The Speaker Guru speaks: Gordon Waters

Questions for the driver designer first and then the Speaker Designer Questions

Since I've actually built a few driver designs for myself (mostly modifying existing frames, custom recones, etc), I'll take a crack at a few of these that apply:

C. Without giving the standard "weakest link" answer, what are the most important speaker components on which to spend money?
Motor design - magnetic gap linearity. Huge gains to be had, in distortion-reduction in areas EASILY audible.

D. What about crossover parts and specialty caps and the like?
Not as much. It's more of a matter of some parts being "different" than "better" in many cases. Sure, there are occasions where a combination of a driver and a better crossover part CAN create better resolution, but often times it's just a matter of the voicing matching better.

E. Can a good speaker be made with low cost parts?
Absolutely. Anyone who doubts this, listen to a Dynaco A25. And that was THIRTY YEARS AGO.

G. So what are the three main differences between a $5 and $100 woofer?
I'll address these together as they're related in my thinking. The main things that make a driver "great" is a balance of damping, a lack of motor distortion, and careful consideration to dispersion characteristics. The first requires a careful matching of materials, adhesives and coatings. The second: good cone design and motor linearity (better top-plate/pole piece design, the use of shorting rings, etc.), the third, a careful consideration to geometry. Sure, it's possible to get this done with "cheap" materials to an extent, but the amount of thought, trial tests and test manufacturing required to optimize a design. It takes a pretty good chunk of money to cover the design costs. Hence, with good drivers, a good proportion of what we're paying for is amortizing the design costs.

H. A lot of speaker manufacturers say they use "custom designs" from Vifa or Seas. Are they really custom or just small changes made to stock drivers?
Could be anything from slightly different combination of stock parts from the OEM to SPECIAL parts the manufacturer provides TO the OEM. I know of several instances, for example, where a speaker maker SUPPLIED speaker cones (from other sources) to Vifa, to assemble into their drivers. Can't get that off the shelf.

I. Other than the standard line about stiffness, what do you have to say about cone material?
Damping. More important: getting the right damping in the right place, is critical.

J. What about tweeters, metal or fabric dome and why?
Either works. When designed right, it's a matter of getting the dome to behave within its useful bandwidth.

K. It seems like expensive woofers are harder to tame than cheap woofers. How come so many expensive woofers are so hard to tame in the crossover?
Many "exotic" drivers seem to be designed with very narrow design perogatives. Seemingly, many high-end driver manufacturers, in an effort to optimize ONE parameter (i.e., say, distortion products), "throw out the baby with the bath water". Meaning the drivers no longer have good damping characteristics. Some high-end drivers are better than this than others, and I hope, as these technologies become more mature, MORE OEM driver manufacturers learn to tame these anomalies...

L. Are there any tip-off specs to know a driver is good or bad on paper?
Not really.

Q. What does the future hold for driver design?
Hopefully, VASTLY higher efficiencies. It'll take something different than the electromagnetic motor or electrostatic technologies (as we know them), but this will be something that will DRASTICALLY improve things.

1. Without giving the standard "weakest link" answer, how important would you rate speakers as component?
The speaker is the actual interface with the listening room; and as such, it is far harder to design properly than amplifiers, source units and other components. We've gotten to the point where it's almost trivially easy to find GOOD source components, but still a dicey game to find speakers that work in a room well. I would say that while it may not be necessary to budget a majority percentage of a system pricing to speakers, it IS necessary to CAREFULLY audition speakers for the task. It will usually wind up being the case, in most instances, that the speakers WILL require a LARGE percentage of the resources in a system to provide performance of equivalent quality of the units driving it.

2. How have speakers changed in the past 25 years?
Mostly in two categories- better materials and better modeling/design tools. Better materials, include not only "exotic" materials (carbon fiber, metal cones, etc), but sometimes seeming SUBTLE changes to WELL-KNOWN materials. For example, some companies are experimenting furiously with BLENDS of fibers mixed in with cellulose/paper pulp to create composite materials that have properties BEYOND what any ONE of the materials can do on its own. Also, the availability of more varied coating materials to apply to cones has expanded the range of damping options for a given material family, which opens up a lot of possibilities. Better modeling/design tools include, but not limited to, computer measurement and simulation programs. It also includes the ready availability of CAD/CAM tools and just a better knowledge of the BASIC MATH behind the fundamental operating modes of loudspeakers, cabinets and environments.

3. Do we have to spend a lot on woofers and tweeters to get good sound?
As always, it depends on the ultimate caliber of the system they are to be used with. But in general, the answer is NO. I've seen MANY designs with seeming "pedestrian" components (including some I've built myself) that have achieved amazing levels of performance. OTOH, once you get to the last degrees of current art (better motor designs, high-performance concentric drivers, and the like), the prices do tend to increase almost exponentially.

4. What are the top three design parameters you use? (or the top three things you worry about most)
1) Dynamic linearity: the ability of the speaker to accurately reproduce SOFT and LOUD sounds without changing the character of the sound in either, relative to the other. This is hard to quantify, sometimes, but taking response plots at different levels can be a starting point.
2) Efficiency, in the sense that in general, IME, the most efficient speaker, all else equal, will be the most life-like sounding speaker. Sometimes it is worth trading off for other virtues, but the less the amp has to work and the less heat the speaker has to dissipate, the better the sound. Usually in ways describes above (dynamic linearity).
3) Good time response: not necessarily completely flat phase response, but a MINIMIZATION of time delay discontinuities/shifts within the bandwidth of a speaker. This has, in my own designs lately, resulted in me using a lot of bass alignment that actually produce response down to 15 Hz (extremely low box tunings, to shift the group delay associated with roll off to a band lower than any normal signal has energy) and the usage of super tweeters to extend the response up to and beyond 40KHz (same effect of reducing group delay by placing it in the octave above normal human hearing roll off or higher).

5. Do you design on measurement or sound?
Both. It's an iterative process. Usually, I start with basic driver measurements on-baffle, and design a crossover topology. I'll listen to see if that topology "behaves itself" and make any appropriate changes. Once it seems to behave itself, I'll tweak component values for the best sound, then re-measure to insure that there are no "glaring faults" anywhere in the bandwidth. If there are, re-tuning re-commences until those problems are solved, and a final listening test will determine final voicing.

6. What to you think is that "special something" is that makes some speakers sound so good?
It's not just one something, it's a couple:
1) Focus: the ability of a speaker to act as a "near point source". This has had me working with coaxial designs lately, augmented by only drivers to extend the extreme low and high end bandwidth. Getting 100Hz-10KHz out of ONE POINT, that's the goal.
2) Dynamic "truthfulness" at ANY listening level, which includes being able to project the "energy" or "body" of musical notes as well as the transient start/stop portions properly.

7. What are your speaker audition strong points and weak points, i.e........, what problems do you find most difficult to pick up in listening tests?
System portability - being able to predict by listening in one room how the speaker will behave in another space. This is very difficult, if not impossible. So it's really paramount to test any design in a WIDE VARIETY of listening spaces. As many as possible including DELIBERATELY compromised environments.

8. Do you have a theory on crossovers or use whatever seems to work?
Depends on what the driver arrangement wants. Minimalist/first order is nice, whenever possible, but it's silly to sacrifice driver safety, distortion, in-room dispersion/energy balance to achieve it as a singular goal.

9. Do we overstress the bottom octave and area above 20K Hz?
Not really. It's amazing how much more "life-like" a speaker that TRULY has UNDISTORTED, LINEAR behavior to below 20 Hz, is in real-life listening. Same with a speaker that's flat (including no phase rotations or time delays) to beyond 20KHz. What's funny is that fixing the top-octave-and-beyond extension can actually, in MANY cases, make perceived BASS QUALITY IMPROVE. It's all about timing.

10. What one piece of advice would you give every new speaker builder?
Don't get hung up on having the best materials immediately. Get a bunch of stuff that's reasonably well behaved and EXPERIMENT. Try different crossover topologies. Try different box dampings. Try different tunings, baffles, combinations of parts. You won't ever know how to make something the BEST it can be until you know WHAT can and CAN'T be done with tuning.

11. When you get cornered at a party by a speaker builder, what is the question they ask most often?
"Where do you get your cabinets?"... there's NEVER enough available cabinet options.

12. Will we ever find the Holy Grail in sound reproduction?
Sure. Maybe. I hope so. It will probably require a completely different architecture that no one's come up with yet, though.

13. So, which is it, tubes or SS?
Both, neither. If you had asked me a year ago I'd probably have said tubes. But I've recently heard some solid-state gear that actually did a MAJORITY of the sonic virtues (openness, staging, dynamics, resolution, etc), better than pretty much ANY tube gear I've heard. It's more about the designer than the component parts.

14. What does the future hold for speaker design?
Hopefully, much more flexible and capable room-integration for sound-reproducers, in a wider variety of environments. Once we get there, it'll be MUCH easier to get "realistic" sound...

speaker icon

Copyright Peter Jay Smith 2005 Return to