Loudspeakers: When Is Good Enough, Enough? Part 3

In this, Part 3 in the series, I'd like my 30+ year journey through speaker design to illustrate how much of the knowledge was gained over long periods, by testing combinations and by following what others were doing to improve their systems. Hopefully it will become apparent that the journey of our industry in audio has been and continues to be an ever-fascinating learning experience…

In Part 1 of this series we discussed my good fortune in being introduced to speaker design when working within a very neutral listening/speaker test room. The neutrality of that environment led to a couple of young engineers experimenting with a statement-style speaker from the get go; the Marantz Imperial 9 tri-amped version. Combined with our neutral room, that single speaker pair allowed us to;

1. Obtain a truly flat frequency response > 30Hz (in room) - 17KHz ± 1.5dB
2. Eliminate the passive crossover and go active (which greatly improves amplifier control, or damping, over a passive system and allows frequency specific system equalization)
3. Understand how power hungry low bass was for an amplifier
4. Hear a speaker almost unfettered by dynamic limitations when working into rooms under 3000 cubic feet (the IEC standard size room upon which THX Ultra standards are based)

Once we had the tri-amped version where we wanted it, we began trying, as closely as possible, to replicate the sound in a passive crossover, "saleable-to-consumers", version of the Imperial 9 tri-amp. The Marantz/Superscope management had briefly looked at our tri-amped version and made two observations; a) since Marantz was predominantly known as an electronics company no one would pay $1000(!) each which is what the tri-amped 9s would have had to retail for (in 1973). And b) the 36"H x 24"W x 16"D proportions of the tri-amp version "looked like a refrigerator" which "no wife would want in her home". So our marching orders were to reconfigure the 9's proportions and "get the price under $500 each."

After a couple of months we were able to get a pretty fair sounding representation of the Imperial 9 tri-amps in a passive version. Each of the four 3.5" midrange drivers were given their own critically damped enclosures and we kept the active version's identical crossover frequencies of 600Hz and 2800Hz using 12dB Butterworth filters on the woofer, high and low pass sections of the mids and on the tweeter (The Linkwitz-Riley crossover alignments did not exist at that time.). Using simple Zobel circuits, we designed a constant impedance-style network in which the impedance only varied between 4 to 8 ohms for the speaker's entire bandwidth. And finally, with the exception of the bottom end which could not be EQ'ed in the passive version, we were able to keep the frequency response to a respectable 40Hz (in-room) - 17KHz ± 1.5DB

Know what? When we directly A-B'ed the tri-amp 9's versus the passive version, it was immediately obvious that the passive version lacked the firm control and damping on the bottom end. And, from roughly 200Hz through 5000Hz, there was definitely a loss of the easy and open articulation on vocals and instrumental harmonics as they transitioned through the crossover regions. This is one of the reasons I've always remained a big fan of bi-amped or tri-amped speakers with electronic crossovers.

When I left Marantz in 1973 I purchased that single pair of tri-amped Imperial 9's. In the years that followed I continued to try to improve their sound. The four big 10" woofers directly coupled via 14-gauge cable, only 2 meters long each, imbued the 9's with great dynamic range. Note also, that the passive version of the 9s had a 94dB/2.83V(1 watt @ 8 ohms)/1 meter sensitivity. So the dynamic range aspect of passive Imperial 9s was never an issue either. (And though I've not heard it personally, it would appear that Audioholics' Prez Gene DellaSala 's four 10" woofer RBH T-2 Reference system must have a similar surfeit of dynamic range also.)

Once I got the 9's home and into a 3000+ cubic foot living room-with-attached-through-an-archway-dining room it became apparent that the woofers and the Marantz Model 250 amplifier used to drive them were both woefully inadequate when driven hard. One hundred twenty-five watts RMS/channel gets used up very quickly when you've got +6 dB boost dialed-in at 30 Hz. That and the fact that I loved to demo the heartbeat off of Pink Floyd's Dark Side of the Moon at full volume. (I actually cracked the window glass in the 1934 beach house I was renting at the time.)

By 1975 I was in full audiophile mode so I made a couple more changes to the tri-amped 9's. First drivers to go were the two 2" phenolic-ring cone tweeters. At the time of their design these tweeters were very good for their price/performance. But soft domes were starting to be introduced which sounded much smoother from the 9's 2800Hz roll-on crossover point, and soft domes went out to 20KHz minus only 2 dB. The only problem with soft domes of that era was that most had fairly high free air resonances (~ 2200 Hz) so it was quite easy to blow a tweeter. Blown tweeters were okay with me because Japan was constantly improving their silk dome designs in those days and it was fun to try out new samples. I eventually ended up using a Foster-Culver (then Culver-Tonegon) design which had been widely copied throughout the Far East. In fact, in the late nineties, I used a much-modified version of this same Japanese silk-dome driver in designing the M1 Active speaker for Alesis Studio Electronics. Staying at roughly $4 over its 25 year life span that particular tweeter is what has made me a die-hard soft-dome devotee. (It's been only recently that hard dome designs have emerged that I believe are genuinely excellent. I'll talk more about these two different design philosophies in part four of this series.)

In 1977 I moved into a house with a very small living room. Not being able to position the speakers well out into the room, and sitting back only about 10 feet, the standing waves generated by the 9's four 10" woofers became very apparent. Fortunately, 1977 was the year AudioControl introduced their first product, the Model 520 equalizer.

A revolutionary product for its time, the 520 had five rotary controls for each channel with frequencies centered on 36Hz, 60Hz, 120Hz, 1KHz and 15.5KHz. It also had an 18dB/octave subsonic filter at 20Hz. With the 520, I found that, though the frequencies were fixed and the filter Q's were wide, I could dial out the overbearing bass and the now way-too-forward midrange to very respectable degree. What an amazingly forward thinking product!

In 1980 I went to work for Kenwood Electronics and was able to purchase a pair of their first so-called " High Speed Direct Drive " amplifiers. The Kenwood L-09M mono amps when rated at 300W RMS at 8 ohms, 20Hz - 20KHz with no more than 0.02% THD. We measured them at 380 watts 20Hz - 20KHz into 8 ohms! The L-09Ms where among the first amplifiers which extolled the virtues of a high damping factor. They were rated as having a damping factor of "200 into an 8 ohm load and 250 into an 8 ohm load without speaker cable". (Seems specsmanship has always ruled. Just how do you connect a speaker to an amplifier without using any cable?) These 47 pound mono behemoths came with their own 1 meter length of wire which consisted of "seven flexible stranded (and insulated) wires of .23mm diameter each".

Marrying the Kenwood L-09Ms to the Imperial 9's two 10" woofers per side solved the long standing problem with the 125 W/channel Marantz Model 250 amp running out of juice so quickly. Now I started blowing up woofers! So I went back to my friends at Foster Culver and they kindly agreed to hand-build some more robust woofers; four-layer coils on a phosphor-bronze formers with high-temp lacquer pushing a stiff, thick paper cone with circumferential decoupling rings. These new 10" drivers did the trick! I could now not only play the system more loudly for longer periods but I could also hear a greater "ease" (there was still headroom left!) in the bass portion of the system when played at moderate to moderately loud levels.

What I learned from this experience is, that for unfettered-sounding dynamic range, you must have more than "just adequate" power pushing woofers whose cones are more than "just adequate" as rigid pistons (thicker, more rigid paper cone was the key in this case). It is only then that you can begin to "hear through" the previous limitations of the system and listen to a more revealing rendition of what is contained on any given recording…

Now I found I had a new problem. The system sounded tight and fast with snappy bass when the AudioControl 520 was out of the system . But the system turned to lazy mush when the 520 was re-inserted. It turns out that the 520 had op-amps in the design which totally obliterated the benefits of the faster design incorporated into the Kenwood "High Speed" L-07CII control pre-amp and L-09M power amps (each of which had discreet components). So the 520 had to go and I was now stuck once again with more exacting room placement as my only means to attenuate those pesky standing waves.

Once I found the op-amp problem in the AudioControl unit I knew what to listen for with other systems. This is how I quickly became aware that these slow slew rate op-amps were exactly the same problem inherent in most dealers' speaker switching systems of the day. That was the truly sad part of this discovery; that most mid-fi dealers were unknowingly degrading the sound of the manufacturers' loudspeakers that they demonstrated!

The example above is one very important instance wherein the tweak dealers of the day who eschewed the use of speaker switchers were correct. Their ears had told them. The tweak dealers also knew early on about proper speaker placement. Many had learned about placement by following the sixties and seventies Stereophile articles of founder and chief tester J. Gordon Holt.

So much for the background on having grown up with a system already possessing great dynamic range. Let's go forward now and see how troublesome it can be to build a system "backwards". That is, start with not enough dynamic range with a loudspeaker/electronics/room combo, and try to go forward from there…

Loudspeakers: When Is Good Enough, Enough? Part 3 - page 2

Setting Up Your Own Sound Stage or Surround Environment

Our mission with the loudspeakers, which will hopefully grace our home for many years, is to determine "when is good enough, enough? Let's see if we can pin down a checklist of what we're trying to accomplish.

1. The dynamic range capability of a speaker system is intimately related to the room's cubic volume and acoustics. It is probably the hardest parameter to know if you've got "enough of" beforehand. You must first get a specific system with all its loudspeakers, into a specific room. A system's dynamic range is directly related to the combined piston area of the drivers, the loudspeaker system's efficiency and the room's size and acoustical properties. There are two similarly related goals here; to play music as loudly as you'll ever want to hear it for a sustained period; or to play a movie as loudly as you want without the system showing any signs of strain or distress. I'm going to tell you up front, that is a very tall order.

Let's take the example of my current home system. Hopefully, the description of my struggle to attain dynamic range with a solid foundation of bass will illustrate at least one example of the obstacles that may be encountered.

Our small home was built in 1948. It has a fairly open floor plan with the living room/home theater system connected to the dining room on the right and a stairwell in the rear which leads down to the bedrooms. To the left of the listening position, a leather loveseat-size sofa, is the front door. The system is set up diagonally with my beloved 40" Mitsubishi (CRT) TV in a corner atop which are mounted the left, center and right channel speakers of a modified Infinity Modulus system. ($1800 SRP for four satellites, a center and a 300-watt 12" subwoofer). The Modulus left and right satellites consist of a single 4" mid-woofer and a ¾" hard dome tweeter and hang outboard of the Mitsubishi monitor on a special bracket that Infinity engineers designed for the Modulus. The center channel consists of two 4" mid-woofers and a ¾" hard dome tweeter.

The six side and rear surround speakers are a different story. The Modulus left, center and right speakers at the front of the system are just barely acceptable to my significant other looks and size-wise so I elected to use three pairs of JBL SoundPoint SP5, 5 ¼" two-way in- walls ($150/pair) for the left and right side surrounds and left and right rear surrounds.

The two left-side two-way systems are wired in series, in a corner, over our front door and facing at 90° angles toward each other. Similarly, the two right-side systems are wired in series, on a three-sided wall column, over our dining room table and facing at 90° angles away from each other. The rear surrounds are 180° from the main left, center, right system and feature the two final 5 ¼" in-wall systems. These two in-walls are wired separately to channels six and seven of my Harman Kardon AVR7200 receiver. These two systems face at 90° angles toward each other. (Refer to the pictures to get a better idea of my layout.)

I use two subwoofers to handle the low frequency chores. Up front, behind the Mitsu TV and equipment rack, there is an Infinity Intermezzo 1.2, a 12" sub sporting 850 watts in a sealed enclosure ($1800 SRP). To the rear, behind a leather chair but 3 feet from either wall because of the stairway is a ported Infinity Interlude IL100S 10", 250-watt subwoofer in maple ($500 SRP). I would have liked to use the much more potent, also ported, 300 watt, 12" Modulus subwoofer in the back but it again did not pass the WAF test. Too big to fit comfortably and black is a definite mismatch to the furniture in that part of the living room.

With our open floor plan I'd say that my system is working into 3000 cubic feet + of space. That figure represents is a rather small 1500 cubic foot living room. The additional 1500 cubic feet is a sum of the stairwell in the rear which goes downstairs, the dining room, a small hallway and my office. Also, there are continuous, single-pane windows which almost completely cover the right-front sidewall of the system and left-rear side windows which cover over 60% of that wall.

I suspected all these single pane windows would absorb lots of bass energy. And they do. When I performed the separate RABOS set-ups for each of these subs I found that the room was so "loose", so absorptive of low frequencies, that the 10" Interlude IL100S sub required no RABOS equalization ! If you think about it and look at the pictures showing where this sub is mounted it may start to make sense. The sub sits sort of out on a precipice, away from most surfaces, much like it would "see" if it were being tested in a full sized, 4pi anechoic chamber. So the three fully clockwise potentiometer positions (out of a possible 21 positions for each control) are: for Frequency; 80Hz, for Attenuation; 0 dB , and for % Width; 49.5%. That last parameter, % Width, says this is a very wide-band "Q" correction. My final point-by-point frequency response for the IL100S indicates 40Hz - 85 Hz, -1dB or 37 Hz - 94 Hz, -3 dB.

For the Infinity Intermezzo 1.2 sub mounted up in the front corner, behind the TV and equipment rack, the RABOS CD test disc ends up indicating that there is only a single, very broadband room peak, like the top half of a sign wave, for the entire 2 ¼ octave bandwidth of 20Hz-100Hz. To tame this very broadband peak the RABOS settings ended up being: for Frequency; 52 Hz ; for Attenuation, -14.1 dB ; for %Width, 49.5%.

These settings yield ±1dB

20 Hz - 74 Hz or ±1 dB @ 20 Hz to -3 dB @100 Hz.

From the very "loose room" equalization settings I'm using (above) I'd characterize the room another way by saying that it seems almost impossible to generate a standing wave. You sure can't feel or hear standing waves anywhere in the living space. And, acoustically anyway, that's not the blessing you might think. What is happening now is that I'm getting only marginal corner bass reinforcement from my front 850-watt Intermezzo sub and virtually no boundary reinforcement for the 250-watt Interlude which is operating out from both walls and in a 90° corner, hanging out above the descending stairwell. This means that my system is depending almost entirely on the piston area of the two subs for all bass information . The room is more bass trap (the loose single pane windows) than bass augmenter (solid walls).

When pushed really hard I have just barely enough bass quantity to match the SPLs of the rest of the system and the quality of the rear mounted, under-powered (and ported ) Interlude 10" sub noticeably deteriorates, sounding uncontrolled and tubby, as volume levels increase.

Fortunately, at the front, for most all but the very loudest soundtrack crescendos, the 850 watt, sealed Intermezzo sub stays composed, sounding quick and always tuneful on both music and effects. This is one heck of a subwoofer!

Now to the front Modulus trio. The left and right satellites as well as the center measure -3dB at around 140 Hz. Styling-dictated, small, sealed cabinets with (relatively) large 4" drivers in sealed enclosures are the culprits here. Match them with subs which do not get above 100Hz (regardless of what the back panel silkscreen says) and I've got a big hole in the response similar to most of the HTIBs on the market.

Play a Dolby Digital DVD which has an 80 Hz rolloff in the software and the hole becomes even bigger (80Hz to 140Hz; almost a full octave!). Is it any wonder then, even backed up with 100 RMS watts/channel from the H-K 7200, that I don't think the front of the system plays loudly enough? I'm missing an incredible amount of primary and second harmonic bass frequencies from 80Hz to 140Hz. It is these frequencies which add a substantive "foundation" of upper bass to satellites. Plus, it's very difficult to blend the satellites of a system such as this with the subwoofers. A frequency gap this large will pinpoint and highlight the subs' location relative to the satellites fairly easily.

The final element of my current system is the six JBL 5 ¼" two-way inwalls which I use as surrounds. Given the 140Hz low end limitations of the three Modulus' and the compromised positioning, low power and lack of control of my rear mounted sub, I'd have to say that the surround speakers and the front 850 watt Intermezzo sub are the only adequate performers of this loudspeaker menagerie.

So, as far as dynamic range is concerned I'm fighting an uphill battle. My first task then is to find a trio of left, center, rights that can get down to 80Hz (or at least 100 Hz) -3dB. Task two is find a small, yet very powerful sub (for the rear of the room) that can actually produce flat linear response from 27Hz -3dB up to 120-140 Hz. Both of these tasks are difficult to accomplish in small enclosures. But neither is impossible.

Perhaps someday I'll construct a Part 4 wherein we'll discuss frequency response, distortion and the materials which go into a speaker. All, either positively or negatively, contribute to the final sound quality. It's just a matter of understanding what particular measurable parameter is affected by which part. The human ear is an incredibly complex yet perceptive instrument. And we continue to understand more each year about the ear-brain relationship with the real world.