crossover redesign.

When you test a netwrok using instruments you need to terminate the output with a resistor equal to the impedance of the driver it is desinged to drive. That wouldusually be 8 Ohms for the tweeter, 13 Ohms for a K55 and 6 in sereis with a 1 mHy inductor for the K33 woofer. In that case you measure across the 6 Ohms resistor. Using 10K Ohms will casue the response to show peaks and dips that are not really there.

Al K.

[sfogg]

Al,

"you need to terminate the output with a resistor equal to the impedance

of the driver it is desinged to drive. That wouldusually be 8 Ohms for

the tweeter, 13 Ohms for a K55 and 6 in sereis with a 1 mHy inductor

for the K33 woofer. In that case you measure across the 6 Ohms

resistor. Using 10K Ohms will casue the response to show peaks and dips

that are not really there."

I meant having it normally terminated to the drivers like always. Just

that when he connects across whatever driver put 10k inline with the

positive to the Behringer. It is a line level component that doesn't

have the different voltage inputs available like in the spectrum

analyzers. The 10k was just to protect the input to his instrument.

Shawn

[Al Klappenberger]

Shawn,

Ok.. I get what you'r dong with the 10K resistor. I really doubt it's need for the line inputs though. I doubt you could do any damage to the instrument that way.

Al K.

[sfogg]

Al,

I'm not doing it that way, just suggesting it for Dean because he is

using the RTA function in his EQ for measuring. That has a lot less

flexibility as to what it can handle on its line level inputs. Just

extra safety to be sure he doesn't damage it as it isn't meant to be

hooked up to speaker level connections.

Shawn

[Al Klappenberger]

Shawn,

Yep.. I agree..

Al K.