Current Trends in multi thousand dollar speakers


Have any of you been paying attention to the current trends in larger multi-woofer speakers that cost multiple thousands of dollars? So that many of you can follow along, i'll use the Legacy Focus 20/20's at $6K, the Piega C8's at $15K and the Aerial 20T's at $23K as points of reference. All of these have been reviewed in Stereophile over the last few months. If you're not familiar with these, all of them are vertical dynamic designs using multiple woofers in vented cabinets.

If you look at the response of of these speakers, they all have very pronounced bass peaks with elevated low frequency plateau's taking place. Of these three, the Legacy's are by far the worst of the bunch. Not only do they diverge from neutrality the most ( +7 dB peak @ 100 Hz ), their elevated bass output or "low frequency plateau" levels out at 40 Hz and at 400 Hz. That is over 3+ octaves of "extra" output that wasn't on the recording. Above 400 Hz, the output levels off with very noticeable rippling slightly above that point in the midrange and multiple large peaks with a dip up in the treble response. Below 40 Hz, the output drops like a rock. The reason that the plateau levels out at 40 Hz is because of the associated sharp roll-off associated with vents below their point of resonance.

To sum things up, this speaker, which Paul Bolin raved about in Stereophile, is anything but "smooth" or "linear" in reproduction. As can be seen in the graphs, there is a very definite "boom & sizzle" type of response taking place here. As a side note, i found that the Legacy Signature III's showed a similar large bass peak centered at appr 100 - 110 Hz, so this would seem to be a consistent design attribute / "house sound" / "family voice" to Legacy speakers.

Moving onto the Piega's, their overall response looks to be measurably smoother than the Legacy's from the midrange on up. As far as bass goes, the Piega's peak occurs at an amplitude of +5 dB's and is centered at appr 85 Hz. Their "bass plateau" is quite wide, actually just as wide as that of the Legacy. Both show the same appr "elevated output" aka "bloat" from about 40 Hz to 400 Hz. Much like the Legacy's, the Piega shows the typical sharp roll-off below 40 Hz due to the output of the vent being out of phase with that of the undamped woofer. Even though both speakers show very similar plateau's and a similar F3 ( -3 dB point ), the Legacy's bass plateau has both a higher peak and a higher average.

Moving up to the $23K price range, we've got the Aerial 20T's. Similar to the Piega's, the Aerial's are reasonably smooth in response from the mids on up with a few low amplitude peaks and dips. Side by side comparisons though, it would appear that the Piega's are a little "flatter".

When it comes to low frequency performance, the Aerial's produced a +5 db peak centered at appr 60 Hz. Of the three speakers mentioned here, the amplitude of the peak is the same of the Piega's ( +5 dB's ), which is much lower ( 66% reduction ) than that of the +7 dB peak of the Legacy's. Even with this 66% reduction of the peak amplitude at resonance compared to the Legacy's, we are still talking about a divurgence of +5 dB's here!!!

As far as the "bass plateau" goes with the 20T's, this speaker is much more linear than either of the above. While the Aerial's also level out at appr 40 Hz and drop like a rock below that point, the upper end of the bass region is MUCH smoother. Whereas the others were contributing added output up to appr 400 Hz, the Aerial's are leveling out at appr 120 Hz or so. In effect, the Aerial's appear to offer the most controlled bass with the least amount of bass colouration. Then again, they are by far the most expensive also.


As far as low frequency extension is concerned, the Aerial's resonance peak is centered the lowest of the three i.e. 60 Hz for the Aerial's vs 85 Hz for the Piega's and 100 Hz for the Legacy. Even though the Aerial's have a resonance that is 25 Hz below that of the Piega's and 40% lower in frequency than the Legacy's, all of their -3 dB points are within a very few Hz of each other. While the graph's aren't completely legible, it appears that the F3 ( -3 dB point ) for all of these speakers are right about 34 - 38 Hz or so. How do such different designs achieve similar F3's? It has to do with the tuning of the vents and the amplitude of the peaks at resonance.

By creating a huge peak at resonance, it takes longer for the amplitude of the signal to fall off. As such, the Legacy's much larger peak at resonance allows it to achieve appr the same F3 on paper that the other designs worked harder to achieve. As such, were the Legacy's designed this way because they like the sound of massive bloat? Were they designed this way so that they could claim a lower F3? Could it be a combo of the two? We'll probably never know.

What does all of this add up to? Judged in comparison to each other and strictly talking about bass linearity, the Aerial looks the best on paper by far. Why just on paper? Because we have to factor in the added gain associated with in-room response. Our ears hear the entire presentation i.e. the speaker and how the speaker loads up / pressurizes & excites the room. As such, what looks the best on paper may not be what you like the most in your room. If you're room is properly set-up, the results on paper and the results in the room should pretty well jive. That is, at least as far as frequency response & linearity go. There are a LOT of other factors going on here though, not to mention personal preference.

What happens if the room isn't properly set up? Compared to anechoic responses, all speakers will have greater output / added extension when placed in an average listening room. While specific speaker placement comes into play in terms of the extension and amount of boost, most rooms will produce maximum ouput somewhere in the 50 - 80 Hz range. Obviously, this varies with the size and shape of the room.

The net effect is that these speakers are going to produce even MORE bass than what they already show in these graphs. Not only are we picking up low frequency output from what is called "room gain" ( "cabin gain" in a vehicle ) by pressurizing the room, we are also going to be exciting the resonances of the room too. All of this adds up to GOBS more "apparent bass". Add in the fact that this bass lacks speed and control* and you've got "bloated, ill-defined thump" running rampant.

Other than that, one has to wonder just how extended the bass response of these designs would be if they didn't have such HUGE peaks? After all, the higher the peak at resonance, the lower the -3 dB point of the speaker appears to be. Do we have to add "bloat" to get extension? How do you get around all of this and still keep good sound? That's easy but it is a completely different subject : )

What i want to know is, what do you folks think about this type of performance at these price levels? Is there anything that we can learn from this? Do we see a specific trend taking place here and in other parts of the audio market? Inquiring minds want to know : ) Sean
>

* vented designs all suffer from a lack of transient response, increased ringing, over-shoot and phase problems. In this respect, a well designed port is typically "more linear" than a passive radiator.
sean
Look at the theoretical frequency response curves vs critical dampness. Look at the frequency response curves that were recorded in Stereophile reviews and it looks like they are critically damped at or higher than 0.7. If you go to R. Harley's book on high end audio you will see a few pages describing these effects. The argument being that a critcally damped speaker of 0.5 has better bass transient response but maybe a leaner sound and no hump. Where something like a critcal damped speaker with a function of 1.0 gives that nice hump at 80-120hz. Just a thought.
Hopefully, some of you folks that are interested in learning will be able to follow along with this.

To try and put things in understandable terms, the "Q" of a speaker relates to the amount of damping at resonance. The speaker with a lower Q has more damping. It is this damping that keeps the speaker from going into oscillation. Higher Q designs go into oscillation both easier and to a greater extent once excited. This is what causes both the higher output at resonance ( bass peaks ) and the poorer transient response. After all, if the driver itself is undamped and resonating, there is nothing to stop it from ringing. Ringing equates to poorer transient response and lack of definition. One note "blurs" into the next.

Think of a low Q speaker as having a lower noise floor i.e. more "inter-transient silence". A bass note will hit, stop on a dime and then play the next note. Once you hear such a system, the difference in definition, separation of notes and "speed" is quite apparent.

Having said that, most people that hear a sealed speaker with a Q of .5 or so think it sounds noticeably lean i.e. just like Clara Peller of Wendy's hamburger fame yelling "WHERE'S THE BASS" : ) For sake of reference, Dunlavy shot for a Q of .5 in his larger designs. AR aka Acoustic Research also used a Q of .5 in some of their larger models "way back when". If you do some research, you'll find that AR and Dunlavy shared many similar design philosophies.

A Q of .7 in a sealed design is much more common and still provides pretty reasonable damping / transient response. This gives you more apparent bass AND more extension without getting "sloppy". Just like the "bass hump" that designers / engineers are building into the ported speakers, the reduced amount of damping at resonance ( higher Q ) allows the speaker to look better on paper i.e. slightly lower F3. Personally, this is the highest Q that i find acceptable in a sealed design. As a side note, .7 to .8 is the "marketable" sound of a sealed speaker i.e. it still has enough bass to attract the "thump happy" folks that buy vented systems while retaining good enough transient response to not annoy those folks that crave "accuracy".

With all of that in mind, one has to take certain factors into consideration when designing a speaker. First of all, the lower Q really DOES have better transient response. Some people find this to sound somewhat "dry" though as the lack of ringing seems to cut the notes short. The truth is that they are just too used to listening to "slop" and need to get re-educated ( both ears and brain ) on the subject. Given that my Father has Legacies and i just went through and re-designed them, i'm going through this with him right now. He can hear that the bass has GOBS more speed and articulation, but he still thinks it sounds "lean". In comparison to the bloat that he had before, it does : )

On top of all of that, the Q of a system changes as the driver heats up. When throttling a speaker system, it would not be uncommon for the Q to start at .5 and climb up to a Q of .6 or possibly even a .7 or so. As such, speakers that start off with a .7 are now at a .8, .9 or possibly a 1.0 under heavy load. While this may lend more "drive" the music, it is also sloppier and less accurate. Then again, if you've had a few "liquid refreshments", you're probably less apt to notice this : )

The "temperature fluctuation" and "Q variance" are a few reasons why some designers shoot for a very low Q to start off with. Using this approach, the speaker system offers very fast / accurate transient response at low to medium levels. When pushed harder, the Q does climb, but not high enough to completely destroy the "speed" & "definition" that the listener is used to at lower volumes. If one started with a Q of .7 and ended up at 1.0 when hitting the throttle, the difference in "bloat" would be more apparent as transient response is now noticeably poorer than if one went from a Q of .5 up to a Q of .7 under load.

There are some speaker designs that utilize VERY high Q's. If i can remember correctly, the big Carver ribbons had a Q of well over 2 !!! While this seems phenomenally high ( it is ), it isn't quite as high as one might think in this specific application. If this were a more "normal" design with the woofers mounted in a box rather than free air, the bass would be attrocious. Due to being free air or "dipolar" in radiation, you get a LOT more cancellation from out of phase reflections. By introducing a HUGE peak at resonance, the bass that would normally be lost / thinned out due to cancellation is somewhat recovered due to having such a big peak. To anyone that has never heard the larger versions of these speakers, they are known for having over-powering bass. Not only is there too much bass, but what is there lacks definition and speed. If one were to take this design and substitute drivers with a lower Q, they would end up with a much better product. If this sounds like some speakers that are currently being marketed, just remember, they didn't copy the Carver's, the Carver's simply served as "inspiration" : )

Drubin: Very few companies market sealed designs. Even if the designer has enough integrity to market a great product, most reviewers and end users wouldn't know what to do with an accurate speaker. Probably the first thing that they would do was complain that it sounded "lean" because they are used to listening to "indistinct bloated thump". Sean
>

*At one point in time, AR had over 32% of the loudspeaker market and only sold sealed designs. To put that into perspective, Bose is currently the largest speaker manufacturer in the world. As far as sales go, Bose products garner appr 13% of all speaker sales made. Granted, there are more people buying speakers today than ever before, but that should tell you how "powerful" AR was in the marketplace at their peak. As a side note, AR "invented" acoustic suspension ( sealed & stuffed ) designs AND they also invented the dome driver ( tweeters and mids ). As a general rule, the average dome is FAR more "linear" or "accurate" than the average cone driver. As has been previously noted though, linearity and accuracy went out the door a long time ago, so AR products fell out of favor. Not coincidentally, this took place at the same appr time that vented designs started to flood the market. Once again, quantity of sound won out over quality of sound. Having said that, the influence of some specific AR designs are highly evident in several different product lines. Believe it or not, one can show direct correlations between specific AR designs and some of Bill Dudleston's Legacy designs. The difference here is that AR actually used cabinets that were of suitable size for their multiple woofers and retained the sealed design.
In my experience, virtually all the points made above are pretty right on. However, I would like note two reasons for the skewed (measurable and/or audible) tone balances.

I believe most tone-balance/cabinet-tuning decisions are made to offset the audible effects of a speaker having serious amounts of phase shift at every frequency, high to low.
Bass boost, no matter how obtained, is most often used to offset the sharp/overly-aggessive sound of a phase-leading treble (or equivalent woofer time lag). Doesn't happen if designers listen to live music often. You also see designers turn down the tweeter, and "pull apart" the tweeter/mid crossover point to introduce a dip (measured) but "not" audible. Until you hear it after 30 CDs... especially the ones that "work out" the crossover range.

The other factor is the type of room- large and open, vs medium and damped, etc. East coast/West coast homes... Fortunately, in hi-end, most expensive speakers can be designed for use in moderately damped rooms of 300 to 700 sq ft with listening distances between 8 and 15 feet, for a seated listener on an average sofa. That's a useful set of constraints, in my experience.

When the phase shift in a speaker is reduced to just a very few degrees of fluctuation thru the main part of the audio band, it becomes MUCH easier to
A) set the amplitude response to sound flat and to measure flat, and
B) for the speaker's tone balance to remain similar in a wide range of positions.

I think it's important to try and state why something "happens". Above it may seem like I know what I'm talking about. But can I justify A) & B) very simply?

Yes-
For A) set the amplitude response flat:
Having excellent phase accuracy/time coherence between drivers means that all the amplitude peaks and troughs from woofer and tweeter line up over each other. You get the amplitude you put in, since nothing is cancelling. No matter what the test tone or the music.
When they are out of phase, what you measure depends on HOW you measure- the type of test signal, particularly. What you hear depends on the tone range and the complexity of the music.

For B) less room dependent:
Since reducing phase shift results in a more temporally compact, succinct presentation of the music's dynamics and tones, then you can hear the speakers more clearly apart from the room's echo. Which is an echo field also then less cluttered.
If a speaker has gross phase shift from the mid down to the bass, as most of the ones discussed above do, it is usually a lag on the order of many milliseconds- which is the distance to the wall behind, or farther. With ringing at the crossover point.
Which all sound like room problems because of the long time delays. "I just moved them out 1/2" and the bass got much better!" It did- but only because of a particular interaction with the room, and your location, exaggerated by that woofer's trailing output. Slow rhythm is the clue it is not the room.

My comments here are primarily directed at dynamic speaker designs. One factor I should mention, which frankly is just as important to "coming up with" an OK tone balance is the drivers:
Multiple mids and one tweeter change tone balance w.r.t each other, as you move away.
Multiple tweeters and multiple mids change relative tone balance as you move away.
So do multiple woofers vs a single mid.
Physics and measurements support this. Apparently, most designers ignore this, looking for some other performance aspect from using multiple drivers. They should tell you the design distance/listener height.

Hope this is useful info. Nice thread.

Do all of you see a trend? Egads! That would indicate speaker design is a mature technology. No way.

Best,
Roy
Green Mountain Audio
This sure is interesting. I wish you guys would name names, not to put down anyone but to help us get a better handle on the sonic implications of the various design choices you are describing.
I am so glad Sean has brought this up. My two cents on the topic are as follows:

Most Audiophiles get obssessed with the notion of high dollar speakers in the pursuit of sonic perfection. This is why it is without a doubt the most traded and talked about component. It is a combination of sonic qualities and aesthetics. What I find most interesting about this group is that compromise in tolerated more as the price of speakers increase. It is truly an odd relationship but one I have witnessed countless times (both among Audiophile freinds and the press). As Sean points out, this is certainly the case with so called "neutral" speakers costing obscene amounts of money ($25K could be used for many more worthwhile purposes than buying loudspeakers!!).
There is really something wrong with the audiophile community when in indivduals are paying crazy prices and are still upgrading and swapping, etc. As usual many getting too involved in listening to the system and NOT the music (this is in fact now an integral part of the hobby - a real shame). For those of you paying these prices for speakers my advice is 1) buy more music!! Spend a lot less on the speakers (you can get great sound for $2000!); 2)ask yourself why Audio magazines and stores do NOT do blind testing? I think we would find widely varying results that would favor less than orthodox choices. In any event, I think many kilo buck audiophiles (of which I was guilty in past lives) should really take stock of things and listen honestly (close your eyes) forget the prestige, and marketing, and "real wood veneers" etc. etc. I most recently tried this with a fellow audiophile who was listening to a pair of his exotic mini monitors...I swapped out his speakers (KEF xQ1) with modified (and yes I can just hear you all laughing) Optimus Pro LX5 speakers. In blind testing using his choice of recordings and Stereophiles test CD, he was truly unable to accept that he was listening to Radio Shack speakers!!! Laugh as you will, the fact is with the veil removed he couldn't accept the sound which moments earlier (under blind conditions) he was marvelling at. Sad.