[jswallac]

There is an expression in economics that there is no free lunch.  The meaning is that there are trade-offs in whatever you do, you can't have it all.  The pros of high efficiency speakers are obvious, they play louder with a given amount of power from your amp.  This becomes more and more important as some low power amps are gaining in popularity.  But what is being given up?  Since there are a lot of pretty cheap high efficiency speakers out there, I assume they are not costlier to build.  And I assume the many manufacturers choosing to build less efficient speakers are being rational.  So why?  Do less efficient speakers have qualities that are not present in their high efficiency counterparts?

[Kevin P]

A very good question.   Also a very intelligent way of phrasing it.

There is no free lunch.   High efficiency speakers have several positive and negative attributes.   For a given driver size you can make it more effiicent by increasing the motor strength (BL) or using a lighter moving mass (mms = cone, VC, former etc..).   The result of doing either of those has the direct consequence of raising the resonant frequency of the driver (higher Fs) or lowering the Qes of the driver.    The result of those changes are that a given driver may become more effiicient but it looses bandwidth when used in the same size enclosure.    

There are three variables that you are dealing with typically.   Box size, effiiency, and bandwidth.   If you give up or gain in one area you have to compensate in one of the other areas.    As a consequence high efficiency speakers tend to have lower bandwidth (they don't play as deep) and use larger enclosures.   They also tend to be more expensive for both of these reasons.

So what are the advantages?   High efficiency speakers use less power to achieve a given output level.   This means they can reach higher SPLs without compression due to VC heating.   As a voicecoil heats the DC resistance goes up and you have non-linear behavior (distortion).   Larger high efficiency drivers are often used in pro audio because they need high SPL levels to fill a given size venue.    A smaller less efficient driver wouldn't stand a chance of delivering the output necessary.   They also tend to need LARGE boxes and multiple drivers to cover a given bandwidth.   In the pro audio world 40Hz is considered deep.

In the home environment you don't need to achieve the same SPL levels to please most people.   It's also much easier to achieve a given SPL in a smaller space when you are located much closer to the speakers.     Size is important and smaller drivers have some benefits in terms of dispersion characteristics and cost.    

We are currently designing a small two-way system and I've been playing extensively with the tradeoff between efficiency and bandwidth.   Our goal is 80Hz with authority in a small enclosure.   The result is that you have to live with an inefficient driver.   No way around it.

[JohnR]

One thing I would venture to say is cheaper high eff speakers tend to be full/wide range drivers. These often have a rising response, and it seems that the eff rating is often quoted at the most optimistic point.

[suits_me]

I may be repeating what has already been implied, but low efficiency speakers don't only get that way by combinations of enclosure size and driver characteristics. Low efficiency speakers can get that way because they have or purposefully introduce a lot of loss through their crossover design, in an effort to pad down response irregularities or various other resonances or baffle effects.

I think this speaks to John R's post.

In other words, another frequent downside to high efficiency designs is often relatively lively boxes and uneven frequency response.

The charts I've seen of the new Triangle line with the semi horn loaded tweeters doesn't look like something I'd like, although a lot of people seem to. Most of the high efficiency, single driver designs definitely are not something I'd wish to live with - and their charts generally look really scary.

[Kevin P]

One thing I would venture to say is cheaper high eff speakers tend to be full/wide range drivers. These often have a rising response, and it seems that the eff rating is often quoted at the most optimistic point.

They are not the only ones who play that game but your right.   When add motor BL it often creates a sloping driver response.   If you are designing an active system you can use EQ to bring up the bottom end.   The problem is that if you look at a frequency distribution of the total energy in music or movies most of the energy is in the 60-400Hz range.   If you have to apply more power to a "high efficiency" driver with EQ on the bottom end (where most of the energy is required) then what are you acomplishing with a higher efficiency driver?   The answer is nothing.   You just added the extra complexity of an active EQ network and a more expensive driver to get the higher BL which requires EQ to get the same response.   You had to put more power into the driver to get the extra SPL and you are still heating the VC the same as if you used a less efficient driver with no EQ.    The driver's efficiency was gained where you don't necessarily need it as it doesn't limit the system at 1K and above.   The efficiency is needed on the bottom end where making the driver more efficient isn't going to help.

[Kevin P]

I may be repeating what has already been implied, but low efficiency speakers don't only get that way by combinations of enclosure size and driver characteristics. Low efficiency speakers can get that way because they have or purposefully introduce a lot of loss through their crossover design, in an effort to pad down response irregularities or various other resonances or baffle effects.

I think this speaks to John R's post.

In other words, another frequent downside to high efficiency designs is often ...

If you use drivers that are smooth in response to begin with then you don't need to do as much in the network.   People who worship high efficiency designs are often the same who like simple crossovers.   You can do a complex network though which has virtually no loss.   In fact with an electrical network you can create a resonant system to increase output over a narrow bandwidth and get an actual gain in output vs. no crossover.   You just give up bandwidth to do it.

[_scotty_]

I can't see any downside to high efficiency when it is achieved via multiple drivers.  You can have your cake and eat it too. Wide bandwidth,lower power requirements,and reduced distortion.  Your driver choices are certainly more limited compared to a more conventional design. The least efficient driver on average seems to be the tweeter. Which imposes restrictions on how efficient
the rest of the driver compliment can be.   I suppose one way around this is to use a line array to achieve an efficiency gain. High efficiency single driver designs seem to bring a lot of undesirable baggage along with their efficiency.
Scotty

[bpape]

The only downside that I can see is when you're listening at lower levels.  The amp/preamp in such a case are running at such low levels that the signal is much closer to the noise floor of the equipment.  With tubed designs, this is even more of an issue as the S/N is generally lower anyway.

Realistically is this enough to worry about?  Maybe not - just a thought.

[Dan Driscoll]

Folks, this is a very interesting discussion and I am learning quite a bit. Please continue.  

[miklorsmith]

Based on no scientific basis but much reading, I offer the following:

High efficiency isn't just single-driver designs.  There are examples of low-eff SD designs such as Martin King's/Bob Brines' MLTL designs, and there are examples of multi-driver high-efficiency designs.  Cascade Audio and Klipsch are two examples.  I have no experience with these.

My first foray into high-efficiency was a pair of Fostex bass reflex cabinets, built around the 206E.  I followed the TNT recipe and tweaked the drivers to the max.  The result was astounding, but not perfect.  Bass was light and the rising response from upper mid to lower treble meant rock music was verboten.  But, the things they did (do) well were so spectacular that I knew I was on the wrong road with my Gallos and sold them.  (series resistance at the positive lead helps temper response, but nowhere near flat)

(apologies to those bored by repetition of my fascinating history)

I had an opportunity to hear some Decware Radials, which did the same cool things as the Fostexes, but better, but still had some rising response issues.  Then I tried the Zu Druids and these are speakers for the ages.  Rising response gone, good bass, and mids like I've never heard.  The supertweeter is super indeed.

Point?  It can be done well.  When done well, it is the best sound I've heard.  However, it must be very difficult to achieve flat response, which turns off most would-be customers.  Great things are said about Avantgarde hornspeakers, very expensive, that are all over 100 db sensitivity and some people wouldn't trade their Klipsches for anything.

System matching becomes very important with hi-eff speakers.  Lower gain amplifiers and preamplifiers are par for the course.  Low system-noise is critical because of noise-floor issues previously mentioned.  Building a System definitely becomes more than adding parts.

My pre and main amp are in the shop right now.  I'm running a discman to a stock Clari-T to the Druids.  I'd still take the sound over just about any multi-way juggernaut I've ever heard.

[JoshK]

Very good question/topic!  I have been reading up a bit about speaker building over the last six months and from what I gathered, everything said here is true.  There is no free lunch, although some fanatics would try to convince you otherwise.

As a side tangent, I really don't get why people design their systems around an amp, rather than the speakers or the total system.  Seems ass backwards to me.  

I think when high efficiency becomes really interesting is like what scotty says, when multiple high efficiency drivers are used.  But like Kevin says the boxes get big.  As an example, one of the more intelligent ideas I have seen with high efficiency speakers is using the "full range" high eff driver as a wide range foundation and complimenting its weaknesses by augmenting its high and low frequencies.  I forget the link off hand but it used a Lowther driver on an OB supplimented with a ribbon tweeter and a real bass driver.

With pro audio amps being cheap and putting out a lot of juice these days it seems like depending on where you choose to change over to the bass foundation it makes more sense to use low efficiency traditional audio subs in a smaller enclosures with low Fs for bass and look at only high eff from midbass on up.

[Kevin P]

I can't see any downside to high efficiency when it is achieved via multiple drivers.  You can have your cake and eat it too. Wide bandwidth,lower power requirements,and reduced distortion.  Your driver choices are certainly more limited compared to a more conventional design. The least efficient driver on average seems to be the tweeter. Which imposes restrictions on how efficient
the rest of the driver compliment can be.   I suppose one way around this is to use a line array to achieve an efficiency gain ...

The downside would be cost and size. You also cannot use multiple drivers up high because you get driver-to-driver interference and comb filtering issues.

I think the best way of achieving a very sensitive design on the top-end is by horn loading the drivers. Horns for bass are HUGE. Horns that cover the midrange up are of a size that is livable. Horns tend to have a ragged response due to all the reflection surfaces in the throat and edges of the horn. You have to live with some of them as a tradeoff for the greater output.

[Kevin P]

The only downside that I can see is when you're listening at lower levels.  The amp/preamp in such a case are running at such low levels that the signal is much closer to the noise floor of the equipment.  With tubed designs, this is even more of an issue as the S/N is generally lower anyway.

Realistically is this enough to worry about?  Maybe not - just a thought.

Usually it's related to ground loops.   With really high efficiency designs it can be a real challange.    The speakers become more difficult to set-up right.

Reading my other post it sounds as though I'm anti-high efficiency.   That is far from the case.   I love horns and one of my favorite speakers is the Edgar Horns.    They have some real problems though.   For one they won't physically fit in my small room.   That is a problem.    The other would be that they cost about $10,000 so they fit my rule of high efficiency = high cost and large speaker.  

[_scotty_]

Kevin wrote:

You also cannot use multiple drivers up high because you get driver-to-driver interference and comb filtering issues.

I have to agree with you here, multiple tweeters are an impossibillity for this reason. This goes back to my comment about tweeter efficiency.
The tweeters efficiency defines the system efficiency. There isn't a plethora
of tweeters to chose from when trying to execute a high efficiency design.
The saying," amplifier power is cheap" as an excuse for an inefficient design
 ignores a basic fact that most of the time most drivers will show adverse effects from the application of very high power everything else being equal. THD and IM usually rise as excursion increases and thermal effects are also present. The inability to accurately reproduce peaks present in the  program material is a real concern and is often downplayed when less efficient speakers are discussed or reviewed.
Scotty

[bpape]

What I'm talking about is a 95db+ efficiency speaker and say a 20-30WPC amp (not a 5WPC $5k flea amp).  Try running that at 75db when you're putting out .1W or less.  The S/N will be awful on ANY tube amp - ground loop or not.  Sure, it's some sweet midrange but is the loss of dynamic range worth it?

[gonefishin]

I'll add my thoughts to the well thought out comments already posted before me.


Downside of HE designs? Plenty of them. Like others have mentioned...size and cost, limited bandwidth and troubles matching the other drivers along with the room and system are among the largest downsides.

Good HE drivers can get expensive. Heck...even some of the (decent performing) lower priced drivers are still quite expensive.

Along with the High E tweeters being a limiting factor...not going with a horn loaded midwoofer can also cost you some efficiency problems. Well, maybe not so much problems...but still at a cost.

If you've got a single amped system using high E drivers...you'll have to come down (at least) to the efficiency of the lowest driver. I say at least because you will lose some efficiency in both the crossover and box.

If you've got a bi-amped system then you may be able to let your midrange/tweeter combo stay a bit higher when using a seperate amp to power your high E midwoofer in a sealed or vented cabinet (which will be a good bit of difference lower than your mid.tweets).

But, as with running a single amp system comes with compromises...so does a bi-amp or tri-amp system. Pick your poison I suppose??? But trying to figure all this out could be a hobby within itself.

Also...High E speakers certainly may not be for everyone. That's not a problem. But...don't judge a driver by it's speaker alone. A good example of this are Altec speakers. I prefer JBL compression drivers but can also appreciate the "Altec sound". But simply using Altec drivers won't necessarily give you the Altec sound.

Think about the drivers as with "regular type" audio drivers. You can have several different sounds using different implementations of the same drivers. This is the same with higher E drivers, but not as apparant, because you don't have the vast quantities of speaker companies using the same drivers.

Most times...A high E speaker company will use it's own drivers. Examples being JBL, Altec, TAD..etc. Personally..I think JBL speakers have their own sound just like Altec does. But...if you take these drivers out...totally change the box that they may be in or the horn attached...along with the big crossover change and you've got a very different sounding speaker. Some charactoristics may remain...but the sound has certainly changed. The best example of this that I can think of are Altecs...take them out of the resonant cabinet, dampen or change the horn and an active crossover and you've got a MUCH different sound. But...you can even just change the crossover or components and end up with still a similar sound (clean) to the active...although it won't have many of the other advantages that an active may bring. But the passive can also have advantages. The point being...don't judge a driver by any certain cover.

  Downsides?  yep...but the upsides are pretty nice;)
    dan

[doug s.]

i am running relatively efficient speakers - 97db/1w/1m.  noise *can* be an issue, but it's not necessarily gonna be one.  and, my preamp has a healthy dose of gain in its own right.  i am glad that, last time i got it modded, i had its gain reduced 6db.  this was nice, even when i was running speakers 86db-90db efficient.

anyway, i have had several different amps on these speakers - a sharp digital amp, modded asl wave-8's, audio mirror 40wpc 6c33c set monoblocs, & a pair of vintage modded stromberg-carlson ap55's.  (these s-c amps, btw, are the ones that arthur salvatore of hi-endaudio.com considers perhaps the best push-pull tube amps ever - even better than the asl hurricanes.)  noise hasn't been an issue, except w/the ap55's, which have *huge* gain.  i'd suspect, that from their performance, they are pushing full power w<0.5v.  but, even w/these amps, the music still comes thru; the noise isn't an issue, when the music is playing.  there's *plenty* of dynamics, even at lower listening levels.  yes, you *can* hear noise thru the speakers from the listening seat, when nothing is being played, so this amp/preamp/speaker combo may not be for everyone.  i suspect that if i end up staying w/these amps on these speakers, i will opt for a pair of rothwell in-line attenuators (-10db) to place between the preamp & the amps.  the hi-pass setting on my outboard x-over w/this amp is -12db, w/the low pass to my subwoofers at 0db!

doug s.

[_scotty_]

I am listening to 96dB efficient speakers with zero gain in my preamp/buffer.
The amp is configured for 26dB of gain with 1volt in yielding 108watts into 8ohms. The speakers are covering  a frequency range of 20Hz to 20kHz and have an inboard passive xover. The system is in a 25x18x7.75 room. The system is all SS and noise is not a problem. Adequate peak loudness for movie soundtracks and music
playback is not an issue. Interestingly enough there is only the slightest hiss audible when your ears are right next to the tweeters. Other users of Tripath
based amps have reported higher levels of noise from their system.
Scotty

[JLM]

High efficiency usually is overstated and colored.  Coloratios due to a variety of issues including:

- Horn loading (compression situations more sensitive, most horns themselves are very poorly braced for the higher pressures)

- Lightweight cones/domes that flex more than heavier/stiffer/less efficiency ones

- Coaxial/whizzer designs commonly found that add doppler distortions

- Use of driver arrays (total yuck from combing, imaging issues)

- Beaming of high frequencies due to front horn loadings or relatively large single drivers

- General compromises made to reach higher efficiencies

As stated above high efficiency speakers usually lack deep bass, IMO a large percentage of owners/fans of high efficiency speakers have completely forgotten what deep bass is.  Note that deep bass from a properly designed horn would be the size of a room/garage and typically involves use of reinforcement concrete.

Background noise from a variety of power and system electronics factors are magnified.  Although drivers are more expensive, most if not all the cost is offset by use of smaller/simplier/cheaper amps, etc.

Advantages of high efficiency speakers are increased dynamics and allowance for simplier/smaller amps that can add purity and even more dynamics over large/complex designs.  Classic pairing of horns with small tube amps has remarkable synergy.  The value of these advantages should not be underestimated.

Modern materials science and design offer hope on many of the above counts.  Computer design allow easier found solutions to make the best use of modern high efficiency drivers with less than ideal frequency responses (the relatively new 8 inch diameter 97 dB/w/m 9 octave range Visaton B200 comes to mind).  Digital EQ/crossovers provide all kinds of signal magic without analog distortions (but I'm not a fan of room EQ).  Digital amps provide incredible grip.  Battery power combined with digital amp efficiency produce extremely quiet backgrounds (check out our own Red Wine Audio products here at AC).

IMO "the best of both worlds" is usually found in compromise.  Moderately efficient speakers (say 90 - 92 dB/w/m) with moderately powered amps (like 20 - 50 wpc) may provide the best balance of low colorations and dynamics for the average audiophile.  Of course there will always be the extremists who want 140 dB in an auditorium with 10 - 40,000 Hz output.  Those folks will require large complicated (and most certainly fidelity compromised) solutions.

[DVV]

When talking about high and low efficiency speakers, it would be good to have at least a general agreement on what is low, and what is high efficiency. That way we all know what we are talking about.

I'm a little surprised that nobody has (so far, at least) mentioned the greatest virtue of high efficiency speakers - microdynamics.

By "high effeciciency", personally I take to be anything at 96 dB/2.83V/8 ohms or above. Really high efficiency speakers of yesteryear, such as those once offered by Altec lansing, at 102 dB/2.83V/8 ohms, seem to be no more in the audio arena (while probably still present in the PA field).

Microdynamics is usually taken to denote very small, very fine details, which somehow seem to be missing from the vast majority of lower efficiency (say 92 dB/2.83V/8 ohms and below) speakers. Because they can be excited to very high levels with relatively small power, they need to have light and fast cones for bass and whatever for mid and treble (cone, dome, planar, etc), plus simpler crossovers with less power loss.

Older folks here will probably remember the sort of revolution JBL caused when they translated their 4312 monitor to the home audio market as Century 100 in 1970 (or thereabouts). They demonstrated what a high quality, high efficiency (95 dB/2.83V/8 ohms) speaker could do in the days when 60 or 70 watts of power were considere extremely powerful. Despite some obvious colorations, that speaker managed to convey a sense of presence and wonderful "color" of the music through its capability to deal with the microdynamics.

A specific example. On a lost MGM record (I have it, it has a number, but cannot be purcahsed anyhwere???), there's a place where with quality headphones you can hear the creaking of the drummer's pedal. On most speakers, it will be there, but it will be muddled, you won't be able to clearly make out what it is because it's sort of submerged under the rest of the sounds. On a high efficiency speaker, your chances of hearing and correctly identifying it are far better (although of course not guaranteed by default).

Over the last 35 years, this is one boon of high efficiency speakers nobody has been able to convince me is of little importance when compared with whatever. Not to even mention others, like the fact that needing less power, they let the amp stay much more in its optimum operating range, reduce chances of overload and/or distortion, etc. To be sure, nothing being perfect, they also have their pitfalls, but I believe these to be more marketing (price-wise) induced than by virtue of the actual design. The most common pitfall being use of low to very low quality drivers and putting it all on the card of simple efficiency.

I just can't pass up this chance - Paul W. Klipsch knew what he was talking about and what he was on about with his own products.

Cheers,
DVV