[konut]

I have compared a dipole speaker with both passive and active crossovers. For 15 years I owned a pair of Magnapan Tympani IDs. This planar dipole came with internal passive crossovers and an outboard passive electronic crossover. I prefered the outboard electronic crossover. The test was not scientific, purely subjective. One of the reasons I bought these speakers was because it offered this option. I had been using active crossovers in PA situations and was firmly convinced of their superiority. I have spoken to David about this and it was his contention that passive crossovers back then are not comparable to the quality of today's designs and parts. He may be right. Or not. David has gotten me to question my former preference. The problem, as I see it, is that few, if any, speaker manufactures make a product where the option to biamp actively is offered, so direct A/B comparisons are not possible. I can understand why companies are reluctant to do this. Most consumers are not interested in active crossovers, as the added complexity and expense is, jusitfiably, daunting. Boutique companies, like David's, have an interest in making the purchase decision, and ownership, of their products as risk free and simple as possible.  The added complexity  places a great burden on the end user to 'get it right' or risk damage for which the manufacturer is not liable. Then there is the self interest of the crossover designers who have a financial interest in promoting passive crossovers. That leaves those of us who are interested in active biamping with few options. Admittedly  the market for speakers with outboard active crossovers is small, so I dont see any change in the situation soon, if ever. I have found a small company that will sell me a speaker without the crossover, at a discount, and will be reporting about it when I compleat the project.

[David Ellis]

the fact is that, for trying to determine whether or not an active x-over is better than a passive x-over, using the nao & the ht3 makes the experiment 100% totally flawed on its face, from the outset. not one statement above, re: my experience, whether or not i've heard these speakers, etc., contradicts this fact.

What you don't know, because you haven't experienced, measured or tested any of these drivers, is how they are similar or different.   What you further don't know is how a dipole is similar or different than a box speaker.   Your assertion about this a/b being completely flawed on its face comes from you lack of knowledge, experience, and exposure to nearly all aspects of this a/b test.  If you would like to maintain your assertion given your lack of knowledge concerning these speakers, okay.

Also, I think the "v" key types a "w" on your keyboard.  You consistently type the word "wariables"?

it's poor science, plain & simple.

I do have some understanding of the scientific method, and do understand the flaws in the Iowa a/b test (i.e. 2 variables).  However, this is the very best test that you or I know about.  There are no better data points for high-end speakers in an active versus passive configuration.

[doug s.]

Quote:
the fact is that, for trying to determine whether or not an active x-over is better than a passive x-over, using the nao & the ht3 makes the experiment 100% totally flawed on its face, from the outset. not one statement above, re: my experience, whether or not i've heard these speakers, etc., contradicts this fact.


What you don't know, because you haven't experienced, measured or tested any of these drivers, is how they are similar or different. What you further don't know is how a dipole is similar or different than a box speaker. Your assertion about this a/b being completely flawed on its face comes from you lack of knowledge, experience, and exposure to nearly all aspects of this a/b test. If you would like to maintain your assertion given your lack of knowledge concerning these speakers, okay.

my assertions come from the flawed testing method.  that's the only knowledge anyone needs here.  the other data, which i freely admit i am not familiar with, is not relevant.  reasonable persons can agree to disagree about this, i suppose.  

Also, I think the "v" key types a "w" on your keyboard. You consistently type the word "wariables"?

ya, i guess i have a weird typewrighter...  

Quote:
it's poor science, plain & simple.


I do have some understanding of the scientific method, and do understand the flaws in the Iowa a/b test (i.e. 2 variables). However, this is the very best test that you or I know about. There are no better data points for high-end speakers in an active versus passive configuration.

you could be correct - it wery well may be the best data point for hi-end speakers in an active versus passive configuration.  yust means there's no walid data points, imo.  

regards,

doug s.

[David Ellis]

Guys,

Please understand that I really thought the active crossover would sound better.  It "should" sound better.  I almost wanted the active projects from John K to sound better.  It had... more of everything.  It had more drivers, more surface area, more complexity.   It was a dipole!  It even had an active crossover!  John K's speaker should have sounded better, considerably better, than Jim Salk's Veractiy 3.

My current decision NOT to implement an active crossover now is purely and simply because they don't sound any better.  Further, why should anybody spend more when they don't need to spend more for their active crossover and multiple amplifiers.  The same holds true for implementing a dipole.  I can't explain it, but Dave Ellis and 25 other guys heard it the same way.

If my goal was to build... the very best big whoppin'est speaker on the planet... I'd still use a passive crossover.

IMO, the only reason to use an active crossover is to customize the bass response of the woofer in the given room to suit listener taste.

Also, really good amplifiers are really friggin expensive.  Further, more than 1 guy has agreed with me (mostly offline) that 1 "better" quality amplifier has much more impact that 2 "lesser" amplifiers.

I am certain there are a varitable plethora of guys who have implemented active crossovers on many speakers, but this is the only a/b example of an extremely refined active crossover (and drivers) being compared with an extremely refined passive crossover (and drivers).  If you know of another test, I'll be glad to read it with avid interest.

[doug s.]

Guys,

Please understand that I really thought the active crossover would sound better.  It "should" sound better.  I almost wanted the active projects from John K to sound better.  It had... more of everything.  It had more drivers, more surface area, more complexity.   It was a dipole!  It even had an active crossover!  John K's speaker should have sounded better, considerably better, than Jim Salk's Veractiy 3.

My current decision NOT to implement an active crossover now is purely and simply because ...

i agree w/what yure sayin' here dave.  and, i certainly agree that one "better" amp is better than two "lesser" amps.  but, i am still waiting to hear from someone who actively crosses over his ht3.     based upon what i have been reading about the deqx, i am pretty sure this will in fact happen in the not too distant future.

in your case, re: your 1801, i don't blame ya for doin' it the way yure doin' it.  if anyone wants to actively cross over & bi-amp yer speakers, it's not costing 'em a lot more by not being able to purchase yer speakers w/o any x-overs...  and yure right - they will be spending a lot more on outboard amps & x-overs anyways...

regards,

doug s.

[bluesky]

For what its' worth I went through a similar discussion with Greg Osborne some time ago.  It was a much, much shorter than this discussion though.  I had just read Rod Elliott's website on biamping and was all fired up to try it on one of Greg's speakers and asked him about it.

Greg Osborne's response was nearly identical to Dave's.  Greg had found that a well designed crossover was the best way to go coupled with a good amplifier.  He felt it was better to spend your available budget on getting the best amplifier you could rather than just getting more amplifiers.

After saying this though I would still like to try biamping!  I can get a 15 watt Class A amplifier kit for $45AUD each with a THD figure of just 0.00006% and I would just like to try it, despite having a reasonable amount of evidence that it probably wouldn't be worth the effort.  It appears to me that a low power amp might be best suited to this task and as the amps are affordable it would be worth a try at least.

Should I ever get around to this exercise I shall certainly post on it.  However in reality I have far too many unfinished projects to consider it just at the moment.

This has been a fascinating thread but I would like a properly set up trial.  In the near future I hope to get two of Dave's kits.  One for my Dad and one for myself, this may present an opportunity to try it out with identical speakers and amps to see what all the froth and bubble is about.  I have a more than sneaking suspicion that Dave will be proved correct in his opinions (should I be able to build it correctly!).

Best wishes to all,

Bluesky

[timbley]

Well, I had hoped that by trying an active crossover myself, I'd have settled this issue, at least in my own mind. I have only determined that the Klipsch RF-7 passive crossover doesn't work nearly as well with an inexpensive digital receiver as the Behringer DCX does to bi-amp with the same inexpensive digital receiver. Getting a second receiver to allow each of the 10" mid-woofers to have it's own amp channel yields further significant improvements.  The amps, crossover, and EQ together cost less than grand.

It may be true that the RF-7 (with it's passive crossover) powered by a well matched, truely high end amp would sound even better than my setup. But that would cost a lot more money. Or maybe there's a good integrated out there for under a grand that would sound better. It's hard for me to believe that.

The biggest problem I'm having now is there are too many choices! I can set the slopes and crossover frequencies all sorts of different ways, practically infinite. I've only tried 48db/octave slopes so far. I've tried crossover frequencies ranging from 1.6khz to 2.2kHz. They all sound good, but different. EQ is a similar  can of worms. If you like to tinker, it's a dream come true. Otherwise, it could be a nightmare. Fortunately, I enjoy messing with it.

[David Ellis]

Hey guys,

Is there an extensive discussion somewhere online about bi-amping and active crossovers?  I would likely to openly pose a question to the "audience" at-large.

[doug s.]

Hey guys,

Is there an extensive discussion somewhere online about bi-amping and active crossovers?  I would likely to openly pose a question to the "audience" at-large.

funny you should mention this; i was yust reading this thread & was thinking about you.     a coupla speaker mfr's seem to be interested in the deqx...

http://www.audiocircle.com/index.php?topic=17232&highlight=

regards,

doug s.

[David Ellis]

funny you should mention this; i was yust reading this thread & was thinking about you.  a coupla speaker mfr's seem to be interested in the deqx...

This makes VERY good sense from a marketing and crossover design perspective.  Certainly the DEQX has considerably hype from in the DIY crowd because it WILL do things better than a mediocre passive crossover.  The DEQX appears both proven, and credible from an internal perspective.  The commentary on the DEQX page is very solid.  The only part of commentary the DEQX folks conveniently omit is the added noise from processing.  There are D/A and A/D converters inside the DEQX, and copious digital circuits.  These will add noise.  Strangely, the quantity of discussion on this is on-par with the quantity of discussion about amplifiers with too much headroom and too many internal parts.  

Futher I received this assertion via email:

Things like DEQX are interesting because they allow for a x-over thathas the benefits of 1st-order, yet with much steeper slopes.

DEQX can do a 300 dB/octave crossover, yet the outputs sum perfectly
like a 1st-order crossover. Phase is corrected by the DSP. You can even
add enough delay to compensate for the different acoustic centers of a
woofer and tweeter, with the high-pass being delayed the proper amount.

This is a poor insight.

The primary advantage of a 1st order crossover is that both drivers (upper and lower) cover the same frequencies for an extended range.  Since an instrument/voice covers a range of frequencies, a 1st order crossover provides a very broad overlap.  The result is a very connected sound between the woofer and tweeter.

Any respectable passive crossover will be correct with respect to phase and amplitude.

This comment was also added.

It, of course, is an active crossover, so you get benefits from that. With the newer Class D amps, you can pack a lot of channels of
amplification in a small package, so going active isn't as big an issue
as it used to be. (I haven't heard any of these new amps, so I don't
know if they're good or not. My gut instinct is that they're not, but
plenty of people are very excited by them.

My hunch is that these amplifiers sound okay, but not great.  I have not pursued an audition with these amplifiers.  A few of my comrades have done this and I am content to trust their ears.  They claim the class D stuff can sound good, but not great.  As such, I am sure a fully tweaked class D amplifier likely sounds on-par with plenty of hifi amps.

Dave

[skrivis]

The primary advantage of a 1st order crossover is that both drivers (upper and lower) cover the same frequencies for an extended range. Since an instrument/voice covers a range of frequencies, a 1st order crossover provides a very broad overlap. The result is a very connected sound between the woofer and tweeter.

Funny, that's normally seen as their one disadvantage. The overlap produces comb-filter effects, IM, and pronounced lobing of the directivity pattern.

Speakers with 1st-order crossovers do tend to sound less like a bunch of drivers in a box, but this may also be due to other factors.

Any respectable passive crossover will be correct with respect to phase and amplitude.

Then only 1st-order crossovers are "respectable." 2nd, 3rd, and 4th-order crossovers do horrible things to the signal, and can't properly pass a transient to save their lives. Even something as well-behaved as the Linkwitz-Riley alignment screws up the phase.

http://www.rane.com/note119.html

Some info from Rane, with the most relevant comment being:
"Once again, Figure 3 shows the idealized nature of the 1st-order case. Here the result of summing the outputs together produces 0o phase shift. Which is to say that the summed amplitude and phase shift of a 1st-order crossover equals that of a piece of wire."

You can't say that about any higher-order crossover. (Unless you're doing something like DEQX does.)


There's also some interesting info at:

http://www.rane.com/note107.html

http://www.rane.com/note147.html


There is disagreement on how important phase "coherency" is, and none of the tests I've seen have been worthwhile. I feel it is very important, and the step function test reveals why most speakers can't pass a transient properly.

I'm not as concerned with power and frequency response, nor with distortion. (They're important, but they don't tell the whole story, and some speakers have very good power and frequency response with reasonably low distortion, yet sound horrible. Bose might be a good example.)

Power response is a tricky one. The claim is that much of the sound a listener perceives is a blend of direct and reflected (from the room boundaries) sound. If the off-axis response of a speaker isn't flat, then the reflected sound won't be either.

Personally, I hope to reduce reflected sound from the room. Also, if you can delay the sound enough, it will be perceived as separate from the original event and your ears can sort it out. Basically, you need a large room that isn't too "live."

Pat McGinty also made a comment something like, "Once you get the transient response right, power and frequency response fall right into place."

[David Ellis]

Funny, that's normally seen as their one disadvantage. The overlap produces comb-filter effects, IM, and pronounced lobing of the directivity pattern

I understand this too, and it is true.  Indeed the issue of a 6db filter has compromises.  Ya' can have yer' cake and eat it too.

Speakers with 1st-order crossovers do tend to sound less like a bunch of drivers in a box, but this may also be due to other factors.

I think this is because of the response overlap.  The drivers sound more connected this way - despite the distortions you mention above.

I think it's great that you bring forth the other bad elements of a 1st order crossovers.  Clearly there is no BEST solution.

Then only 1st-order crossovers are "respectable." 2nd, 3rd, and 4th-order crossovers do horrible things to the signal, and can't properly pass a transient to save their lives. Even something as well-behaved as the Linkwitz-Riley alignment screws up the phase.

? I am not so sure.  It looks to me that the 2nd order filter also matches phase quite nicely.  Am I missing something?

Or, are you simply referring to the overall phase delay from the bottom to the top of the response?

There is disagreement on how important phase "coherency" is, and none of the tests I've seen have been worthwhile. I feel it is very important, and the step function test reveals why most speakers can't pass a transient properly.

All of the tests I have read have some value.  Maybe you should conduct the perfect test?

I'm not as concerned with power and frequency response, nor with distortion. (They're important, but they don't tell the whole story,

 

IMO, there are snippets of truth everywhere, you are right.  Nothing tells the whole story.  Actually, the only thing that tells the whole story is... the whole store.  I suppose the real problem is that nobody knows the whole story.  I suppose S.L. or Joe D' might come close, but these gentlemen probably lack some aspects of information too.

Recently I have growing interest in spider technology.  A good EE told me that the spider is the most difficult part of a driver go get correct.  I think he was correct.  

I have a very cheap looking paper cone seas driver with what appears to be a very nice spider.  It appears to be the same make of spider as the W18, but with many more ridges.  The paper cone is fairly stiff too.  The midrange of this SEAS driver is amazingly good.  I attribute most of this quality to the spider.

As an aside, guys who understand the intimate chemical and physical properties of a good spider aren't willing to share their knowledge openly.  This is simply because doing so would move their meal-ticket to China.  So, the limits of my knowedge are very finite.

Personally, I hope to reduce reflected sound from the room. Also, if you can delay the sound enough, it will be perceived as separate from the original event and your ears can sort it out. Basically, you need a large room that isn't too "live."

Yep, this is tricky.  Tall walls are flimsy.  Big rooms sound clean, but generally lack dynamics.  Sheetrock has an upleasant resonance (wood sounds better IMO).  Absorption SHOULD be placed correctly.  Diffraction panels should be used too.  Seeking perfection in all of these things contributes to a very unhappy bride    .

Power response is a tricky one. The claim is that much of the sound a listener perceives is a blend of direct and reflected (from the room boundaries) sound. If the off-axis response of a speaker isn't flat, then the reflected sound won't be either.

Yep, also true   The in-room response IS critical.  Below 300hz most rooms get very screwy.  The DEQX can mitigate this, but... at what location in the room.  Rooms will have modal pile-ups scattered about the room.  Fortunately many rooms are fairly flat above 300hz.  

There are a littany of issues in hifi.  I think this is what makes my hungry mind satisfied.  I'll probably never learn all of the issues.

Your insights are very keen.