Your advice to speakers designers


What would it be?
I'd say - instead of building great furniture that also happens to sound good give us great sounding speakers that also happen to be acceptable furniture.
inna
Thank you, Ivan.  Great points you've made in this thread, by the way...

@inna "Though every element and the interaction of them is important, my understanding is that drivers are the single most important element. I don't want to pay thousands for nice finish, give me great drivers first, nice finish can be optional for those...whatever you want to call them."

You have a most reasonable position.  Few people think about crossovers, and that actually drove my point.

In fact, these elements divide and feed the signal to the drivers.  Not that it's my rallying cry, but it's a bit like the source is the most important element in the chain.  The drivers can only produce what gets fed to them.  The crossover determines that.  Because these components live inside the box, and don't have the visual impact of the cabinetry / aesthetics or the drivers, we presume the builder has that part of the loudspeaker equation nailed down.  Seeing more familiar, high-quality, or exotic components (wire, resistors, capacitors, coils) further puts our minds at ease.

In reality, high-level parts quality actually provides little indication the designer has any handle on crossover theory.  It's really something few people understand, and that's why I suggested folks take just half the time spent on their cabinets and learn it.

Anecdote: A friend builds a line loudspeakers that always sounded "distinct" to me.  Initially, I wondered if that reflected the midrange driver.  After building my own speakers with that part, and hearing other products that used this driver through its growing popularity, that characteristic at all.  Over the years, the iterations of his speakers piled up.  One day, for whatever reason, he showed me the schematic of a crossover.  I felt happy to see it featured a simple design, but my eye immediately fell to a part that looked off by a factor of 10.  I began to discuss it with him, and despite his background as an engineer, instantly saw his discomfort with the subject.  He told me he had no idea about the different topologies.  A mutual friend (a technician, not an engineer) gave him those values way back, he'd employed them in every speaker he built, and would never mess with them.  Even as I let him know somehow the decimal point must have got moved over one place, and that he should simply (how much effort and cost would it cost him?) listen to a pair with my suggested value, he wanted no parts of it.  But in fact, that perfectly explained the signature sound of his products.

Unfortunately, most of the folks who build loudspeakers I've met have nothing close to his acumen with mathematics, engineering, and the like.  That's not to say people don't exist who can explain the what, why, and how of their crossover.  But as the hobby has shifted to a more mom and pop type of business, a quick conversation with most people makes it clear cabinets and drivers get the lion's share of attention
So, what kind of questions should I ask a speaker designer to see if he has a good idea of crossovers and their implementation?

"So, what kind of questions should I ask a speaker designer to see if he has a good idea of crossovers and their implementation?"

In my opinion there is no "secret handshake" by which a particularly good crossover designer can be picked out in a crowd.

Personally, I don’t think I could begin to evaluate a crossover designer’s work just by asking him a few questions. And if I had to rely on questions, they would be open-ended, like, "Would you mind sharing with me as much as you are comfortable about what you do, and how you do it, and why?" If he talks a good game, well I guess that’s a start.

Far better of course would be actually listening to his speakers. If they sound good, they probably are good, and that would include the crossover.

The one thing I would caution against is, embracing an exclusive notion of what the "ideal crossover" should be and using that as a yardstick to judge speaker designs from a distance. 

Duke

dealer/manufacturer/crossover designer

Duke, my hypothetical questions would not be of someone looking for particular speakers but of someone interested in how speakers can be built and what it takes to make them sound good.
I don't know a thing about crossovers but I am always suspicious of speakers with multiple drivers. Would you say that three is enough even if the room is big and you sometimes play loudly?

"I am always suspicious of speakers with multiple drivers."

Can you explain what you mean by this?

"Would you say that three is enough even if the room is big and you sometimes play loudly?"

I’m not really inclined to single out the number of drivers as something to focus on. I would be more interested in what the designer is trying to do, and how he goes about doing it, which in turn may or may not call for a particular number of drivers. I suppose there is some correlation between number of drivers and how loud a system will go, but the specifics of a given design matter far, far more than any generalities we might make regarding numbers of drivers.

Let me give an example of what I mean, pardon me from drawing from my own experience but at least this way it isn’t some arbitrary hypothetical: In my opinion, two worthwhile design goals are, good re-creation of the timbre of instruments, and good dynamic contrast. That’s "what I’m trying to do."

The "how" part is, I use driver combinations that a) give me a radiation pattern that is fairly directional and fairly uniform over most of the spectrum; and that b) have about 10 dB of excess thermal power handling capability above whatever in-room peaks we would normally expect. To get more specific, I use prosound drivers, with the high-frequency section employing a constant-directivity waveguide. The woofer crosses over to the waveguide at the frequency where its radiation pattern has narrowed to match that of the waveguide. (A "waveguide", in this context, is a type of horn, designed with pattern control and low coloration as the priorities, rather than acoustic amplification.)

The "why" part is, getting the reverberant field to have the same spectral balance as the first-arrival sound supports good timbre; and if the drivers are only seeing about 10% of their rated power on peaks, they will usually have negligible thermal compression, so the dynamic contrast on the recording is preserved. These aren’t the only things I care about of course, but they’re high on the list.

How many drivers does it take to do what I just described? In this case it would take two to do it pretty well and four to do it even better, and I’ve been known to use more than that, but what would the number of drivers tell you if you had no idea what they were or what I was trying to do?  My point being, it's the SPECIFICS of a given design that matter most. 

Duke