Choosing the Right Speakers
Replacing Your Speakers to Maximize Your Converted Radio's Performance -
Become Your Own Audio Expert to Make More Educated Choices
Before we begin on this subject, we first need to clarify what the term "watts" means - and the answer is that the term "watts"
has no tangible meaning until we have a little more information to work with.   There is more than one rating system used to
advertise a given amplifier or speaker's wattage rating, and a few of these rating systems can be
extremely deceptive.  For
instance, an obscure audio manufacturer can show the rating on the box as being simply "1000 watts", when that speaker
could not withstand  as little as 100 "true, honest watts" over the entire human hearing range for extended periods of time.

Back in the 60's and 70's, virtually every manufacturer of quality, high-end stereo equipment (such as Pioneer, Kenwood,
Sansui, etc.) rated everything in terms of
RMS watts, always including an entire page of every measurable specification.  By
using a
different rating system, manufacturers of budget family gear could literally double the figure and call it peak power.
Another handy trick was that the number could represent the
combined output of both channels without mentioning the fact.

A third rating system, thankfully used very seldom and NEVER by any well known or longstanding brands, is  
IPP which
stands for Instantaneous Peak Power, is always some huge number and pretty much means what it implies.  IPP power is
extremely misleading and probably should be avoided, and anything stated simply in "watts" should be checked into further.

Peak power is no longer considered the mark of a lower quality product due to its persistent use over the decades, and the
public's tendency to be drawn to higher apparent numbers has led even the best manufacturers to often use "peak" power
in their specifications.  Thus, it is important to remember the difference between RMS and Peak power with a simple formula:

Cut "Peak" in half to arrive at RMS, and regard RMS as true, honest watts.  Do not convert RMS.

From an engineering standpoint it's not even close.  But because of the human ear's logarithmic response to power levels
(which we'll get into shortly), it's close enough to be the accepted conversion in the everyday working world.  More than half
of all guitars amplifiers ever made depict their output wattage in the model number, either in RMS or Peak watts, and it's well
understood in the industry that the "ElectroShockRockKnockTheSocks 100" will be either 100 watts Peak or 50 watts RMS.

If you're interested in the actual engineering formula to compare and compute RMS and Peak power, here it is:

RMS watts = Peak watts x 0.707         or          Peak watts = RMS x 1.414

Frequency Response (or Range)

Ideally, the human hearing range spans from roughly 20 Hz to 20 kHz.  The nomenclature "Hz" is an abbreviation for "Hertz",
also often called "cycles per second" - or more fun, VIBRATIONS per second.  "kHz" means "kilohertz" - that is, 1000 cycles.

A young human ear, undamaged by exposure to damaging noise levels, is thus considered to be 20 to 20,000 Hertz (or cycles).
On the low end, 20 Hz is more felt than heard.  We've all heard AC hum, which is at 60 Hz, or "ripple" - its first harmonic at 120 Hz.
On the high end,  most all information above the highest piano note (about 4 kHz) is harmonics, the specifics of which give each
instrument, voice and sound their own flavors - with the extreme high harmonics responsible for brightness and overall clarity.

The low end of 20 Hz is, for all practical purposes, nothing but BASS - with the high end of 20 kHz being nothing but TREBLE.
In between, there is an infinite span of frequencies generally called the "midrange", with some audio discussions (particularly
in the live performance environment) involving the more discerning terms "low mids", "mid mids" and "high mids" depending
on whether the frequencies are right around the accepted center frequency of 1 kHz, or lean more towards the low or high end.

"Wait a minute - how can 1000 Hz be the 'center' frequency between 20 and 20,000 Hz ???"

In one simple phrase, because the human ear sucks as a sound pickup device.  The ear responds to frequencies not in a linear,
straight-line fashion, but on a LOGARITHMIC scale.   Here is a typical frequency response curve which allows you to see it:
Power Handling Capability
This screenshot shows how the
frequencies are spaced apart on the
low end, with one inch representing
little more than 100 Hz of range but
covering thousands of cycles on the
high end.  This lopsided way of
displaying the human hearing range
results in 1 kHz being just past center.

The TOP of the shading shows what
we call "flat" response, meaning that
the signal has equal intensity at all
frequencies.  Manufacturers of audio
gear strive for this theoretically ideal
response curve and usually achieve
it within plus or minus 3 dB.  3 dB or
less is accepted by the industry as
being too small a deviation to be
objectionable to the human ear.
<<<  Bass                Midrange               Treble >>>                
Maximum  Intensity
Zero Intensity
Why Frequency Response is Plotted Using the Logarithmic vs. Linear Scale
Below are two plots of the exact same "pitch inventory" - that is, a sampling of all frequencies occurring during one of my
own songs over the entire duration of the song, with the frequencies'  vertical positions showing their relative intensity to
each other.  Note how the
linear scale on the right, despite displaying the "correct" position of the 1 kHz center frequency,
results in most of the truly useful information - that is, the range that we consciously hear in the form of actual fundamental
instrument and vocal pitches - horribly bunched on the left (bass) side with the highest piano note "looking" more like bass.
Also, note how most of the fine detail between 20 Hz and 1 kHz is almost completely
lost when shown on the linear scale:
The black vertical line
represents the human
center frequency
The black vertical line
represents the human
center frequency
In the above examples, it becomes evident that the LOG(arithmic) scale is much more useful in plotting frequency response.

But on the other hand, it also shows how much more range beyond the fundamental pitches (what
the ear perceives as "useful" information, the actual instrument pitches) exists on the linear scale -
the scale that often makes more sense and is in fact "more true" in cases on the engineering level.

Using the linear scale, it can be seen that achieving a high-end response of 20 kHz is more difficult
and thus more expensive when considering production cost which necessarily translates to PRICE.
Thus, a basic understanding of frequency response "busts"  prices that seem too good to be true.

A theoretically "perfect" response curve is "flat" from 20 Hz to 20 kHz - a curve which unfortunately
is only possible under ideal conditions - such as in the case of a high-end HOME speaker which
can be made as large as needed to reproduce the lowest notes with intensities equal to the mids
and highest frequencies.  But the good news is that we don't necessarily always
desire a perfect
response.  For instance, there is VERY little musical information between 20 and 30 Hz, but this is
where most offending turntable rumble resides.  On the high end, things such as tape hiss and a
little of the undesirable static noise occurs above 10 kHz where there isn't much musical stuff.

When it comes to music listening, there is ALWAYS an unavoidable tradeoff between enjoying the
full possible musical range and the suppression of noise and other elements we DON'T want.  In
the high-end professional environment - especially the live performance environment where it is
NECESSARY to selectively reduce certain frequencies (even some of the "good stuff") in order to
boost the vocal mics above the instruments while also preventing screaming feedback, this can be
done using a 31 band graphic equalizer which allows the user to "slice" the overall range into 31
bands (compared to the broad, 2-band "bass and treble" controls) - literally, about every 3 notes.

Since this level of selectivity is generally impractical inside a car, we can actually USE a speaker's
inherent deficiencies - that is, "peaks and valleys" in the visual response curve - to fine tune our
selection of a given speaker's response to match our individual musical taste as close as possible.
We all know that a large "woofer" is much better at reproducing the low end, and that a "tweeter"
will only effectively reproduce the high end - thus the usual separation into individual low and high
transducers in the case of a
coaxial speaker, or into low, mids and highs in a triaxial speaker.

This being stated, now we can analyze the typical pitch ranges of the various instruments, voices
and noises to get a better idea of what we would tend to seek as individuals in a car speaker.  To
make these ranges easier to visualize, I have plotted all these elements on a blank response curve.

Note:  Disregard the vertical position of any particular range - in this case, vertical positions are ONLY used to
allow for the necessary overlap needed in order to display ALL ranges on a single plot for easy visualization.
The black vertical line
represents the human
center frequency
Bass guitar
Drums (including the deep, percussive "thud" of the kick drum and the shimmering high overtones of the cymbals)
Acoustic guitar (played folk style)
Distorted "rock" electric guitar including screaming high lead solos
Piano, fundamantal notes only (pitches of all 88 keys)
Male vocals (easy listening)
Male vocals ("hard rock")
Female vocals
Capabilities of electronic keyboards, especially synthesizers
Perceived pitches + harmonics of many voices and instruments
Harmonics only (brightness)
Now that we have more familiarity with the frequency ranges of what we're listening to, we can determine the optimum response
for our own particular tastes.  Myself, I would rather have the full range and tolerate the noise and hiss.  I don't care too much for
bass that rattles windows 3 block away, gravitating instead to a pronounced peak in the low mids which still brings out the bass
but doesn't drive the nerves crazy.  Having played rock guitar most of my life, I know what it sounds like and don't need it jammed
down my throat.  Thus a slight "dip" in response around the 1 kHz range, for me, tends to smooth out its irritating hard edge.

Other listeners cringe at the slightest noise, especially in the treble range, and in fact don't like much treble at all.  For this type
of listener, a speaker with a stronger emphasis in the bass and low mids might be the best choice.  Older listeners usually
have reduced treble response due to hearing loss (which always affects the high frequencies first), and might need more treble
than a younger listener would.   I like to "feel" the bass more than "hear" it, and thus enjoy a 30 Hz peak with a 100 Hz dip.

Your preferences will be much different than mine in many cases.  If your favorite band is AC/DC, you might prefer a 1 kHz PEAK
in this range instead of a dip - this emphasizes the harsh nature of Angus Young's guitar work, as well as Brian Johnson's voice.
Whichever sounds in music you like the most will probably also be the ones you wish to emphasize.   Taking some time to analyze
the various frequency ranges, then combining that knowledge with your known preferences will make you a smarter shopper.

If everyone had the same preferences in regard to sound, we wouldn't need tone controls !

This is why no one can really tell you which speakers will sound best to your ears; we can only help you make an informed choice.

Other factors to consider - basic facts about audio in general

Small things naturally move at high frequencies, and large things naturally move more slowly.  Because of this, it is impossible
for a single speaker to reproduce all frequencies effectively.  However, large speakers can reproduce treble MUCH better than
a small speaker can reproduce bass.  Thus, one larger speaker is usually better than two smaller speakers if forced to choose.
NOTE - this applies to so-called "full range" speakers, not a dedicated woofer which is designed to reject high frequencies.

Bass is nondirectional, meaning you're going to hear it just fine regardless of the speaker's physical location.  In fact, you can
often hear bass BETTER from a distance than from directly in front of the speaker - just ask any professional bass player.
The boom-boom teens aren't intentionally disturbing the peace in the entire neighborhood - they just can't hear it up close.

Highs are VERY DIRECTIONAL and will be heard best on-axis (directly in front of the speaker).  Because of this phenomenon,
most of the treble from door and kick panel speakers won't reach the ears since they're shooting the signal towards your feet.
Since there isn't room for a large speaker in these areas, bass response won't be very good either.  Door and kick panel speakers
are okay for adding ambiance and expanding spatiality, but generally don't make for very good listening by themselves.

Since the sound of a speaker is usually dependent on its enclosure, the passenger compartment of the car must also be
considered.  In most cars, the passenger compartment does an excellent job of reinforcing the "high bass" and low mids.
The windshield and rear window make excellent deflectors, bouncing signals from upfacing speakers right into your ears.

Outdoor environments, lacking nearby reflective surfaces, can suck up most of the output power before it reaches your ears.
Convertibles will benefit greatly from the increased power of a converted radio - especially improved bass response.

In Summary

1.  A converted radio is capable of 45 watts "peak", or 22 watts RMS, per each speaker.  All
speakers should be capable of handling 30 to 50 watts RMS for plenty of safety margin.  

2.  Converted radios will work with any speaker impedance between 4 and 16 ohms.

3.  Study sound and frequency response to determine your individual preferences.

4.  If you're planning to install a
pair of speakers to replace a single speaker (such as
two rear speakers for the right side signal to complement a single dash speaker to
reproduce the left side signal), refer to the following formulae for proper impedance
matching so that one side is not significantly louder than the other.  These formulae
also apply if you're changing the speaker setup to work with the original stereo radio:

Connecting two speakers in series DOUBLES the overall impedance.  For instance, two
4 ohm speakers in series presents a load of 8 ohms to the amplifier output.  This is safe.

Connecting two speakers in parallel HALVES the overall impedance.  Two 4 ohm
speakers in parallel presents a load of only TWO ohms which can be harmful to the
amplifier.  In this case, a resistor should be placed in series with the paralleled pair
in order to add a safe amount of resistance OR simply use 8 ohm speakers if available.   

Diagrams of each arrangement are below, and proper polarity should be observed to
ensure correct speaker
phasing.  A brief discussion on phasing is below the diagrams.
Call  928-533-9666 (Paying Repair Jobs ONLY)
+                           +
_                                  _
4 Ohm
4 Ohm
Series Connection
8 ohms
2 ohms
4 Ohm
4 Ohm
Parallel Connection - Dangerous
4 ohms
8 Ohm
8 Ohm
Parallel Connection - Safe
4 ohms
4 Ohm
4 Ohm
Parallel Connection - Safe
2 Ohm
20 watt
What is Phasing ?
The easiest and fastest way to understand phasing is to start with a couple phrases most folks into audio have heard before,
the terms "in phase" and "out of phase".  When a pair (or quad set) of speakers is connected properly as shown in the above
diagrams, the speakers are said to be "in phase" with each other.  By reversing the connections to
only one of the speakers,
they are then said to be "out of phase".   Note that this concept applies BOTH to the single-channel examples above AND
the two speakers in a stereo pair (also to a quad set of speakers).  Improper phasing actually
removes some of the music.

There is no such thing as "absolute" phase - and for this discussion, phase only exists in relation to something else, such
as the phase relationship of an amplifier's output when compared to its input, or the phase of one signal in regard to another.
When two speakers are properly connected "in phase", both woofer cones move outward and inward together.  When they
are connected "out of phase", one woofer cone is moving OUT while the other is moving IN.  Thus, anything common to
both speakers is theoretically cancelled out.  Regardless of the type of music or how creatively it is mixed in the stereo field,
the bass and lead vocals are usually centered.  Karaoke machines in fact use this principle as an aid to removing vocals.

Many of the cheaper stereos manufactured in the 80's, especially boom boxes whose speakers were too close together
to provide much separation, featured an effect called "Stereo Wide", "Stereo Expander", etc.   This feature depended on -
you guessed it -
phase cancellation to create this effect as it removed everything except that was NOT common to both
speakers.   This is the
last  thing an audiophile wanted, and consequently the feature was never offered on high-end gear.

Since the bass is almost always common to both channels, the most noticeable effect of improper phasing is a lack of bass.
To visualize the concept of phase shift, imagine one signal going straight into a stereo receiver, and the same signal going
through 10,000 miles of cable before reaching the receiver.   Even at the speed of light, the signal going through the cable
lag the direct signal due to the distance it must travel.  Stated differently, the direct signal will lead the "cable" signal.

These signals are now undergoing a change of phase in relation to each other, or a
phase shift.  If they are heard together,
one will partially interfere with the other resulting in some reduction of volume along with a somewhat "weird" sound.  A
phase shift can be anywhere from 0 to 360 degrees - with 0 degrees meaning "in phase" and 360 degrees being so "out
of phase" that it has actually become perfectly in phase with the next cycle per second in the frequency spectrum.

Phase shift is not always a bad thing, and is often used to create some cool effects.  Ernie Isley used a phase shifter
extensively on his lead guitar solos; listen to the Isleys' "That Lady" (studio version) to hear it in action.  In this song, the
phase shifter is constantly sweeping the range from "in" to "out of phase" every few seconds, and the resultant sound
is beautiful.   To add to the richness and soul of the phasing, Ernie first ran his guitar through a massively distorted fuzz
pedal before the phase shifter, so that the phaser also had signals high in the upper harmonic ranges to work with.

Such effects can be heavenly for one instrument, but not good for the overall sound. To hear phase shifting effects
as applied to the entire mix, check out  "Listen to the Music" by the Doobie Brothers and listen for an unmistakable
change in the overall sound during the "Lazy flowin' river" interlude after the second chorus.  Parts of it sound like
they're playing under water, the drums seem to trigger the sounds of different  jets flying overhead, and you feel like
you're spinning and being drawn into a vacuum just before they pull you out for the final chorus.  VERY cool effect -
but certainly not something you want to hear constantly and as a fault of the system.  Incidentally, these examples
depict a
sweeping phase shift whereas a "fault" phase shift will remain at some fixed point along that sweep range.

The effect heard in "Listen to the Music" is actually a much deeper phase effect called "flanging" which was discovered
purely by accident years before in some obscure studio while playing back the recordings of some obscure band.  An
engineer, while listening to two synchronized tape machines playing the same song, accidentally leaned on the outer
flange of the reel on one of the machines, slowing it down a bit and causing its signal to LAG that of the other machine
thus producing a phase shift between the two signals.  Varying pressure on the flange produced different degrees of shift.
NOTE:  All diagrams above are "per channel" and depict the connections to one channel only.
Where to Purchase Car Speakers

1st Choice: Greg at  SM Electro Tech offers a NICE line of reproduction speakers for both original and conversion radios
( Greg is also very knowledgeable about classic cars, original AND converted radios.  If his information differs from mine, listen to HIM )
I've done business with  Parts Express  since the early 80's, and they stock a HUGE line of non-reproduction speakers
Another great company to work with, also carrying a full line of speakers for converted radios, is  MCM Electronics
Quick reminder:  Speakers used with a CONVERTED radio should handle at least 30 watts RMS, or 60 watts "peak"
Click Here For My Phone Number
LOGARITHMIC PLOT - more sensible to the ear
LINEAR PLOT - more uniform mathematically
This chart illustrates a LOGARITHMIC, not linear scale