Purposeful Speaker
Design

During my meanwhile 20 years + speaker voyage, which is still ongoing, the building action was and is naturally always accompanied by learning. Meanwhile I have read a considerable amount of text books, magazines, web pages, papers and forum discussions. However, I have found so far that  these media (except for a few books to a certain extent) fail to put the knowledge fragments, they deliver into a closed and logical form to support
goal-oriented HiFi speaker design.

The following is meant as such a framework that puts required knowledge about speaker building into relation and order. It could be the famous 30,000 feet / 10.000m view from above if you want or a “how to” guide. The accompanying bullets and text are not intended to cover and explain all aspects of speaker building. I only try to mention the most important and obvious points as examples or things, which are not further mentioned in any literature. If you find something that is missing in the lists below you will be able anyway to put it into the correct “category”/stage of the project yourself later.
To some people this may all be obvious but I can as well envision this being helpful to others.

First comes the…

1. Application: What do you want to achieve and where ?

Speakers, their dispersion patterns to be precise, and room are inseparable. They build a unity, even if you are going to use a speaker that only minimally engages the room like a horn/waveguide based speaker, which is typically highly directional. It is possible, though, to build speakers, which play well in many rooms.

  • So, what are the acoustical properties of the room, e.g. big, small, live / dull (reverberation), standing waves ? A room that works fine during normal conversations will work fine for speakers. If you feel this is different in your room then it is not the room’s fault but the speakers are to be blamed. Room treatment can be as easy as a carpet, drapes/curtains, couch, armchair, bookshelf and even plants. Complicated and difficult to interpret measurements are not really required. Just clap your hands in the middle of the room and listen how the sound decays. That is easy, cheap and revealing.
  • Check if you should or can rearrange your furniture.
  • Is the room symmetrical or not ?
  • What is the listening distance and the distance to boundaries: walls, ceiling and “floor”? Floor here also means that you possibly have a room with no single and even level or you plan desktop speakers.
  • How many listeners are there ? Is this only for you ? For you and your partner ? Or is there even a bigger crowd to entertain ?
  • Make a simple drawing where the main reflections of the sound arrive from at the listening spot and check which ones are desired, ineffective or even bad.
  • Do you know how your speakers should sound like (and please, “good” is not enough) ? Do you want to create a certain sensation ?
    What is your goal ?
    To get an opinion maybe go out and audition as many different speakers as possible. But watch out in what rooms you are test driving speakers. They could be much bigger than your room and have very different acoustical properties like e.g. hotel or conference rooms at an audio show.
    I think ultimately the attribute “naturalness” is more or less subconsciously on everybody’s agenda (except for party speakers maybe). In this relation it is also strongly recommended to familiarize yourself with natural sounds like unamplified concerts. It really trains your listening capabilities and can even help you to find out what your speaker should be able to deliver (physics apply). Also here: the venue is not your listening room but you may be surprised how much of it can be conveyed into your listening room with the “right” speakers (and good recording techniques).
  • Psychoacoustics in total play a vital role from the first thought about speakers and onwards. Some good books actually start with this very chapter, how come ?

Up to this point, there might be some items a beginner typically fades out because of complexity, especially psychoacoustics. That is fine, one step at a time. But once you've gained experience by playing and trying, don't forget to come back to these topics. Otherwise, you will be playing a lifelong or finally give up.
Joe Jackson had it right many years ago: You Can't Get What You Want (Till You Know What You Want ...
Thanks, Joe !

2. Implementation: How does 1. translate into speaker design ?

This is the core question.
As a result most if not all of your knowledge and experience is required to answer it. Be prepared to constantly learn more about what relates to question number 2. What do you need to achieve your goals determined in 1. ?

  • Should it be sub/sat, big box, small box, no box ? If box, vented or closed ?
  • Which driver(s) with their different technologies and parameters can/should be used in what environment/application ?
  • Structural sizes matter a lot compared to wavelength so there will always be some sort of diffraction, which you should understand as well as resonances.
  • Good speakers don’t have to be expensive but you will naturally look at your budget. If you have a choice maybe you want to consider in which part of the world you are going to spend your money.
  • Learn about distortion and how to avoid it.
  • Are there several possible ways ? Take a closer look and weigh options against each other.
  • Crossover design as such belongs here as well as the question whether a passive, active or hybrid design approach is indicated.
  • What radiation pattern might be the best, wide dispersion, medium or narrow ? Constant directivity or controlled (mixed dispersion types with smooth transitions) ? This is probably the most important topic.

Everything here counts because every single aspect has its pros and cons. On the other hand not everything is equally important. There is this saying about “the vital few and the important others”. Try to get to know the vital few. I think the radiation pattern/polar response is one of them. Compromises can then be made with the important others if required. If something is unimportant, then you cannot really talk about a compromise, can you.

Then, and only then (!) comes…

3. The quest for suiting drivers

I have made this a dedicated step because there are so many people, who start a speaker project here. Often times you see questions like:
Is driver A or B of company C (or D) the better choice ?
I have seen people asking for comparisons of dome tweeters, ring radiators and ribbon tweeters or likewise for other drivers without specifying the application. So, the better choice for what ? See 1 and 2. These driver examples may be three totally different worlds ! Their dispersion characteristics have to fit to the next bigger and/or smaller transducer and the whole concept.
If the required drivers are rare then there is a very close interaction between 2 and 3. You might even find that what you need does not exist. So either wait, find a workaround or a make a compromise. If you choose to wait, double-check physics and make sure you don't expect the impossible. Otherwise you wait until the cows come home.
Once you are experienced you can start here. But then the story goes like: Hey, with this driver I can build such and that. But that means you have already gone through 1. and 2. in the past, maybe even multiple times.

Finally, you start…

4. Building and testing (verification and validation)

If you are lucky you can simulate what you are building beforehand but do not attempt building speakers without being able to measure unless you build according to somebody else’s specs and plans. Building speakers is not digital, i.e. it works vs. it does not. You will always hear “something” at the end of a day. I have heard people saying that their ears are their measurement device. Ok…so where is the scale then, which is inherent to every measurement device ? It’s like with your eyes: You cannot say precisely how long a piece of wood is just by looking at it. You might come very close but not close enough and estimating is not repeatable with constant “precision”.
Furthermore, first build it to work, “nice” comes later. Forums are full of gorgeous and great looking speakers but often no single measurement is shown. So are we looking at a working speaker or good craftsmanship ? “Nice” normally also cost additional time and money, which you are going to waste if the speaker does not work.
Once you have a proper set of measurements and you are able to interpret them you will sooner or later see automatically, what can be made “nicer” and what has to stay as it is. In my opinion form follows function but my wife does not always agree and I am not sure about your significant other. Children and pets might even require additional thoughts.

  • So, what do you hear in your room, finally ?
  • Is there coloration or anything that disturbs you ?
  • Stereo is all about phantom imaging. How is the imaging/stage/auditory scene/you name it ?
  • Does it sound as specified in 1. ?
  • How does your brain and body react upon the stimulus ? Very often the first impression is the right one. But not always because hearing is very closely coupled with adaption and you might be coming from a “wrong adaption”. This bullet is especially important if you have the property “naturalness” listed as a requirement.
    Take (mental) notes of the first impression and how it changes over time and with program material. This experience is elementary and it provides answers of paramount importance to your design. There is only one “first time”. A long break, though, e.g. vacation in a different environment with different auditory stimuli can bring back the first time impression once you’ve returned home.
    As a close second, the long term experience counts. Does it get annoying over time, is there listening fatigue or are you missing something completely ?

 

5. Close the gap (reiteration)

If there are deviations from your expectations and wishes then go back to 1. to take another look. If not, congratulations !
The more experienced you become the faster you can theoretically run through the “program”, unless you spend more time with new little details you have learned about.

 

Closing comments

Compared to other hobbies loudspeaker business is progressing relatively slowly. That means drivers of regarded companies are good for a long time (5, 7, 10 or even more years). In this light experience counts a lot because this knowledge will not be obsolete tomorrow. This also means you don't need to hurry to learn. Enjoy what you are doing and don’t be afraid of the wealth of topics to cover !
And no matter how or what your speakers will be, a “uniform” (absence of dips and peaks, although dips are less severe) on axis response and off axis behavior/radiation pattern have to be on your priority list. Not more and not less. The rest is determined in 1. and 2.
It is always the speaker in its entirety and its interaction with the room that is responsible for audible impressions. It is not a single driver with its specific technology that makes a speaker good or bad to begin with and not the room alone as already stated above.
I think building speakers is not an art. It is science but it is darn complex with many variables and even unknowns. And you might as well see that some nitty gritty details can be overlooked easily. Sometimes a miss can be compensated for and sometimes it means that another lesson has been learned. You may also find one day that being a speaker builder feels a bit like a juggler with many items in the air. Just make sure you don’t drop any of the vital ones. Maybe that is the art aspect in this game.
There is no “free lunch”, all the learning is up to you but hopefully this “big picture” helps you with your project(s).

Last updated 03-Feb-2019