The Speaker Guru speaks: Jeff Bagby

Speaker Designer Questions

Jeff Bagby is a twenty-year engineer in the automotive industry. He is Senior Product Quality Engineer involved in the design, testing, and assembly of automatic transmissions. He has also been a speaker builder for 25 years and has built and sold almost 100 hundred sets of speakers in that time, including setting up a recording studio. He is the creator of the MS Excel based "Passive Crossover Designer" crossover program, as well as several other speaker design related software applications, and has written numerous tips in the area of crossover design. He lives in Indiana with his wife Chris and son David.

1. Without giving the standard "weakest link" answer, how important would you rate speakers as component?
First, speakers are the final element that convert the electrical signal into the acoustic one that we hear. And second, there is far more variation in the performance of different loudspeakers than any other component in the chain. Because of this, I think most will agree that they are among the most important, if not THE most important component you will own in your audio system.

2. How have speakers changed in the past 25 years?
In some ways very little, but there have been some small revolutions and breakthroughs. If you look back at the best speakers available in 1980 you will find that they very closely resemble good speakers today, and will still sound very good. The advances have come in the area of computer modeling and the ability to measure things we could not adequately measure before. These advances have resulted in changes in driver design that have significantly lowered distortion compared to drivers from 25 years ago. Consider the shorting rings in motors and the high X-max drivers available today as examples. There have also been quite a few advances in cone material in that time as well. Our understanding of what it takes to produce good sound has is also much clearer. In 1980 very few commercial speakers implemented what we call Baffle Step Compensation, and few took into consideration the off-axis response and crossover lobing in the design. The advent of powerful, inexpensive measurement and modeling software, along with powerful, inexpensive computers, have opened the door for much more refined designs.

3. Do we have to spend a lot on woofers and tweeters to get good sound?
No, I don't think so. A very satisfying system could be built using modest drivers if the proper care is put into the design. There are also some very nice twenty dollar midbasses and tweeters out there to work with too. I have built some inexpensive systems that simply amazed me when they were finished when I considered how little was invested in them to start with. Of course, they had their limitation too. It is all a matter of what level of performance you want from your system. You cannot achieve the highest levels of performance, ie; low distortion, high output, extended frequency response, excellent clarity etc. without using the better drivers available. But keep in mind that the law of diminishing returns applies to speaker design as much as it does anything else.

4. What are the top three design parameters you use? (or the top three things you worry about most)
For me, in designing a speaker, I find the top three design parameters to be the ones I find the most audible in establishing the character of that particular loudspeaker. Those three would be:

There are many other things that I take into consideration as well, and a good design is a combination of many different things, but you asked for my top three. As far as what I desire from a speaker the top three would be the ability to play loud and still sound satisfying while giving excellent bass extension at the same time. This combination is not the easiest, or at least the cheapest, to achieve.

5. Do you design on measurement or sound?
I always begin with measurements and computer modeling. I typically use software that I have written myself along with the pro version of CALSOD, which I have a lot of experience with. I will optimize and adjust circuit topologies until I get a design that looks good on the computer. Next I build the crossover and listen for a while. I have several tracks that I tend to use often when "voicing" a crossover. This process involves making small adjustments to the crossover with the speaker in my listening room, followed by some measurements (often with an RTA) and carrying each revision back to CALSOD to see what the simulation shows, and then listening some more. I do this until I feel it is a completed design based on first what I hear, but also what I measure and what the computer shows as well.

6. What to you think is that "special something" is that makes some speakers sound so good?
Well first, I think they have to get those three qualities right that I discussed above in number 4. Beyond that there are many things that combine. Sonically, I think it is a clear, accurate midrange. To get this an open baffle or a good, solid cabinet is paramount. Think of the best sounding commercial designs and one of the things they have in common is their robust cabinet construction. Special care to get the crossover right is another "something" that sets some designs apart. Choosing crossover points and drivers wisely, i.e., keeping drivers in their optimum operating range is an important design decision. But overall our impression of any speaker is dominated by its on axis (our listening axis, that is) frequency response. This first arrival information establishes the main character of the speaker for us. Everything else in simply layered upon that.

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?
That's a good question, and a difficult one to answer. I think the hardest thing to discern about a design is the bass quality because it is so room dependent. I have heard speakers that sounded like they had the correct bass balance in my room sound altogether different in another person's room. The thing I place the greatest emphasis on when I audition a speaker is the overall tonal balance from top to bottom, how integrated does it sound (driver to driver), and how natural do voices sound. I feel I have a good (trained?) ear for what a natural human voice should sound like, as well as for how balanced a speaker is top to mid to bottom in perceived tonal balance.

8. Do you have a theory on crossovers or use whatever seems to work?
Nothing like saying "I only use fourth order L-R crossovers" or anything like that. I use what it takes to optimize a response to my target, but I do tend to attempt to do it with the fewest number of reactive parts as possible. I use the optimizer in CALSOD to help me accomplish this goal and reduce the parts counts when possible. It is easiest to design a speaker using an acoustic fourth order L-R crossover, but if the speaker can do it, I prefer second order slopes with time aligned drivers if possible, and even prefer the idea of minimum phase systems beyond that. I tend to avoid higher order electrical crossovers unless the driver requires it (like with a ribbon tweeter). I prefer parallel crossovers over series crossovers because the options are so much broader and the opportunities for success much greater. But, the fact is, I have heard both excellent and poor speakers using all different types of crossovers.

9. Do we overstress the bottom octave and area above 20K Hz?
Yes and no. I think it depends on your goals and the type of music you like to listen to, as well as the system you are playing it over. Some people really like single driver full-range speakers that have a bandwidth of, say, 50Hz to 15kHz, and pretty much nothing on either side of that. If you have a low powered amp and like vocal-rich chamber music the speaker for you may be very different than the person who likes loud rock or symphonic music and uses their system for home theater as well. Personally, I want the foundation that deep bass extension brings, but I have heard satisfying music without it as well. People will often make the comment that many of us can no longer hear to 20kHz, so why does it matter? But I think even they would notice the difference between a system that rolls off at 10kHz and one at 20kHz. I don't see much use in pushing it much beyond this frequency though.

10. What one piece of advice would you give every new speaker builder?
Do your homework. Research. There are a lot of great sources of information out there. Do yourself a favor and study it. Ask questions on discussion forums, and then play with modeling software. It can be a great learning tool. The better prepared you are before the project begins the more likely the project will turn out to a success for you.

11. When you get cornered at a party by a speaker builder, what is the question they ask most often?
First, I have never been to a "normal party" where I met another speaker builder. If you are talking about a gathering of speaker builders then the most common questions would be "Well, be honest, how did my speakers really sound"? (All of us would like a little validation from someone else.) I also get asked about what I think of the sound of expensive capacitors and do I use them, etc? As for me the question I would like to ask is if I can ride in John K's Ferrari?

12. Will we ever find the Holy Grail in sound reproduction?
I have no idea what it would be if we were to begin looking for it. Reproduced sound will never sound like live sound. But we can come close to fooling ourselves into thinking it does sometimes. Reproduced sound is illusory, and some illusions are better than others, but it will never be perfect. I think, however, that we will continue to incrementally improve the illusion of reality, and eventually, it may become very realistic.

13. So, which is it, tubes or SS?
I only have experience with solid state, but I would like to gain some experience with tubes some day.

14. What does the future hold for speaker design?
I think the near future is in the growth and continued development of digital processing. There are so many things that can be done digitally that we can barely imagine them all - equalization, phase correction, time alignment, crossover optimization, etc. I think you will see more and more multi-amped, multi-way, digitally processed speaker systems in the near future that offer a level of performance unattainable any other way.

speaker icon

Copyright: Peter J. Smith 2004 Return to