[MemphisJim]

Hi Dave-

I'm wondering if you have contemplated doing or have done/ messed with what is listed in the following link. It looks like it would take changes to the xover, which is way beyond me.

http://sound.westhost.com/bi-amp.htm

Jim

[David Ellis]

Good question.

I read this article thoroughly about 3 years ago and Rod's sentiment very well.

Bi-Amping doesn't necessaily require crossover changes - unless an active crossover is implemented ( very ugly).  Simple bi-amping just separates the signal after the preamp and applies 2 amplifiers.  One is for the tweeter and the other is for the woofer.  Assuming the same gain, a passive crossover can be used.

I agree with Rod Elliot on many things.  This isn't one of them.  There is indeed some theoretical advantages, but I can't hear them.  I also had a friend in Seattle a/b test and active versus passive crossover.   The active crossover obvviously required bi-amping.  The result - no audible impact.
I am not qualified to debate this issue technically.  I can only comment that in my experience, bi-amping is not financially wise.  In one of these experiments, my Bryston 3B-ST easily beat a pair of Marantz PM17 amplifiers.

[ultrachrome]

I have a small collection of amps and I'd like to play with biamping, maybe using a SS amp for the woofer, and a triode wired ST-70 for the tweeter.

I don't plan to get rid of the passive crossover, but how would I match the output of two different amps?

I have a pair of ESP L-R boards.   Could I just set the HP/LP active crossover frequencies a couple octaves past the passive crossover frequency.  For example, set the active low pass for 5kHz and the active high pass for 500Hz.  Or is there a much simpler way to do this?

[konut]

If you're separating the LP and HP of a passive crossover and feeding from two different amps, its still not true biamping. No wonder it doesnt sound any better. Its only biamping if your using an active, or passive, line level crossover. The passive line level crossover will have an insertion loss. The active will have unity gain. All the active line level crossovers I know of have level controls to achieve balance between the highs and lows. To use the passive crossover ,separated, and using two amps to feed highs and lows kind of defeats the whole purpose as your still sending a full range signal from the amp, wasting power. Those passive components cause some level of degradation. An active, properly implemented, will have less degradation. EVERY speaker that I've biamped using an active, or passive, line level crossover, has sounded better. Not even close.

[ultrachrome]

I'm just trying to figure out how to biamp through the passive crossover and adjust for the differences in gain that will likely exist if I use two different types of amplifiers for top and bottom.

I don't necessarily want to replace the crossover with an active one but rather be able to provide more power to the speaker and experiment with different amps.

I think I'm just going to try and make an in-line attenuator to put in front of the amp with the greater gain with some cheap 10k alps pots from RS.

[David Ellis]

There appear to be three subjects being discussed here.  I will try to address all of them.  

1.  Active crossovers.  These have some theoretical advantages, but with the 1801 (and most systems) there is more lost than gained IMO.  An active crossovers assumes no need for baffle step compensation, and perfect driver response.  So, in an indeal situation (i.e. an impossible situation) a "text book"  mathematical crossover will indeed divide the signal between the amplifiers and work better.  Given the 1801, the W18 driver is far from flat and there is need for baffle step compensation.

2.  Bi-Amping.  This assumes the signal will be split after the preamp and 2 amplifiers will be driven.  The output power amp signals will then be driven to the tweeter and woofer respectively.  I have heard this on a few occassions and understand the theory.  The setting was 2 Marantz PM17 amplifiers versus 1 Marantz PM17.  I heard extremely little/no impact.

3.  Quality amplifiers.  Better sounding amplifiers are... simply better. I do NOT find that 2 cheaper amplifiers will equal the sound of 1 more expensive amplifier.

I believe that MOST efforts by hobbyist guys in active crossovers and multiple amps are accomplished by guys with no ability to measure and verify their results.  I also believe these projects achieve far less than ideal results.

I heard Jim Salks  3-way compared to John K's NAO (active & Biamped) and heard differences between the speakers, but no advantage to bi-amping.

Ed West in Washington state engineered a "gizmo box" and compared an optimized active crossover and multiple amplifiers with a passive crossover with respectable drivers.  He and his audience found no difference in sound quality.  

I understand most of the relevant theory herein and have read Rod Elliot's (Elliot Sound Products) comments about Bi-amping.  These comments are very compelling, but I disagree the impact.

So, my summary and recommendation is:

1.  Given an optimized passive crossover, an active crossover and multiple amplifiers will only cost more $$$.

2.  If you wish to putter with amplifiers, simply buy/build a better one.  This will provided the best bang for the $.

From Ohio...

[konut]

I would agree wholeheartedly that active biamping is more expensive. I prefer to do my baffle step correction in the line level domain where it can be more precise(graphic or parametric equalizer) and can be easily adjusted. I know the standard objection to this is that equalization introduces phase errors but if you stop and think about it, your passive crossover does the exact same thing at a power level(watts). IF DONE CORRECTLY line level adjustment is less distorted and has less artifacts than power level equalization. Rod Elliot's article is spot on. I also agree that biamping using the passive crossover, split, is a waste of time and effort with no real gains whatsoever. As far as amps are concerned, I assume one wouldnt want a deficient amp anywhere in any system. Measuring a system is not as difficult as you'd think. In addition to discs with test tones on them, there's software out there that can be used to hook up an output from a computer to input test tones through a system. I use Tone Generator X by Okkibokki Software(freeware) in a 600mhz iBook  through an Edirol UA-3 USB interface. An easily accessible RS SPL meter is the only other component you'd need.

[David Ellis]

...

I just accomplished a brief re-read of the ESP article.  I might have missed something the first time, or maybe there is an additional piece of information added.  Either way, there is a comment at the beginning of this article that sheds new light on Rod's comments.

If you are in a position to spend $25,000 or more for a pair of speakers, then this is approaching the "cost no object" arena

In this regard, given some knowledge and squeeking the last little bit of resolution from a system... sure, bi-amping would seem to eek that last little bit of something (?) from a system.

I believe, however, that we disagree here:

Measuring a system is not as difficult as you'd think. In addition to discs with test tones on them, there's software out there that can be used to hook up an output from a computer to input test tones through a system. I use Tone Generator X by Okkibokki Software(freeware) in a 600mhz iBook through an Edirol UA-3 USB interface. An easily accessible RS SPL meter is the only other component you'd need.

I have a radio shack spl meter, and have plotted some response curves in room.  I have also attempted to ascertain the phase of a speaker knowing the crossover frequency and reversing the phase of the drivers.  This is EXTREMELY difficult to measure using test tones.  I dare say that obtaining optimal results using test tones and a Radio Shack spl meter is... impossible.  I have tried this.  It can work okay, but it's always better when using real measurement gear with a gated response.

The room modes and reflections will swamp phase and really muddy the response.  Plotting a set of in-room responses would be a nightmare.  Aside the the RS spl meter being inacurate, the time involved is extreme.  I plotted about 20 points of data on a graph, and this took me about 10 minutes.  It gave me some idea about the response, but the phase remained invisible.  In contrast.  I can do a gated time response in about 45 seconds that has about 1000 points that fully reveals the phase.  After about 100-200 uses of this method, with listening and modification intermixed, a prototype crossover is generally acceptable.  Obtaining these same results with a RS spl meter would take.... way too long - and they'd still be only mediocre IMO.

Surely using a RS spl meter and test tones are better than nothing, but the very best results will be obtained using good measurement equipment.  

I must admit only a brief experience with a RS SPL meter.  I learned very quickly that it was grossly inferior to real measurement gear.  There is, however, another fine gent who spent years with spl meters before using some real measurement gear - Rick Craig.  Rick has done some very good crossover work lately, and has a plethora of experience with spl meters and test tones.  For a 2nd opinion on this matter, you could certainly query Rick.  

IF DONE CORRECTLY line level adjustment is less distorted and has less artifacts than power level equalization.

We disagree, but that's okay.  I think both methods of filtering/equalization produce the same level of audible distortion.  However, my only good a/b test in this matter was with John K's NAO loudspeaker.  I think that John K is darn smart, and did the job correctly.  What do you think?

http://www.geocities.com/kreskovs/John1.html

Dave

[konut]

I find it curious that you cite a partial sentence on the ESP page in question, and use that to buttress your opinion, yet in the very same paragraph there are these sentences. "Biamping is not a simple tweak, and is not to be taken lightly.  Make no mistake though,  its application will improve almost any loudspeaker available, with very few exceptions." And can certainly be implimented for very much less than $25,000. As far as measurement is concerned perhaps I wasnt clear. It was not my intention to intimate that the simple measurement technique I had outlined was suitable for measurement of phase, but rather to integrate the levels of the drivers under measurement and to then equalize to a suitable flat response. As the phase of the drivers is a function of the crossover(active or passive) design, it would be known in advance just what the drivers were doing in relation to each other. BTW,  the active electronic crossovers  I consider useful for home use are borrowed from the sound reinforcement field and utilize a 4th order topology giving in phase alignment to the drivers.
        Mr. Kreskovsky is indeed a valuable resource of information and I thank you for linking that page of facinating information. Its puzzling, though, that he cites many times that he uses line level, both analogue and digital, crossovers in lieu of passives. Perhaps I missed it but there is no mention of an A/B test of line level vs passive crossovers. As to the NAO loudspeaker in question, I believe one of the primary design goals was a so called 'transient perfect' response. While I'm confident that he achieved this criterion, I question weather the dipole dispersion of the open baffle design could swamp any effect that the differences between the passive and active crossover would make. Now dont take this too seriously. I used a pair of Magnapan Tympany 1D's for many years and found the sound to be very exciting and engaging, but ultimately I grew tired of the limited sweet spot, and the difficulty of finding a perfect placement in a given room for optimum soundstaging. I ultimately came to the conclusion that a controled dispersion was preferable in both flexibility and soundstaging.
         The overwhelming majority of your customers would probably not be interested in active biamplification because, as Mr. Elliot says, its not simple and is not to be taken lightly. Also, its not within the capability of most consumers to understand and impliment. I'm sure your products sound great and, like any businessman, must be designed and made with certain price point considerations. My only disagreements with you is your blanket statement about active crossovers being 'very ugly' (whatever that means)  and of 'no audible impact'. Perhaps in the situation(s) you've encountered them, but certainly not for most of us that have experienced them to great effect.

[timbley]

Interesting topic. I went ahead and tried an active crossover because I was curious about this subject. So far, I am of the impression it really makes a significant improvement. I will at some point put the passive crossovers back in and have another listen. It's not a cut and dry comparison because I'm using much steeper slopes than the passive networks have, and I'm applying my own EQ wih the DEQ2496, vs the resonant peak filter used in the speaker, which I don't know the exact parameters of. One of the advantages of a digital crossover is that you can use 48db / octave slopes just as easily as 12 db/ octave. And, you can change the crossover frequency while listening to music. It's obvious to me that the EQ employed in the passive crossover wasn't all that sophisticated. There's a circuit to take care of the tweeter's peak, and that's about all I'm seeing.

One thing I have noticed is that some things really don't sound significantly better, like the radio right now. It sounds as flat and bland as it did with the passives.

[David Ellis]

Perhaps I missed it but there is no mention of an A/B test of line level vs passive crossovers.

There is no mention of any a/b test on John's page.

At the Iowa DIY 2004 gathering I was able to listen to Jim Salk's Veracity 3 (passive) and John K's NAO (active).  Both speakers sound extremely good, but the sound quality was equal IMO.  

There are two other comments that are very valid in many cases.

I had outlined was suitable for measurement of phase, but rather to integrate the levels of the drivers under measurement and to then equalize to a suitable flat response.

Interesting topic. I went ahead and tried an active crossover because I was curious about this subject. So far, I am of the impression it really makes a significant improvement.

In may commercial cases this will likely be VERY true.  I believe this is simply due to stamped driver parts, off-center, non-concentric, out-of-round etc etc parts in the driver motor and no/poor QC at the end of the assembly line.  It is might be due to fading electrolytic capacitors in the crossover.  Basically, I think the quality of crossover implementation and quality control in commercial loudspeakers is... junk.  However, I cannot back this up with facts, and I will not back this up with measurements.  This would land Dave Ellis is a court room, and I simply don't have the time for this.  

Dennis Murphy has measured and posted a few response curves from commercial loudspeakers www.murphyblaster.com .  I believe his commercial measurements were always sub-par.  Dennis can do this simply because he IS the audio enforcement "CZAR" in Washington D.C. .  In this regard, I have no problem believing that you might indeed iimprove upon many commercial designs.

Some folklore... my memory abou this story is a bit clouded, and my source for this information is not 1st hand.  Nonetheless, I do believe this story true, and message valid.  A few years ago there was a great guy who built small speaker called the Spica.  This speaker was very well appreciated and well engineered.  Maybe 1000 of these speakers were sold.  The "one man show" was very successful, and someone else decided to purchase the rights to this speaker for a fair sum of money.  I think $50k to $150k seems about right, but my memory is a bit foggy.  The builder agreed to continue his work on the project under the supervision of the new product owner.  Anyhow, the new owner decided the European drivers were too expensive, and sourced "different" some drivers from China.  The new drivers were not consistent, measured poor, and sounded poor.  The endeavor towards lesser value and more profit failed.  The builder was very upset with the poor quality drivers because he was concerned with quality.  This happened maybe 8-10 years ago.

Recently some drivers from China (i.e. Usher) are much better, but IMO continue to lag behind the very best drivers (i.e. Accuton, SEAS Excel & Scanspeak, Skaaning) from Europe.  I believe there are  a few reasons for this.

1.  The very best drivers require machined parts and tight tolerances.

2.  The very best drivers require extensive R&D by engineers with many years of experience.

3.  The very best drivers require "custom built" parts (i.e. surround & spider) to obtain the very best sound quality.

I believe the folks in China (with VERY few exceptions) are NOT implementing these things.

Further, it's my opinion that any commercial hifi loudspeaker in the USA that retails for less than $1k uses drivers & parts from China.  

I mention these things to illustrate my view of the commercial loudspeaker industry.  This isn't disheartening, it's business.  Making a profit in hifi is a very difficult task.  In this day, marketing is much more important than sound quality because... most folks (even audiphiles)... seldom listen to live unamplified music.  

back on track...

One thing I have noticed is that some things really don't sound significantly better, like the radio right now. It sounds as flat and bland as it did with the passives.

Yep, the source will be the weakest link in this chain.

It's obvious to me that the EQ employed in the passive crossover wasn't all that sophisticated.

Yep, I think you are right on target.  Crossover parts & design cost money.

BTW, the active electronic crossovers I consider useful for home use are borrowed from the sound reinforcement field and utilize a 4th order topology giving in phase alignment to the drivers.

4th order LR will come close, but the phase of the drivers also depends on their acoustic centers.

And, you can change the crossover frequency while listening to music.

This is very significant - good on ya'!  A good ear can certainly discern quality and correct presentation with music.  For me, this took a some time at the symphony and time at the measurement bench.  It didn't come easily, but my ears are now fairly keen.

Whew, that was long.

[konut]

Its obvious that you have your mind made up on this subject. Nevermind that there are alot of very well respected speaker companies who both design and manufacture their drivers and speakers, such as JBL, Radian, McCauley, EAW, Dynaudio, Tannoy, Electrovoice, Genelec  and other manufacturers, that use the high quality drivers you cite, that build speakers that employ both passive crossovers and bypass circuitry to remove the passive completely for use with active crossovers. We get annecdotes about listening to two different speakers in less than ideal conditions and a defunct manufacture that has a slight tangential stretch  to illustrate your less than on topic point. Cant you just admit that MAYBE you dont have enough experience with the active biamping to make a DEFINITIVE statement on the subject?

[David Ellis]

Its obvious that you have your mind made up on this subject.

 

For now, yes.  I heard possibly the best impelementation of an active crossover (the NAO) and possibly the best implementation of a passive crossover (the Veracity 3), and there was no significant impact in sound quality.

JBL, Radian, McCauley, EAW, Dynaudio, Tannoy, Electrovoice, Genelec and other manufacturers

I have little/no knowledge of Electrovoice or EAW, but have some feedback about the other products relative to the 1801.  Probably the most profound feedback came from a gent that spent his life in recordings studio's that said the 1801 midrange and highs were much cleaner than some mega (50k$?) Genalec monitors.  I have also heard 4-5 studio monitors with some variety of active crossover or integral plate amp.  They were less than impressive, but these were not high $$ speakers.  Hm, have you ever heard the Veracity 3, NAO, or the Sigfried Linkwitz Orion? IMO, these are the best speakers on the planet.

IMO, the very best speakers on the planet have drivers with stiff cones and well refined suspension and motor structures.  To my knowledge, none of the products you mention occupy this category.  They occupy  the other 99% of products with flexible cones.  These are extremely well suited for creating recording masters and mixing for the other 99% of people who use similar coned speakers.  I am confident the manufacturers you mention make very good speakers.  Certainly JBL pioneered the road towards extremely good motor structures in the 1970s.  The motor structures and tolerances of these drivers continue to be 1st rate.

Cant you just admit that MAYBE you dont have enough experience with the active biamping to make a DEFINITIVE statement on the subject?

All things expressed here are my opinion only.  You are certainly welcome to disagree.  I have no problem with this.  Heck, I sincerely hope other folks feel welcome to disagree.  Certainly there are many points of data that seem to point towards many different conclusions.  I could certainly point to most/many hifi loudspeakers manufacturers who don't use bi-amping in their $15k + speakers.  But... "Nevermind that there are alot of very well respected speaker companies who both design and manufacture their drivers and speakers using passive crossovers with exemplary results" such as Avalon, Kharma, Joeseph Audio, Burmester...

I do have another thought.  It appears most of the commercial folks you mention are focused on high spl applications.  As such, maybe any advantages are rooted in amplifier power handling ability.  Mabye the extreme power requirement in pro-sound is better suited for bi-amping??  I do know that when I a/b compared the above mentioned systems that about 3-5 watts (average) were being pumped through the speakers.  This is typical for most home users.  Certainly this is NOT typical for most pro audio applications.  What do you think?

A summary of my opinion on the matter of bi-amping to this point is:

Bi-amping is a very good theoretical proposition, but I can't hear it.

Lastly... I really don't understand what you mean by this.  What "annecdotes", what "less than ideal conditions", what "tangential stretch"?  Please explain.  I think (?) that I understand your target, but would rather understand clearly before proceeding.  Is it my rant about drivers from China?

We get annecdotes about listening to two different speakers in less than ideal conditions and a defunct manufacture that has a slight tangential stretch to illustrate your less than on topic point.

[JoshK]

I heard possibly the best impelementation of an active crossover (the NAO)

The problem with that statement is that the crossover isn't active where I would think it matters most, the mid/tweet.  It is only active between mid/bass.  Yes this frees up some juice for the mid/tweet but it doesn't alleviate the passive xo in the range where human hearing is most sensitive.  

Excuse my intrustion, but I think if one is to evaluate active vs. passive a better example needs to be done.  I don't personally have an opinion on the subject, just the testing criteria.

[konut]

Ok, I think I see the problem here. When I look for a loudspeaker, one of my requirements is that it should be able to reproduce the full frequency spectrum at realistic volume levels. Realistic meaning just the other side of 120db   http://www.zainea.com/Dynamic%20range.htm  You might find this a little excessive but every once in awhile I like to enjoy the full experience of real dynamics. What is the maximum  rated SPL of your speakers?

[David Ellis]

I skimmed the article you linked, and I don't think this is the source for our disagreement.  I am familiar with the discussion concerning classical music dynamic peaks and they are indeed extreme.  My speakers will easily produce symphony spl in-room with no compression, and then more when needed.  On a few occasions I have sat 15-20' behind the condcutor.  On one occasion I sat about 5' on behind the concuctor, but this was in a church/gym and there were only about 30 instruments.  The 1801 doesn't have any problem with this level of spl.

In my above comments ( a couple posts up) I thought you might have desired bar-room/rock concert spl.  I thought this might have been your basis for any bi-amping advantage.  Hm, maybe not.  

I must admit knowing extremely little about how pro sound systems are arranged for chest-pounding spl for several hours.  Are these systems commonly setup with bi-amping for the different drivers?

Since most folks are unfamiliar with symphony spl, I find a Briggs & Stratton reference more universally understood http://www.ellisaudio.com/soundpressure.htm .  I realize this might seem a bit simple, but most folks understand this level of spl.

I must admit there IS something special in the bass region that the 1801 simply doesn't do that happens with a big 3-way.  Sure, the 1801 makes a bass guitar sound like a bass guitar, and a tympani sound like a sympani sound.  It does okay with a Stravinski bass drum whack, but a solid kick drum hit is very dynamic and fast.  There is nothing like a big whoppin 15"-18" driver for this.  The dynamics of those bigger woofers is very real.  Listening to a kick drum on the 1801 at higher spl really pushed the limits of the very modest 7" W18 driver.

Oh, you never commented on the "tangential anectodal stretch" issue.  Could you explain?

[konut]

Active biamplification has little or nothing to do with sub standard or out of spec drivers.
Its somewhat ironic you use Linkwitz as a reference point as this sentence appears on his discussion of the crossovers for the Orion, "The amplifier takes maximum control over the motion of the speaker cone which gives a greater sense of clarity and dynamism compared to a passive dividing network between amplifier and driver". He goes on much the same as on the ESP page as to the benifits of active crossovers. I dont think Linkwitz uses substandard divers, or amps
Perhaps if you used a competent subwoofer, of which there are several, actively crossed over, you might get that extra octave of bass that the 1801 doesnt seem to manage. Actively crossed over in this way, you may be surprised that your 1801 will be even more dynamic than they already are.

[David Ellis]

I agree with the theory presented by Mr. Linkwitz, ESP and ,many others, but I just don't hear it.  

Perhaps if you used a competent subwoofer, of which there are several, actively crossed over...

Unloading the bass from the W18 clears the midrange, but not because of bi-amping.  A good actively crossed dual "subwoofer" setup is on-par with a good passive crossover that will also filter the lower hz.

Active biamplification has little or nothing to do with sub standard or out of spec drivers.

I think you missed my point.  A quick summary....  With average quality drivers and average quality crossover parts & implementation, I find it very easy to believe that a guy with an active crossover and an equalizer could do better.  

There remain a few questions that you have not clearly addressed.

1.  I think that John K is darn smart, and did the job correctly. What do you think?

2.  Hm, have you ever heard the Veracity 3, NAO, or the Sigfried Linkwitz Orion?

He goes on much the same as on the ESP page as to the benifits of active crossovers. I dont think Linkwitz uses substandard divers, or amps

I totally agree with the theory.  It makes sense.  Indeed Mr. Linkwitz uses very good source gear.

I can not and will not argue with the theoretical assertions of active crossover proponents.  These assertions make extremely good sense to me.  I agree with them, but can't hear the impact.  I was in the majority of folks at the IOWA DIY gathering.  At the end of the day all, after the tabulations, the passive crossover was deemed slightly better. ?

[EProvenzano]

Hello,

I'd like to quickly comment on my experience with bi-amping my monitors.
First, I'll admit I own a pair of Dave's 1801's and you might consider the following biased.

I bi-amped my 1801s using an Odyssey Stratos for the woofers, Jolida 801A for the tweeters and Marchand XM44 handling the XO duties.  I did this only because I had an opportunity, and was curious what all the fuss was about.

In brief, the results were very under-whelming. The speakers always sounded unnatural and I could never notice any of the advantages I read about.
I admit that my test period was short at 14 days, and I have no proper measuring implements. In my opinion, it would take significant time and some measuring equipment to do the job right.  I did experiment with various xo points and slopes, some symetrical, and others asymetrical, using 12, 24, and 48 db/octave filters.

Please let me know if I could ellaborate for anyones benefit.  My opinion is only a useful data point for those who might consider active bi-amping with minimal implements. In that case, I'd suggest against it.

Best regards,
EP

[bluesky]

Hi Folks

I am following this thread with great interest because of the amps that I currently have, or building at the moment.  

The first amp which is completed, is an Aksa 55 watt Nirvana, which I hope to upgrade to Nirvana Plus status in the not too distant future and which I know would work very well with the 1801's.

The second amp is a 15 watt Class A amp (a Silicon Chip design, an Australian Electronics magazine) which has an amazing 0.00006% THD figure and is reputedly a very fine sounding amp and the kit is relatively inexpensive.   I have built the PCB's for this but I was giving thought to biamping because of their relatively low power and making a second amp would be reasonably affordable and, I would of course use Rod Elliott's PCB's if I were to pursue this course.

I would appreciate any advice on this one although it would be reasonable to assume one side would advocate biamping and the other not.

Dave, do you think the 15 watt Class A amp could work effectively with 1801's despite their low power rating.  I have read that seem to "punch above their weight" in listening tests.

Regards

Bluesky
Brisbane, Australia