In the world of high-end audio, debates can get heated, and opinions are often held with religious fervor. I’ve been in this industry for half a century, designing and building audio equipment, and I’ve seen my fair share of controversies. But there’s one day that stands out in my memory, a day when theory and practice, measurements and perception, came head-to-head in my own listening room.
For years, a certain individual – who shall remain nameless out of respect – had been challenging me on my YouTube channel and throughout various industry forums. His worldview was firmly rooted in two beliefs: first, that higher sample rates in digital audio were not audible, and second, that everything we hear regarding differences in audio equipment can be measured and quantified. As someone who has spent decades fine-tuning equipment by ear, I vehemently disagreed with both assertions. Don’t get me wrong – I wish it were that simple. Our design lives would be so much easier.
If we could just hook up a piece of equipment to a computer or measuring device and have it tell us exactly how it would sound. But in my experience, that’s never been the case.
One day, this gentleman showed up at our facility, ready for a showdown. He was prepared to be proven right or wrong, and I was eager to demonstrate what my ears had been telling me for years. We set up a series of double-blind listening tests, first comparing different-sample-rate digital files, and then different amplifiers. To my satisfaction (but not surprise), my challenger correctly and consistently identified the differences in both tests. I thought, Finally, we can put this debate to rest. But what happened next left me scratching my head. Despite clearly hearing the differences, he declared that he was hardly going to change his public stance. His reputation, he explained, was built on these beliefs, and he wasn’t about to overturn them based on a single listening session.
I was speechless. Here was irrefutable proof, experienced firsthand, and yet it wasn’t enough to sway deeply held beliefs. It was a stark reminder of how entrenched we can become in our positions, even in the face of contrary evidence.
This experience underscores a crucial point in the world of audio: while measurements are indeed critical in the design process, they are only one tool in the chain. They provide invaluable data, help us troubleshoot issues, and guide our designs. But in the end, you have to listen. The human ear is an incredibly sophisticated instrument, capable of detecting nuances that our current measurement techniques might miss. It’s not that measurements are wrong – far from it, they’re essential. But they don’t tell the whole story.
In the following sections, we’ll dive deep into the world of audio measurements. We’ll explore what they can tell us, how to interpret them, and importantly, what their limitations are. We’ll look at frequency response curves, distortion measurements, noise floors, and more. But as we do so, let’s keep in mind that lesson from my listening room showdown: measurements are a means to an end, not the end itself.
Total Harmonic Distortion (THD)
Among the most frequently encountered specifications is total harmonic distortion (THD). THD is a common specification you’ll see when looking at audio equipment. It’s a way to measure how much an audio device changes the sound that goes through it. Let’s break it down in simple terms.
When we play a musical note through an audio device, we want it to come out sounding just like it went in. But sometimes, the device adds extra frequencies that weren’t there before. These extras are called harmonics, and they’re mathematically related to the original note, but they’re not all the same note at higher pitches. For example, if you play middle C on a piano, the harmonics would include higher C notes, but also other notes like G and E. These added harmonics can change the timbre of the sound. Even-numbered harmonics (2nd, 4th, 6th) tend to sound more harmonious, like they “belong together,” while odd- numbered harmonics (3rd, 5th, 7th) can sound harsher, as if they’re clashing. The mix of these added harmonics determines how we perceive any distortion in the sound.
THD measures how much of these extra harmonics get added. It’s usually shown as a percentage. If you see a THD of 0.1%, that means 99.9% of the sound coming out is the original signal, and the 0.1% is these added harmonics.
In the world of high-end audio, we often see very low THD numbers. Some equipment boasts THD as low as 0.001% or even less. That’s incredibly small! But here’s an interesting thing: most people can’t hear distortion when it’s below about 0.1%. Our ears just aren’t that sensitive. So while it’s nice to see those super-low numbers, the difference between 0.1% and 0.001% THD probably won’t be noticeable to most listeners.
So why do we care about THD? Well, when it gets high enough, it can change how music sounds. It might make instruments sound different or add a harsh edge to the sound. In extreme cases, it can even make listening tiring over time.
But there’s more to the story. Even when THD numbers are below what we can directly hear, they can be indicators of other things going on in the circuit that do affect the sound. For example, a slightly higher THD might hint at a circuit design that’s more musical or natural sounding, even if we can’t pinpoint the distortion itself.
On the flip side, chasing extremely low THD numbers can sometimes lead to problems. Designers might use techniques like excessive negative feedback to achieve those low numbers. While this looks great on paper, it can introduce other issues that we can hear, like a harsh or sterile sound quality.
That’s why we try to keep THD reasonably low in high-fidelity audio equipment. We want to reproduce the original sound faithfully, but not at the expense of overall musicality. Once we’re below that 0.1% mark, other factors often become more important in determining the overall quality of the sound. Remember, THD is just one piece of the puzzle when it comes to audio quality. It’s a useful measure, but it’s not the whole story. When you’re looking at audio equipment, consider THD alongside other factors to get a complete picture of how it might perform. Sometimes, a device with slightly higher THD on paper might actually sound better in practice. Trust your ears as much as the numbers.
Intermodulation Distortion (IMD)
Intermodulation distortion, or IMD, is another important measure of audio equipment performance that’s often overlooked in favor of THD. But in many ways, IMD tells us more about how a device might actually sound.
IMD happens when two or more different frequencies interact in a device to create new, unwanted frequencies. Unlike harmonic distortion, which adds frequencies related to the original tones, IMD produces frequencies that aren’t musically related to the input. This can lead to a muddy, harsh, or confused sound.
Here’s a simple way to think about it: Imagine you’re listening to a duet – a singer and a guitar. With high IMD, you might hear extra notes that neither the voice nor the guitar are actually producing. These extra tones can clash with the music, making it sound off or unpleasant.
We measure IMD as a percentage, similar to THD. But here’s the key difference: IMD is generally more audible than THD at similar levels. While you might not notice 0.1% THD, you could very well hear 0.1% IMD. It tends to sound bad in a way that catches our attention. What’s more, the presence of IMD often points to other issues in the circuit. It’s like a canary in a coal mine – if you’re seeing significant IMD, there’s likely something else going on that could affect sound quality.
That’s why, in many cases, IMD can be a more useful measure than THD when evaluating audio equipment. Low IMD usually correlates with a clean, non-fatiguing sound. High IMD often means muddled or harsh reproduction. But as always in audio, it’s not just about the numbers. Some designs might measure slightly higher IMD but still sound great due to how that distortion is distributed, while some very low IMD designs might sound sterile or lifeless.
Signal-to-Noise Ratio (SNR)
Another critical specification is the signal-to-noise ratio (SNR). This compares the level of a desired signal to the level of background noise. Usually measured in decibels (dB), higher numbers indicate better performance. An SNR of 100 dB, for example, means that the desired signal is 100 dB louder than the noise floor. For high-fidelity audio, look for SNR values of 90 dB or higher. However, remember that even the quietest listening rooms rarely have a noise floor below about 30 dB, so ultra-high SNR figures may not translate to audible improvements in real-world conditions.
A visual representation of different signal-to-noise ratios as applied to an image. The same kind of thing happens with audio signals. Courtesy of Wikimedia Commons/Dtrx.
Frequency Response (FR)
Frequency response is another key specification. It indicates the range of frequencies a device can reproduce at a consistent level. Typically expressed as a range followed by a variance – for example, “20 Hz – 20 kHz, ±3 dB” – this means the device can reproduce all frequencies between 20 Hz and 20 kHz within a 6 dB range (3 dB above or below the midpoint). While the human hearing range is generally considered to be 20 Hz – 20 kHz, many of us audiophiles prefer equipment with wider frequency response, arguing that it contributes to a more accurate reproduction of transients and harmonics (since the ear is very sensitive to phase issues that inevitably arise with limited-bandwidth equipment).
When interpreting these specifications, context is crucial. A specification that’s excellent for one type of equipment might be mediocre for another. Always compare specs within the same category of equipment. Moreover, remember that specs are typically measured under ideal conditions; real-world performance may differ. While specs are useful, they don’t tell the whole story. Trust your ears and, if possible, audition equipment before making a decision.
A loudspeaker frequency response graph. Courtesy of Wikimedia Commons/JPRoche.
Slew Rate
Slew rate is a measure of how quickly an amplifier can change its output voltage in response to a rapid change in input. In simpler terms, it’s the amplifier’s speed limit – how fast it can react to sudden changes in the music signal.
The reason slew rate matters is that music, especially complex recordings with transients like drum hits or plucked strings, can change very rapidly. An amplifier with a low slew rate might not be able to keep up with these quick changes, leading to a form of distortion where the output signal lags behind the input.
For listeners, insufficient slew rate can manifest as a loss of detail, particularly in the high frequencies, or a smearing of transients. It can make fast, dynamic music sound dull or less impactful. In extreme cases, it can lead to audible distortion.
From a technical standpoint, slew rate is typically measured in volts per microsecond (V/μs). It’s determined by the amplifier’s internal compensation capacitance and the current available to charge this capacitance. The formula for slew rate is SR = I/C, where I is the maximum current avail‐ able to charge the compensation capacitor, and C is the value of that capacitor. In practice, achieving a high slew rate often involves trade-offs with other performance parameters, such as stability and noise.
Slewing-Induced Distortion (SID)
SID, or slewing-induced distortion, is a form of distortion that occurs when an amplifier is pushed beyond its slew rate capabilities. It’s what happens when the amplifier can’t keep up with rapid changes in the input signal. SID exists because real-world amplifiers have limits to how quickly they can respond to input changes. When these limits are exceeded, the output signal becomes distorted, no longer accurately representing the input.
For listeners, SID can manifest as a harsh, gritty sound, particularly noticeable on high-frequency content or fast transients. It can make music sound less natural and more fatiguing to listen to over extended periods.
Technically, SID occurs when the rate of change of the input signal exceeds the amplifier’s slew rate. It’s often associated with a phenomenon called “hole punching” in the output waveform, where the amplifier’s output remains constant for a brief period while it catches up to the input. The severity of SID is related to both the slew rate of the amplifier and the frequency content of the input signal. It can be measured by comparing the amplifier’s output to a theoretically perfect version of the input signal and analyzing the differences.
Transient Intermodulation Distortion (TIM)
Transient intermodulation distortion is a form of distortion that occurs when an amplifier’s negative feedback loop can’t react quickly enough to correct errors in the output signal, particularly for high-frequency transients. TIM exists because many amplifiers use negative feedback to reduce distortion and improve performance. However, this feedback takes time to work, and if the input signal changes too quickly, the feedback can’t keep up, leading to distortion.
For listeners, TIM can result in a harsh, edgy sound, particularly noticeable on complex, dynamic music. It can make recordings sound less natural and more fatiguing over time.
From a technical perspective, TIM is closely related to slew rate limitations, but is specifically associated with the interaction between high-frequency signals and the amplifier’s feedback loop. It’s often measured using a complex test signal that combines a high-frequency tone with a lower- frequency one. The intermodulation products that result from this test can provide insight into an amplifier’s TIM performance. Minimizing TIM often involves careful design of the amplifier’s open-loop response and feedback network.
Multitone Test
The multitone or 32-tone test is a sophisticated method for evaluating an audio system’s performance under conditions that more closely resemble real-world music signals than traditional single- or dual-tone tests. This test exists because most music is far more complex than simple sine waves. By using multiple tones simultaneously, we can better understand how a system performs with complex, dynamic signals. For listeners, the results of a multitone test can indicate how well a system will handle real music, particularly in terms of clarity, detail, and the ability to separate different instruments and voices.
Technically, the test typically uses 32 different sine waves, spaced logarithmically across the audio spectrum. These tones are played simultaneously, creating a signal that’s much more spectrally dense than traditional test signals. The output is then analyzed, usually using FFT (fast Fourier transform) techniques, to see how accurately the system reproduces each tone and what intermodulation products are created. This can reveal non-linearities and other issues that might not be apparent with simpler test signals.
Crosstalk
Crosstalk refers to the unwanted leakage of a signal from one channel into another in a multi-channel system, such as stereo or surround sound. Crosstalk exists because it’s challenging to completely isolate different channels in real-world audio systems. Small amounts of signal can leak between channels due to electromagnetic interference, poor shielding, or shared ground paths. For listeners, excessive crosstalk can reduce stereo separation, making the soundstage less defined and potentially collapsing the three-dimensional image that good stereo reproduction can create.
In technical terms, crosstalk is typically measured in decibels (dB) and represents the difference in level between the desired signal in one channel and the unwanted leakage of that signal into another channel. It’s often frequency-dependent, with higher frequencies more prone to crosstalk due to their shorter wavelengths.
Measuring crosstalk involves sending a signal to one channel while measuring the output on the other channel(s). Good designs aim for crosstalk figures of -80 dB or better across the audible frequency range.
Square Wave Test
The square wave test is a method of evaluating an audio system’s performance by inputting a square wave and analyzing the output. This test exists because square waves are rich in harmonic content and can quickly reveal a system’s behavior across a wide range of frequencies. They’re particularly good at showing an amplifier’s ability to handle rapid transitions and maintain stability. For listeners, the results of a square wave test can indicate how well a system will handle transients and maintain accuracy across the frequency spectrum. Poor square wave response often correlates with less precise imaging and a less natural sound.
Technically, a perfect square wave contains the fundamental frequency plus all odd harmonics, theoretically extending to infinity. Real-world systems can’t reproduce an ideal square wave, but how they deviate from the ideal can be very informative. Engineers look at characteristics like rise time, overshoot, ringing, and tilt in the reproduced square wave. These can indicate issues with frequency response, phase response, and stability. The test is typically performed at several frequencies to evaluate performance across the audio band.
THD+N (SINAD)
The SINAD (signal-to-noise and distortion) measurement exists to quantify how much an audio system distorts the signal and how much noise it adds. It’s a comprehensive measure of an audio system’s fidelity. For listeners, lower THD+N generally correlates with cleaner, more accurate sound reproduction. However, it’s worth noting that not all distortion is equally audible, and some types of distortion (like low-order harmonics) may even be perceived as pleasing in small amounts.
From a technical standpoint, THD+N is typically expressed as a percentage or in decibels. It’s measured by inputting a pure sine wave and analyzing the output. The measurement includes both harmonic distortion products (integer multiples of the input frequency) and any noise added by the system. The formula is:
THD+N = (√(Vrms² - V₁²)) / Vrms
…where Vrms is the RMS voltage of the entire output signal, and V₁ is the RMS voltage of the fundamental frequency.
When expressed as SINAD, it’s calculated as:
SINAD = 20 * log₁₀ (V₁ / √(Vrms² - V₁²))
Modern high-end audio equipment often achieves THD+N figures below 0.1% or SINADs greater than 60 dB, but these measurements alone don’t tell the whole story of how a device will sound.
0 comments