Audio Reproduction Systems

By William "Bill" Yaple | Posted September 30, 2012 at 12:35AM

Ever wanted to know all about audio reproduction, but were afraid to ask? Do Watts and SPL sound like terms normal people would never use in public? You're in luck! The head audio geek at TargetPC, or Bill as I'm known to some, has decided to put fingers to the keyboard to demystify the wonderfully confusing world of sound. Why not follow along as I tackle a few difficult concepts and simplify manufacturer specifications that leave most of the world reaching for an air sickness bag.

Terminology

What would technology be without its beloved technobabble? Specifications, or geakspeak to many, is just the verbage that must be used when conveying complex ideas so all is clear between engineers. Inner audio circles label their language as "dB." In my 20 years in professional audio, I've actually had people come up to me and spout, "Yeah man, just got a hot new 24 channel console, let's talk dB." Without touching on the social ramifications of such cryptic communication, here's a very condensed list of words and abbreviations to get anyone noticed in a crowd. Use these sparingly as prolonged exposure can do odd things, like increase the distance between you and your spouse or worse, attract more dB'ers.

 
 
 
Term
Definition
dB
Decibel or 1/10th of a Bell
SPL
Sound Pressure Level(measured in dB)
Volt
Unit of Pressure (denoted as E or V)
Watt
Unit of Power (denoted as P or W)
Amp
Unit of Current (Amperage, denoted as I)
THD
Total Harmonic Distortion (a percentage)
S/N
Signal to Noise Ratio (measured in dB)
FR
Frequency Response (measured in Hz & dB)
Hz
Hertz (frequency or recurrence)
Z
Impedance (measured in Ohms)


A few modifiers are necessary as well. If I want one thousand of something, I'd say "kilo" as in kilohertz (1000 Hz). If I'd want one thousandth of something, I'd say "milli" as in millivolt (mV). Glance at a wider range if you prefer; this list is far from exhaustive.

Abbreviation
Fraction or Multiplier
p
pico (0.000000000001)
u
micro (0.000001)

m

milli (0.001)
k
kilo (1,000)
M
Mega (1,000,000)
G
Giga (1,000,000,000,000)


Amplifiers

The last time you requested someone to "turn it up," you could have said, "hey man, amplify that sound." An amplifier is a set of components designed to make something "more" or louder if you prefer. Similar to an automobile accelerator pedal, punching it equates to crankin' the knob off. So what's important when browsing your local electronic retailer? Is more always better?

Power Output

Watts are easy as PIE. No, not the fork and plate kind, the physics kind. Power (in Watts) equals current (I or A) times voltage (E or V). Since the formula is a linear ratio, increasing the power would mean an increase in voltage and/or current. All power amplifiers generate distortion, which are errors created when boosting the signal from a lower state to a higher state.

Distortion

Typically measured as a percentage, THD refers to what fraction of the output is not exactly like a larger mirror image of the original. For example, 1% THD would suggest that 1 part out of every 100 contained amplification errors. Obviously, a larger amount of errors would raise the THD and the unlistenability of the sound signal.

For most listeners, 3% THD is the threshold whereby distortion becomes noticable with most types of music. At the 10% level hard clipping, which manifests itself with audible cracks and screeches usually provokes the hand to fly up and turn the knob down to something less offensive. Barely audible even with test tones, (and who listens to test tones) is distortion in the 0.05-0.20% range.

Amplifier Classifications

This is a toughie. There are two primary modes or types of audio amplification--class A and class AB. While I would love nothing better than to talk dB about these two types for the next twenty years, the gist is that class A amps are expensive, sound great, very inefficient, and impractical. And I love to design and build them. Class AB amps are inexpensive, good sounding, moderately efficient, and practical. Probably 99.9999% of all amps sold today are based on a class AB design. We shall not make mention of the dreaded economy class B amps. They cost pennies to make, sound rather nauseating and have no business in audio reproduction.

Quantity vs. Quality

Hop in a brand spankin' new Ford Mustang Cobra. Feel the raw V8 power. Feel the pressure in the small of your back when tromping on the accelerator. Fly around the curves with all four tires screaming in protest as they approach their wipeout limits. Raw, unrefined power, and a lot of it. Now, slip into a BMW M5 sedan. Enjoy the solidity and confidence at 150 mph as only a BMW can deliver. Gracefully swoosh around turns without spilling your latte. Roll down your window and snobbishly ask for Grey Poupon.

Class A amps deplete your wallet like the M5 but once the sheer quality invades the senses, anything less is suddenly slumming it. Class AB is by no means low rent. Properly executed these designs can be very musical without breaking the bank. However, all this drivel about amps does us no good if they aren't connected to anything. Since the human ear can't hear the electrons in amplifiers, hooking some wire from them to a device capable of moving air molecules seems like the best course of action.

Speakers

The final or "back end" of a sound system is the actual reproducer itself. Something must move air for the human ear to perceive and interpret sound. Existing in two primary forms are driver elements, active (electrostatic) and passive. Someday, a direct connection to the brain via a neural interface might just be possible, but none are available as of yet.

The electrostatic element is usually found in large, costly speakers such as Quad or Martin-Logan. These are speakers that require high voltage charge, courtesy of your 120V wall outlet, to operate. No, they do not have internal power amplifiers within their boxes, think of it like a charge on a capacitor. The passive element is the most common and prolific element. It consists of a driver that needs no external power other than the amplifier signal. Passive sub-woofer drivers, such as those found in Vandersteens are not covered in this article.

Power Input

Speakers have limits as to how much juice can be put to their inductive coils before they physically bang against their end stops. A loud crack is heard when this happens and it's not recommended that anyone try this to any speaker you plan to listen to in the near future. Rest assured that when I spy "30W" stamped on the back of a driver, that is an absolute limit (within certain tolerances of course).

Quality vs. Quantity

Oh my, this is the part that I must refrain from typing 10 zillion pages. Differences exist in speakers that are so large that they can be heard in seconds--especially at the sub $150 level. Also, ridiculous specs like 1000W PMPO (peak music power output) boggle the mind. Yeah, like you should believe that some wall wart labeled as 10W can pump out 100 times that to your shiny new $49 boomers. PMPO is not a verifiable spec in any way shape or form, that's why over 20 years ago, the FTC developed strict standards for amplifier claims.

Because 99% of all computer speakers are BYOA (bring your own amps), power ratings are essentially meaningless. Since one can't be evaluated without the other, only the end result (total output) holds any significance. But do you need something that can make your pants/skirts flap? Some of the highest quality sound can be had with less power than you might think. It all depends on the total efficiency of the system, not just a wattage printed on a box side panel.

Efficiency

This is where SPL rears its head. Speakers have a particular efficiency. When a set amount of power, say 1 watt, is fed to the device, output is measured at a certain distance, say 1 meter, and SPL can be calculated. If a certain midrange produces 90dB SPL @ 1W @ 1m, then we have a known reference from which to base further calculations.

But most computer speakers have their own internal amplifiers, so individual efficiencies can't be measured easily. Total SPL can be measured and these output specs are becoming the defacto standard--you can find max SPL's quoted on an increasing number of speakers boxes. Here's a quickie guide to how loud is loud; every 10dB increase is a doubling of perceived loudness.

Sound Level
Loudness
0dB
Threshold of perceivable sound
40dB
Quiet country night, breath sounds
60dB
Smooth running car engine
70dB
Average conversation
80dB
Background music
90dB
Average music listening level
100dB
Loud music
110dB
Most rock/pop concerts
120dB
Threshold of pain over time (e.g. hours)
130dB
Rapid hearing loss over time (e.g. minutes)
140dB
"WHAT! I CAN'T HEAR YOU!" <---deaf in seconds
150dB
Heart attack possible

 

If you enjoy your hearing, don't subject yourself to more than 110dB for an hour, some say even less. If the speakers you just purchased actually reproduce sustained levels of 110dB, please set an egg timer or something. Ringing ears and headaches are not cool.

Subjective Listening Tests

The conundrum: who's right? With an ever growing base of review sites, who's subjective opinion matters? With other hardware like motherboards and chipsets, all is needed is objective data and practical tests. Very little is actually subjective. It's impossible to review speakers without amps and a sound source. Therefore, the believability of the reviewer must come directly under fire. The crux of the argument is this: only those with studio recording experience can really know what certain instruments (including the human voice) sound like.

All my subjective listening sessions include a wide variety of tunes to be sure, but they also included several selections that I personally multi-tracked using my own analog/digital gear. I still periodically master or re-master various types of radio/TV ads and personal material when the situation arises. Because tube microphone pre-amps can run into the $1000's, I designed and built my own hybrid class A, no feedback pre-amp in 1996.

Merely saying, "I cranked 'em up with Q3 'til my ears bled" doesn't hold much water. What does that mean? Were the test units just loud? Measurements can, most of the time, shed some light as to why a particular set sounds weird in certain areas. On a rare occasion, tests cannot reveal why a speaker sounds good or bad. Transient analysis is still under development and besides, some of that high end evaluation gear is only available to the high end mags, like Stereophile. At $25K and up, I won't ever have the golden opportunity to even touch one...

If you glean nothing else from this article, remember the following: listen extensively for yourself. After all, whether I recommend something or not, you may have vastly different requirements. I only get excited when I hear "audio truth" or accuracy and emotion in the reproduction chain. I have one overriding rule: the midrange (200Hz-3kHz) must be neutral sounding. If voices and the vast majority of woodwind (unamplified) instruments sound like other instruments, I personally can't stand to listen for long.

Objective Tests

These are relatively easy to determine and involve evaluating parameters such as:

  • Maximum SPL
  • Frequency response (including low frequency extension and crossover points)
    Frequency response and SPL are measured with a dB meter of known calibration and accuracy
  • Power input (from the supply)
    Power can be directly measured by P = IxE (current times voltage)
  • Power output (amplifier RMS, peak)
  • Impedance
    Impedance is directly measured by Z = E/I (voltage divided by current)
    **(For you EE's out there, I'm ignoring capacitive and inductive reactance for reasons of simplicity)**
  • Signal to noise
    Signal to Noise is given by dB = 20 x Log (V2/V1)

How Much Power Do I Need?

This is such a common question, I felt compelled to address it in it's own category. The bottom line is, it doesn't matter. Let me give you an example:

 
Rated Amp Power
Speaker Efficiency
Max SPL @ 0.5m
Speaker Set #1
1 W/ch
100dB SPL 1W/1m
112dB
Speaker Set #2
10 W/ch
90dB SPL 1W/1m
112dB
Speaker Set #3
100 W/ch
80dB SPL 1W/1m
112dB



Notice the power jumps by a factor of 100 from set #1 to set #3. Notice the total output (volume) stays the same. Why would a manufacturer sell a set like #3? In order to "fix" problematic frequency responses or most likely to add some funky low end (below 100Hz) eq curve, more power is required. If a woofer is 10dB down at 50Hz and the maker wants it to be effectively flat (0dB) then all they need do is put a 10dB boost in the amps and voila'--flat to 50Hz is now possible! But, that means the power output surges by a factor of ten (10dB = Power x 10).

Set #1 isn't exactly the cat's meow either. Locating drivers that sound musical and are incredibly efficient are few and far between. So the best and most common solution sits squarely in the middle: set #2. With a moderate amount of power and efficiency, the designer achieves a balanced design. Large amps and transformers aren't required and neither are incredibly high SPL drivers.

Conclusion

If it's simply not feasible to listen to a speaker setup with your own ears, then you must find reviews that convey not only the objective but also subjective qualities of the set of interest. Those genuinely concerned about great, long term sound quality, or what little of it there can be at the $50-$200 price level, can rest assured that my reviews at TargetPC are the most thorough on the web. No one else cares so deeply about uncovering manufacturer defects, power output claims, and overall listenability beyond the "makes your teeth chatter because they are sooo loud" group. Have I pissed off manufacturers? You bet. But dedication to truth in advertising is my primary goal. To that end, a few makers have contacted me specifically asking for more in depth reviews. Visit us often to see what develops in the very near future.


Latest news

Latest reviews