What does Bob Moog thing about loudspeakers?

High efficiency, big fat dynamic range, wide bandwidth. Big dynamic range is realized through correct design of the driver’s motor system and the loudspeakers impedance match ratio to the room. In the early days of audio, inefficient loudspeaker drivers were made much more efficient through the use of horns. Horns are acoustic transformers, able to couple a high acoustic impedance like a dynamic driver, to the relatively low impedance of our atmosphere thereby radically improving power transfer. Western Electric and others were getting 50% efficiency with their monumental horn systems designed for the early theaters and yet hear we are nearly a hundred years latter happy with 1% efficiency in our homes.... But theatrical perfection and horns don’t really fit in normal homes, and even if they did there would still be fidelity issues, even in the largest of living rooms. Since horns are bandwidth limiting, only able to cover a few octaves without introducing high levels of distortion, an array of specific horns must be built to adequately cover the musical scale. The most basic is a two horn system, bass/midrange and midrange/treble. Large systems able to cover the entire musical scale were made up of five horns or more. Bandwidth optimization and splitting the music also resulted in timing issues and the introduction of the loudspeaker crossover network. The two technologies did allow for some improvement in fidelity back in the day as related to home playback. So, after roughly eighty years of band-aids and increasing complexity, Zu is continuing the search for real solutions for underlying problems. through researching the work of the pioneering giants, application of modern physics, innovation, the use of a highly developed materials pallet and making combination and assemblies repeatable through high-precision manufacturing.

Others on this similar path include the legendary Bob Moog—engineer with an ear. His comments are all relevant and applicable though in the context of what makes a good loudspeaker for plugging your Moog synthesizer into. His bottom line: “Use spec sheets as a guide, but rely primarily on your ear."

First, let's talk about acoustic power output. What we want out of a speaker is sound (acoustic power). As converters of electrical power, loudspeakers are generally inefficient. Most of the high-price juice going into a speaker cabinet winds up as heat instead of sound. Good, wide-range speaker systems typically have an efficiency of from 1/2 to 5%. That is, 100 watts of amplifier power may yield 1/2 to 5 acoustic watts of sound power. How loud is one acoustic watt? Well, a premium home music bookshelf speaker will generally burn out before it produces one acoustic watt continuously. A typical full-size professional studio monitor will produce three acoustic watts at rated power. And a supergroup's stadium sound reinforcement rig may produce a total of 100 to 1,000 acoustic watts, wide open. In a typical club environment, a speaker emitting one acoustic watt will produce a sound pressure of around 110 dB SPL onstage. Ten acoustic watts will produce an additional 10dB, which, as the textbooks say, is loud enough to hurt.
How about frequency response? The lowest F on a bass guitar is 42Hz, the C below that is 32Hz, and low A on an acoustic piano is 27-1/2Hz. Below 60Hz or so, every hertz of response is a significant addition in speaker system size, weight, and price. One person can move a moderately efficient speaker cabinet with a low-frequency cutoff of 45Hz, but it will take two people to handle a 30Hz cabinet of the same moderate efficiency. To my ears, response to 45Hz is necessary for a good "commercial" sound, while a 30 or 32Hz low-frequency cutoff adds a fullness that is sure nice to have. At the high end, you can hear the difference between 12kHz and 15kHz. A 12kHz high-frequency cutoff (and flat response below) gives a smooth, sparkly quality to bright timbres; extending the response to 15kHz adds a touch of brilliance and tinkle that can be significant in recording, or in small clubs.
Distortion becomes important when more than one pitch is played through the sound system. If you're feeding two or more keyboard instruments through the same speakers, you will have to be concerned about distortion. Unfortunately, speaker distortion characteristics are not given on spec sheets. You'll have to listen for yourself. It's generally audible when loud, low notes are played. To test for distortion, play a loud bass note along with a midrange chord. Speaker distortion will produce a "muddiness" that arises from sum and difference frequencies generated by the distortion component. In general (but not always), high-efficiency speaker systems and large speakers distort less than low efficiency speaker systems and small speakers.
Without a doubt, a typical good 100-watt guitar amp has the efficiency and stamina to put out a few acoustic watts. However, its frequency response and distortion characteristics are optimized for guitar: no significant response below 100Hz, a broad spectral "hole" around 500Hz, and sharply rising response above 1kHz, with some "warm" (low order) distortion. Guitar amps, therefore, are generally not suited for synthesizer sound reproduction. Similarly, most PA. systems are designed to make the human voice sound good. The P.A. frequency response (determined largely by the speakers) generally has a broad peak in the "presence" region of the spectrum (2–3kHz) and decidedly weak bass. Professional studio monitors, on the other hand, have more-than-adequate frequency response distortion characteristics, but often lack the stamina to produce loud, sustained, steady tones without over-heating. This is doubly true for some music speaker systems. 
Keyboard amplifiers come closest to meeting our power, frequency response, and distortion requirements. High-frequency response is sometimes a problem. Many keyboard amplifier-speaker systems are designed primarily for tone-wheel organs, electric pianos, and similar instruments with little harmonic content. Such systems rarely have adequate high-frequency response for synthesizers. However a keyboard amplifier-speaker system with good speaker response to 12kHz or so is likely to meet all of our requirements for synthesizer sound production. If the speaker system itself is efficient, a 50– to 100–watt power amplifier will produce 2–5 acoustic watts, which is plenty for rehearsal or club work, while a 200- to 400-watt power amplifier will produce upwards of 20–25 acoustic watts, which is adequate for 95% of indoor gigs. When selecting an amplifier-speaker system for your synthesizer, it is a good idea to pick a few speaker systems that are efficient (that is, they sound just plain loud when fed with the output of a modest power amplifier) and then select the speaker from that group that sounds the smoothest and fattest. Use spec sheets as a guide, but rely primarily on your ear. —Bob Moog