We’ve talked about audio component and speaker distortion a great deal. The reason for this is that distortion is one of the critical factors in quantifying the quality of an audio system. Components that add a lot of distortion take away from the realism of the listening experience, making the music blurred, muddy and unnatural. This article will follow an audio signal through the typical car audio system, from the source unit to the speakers’ output, to explain how unwanted information is added to your music. We aim to emphasize the importance of choosing high-quality audio components at every stage to get the most from your stereo upgrade.

As our virtual audio test source, let’s imagine an audio track with three sine waves. The first is at 50 Hz, the second at 1 kHz and the third at 5 kHz. I’ve chosen this approach so you’ll be able to “see” how things change as we work through the system.

As you can see above, we have three unique peaks with little to nothing between them. This is as pure as our audio signal will ever be. The sine waves are “recorded” at -10 dB FS, so we should have enough room to add harmonics without clipping the signal.

The System: Source Unit, DSP, Amplifiers and Speakers

For this example, we’ll use a typical consumer-grade source unit from a name-brand company like Sony, Kenwood, Alpine or Pioneer. These radios typically add about 0.1% distortion on the preamp output unless they are a premium Mobile ES or eXcelon XR model. This rating of 0.1% means that for every piece of frequency content, information is added at a level of -60 dB at different harmonics. Most radios I’ve tested add second-order harmonics at this level, so to keep the example simple, we’ll only consider that.

So, let’s add second-order harmonics to our test track. The third harmonic of 50 hertz is 150 hertz. The third of 1 kHz is 3 kHz, and the third of 5 kHz is 15 kHz. So, we’ll add sine waves at -70 dB FS to the track – which is 60 dB less than the original 50-, 1,000- and 5,000-hertz signals.

Now we can see how there is some low-level content added at three times each of the fundamental frequencies. Remember, in real music, we aren’t dealing with pure tones but with bands of audio information. So, everywhere there is audio information, the product adds harmonics to the output.

Signal Processing

The next step in the audio food chain is to pass the signal through a processor. Once again, our goal is to explain the concept here, so we will keep things unrealistically simple and ignore any equalization or filtering. We’ll consider the harmonics added by a typical good-quality processor. Since this system isn’t using top-of-the-line gear, we’ll use an affordable six-input, six-output processor as the model. I’ve tested several of these, and they add about 0.05% THD to the signal. That value equates to another round of harmonics added to the output at a level of -80 dB.

Since the processor doesn’t know the difference between the original audio source and the information that includes harmonic distortion, it adds harmonics to the harmonics. Think of this like compound interest on a car loan or mortgage. In this case, those harmonics are below the background noise level in the audio track, so they are invisible.

The third-order harmonics have now increased to an absolute level of -67 dB FS from their -70, thanks to our digital signal processor’s third-order distortion characteristics.

System Amplification

Here’s where we need to make some assumptions about how the system is being used. Amplifiers don’t have linear distortion characteristics. They add more harmonics at low levels and less when being pushed hard. For this example, we will go right up the middle and assume we are listening to the system at a moderately loud playback level. Given the typical speakers in a vehicle and their reference efficiency, let’s say we’re enjoying an average playback level of around 90 dB SPL. Believe it or not, your amplifier is likely only producing 1 or 2 watts of power, assuming you haven’t cranked up the bass control on the radio. Most consumers and audio enthusiasts underestimate the importance of clean audio amplification and low output levels. This is why BestCarAudio.com measures amplifier THD+N and S/N ratio at 2 volts, as this equates to 1 watt of power into a 4-ohm speaker.

A modern, affordable Class-D amplifier adds about 0.01% THD, but it’s often reasonably equal between first and second harmonics. We don’t want to overload the concept, so we’ll add -83 dB FS frequency content at first- and second-order spacing to our audio signal. Please don’t fret; this will make sense when you see it.

The third-order tones are now above -70 dB FS, making them only 60 dB SPL quieter than our original signal. This isn’t loud but represents a total harmonic distortion level of 0.1%. Imagine if any component in the chain were of poor quality. We’ve measured ultra-compact amplifiers that are much worse than our example. Stacking another 10 dB of harmonic distortion on top of the existing audio content is almost easy.

Speaker Distortion

Sadly, we’ve reached the weakest link in the audio system in terms of adding distortion to our original signal. We’ve just started characterizing speakers’ harmonic distortion characteristics, so officially, we don’t have a massive database of information to pull from. For this example, let’s say that the vehicle has a set of upgraded speakers in the front and back – something like a 6.5-inch coaxial in both locations. We know that every speaker increases the amount of distortion it adds as cone excursion increases. Since reproducing bass frequencies requires that we move a lot of air, that’s where we’ll run into the most significant distortion issues.

We’ve already decided we’re at a playback level of around 90 dB SPL. Based on our testing and some educated forecasting, let’s add 1% even- and odd-order harmonic distortion at the bass frequencies. We’ll add 0.3% at the midrange and 0.1% at high frequencies. These equate to -50 dB FS tones at 100 and 150 hertz, -60.45 dB FS tones at 2 kHz and 3 kHz and -70 dB FS tones at 10 kHz and 15 kHz. Let’s see what that looks like.

Distortion Adds Something from Nothing

Let’s analyze this final signal. We have 100- and 150-hertz harmonics in the bass region at levels of -50 and -49 dB FS relative to our -10 dB FS source. That’s about -36.5 dB of distortion or 1.5%. For the -10 dB FS midrange tone, we have harmonics at -60 dB and -67 dB, which work out to a total harmonic distortion level of -45.47 dB or 0.423%. Finally, for our high-frequency information, the harmonics are at -58.5 and -52.3 relative to the original signal, for a total distortion of -51.4 dB or 0.27%.

What Are Our Takeaways?

We first need to realize that distortion is added at all volume levels, not just when a component is pushed beyond its linear operating range. Second, speakers are notorious for adding large amounts of distortion. Buying speakers with distortion-reducing technologies like shorting rings and copper or aluminum T-yoke caps can make a massive difference to your audio system. Lastly, this is a very simplified example of how distortion works. We ignored the first- and third-order distortion from the head unit and processor, along with the intermodulation distortion created between these fundamentals and harmonics. More importantly, we ignored that music is full of frequency content and not just three bands. Imagine these exact multiples of every frequency between 50 and 5 kHz.

Higher-quality audio components and speakers can quickly improve distortion performance by an order of magnitude or two, dropping it by 10 or 20 dB overall. Designing an audio system with a subwoofer will alleviate the need for a small speaker to try to reproduce bass frequencies for another significant reduction in unwanted harmonic information.

We also didn’t talk about background noise at all. A source unit like this might have a signal-to-noise ratio of 73 dB. The DSP has an S/N ratio of 78 dB, and the amp is likely in the 88 dB range. All this noise adds up and can become audible between tracks or during quiet passages in your music.

When it’s time to upgrade your vehicle’s audio system, drop by a local specialty mobile enhancement retailer and audition some options that will work with your application. The time you invest in choosing the best-performing solutions for the money will reward you with hours of great listening.

This article is written and produced by the team at www.BestCarAudio.com. Reproduction or use of any kind is prohibited without the express written permission of 1sixty8 media.

Picking new speakers for your home, a new set of headphones or upgrades for your car audio system should all involve the same auditioning process. Take two or three of your favorite songs on a memory stick to a local mobile enhancement retailer and audition the speakers under controlled conditions. We can’t count the number of times people have purchased speakers based on a brand’s perception of quality only to hear others that sound significantly better for equal or sometimes dramatically less money.

We’ve covered the process of auditioning speakers in the past. In this article, we’ll look at what makes speakers that look somewhat similar sound so different.

For this discussion, let’s look at the differences between 6.5-inch woofers that you’d find in a component set. The same design differences apply to many coaxial speakers and even to subwoofers.

Speaker Sizes

You’d think that within a specific speaker size class, the effective cone area of a driver would be pretty consistent. The specification that describes the effective cone area is called Sd and is typically specified in square centimeters, though the official standard is square meters. Many entry-level or high-excursion 6.5-inch drivers have an effective area specification of around 120 square centimeters. Those designs that have been optimized to maximize surface area might be above 140 square centimeters. That’s 17% more cone area that fits in the same application.

In terms of efficiency and low-frequency output, more area is better. The drawback of a larger cone is that it becomes directional at a lower frequency and necessitates a tweeter that can play loudly at frequencies below 2 kilohertz without producing a lot of distortion. Purely from an effective cone area standpoint, you can imagine that different driver designs sound unique, and more so when listened to off-axis.

Cone Excursion Capability and Power Handling

If you want to listen to your music at high volume levels, you need a driver that’s designed to be reliable and can move a lot of air. Without getting overly complicated, the length of the voice coil in relation to the height of the motor structure’s top plate determines how far the cone can move forward or rearward linearly. This specification is known as Xmax. It’s calculated by subtracting the top plate’s height from the voice coil’s height, then dividing by two. The suspension design also plays into how linearly the driver operates, but we’ll skip that for the moment. A basic OE replacement speaker might move forward and rearward 2 or 3 millimeters in each direction. A mid-level driver that can play much louder might reach up to 5 millimeters in each direction. The most premium designs offer more than 8 millimeters of excursion (in each direction) and often outperform larger drivers with lesser designs.

Of course, to make a speaker cone move greater distances, an amplifier needs to feed it significant amounts of power. As speakers are notoriously inefficient, much of the energy they receive is converted to heat in the voice coil winding. To increase power handling, larger voice coil formers are necessary. A typical replacement or basic upgrade speaker might have a voice coil with a diameter of 25 centimeters or about an inch. These drivers can often handle up to 75 or 80 watts of power if the winding is relatively long. If it’s short, power handling is usually down around 50 watts.

Better drivers will use larger voice coils in the 38-mm or 1.4-inch range. Power handling on these drivers jumps to around 100 to 125 watts, depending on the rating and testing method. Finally, the most custom designs might use a 51-millimeter former for the most durability at extreme operating levels. Oddly, the companies using these designs seem conservative, with their power ratings at about 100 to 150 watts continuous.

Cone Materials

If ever there was a topic that confused consumers, it would be the benefits and drawbacks of different woofer cone materials. Paper, plastic, carbon fiber, aluminum, layered composites, woven composites and all manner of in-between designs are prevalent in the top brands. Is one better than another? Some might excel in some frequency ranges while performing poorly in others.

The goal of the woofer cone is to move forward and rearward linearly without resonating. Rigid cones that aren’t well-damped tend to get very excited at higher frequencies and can cause harshness in the upper midrange. They are all reasonably similar at lower frequencies, except for how their mass works with the suspension and motor design to affect bass reproduction. If you see or hear claims of “tighter bass” based on a suggestion of improved cone material, someone doesn’t understand speaker design.

We should talk about dust caps and surrounds as an extension of cone materials. These components exhibit the same distortion-causing resonance issues as a poorly designed cone. These parts aren’t afterthoughts, and their design and selection are paramount to the proper operation of a speaker.

Motor and Suspension Design

Perhaps the most significant factor of loudspeaker sound quality is the motor’s design and the selection of suspension components. As an extension of our discussion of voice coil geometry and excursion limits, how the suspension behaves at extreme drive levels can effectively determine the sound quality of a speaker. Cupped spiders or those with linear compliance curves can result in significant distortion at lower frequencies and high excursion levels. Distortion will occur if more electrical input doesn’t equate to perfectly symmetrical or a proportional increase in cone travel. I’ve measured high-efficiency drivers that produced more output at 160 Hz than 80 Hz when driven with an 80-hertz sine wave. That’s right; the source information didn’t contain any audio at 160 hertz.

Consistent voice coil inductance based on cone position is also an important issue. When the voice coil moves forward, the T-yoke occupies less of it. When it moves rearward, more of the coil surrounds the T-yoke. This not only changes the inductance of the driver but its perceived frequency response.

The result, in extreme cases, is akin to listening to your voice when speaking through the blades of a moving fan. More high-frequency information is produced when bass information moves the cone outward and less when the cone moves inward. Features like aluminum and copper shorting rings in the motor and copper caps or shields on the T-yoke can help reduce this phenomenon.

Another factor that plays a huge role in the understanding of speaker quality is the stiffness of the suspension. A “tight” or inflexible driver typically has a higher Qms (mechanical Q) compared to a very soft one. This results in the driver being overdamped, which causes it to ring and resonate after the signal has stopped. It’s like flicking one of those spring door stops you’d find on the baseboard at home. Mathematically, perfect damping occurs when a driver in its enclosure has a Qtc (total system Q) of 0.5. At this value, the transient response is considered perfect. This comes at the expense of some output in the midbass region. A total system Q of 0.707 is called a Butterworth response, and it exhibits flat frequency response above the resonant frequency with acceptable time-based performance.

Systems with Qtc values around 1.0 are often described as warm as more upper bass information is produced. However, this comes with a significant increase in system distortion and a lack of what is described as “cone control.” It might be fun, but it’s not technically accurate.

Why You Need to Audition Car Audio Speakers

No two speaker designs are going to sound the same. Some drivers are optimized for efficiency to serve as original-equipment replacement speakers that will work well with a factory-installed or low-power radio. Other drivers are designed to handle significant amounts of power and produce a generous helping of bass at the expense of upper-frequency output.

Efficiency, frequency response, distortion characteristics, directivity, Q-factor and much more change how a speaker sounds. To choose an upgrade that will work well in your audio system, audition the drivers you have in mind under conditions that are as controlled as possible. Looking at graphs and specifications can, if you have years of experience understanding how the information affects performance, tell you something about the driver.

Still, none of that characterizes non-linearities that cause distortion. No car audio companies share that information publicly. As such, you must train your ears to pick up issues affecting performance. Take your favorite music to a local specialty mobile enhancement retailer and start listening. Give the volume on the source unit a good crank and get the speakers working so you’ll know what to expect.

Once you’ve established a baseline for quality, listen to even more speakers. When you can pick out the differences, choose the driver that’s the most accurate in all regards for your vehicle – you’ll be happy you did.

This article is written and produced by the team at www.BestCarAudio.com. Reproduction or use of any kind is prohibited without the express written permission of 1sixty8 media.

Up to this point, we’ve explained the difference in performance between entry level, poorly designed and premium car audio amplifiers. We hope you’ve found this informative, and now it’s time we took a close look at car audio speakers. No car audio component is more crucial than speakers for reproducing music with accuracy and clarity.

This series of articles will analyze the impedance, frequency response, output capability and distortion characteristics of different car audio speakers. The goal is to give those of you who want to upgrade the clarity and performance of your audio system a clear correlation between design features, specifications and, ultimately, performance.

Factory-Installed Honda Civic Speaker

I have a set of door speakers from a Honda Civic for our first subject. This is a woofer (no tweeter) with an effective cone diameter of 125.5 millimeters measured from the middle of the surround on one side of the driver to the center on the other side. The cone is made from a woven yellow fiber which could be of glass or aramid composition. The dust cap is formed from soft textile but is much less rigid. The speaker has a rubber surround, which lasts longer than foam.

Mechanically, the speaker has a relatively small-diameter flat linear spider bonded to a 1-inch voice coil former. There’s no cooling vent on the rear of the magnet or venting under the spider mounting ledge. The basket is formed from injection-molded, glass fiber-reinforced polycarbonate and has six deeply reinforced spokes. As is typical for an OEM speaker, the mounting flange includes a built-in spacer with an integrated gasket that will bring the speaker out near the grille in the interior door trim panel. Overall, aside from a small voice coil and lack of cooling technologies, the design offers nothing of significance to complain about.

Measuring Thiele/Small Parameters

Every speaker of every size can have its low-frequency characteristics modeled by a set of measurements and values summarized as Thiele/Small parameters. These measurements can be used with enclosure simulation software to predict how the driver will behave in an enclosure.

The Thiele/Small parameters quantify the driver’s suspension compliance, resonant frequency, mechanical Q, electrical Q and motor force. The information does not describe any nonlinearities in the suspension or magnetic fields or the excursion limits of the design. Far too many amateur audio enthusiasts think you can quantify the low-frequency sound quality of a speaker using enclosure simulation with Thiele/Small parameters. You can’t.

I’ll use my Clio Pocket with the added mass process to measure this information for the Honda speaker.

Is there anything we can discern in terms of performance from the measured Thiele/Small parameters? The first thing we see is that the driver has a relatively high total Q (Qts) of 0.69. This will add a little resonant bump in output in the lower midbass region. It’s likely a good design trade-off for a speaker designed to be used without a subwoofer, as it will add a touch of warmth to the sound. However, in absolute terms, this will be a bit of unwanted distortion. Lastly, the predicted efficiency is relatively high at 89.04 dB SPL when driven with 1 watt of power and measured at 1 meter. This is also normal for an OEM speaker as they trade low-frequency output for increased output at higher frequencies. The ~10-gram moving mass supports this theory.

Let’s look at what the BassBox Pro enclosure simulation software predicts this driver will do in our 3-cubic-foot test enclosures. I chose this volume as it’s typically large enough to have minimal effect on the driver’s performance and should simulate how the speaker will behave in a door or rear parcel shelf.

As you can see from the graph above, this is more of a midrange driver than a woofer. I guessed at the 30-watt power handling based on the diminutive size of the voice coil and lack of cooling features. In terms of predictions, the driver has a -3 dB frequency of 98 hertz and would greatly benefit from being used with a subwoofer.

Measuring Driver Impedance

Part of measuring Thiele/Small parameters is to make a series of impedance sweeps. Impedance is the opposition to the flow of alternating current (AC) signals. As you can see from the graph below, the driver has a fairly tall, narrow peak around its resonant frequency of 74.7 hertz. You can also see the increase in inductance at higher frequencies as the upward trend to the right.

We can see something else in this graph. Something has caused a noticeable resonant peak at about 700 to 800 Hz, and there are additional wiggles in the response at 2.4, 3.7 and 5.2 kHz. These are likely caused by the cone, dust cap or surround resonating. We’ll see if any of these translate into quantifiable distortion in the acoustic measurements.

Speaker Acoustic Measurements

With the driver loaded into my 3-cubic-foot test enclosure, I placed it on the floor of my lab. The microphone from the Clio Pocket is 1 yard above the top edge of the cone, where it meets the surround. We’ll use this position for all speakers going forward. We’ll begin the testing by taking frequency response measurements at increasing drive levels. While there is no specific standard, we’ll clone what Vance Dickason uses in his transducer tests in Voice Coil magazine with 0.3, 1, 3, 6, 10 and 15 volts. It’s doubtful that the driver will remain linear in output at the 10- and 15-volt levels as those values equate to 25 and 56 watts of power into a 4-ohm load. I will add a 2-volt measurement that equates to 1 watt into a 4-ohm load.

Before we get into the analysis of the speaker, we need to understand a few things about the measurements. First, the information below 30 Hz can be ignored. There is no output of 100 dB SPL at 10 Hz. Second, the dip at 130 Hz is a reflection in the room. It can be ignored as well. We know this is an acoustic cancellation because there is no dip or peak in the impedance or distortion curves. Sorry, I don’t happen to have an anechoic chamber at my disposal. In the meantime, I’ll continue to purchase lottery tickets!

Well, here’s our first look at the Honda speaker. From 160 Hz through to 1.5 kHz, the response is adequately flat given the non-anechoic characteristics of my lab. From 1.5 through to 5.5 kHz, there is a bump in the output of about 6 dB.

The black trace lower in the graph is the total harmonic distortion (THD) measured by the Clio. Let’s look at a few frequencies and make some percentage distortion calculations. From 200 through to 400 Hz, the harmonic distortion is -49 dB, equating to 0.35% THD. At 80 Hz, distortion is at 1.5%, and the significant bump in distortion around 1.3 kHz represents approximately 0.89% distortion.

Let’s sweep it again with a little more voltage – this time, the signal generator is set to 1 volt RMS.

The first thing to observe at this higher drive level is that the output increases linearly. All frequencies are roughly 10 dB louder. This is good because neither the suspension compliance nor the motor force has become a limiting factor. Something is happening up at 4.5 kHz that’s caused a bump in the distortion curve. Overall, though, it’s not too bad for this roughly 0.25-watt playback level.

Let’s bump things up to 3 volts.

In terms of frequency response, things remain nice and linear. All frequencies are once again about 10 dB louder. What isn’t so good is the harmonic distortion characteristics. A bump appears between 700 and 900 Hz at almost 2% distortion. This would be audible if not buried with other audio information. Distortion in the bass frequencies, 70 Hz, is over 3%. This 3-volt drive level equates to roughly 2.25 watts of power for a nominal 4-ohm speaker.

OK, how about 6 volts from the function generator for the next sweep?

A drive level of 6 volts is roughly 9 watts of power into a 4-ohm load. The graph above shows that distortion at all frequencies has increased by more than the increase in fundamental output. For example, when driven with 3 volts at 900 Hz, the THD was around 2%. Now, with 6 volts, the distortion has increased to 3%. Remember that bump we saw in the impedance graph around 800 Hz? Well, now it’s back as a peak in the distortion graph. You’d be surprised what you can learn from impedance graphs.

Last but not least, let’s feed this driver with a 10-volt sweep that equates to about 25 watts of power.

Though we only picked up about 3 dB more output, the distortion has increased significantly. We have 7% distortion at 800 Hz and over 3.5% at 200 Hz. If we look down in the bass region, 80 Hz is at about 10% total harmonic distortion. In short, this speaker would sound pretty bad when driven with much more than 10 to 15 watts of power and would be screaming at 25 watts.

Better Speakers Offer Better Performance

In terms of establishing a foundation for our measurements and speaker comparisons, we’ll stop here. This article will serve as a benchmark for what looked like a reasonable quality OEM speaker. We’ll test some speakers that might be better and some that might be worse over the next few months. This information should allow us to develop a correlation between design features and performance. In the meantime, if you’re shopping for new car audio speakers, drop by your local specialty mobile enhancement retailer to audition some options for your vehicle.

This article is written and produced by the team at www.BestCarAudio.com. Reproduction or use of any kind is prohibited without the express written permission of 1sixty8 media.