What makes a speaker good?

Building a rigid cabinet that does not resonate at some frequencies is a top priority for good sound. A really strong rigid cabinet made from relatively strong inert materials can be very expensive. There are companies that even go to the extreme of machining cabinets from solid blocks of aluminum. Some cabinets alone cost thousands of dollars.

Crossover capacitors and inductors also work best if the values are precise and the dielectric material is suitable. An inexpensive capacitor with a 10% tolerance can cost less than a dollar; a precision capacitor with a 1% tolerance and a high-quality dielectric can cost $30. Inductors can be even more expensive.

And then there are the speakers or drivers themselves.

There are tweeters that cost from $2 to $400, and the better ones can give more power output with lower distortion. They are made in many designs and types, each with its own design advantages and drawbacks.

There are midrange and bass drivers that have cones made from many different materials that have widely varying costs, each of which have different rigidities and weights. Low weight and high rigidity are the goals. The magnet structure can be made from different grades of magnetic iron alloy and this gives different magnetic strengths.

All of the above can be manufactured with varying degrees of precision and quality control, which are also major cost factors.

The bottom line is that every speaker designer wants low distortion, high sensitivity, and a product that will sell for a particular price. He has a lot of conflicting design choices to make. That is why speaker distortion, quality and prices are all over the place.

The consumer, of course, also wants the lowest distortion for his dollar, and will have varying priorities regarding speaker size, cost, appearance, and type. The trick is to find out which speaker models deliver the best performance for the buck in each price range and type.

Good sound simply means low distortion, and speakers, amplifiers, and signal sources all contribute some. Achieving low distortion in any component of the audio system involves cost decisions by engineers that affect the price of the product and the resulting sound quality.

Good post.

+1 :-)

It's all in the crossover design. Drivers that are properly integrated to operate well within their ideal range(s). A speaker consisting of $10 drivers will sound good when careful consideration is given to the crossover design. When you start talking about ultimate fidelity, everything else comes into play.

It's not ALL in the crossover design. Like any product honestly, it's a total approach - considering all the many variables involved.

Speaker design takes a person with the ability to combine science and art. During the design phase, it's almost pure science. Choosing the best drivers, developing the best cabinet, all with the final implementation in mind. The drivers, cabinet, and components used will determine the performance ceiling.

But once it's built and time for voicing, that's almost pure art. Those who voice with measurements as the driving factor are missing the boat. I can put together crossovers for a speaker in 10 different ways that would all measure flat, but only a few will have that 'magic'. But no matter how 'perfect' the crossover, the parts used simply cannot perform above beyond their own inherent limits.

In the end though it's like asking what's the best wine? There's no answer to that ... only what is best for you. Sound is just a sense, like taste afterall. But like wine drinkers, the more educated you become, the more you appreciate the improvements brought by speakers designed from the full approach and without compromise.

Great explanation

It always amazes me that people think that "technical" means all science, when "techne" the Greek term that technical comes from, denotes skill that includes art--that's often the difference in the craft of a competent designer vs a great one.

You're on the quest for ultimate fidelity. I'm simply talking about "what makes a speaker good"

If it eats its green vegetables.

This is probably stating the obvious but since I didn't see the obvious response provided, I will. Characteristics of a good sounding speaker in no particular order...

Smooth and relatively flat frequency response over a wide bandwidth. Relatively, because there is some room for voicing and preference, but large or erratic deviations are almost universally disliked. Bandwidth can be argued but maintaining control from perhaps 30Hz to 15kHz is probably a minimum.

Good power response in the horizontal plane. Probably the biggest overlooked factor in speaker design. Opinions vary, but mine says to maintain dispersion pattern from at least 100Hz to 3000Hz. Lower is better, down to lowest room mode ideally, but very difficult. Some say pattern needs to only hold down to a few hundred Hz. I disagree as this is still in the critical vocal range. Up to a few khz is better but not as critical as we are less sensitive to it.

Well behaved vertical off axis response. Not as important as the horizontal but obviously still contributes to total power response into the room. Large driver spacing with harsh lobing is a no no.

Low distortions of many types. This includes power compression. Not only does the speaker need to have distortion low at normal volume levels, but it must maintain composure at reference levels across the design bandwidth. It does little good if your speaker is ruler flat from 10 to 25000Hz if it can only maintain that response to 70db.

The relative importance of these is where serious design experience and copromise enter. Also notice I didn't say which power response is best or how to achieve it. I've heard great omnipoles and great dipoles. I also didn't mention crossover types, parts, or complexity. I've heard great speakers with expensive xo components, cheap components, and active components. Some choices obviously make it easier or more difficult to reach the necessary performance though.

Thanks for these contributions. Excellent replies all.

I believe good loudspeakers are designed for accuracy, the ability for the loudspeaker to deliver a faithful reproduction of the source material. An accurate loudspeaker is generally preferred over a loudspeaker that is designed to have a particular sound characteristic, or one that is not capable of accuracy.

How can we know what is accurate? It is simple, when you hear a sound in nature, not a reproduction from an original, by definition it is accurate. A clap of the hands, a car driving by, your child’s voice, these are all familiar sounds. Take these same sounds and reproduce them through a loudspeaker that has been designed to have a certain sound characteristic, or one that is not capable of accuracy, and those familiar sounds will now sound different.

Every second of our lives we are listening to various sounds, most of which we consciously filter out and ignore, yet subconsciously we are analyzing each and every wavelength. We rarely pay attention to the quality of a sound when it occurs in nature, instead, we focus on the content of what we are hearing.

When listening to loudspeakers, we pay as much attention to the quality of the sound being reproduced as we do to the content, perhaps even more so. Our subconscious mind knows how something should sound, and eventually the inaccuracies of a colored, or incapable loudspeaker become readily apparent, and often become a source of annoyance.

The most visually indicative representation of an accurate loudspeaker is a “flat on-axis frequency response,” but this is really only a small fraction of what goes into designing an accurate loudspeaker. Linear, and consistent off-axis response, smooth sound power response free of sharp peaks and dips, low distortion (both harmonic and IM), phase integrity, accurate impulse response free of ringing, a clean spectral decay, and a host of other considerations.

In nature, a sound is reproduced without influences from loudspeaker characteristics such as cabinet resonance, driver and crossover distortions, diffraction and baffle reflections, phase anomalies etc. These inaccuracies are audible, and smear sound quality, making a loudspeaker sound like what it is, a reproduction.

With this in mind, design goals for a more accurate speaker require less cabinet resonance, less distortion, less diffraction and reflections, and fewer phase errors. A major reduction of all the negative characteristics the loudspeaker itself imparts on the source material.

This is an edited version (to remove brand names) of what a well known designer has posted on his web site, and information I thought pertinent to the OPs question. I also happen to agree with the content. I hope the author doesn't mind the liberties I've taken with his original posts.

Jay

Found these:

Identifying Legitimately High Fidelity Loudspeakers: Myths & Facts about Crossovers
http://www.audioholics.com/education...ics/crossovers
http://www.audioholics.com/education...ils-capacitors
http://www.audioholics.com/education...p-compensation
http://www.audioholics.com/education...room-interface
http://www.audioholics.com/education...ers-conclusion

Identifying Legitimately High Fidelity Loudspeakers - the Economics of Cost Cutting
http://www.audioholics.com/education...y-loudspeakers
http://www.audioholics.com/education...de-loudspeaker

Identifying Legitimately High Fidelity Loudspeakers: Myths & Facts about Cabinets
http://www.audioholics.com/education...eaker-cabinets
http://www.audioholics.com/education...-face-stuffing



good reading all.