HIFI Diary: A Detailed Science Guide on HIFI Cables – The Analog Cable Edition

Foreword

For a very long time, the HIFI community has been engaged in fierce debates over cables; the argument between “cables matter” and “cables are useless” has raged for decades without a clear winner. In the course of my own reviewing, I have, in fact, wrestled with this dilemma more than once. Yet in the vast majority of cases, I firmly recognize that cables do alter a system’s timbre and improve its overall performance. In what follows, I will attempt to explain the secrets of HIFI cables in the most accessible way possible. This article is very long and may seem rather dry, but I believe it is a truly worthwhile read for any HIFI enthusiast.

Please note that this science guide is presented purely for knowledge sharing. The brands and models mentioned in the text serve only as illustrative examples and do not carry any recommendation. Given the limitations of my knowledge, this text can almost be assumed to be 100% erroneous, so corrections from experienced experts are most welcome. Additionally, I have long been writing HIFI science articles. To read this piece, it is advisable to possess at least a basic understanding of system clocks. I have also previously written several articles analyzing cables:

HIFI Diary: Starting from SATA Data Cables, Discussing the Chaos of DIY Cables – RainLain
HIFI Diary: Selection and Discussion of Ethernet Cables – RainLain
HIFI Diary: A Comparison of Solders – RainLain
HIFI Diary: How USB Cables Affect Sound Quality – RainLain

Main Text

If power cables are the physical interface for energy supply, and digital cables are the RF conduits for timing information, then analog signal cables are the one link in the entire system whose sole object of transmission is the music signal itself. What they carry is neither encoded “0s” and “1s,” nor rectified 50 Hz mains-frequency pulses, but rather the music that, after digital-to-analog conversion, has been restored to a continuous voltage waveform. These signals include: the faint phono cartridge signal from a turntable, the standard 2V RMS line-level signal from a DAC, and the analog voltage signal between preamplifier and power amplifier that carries all the dynamics and detail.

Precisely for this reason, the correlation between the physical properties of analog signal cables and subjective listening impressions is far more direct than with power cables or digital cables. They are more easily perceived, and also more liable to spawn all manner of mystical theories and pseudoscience. But in truth, no type of cable obeys the predictions of classical electromagnetism and materials science more rigorously than the analog signal cable. We shall explain—from the four dimensions of capacitance, inductance, shielding structure, and conductor material—why it can, and indeed must, change the sound of a system.

I. Capacitance: Its Effect on High Frequencies and Imaging

Of all the parameters discussed in relation to analog signal cables, none is more fundamental yet more overlooked by enthusiasts than capacitance per unit length. An analog signal cable consists of a center conductor (or a pair of conductors) and a shield (or return conductor), separated by an insulating dielectric. Physically, this constitutes exactly a shunt capacitance distributed continuously along the length of the cable.

This parasitic capacitance, in series with the output impedance of the upstream device, forms a passive low-pass filter. Its -3 dB cutoff frequency is determined by the following formula:

f₋₃ = 1 / (2π · Rₒᵤ · Cᵢₑ)

where Rₒᵤ is the output impedance of the source device (originating from the upstream equipment connected to the cable), and C_total is the total shunt capacitance of the cable (originating from the cable itself).

The critical point is that the output impedance of different sources varies enormously, and enthusiasts often overlook this variable. A typical modern DAC (Digital-to-Analog Converter) or preamplifier may have an RCA output impedance as low as 100 Ω, or even under 50 Ω, whereas certain traditional tube preamplifiers, passive preamplifiers, or phono stages can have output impedances as high as several hundred or even several thousand ohms.

For example: take an RCA signal cable 1 meter in length, with a unit capacitance of 100 pF/m (a very common value for ordinary cables); its total capacitance is 100 pF. If connected to a DAC with an output impedance of 100 Ω, the -3 dB cutoff frequency is roughly 15.9 MHz—far beyond the audio band, and thus has virtually no effect on the sound. But if connected to a tube preamplifier with an output impedance of 2 kΩ, the -3 dB point of the exact same cable plummets to about 795 kHz, and its phase response begins to exhibit measurable deviation below 20 kHz (the range of human hearing). Even more critically, the capacitance of certain phono cables or specialty construction cables can reach several hundred pF/m, and the drive conditions for certain high-output-impedance cartridges (such as MM cartridges with improperly matched loading) are even worse. In such scenarios, high-frequency roll-off and phase distortion will significantly intrude into the audible band.

Let us expand on this section further, as many readers may still be utterly baffled upon reaching this point. You may well have this question: the cutoff frequency is already 795 kHz—by any measure far beyond the human ear’s 20 kHz hearing limit. So how can a cutoff frequency of 795 kHz cause an audiophile to perceive that “the highs have become darker,” “imaging has deteriorated,” and “the soundstage has shrunk”?

We first need to clarify: our requirement is not to regulate signals above 20 kHz, but to ensure phase integrity below 20 kHz. -3 dB is merely the engineering metric that marks the “turning point” of this filter, but the filter’s effect on the signal begins well below the cutoff frequency.

1. The Filter’s Attenuation Does Not Occur Suddenly

The amplitude-frequency response curve of a first-order RC low-pass filter (precisely the circuit formed by cable capacitance and output impedance) does not plummet vertically at f₋₃. It is a smooth curve that descends extremely gradually starting from very low frequencies; the closer to f₋₃, the more pronounced the attenuation becomes, eventually continuing to drop at a slope of -20 dB per decade.

The table below shows the amplitude attenuation at various frequencies for a system with a cutoff frequency f₋₃ = 795 kHz:

Frequency	Amplitude Attenuation	Within the Audible Range?
10 Hz	Close to 0	—
1 kHz	~0.0000008 dB	Physically immeasurable
10 kHz	~0.00008 dB	Completely inaudible
20 kHz	~0.0003 dB	Amplitude attenuation inaudible, but note—phase shift has already begun to accumulate
100 kHz	~0.008 dB	Beyond human hearing bandwidth
795 kHz	-3 dB	The cutoff frequency itself

As you can see, at 20 kHz, the amplitude attenuation is only about 0.0003 dB—extremely difficult to measure with any instrument, and absolutely impossible for the human ear to hear a direct loudness change. If the darkening of high frequencies were solely a matter of amplitude, then a cutoff frequency of 795 kHz should indeed be completely inaudible. But in reality, a crucial piece of information is missing here: phase shift.

2. Phase Shift Begins to Corrode the Signal Far Below f₋₃

For the same first-order RC filter, the phase response formula is:

Phase shift Φ = -arctan(f / f₋₃)

This tells us: when the signal frequency reaches the cutoff frequency, the phase lag is exactly -45°; when the frequency is far below the cutoff frequency, the phase shift is extremely small but not zero, and it increases in a roughly linear fashion with increasing frequency. Let us calculate again for the system with f₋₃ = 795 kHz:

Frequency	Phase Shift
1 kHz	~ -0.07°
10 kHz	~ -0.72°
20 kHz	~ -1.44°
100 kHz	~ -7.2°
795 kHz	-45°

At 20 kHz, the amplitude attenuation is negligible, but the phase already lags by approximately 1.44°. This 1.44°, viewed in isolation, also seems very small, but in practice, it has fully entered the realm of being audible:

First, there are two analog cables—left and right channels—and relative phase differences exist between them. One of the core mechanisms of stereo localization (soundstage/imaging) is the binaural detection of interaural phase differences at very high frequencies. If the capacitances of the left and right channel cables differ slightly (e.g., one is 102 pF, the other 98 pF), the cutoff frequencies of the two cables will differ slightly, causing the phase shifts at 20 kHz to be different. This relative phase difference between channels directly leads to the wandering and blurring of high-frequency sonic images.

Second, consider the relative timing of high-frequency overtones within a complex waveform. The sound of a triangle strike may have a fundamental frequency of only a few kHz, but the overtone series that determines its “metallic character” and “shimmer” can extend well beyond several tens of kHz. In a cable with phase shift, the high-frequency overtones experience a time delay relative to the fundamental. Even if the amount of delay is extremely small (far less than 1 millisecond), it alters the phase structure within the overtone series, thereby changing the timbre of the instrument. The human ear is exquisitely sensitive to this.

3. Understanding Amplitude and Phase Together

Let us summarize. When we commonly speak of “hearing the highs darken” or “overtones disappearing,” it is almost never because the energy at 20 kHz has actually been shaved off by a few dB. The true physical process is:

Cable capacitance + Output impedance → Low-pass filtering effect → Phase response below 20 kHz slightly distorted → Interchannel phase mismatch → Spatial imaging at very high frequencies disperses, temporal relationships within the overtone series are disrupted → The auditory center interprets this loss of information as “dark,” “veiled,” “loss of air.”

It is not true “darkening” (the tweeter is not broken; the frequency response curve is practically flat below 20 kHz—simply put, frequency response measurements cannot detect it); rather, the structural integrity of the high-frequency information has been degraded. It is akin to a photograph whose contrast and sharpness are normal, but if all the fine texture is slightly blurred, you will feel the photo is “not transparent enough”—even though its brightness has not decreased.它的亮度并没有降低。

4. Where Does the “Safe Cutoff Frequency ≥ 100 kHz” Come From?

This is an empirical rule of thumb in engineering. The goal is not to guarantee that the -3 dB point itself lies above the audible band, but to guarantee that the phase shift within 20 kHz is suppressed to a practically negligible level.

Let us calculate the phase shift at 20 kHz for several typical cutoff frequencies:

Cutoff Frequency f₋₃	Phase Shift at 20 kHz	Engineering Assessment
795 kHz	~ -1.44°	Critical zone: possibly detectable by astute listeners under extreme conditions
500 kHz	~ -2.29°	Caution zone: optimization recommended
200 kHz	~ -5.7°	Not recommended: clear phase damage already present
100 kHz	~ -11.3°	This is the “minimum acceptable limit” suggested in some reference materials
50 kHz	~ -21.8°	High-frequency phase severely distorted
20 kHz	-45°	The cutoff frequency itself; amplitude attenuation has already reached -3 dB

My earlier recommendation of 100 kHz as a reference target is a conservative suggestion that includes an engineering margin; it is not an absolute physical boundary. Some enthusiasts with extremely acute hearing and highly resolving systems may perceive a difference even in a 795 kHz scenario; whereas other systems or listeners may not hear it even at 200 kHz. But if the cutoff frequency has already fallen to around 100 kHz or lower, the phase shift at 20 kHz has accumulated to over 10°—for the vast majority of high-quality systems (and for enthusiasts who have not yet gone deaf), this is already an audible region of degradation.

From this, a practical rule is derived: Analog signal cables cannot be discussed in isolation from the upstream output impedance, nor can the cable’s property as a total shunt capacitance be ignored, and still less can the difference between the L and R cables be overlooked. It is not an independent accessory, but rather forms an electrical system together with the preamplifier/source. However, for the vast majority of modern systems (DAC → active preamplifier/headphone amplifier/power amplifier), the output stage is generally low output impedance. The areas requiring greater attention are tube preamplifiers, passive preamplifiers, vintage equipment, or phono stages; for these devices, it is best to first calculate the capacitance upper limit and then select cables based on specifications.

II. Inductance: Its Effect on Dynamics

If capacitance affects high-frequency extension, then the parasitic series inductance of an analog signal cable affects transient response and phase fidelity.

1. Inductance Resists Change

The signal current must flow through a complete loop: the center conductor carries the signal current to the load, while the shield (or another conductor) carries the return current. The physical geometric relationship between these two determines the inductance of the entire loop. The farther apart they are, and the larger the enclosed loop area, the higher the loop inductance. Conversely, when the signal and return conductors are tightly coupled, the loop inductance is minimized.

The steepest transients in a music signal—such as the attack of a hard percussive strike, the eruption of a full brass ensemble, a strong string pizzicato—correspond to extremely high voltage slew rates. For a standard 2V RMS line-level signal, the peak voltage V_peak ≈ 2.83 V.

To calculate the rate of current change, we also need to know the input impedance of the downstream device, R_load. The input impedance determines how much current the signal source must drive.

Typical Scenario Assumptions:

Signal source peak output voltage: V_peak ≈ 2.83 V

Downstream power amplifier input impedance (RCA): common values are 10 kΩ, 20 kΩ, or 47 kΩ. Let us take a commonly seen, conservative value of R_load = 10 kΩ.

Rise time of a transient signal, t_rise: defined as the time required for a sharp attack to rise from 10% to 90% of its peak value. Extremely steep percussive transients in high-quality recordings can have a t_rise as short as 5–10 microseconds (μs). Let us take a representative value of t_rise = 5 μs = 5×10⁻⁶ seconds.

Calculate the peak current I_peak:

I_peak = V_peak / R_load = 2.83 V / 10,000 Ω ≈ 0.283 mA = 2.83×10⁻⁴ A

This appears to be a very small current. But the key is its rate of change—the current transitions from near zero to 0.283 mA within 5 microseconds.

Calculate the rate of current change ΔI/Δt:

ΔI/Δt ≈ I_peak / t_rise = (2.83×10⁻⁴ A) / (5×10⁻⁶ s) = 56.6 A/s

This value itself is not large, but please note: when the downstream input impedance is lower, ΔI/Δt will increase proportionally.

If the downstream device is professional equipment with a 600 Ω input, the situation changes completely:

I_peak = 2.83 V / 600 Ω ≈ 4.72 mA
ΔI/Δt ≈ (4.72×10⁻³) / (5×10⁻⁶) ≈ 944 A/s

This is more than 16 times higher than the 10 kΩ scenario. The influence of inductance is dramatically amplified in this moment.

2. The Induced Voltage Generated by Inductance

Now let us use V = L × (ΔI/Δt) to estimate the counter-voltage generated across the inductor.

First, a reference for the typical parasitic series inductance of an ordinary quality RCA signal cable:

For a 1-meter coaxial RCA cable, the typical loop inductance value is about 0.2–0.5 μH (200–500 nH), depending on braid density and geometric symmetry. Let us take a median value of L ≈ 0.3 μH = 3×10⁻⁷ H.

Scenario 1: Downstream input impedance 10 kΩ

V_inductor = L × (ΔI/Δt) = (3×10⁻⁷) × 56.6 ≈ 1.7×10⁻⁵ V = 0.017 mV

This induced voltage corresponds to approximately 0.0006% of the signal peak (2.83 V)—truly very minute.

Scenario 2: Downstream input impedance 600 Ω

V_inductor = (3×10⁻⁷) × 944 ≈ 2.83×10⁻⁴ V = 0.283 mV

Relative to the 2.83 V signal peak, this represents about 0.01%, or -80 dB. Still minute, but for a high-resolution system with a dynamic range exceeding 120 dB, this magnitude has already entered the threshold of audibility.

Scenario 3: Longer cable (3 meters), higher inductance (1.5 μH), 600 Ω load

V_inductor = (1.5×10⁻⁶) × 944 ≈ 1.42×10⁻³ V = 1.42 mV
Represents approximately 0.05%, or -66 dB.

At this point, the instantaneous voltage loss caused by inductance has reached the millivolt level. The lethality of this loss lies not in the absolute magnitude, but in its frequency dependence—it specifically targets the rapidly changing parts (high-frequency transients), while exerting virtually no suppression on slow, low-frequency signals.

Therefore, in scenarios requiring long cable runs (e.g., exceeding 3 meters), we typically recommend using balanced analog cables (XLR) with a twisted-pair structure. The positive and negative signal lines are tightly twisted together, the loop area is extremely small, inductance is low, and the ability to reject common-mode interference from external magnetic fields is strong. For single-ended analog cables (RCA) of coaxial construction, the signal is transmitted by the center conductor, while the return current flows along the inner wall of the shield. If the braid density and structural symmetry are poor, slightly higher loop inductance may result.

The effect of inductance on analog signals does not primarily manifest as high-frequency roll-off in the frequency response (that is more commonly seen as skin-effect loss at very high frequencies), but rather as a phase lag during transient current changes. When a steep transient suddenly appears in a music signal (e.g., the attack of a percussion instrument, the eruptive moment of a brass ensemble), the instantaneous rate of current change in the cable is extremely large. The cable’s own inductance resists this change, causing the arrival time of high-frequency transient components to be slightly delayed relative to low-frequency components. In a multi-way system, if the loop inductances of the left and right channel cables are mismatched, this will directly lead to wandering and blurring of the stereo image.

This is also why well-designed analog cables, whether coaxial or twisted-pair balanced, place extreme emphasis on precise consistency of conductor spacing and geometric symmetry. It is not for aesthetics, but to reduce the loop inductance to the lowest physically permissible value and to ensure strictly matched parameters between the two channel cables.

3. Phase Lag

电The capacitive low-pass effect is a frequency-domain roll-off, whereas the inductive effect is essentially a time-domain delay. It does not significantly attenuate high-frequency energy but causes the arrival time of high-frequency transient components to be delayed relative to low-frequency components.

In an LR circuit consisting of an inductor and resistor in series, the time constant is:

τ = L / R

Here, R is the total resistance of the loop, including the source output impedance, the cable’s own DC resistance, and the load input impedance in their parallel equivalent. Taking a simplified estimate, assume the total loop resistance R_loop is primarily determined by the load input impedance (when 10 kΩ).

For L = 0.3 μH, R = 10 kΩ:

τ = (3×10⁻⁷) / 10,000 = 3×10⁻¹¹ seconds = 30 picoseconds

This magnitude is far below the temporal precision distinguishable by the human ear; a 30 ps delay cannot be directly heard.

But when the load is 600 Ω and the cable inductance is higher:

τ = (1.5×10⁻⁶) / 600 ≈ 2.5×10⁻⁹ seconds = 2.5 nanoseconds

2.5 nanoseconds itself is still a single delay that the human ear cannot directly resolve. However, the effect of inductance is not a simple “overall delay”—it is a frequency-selective delay. In a complex music signal, this minute frequency-dependent phase shift rearranges the relative timing of different frequency components across the entire waveform. Once differences occur between the left and right channels, the same problem I discussed in the capacitance section arises: the wandering and blurring of the stereo image.

Calculation Example of Left/Right Channel Inductance Mismatch:
Assume left channel cable L_left = 0.30 μH, right channel L_right = 0.35 μH (a difference of only 0.05 μH, entirely possible within manufacturing tolerance), downstream load 600 Ω:

Δτ = (0.35 – 0.30)×10⁻⁶ / 600 ≈ 8.3×10⁻¹¹ seconds = 83 picoseconds

An absolute time difference of 83 picoseconds cannot be perceived by the human ear as a discrete echo. But the central mechanism of stereo localization is the precise detection of interaural time differences (ITD)—in a real sound field, a sound source deviating 1° from straight ahead produces an interaural time difference on the order of only about 10 microseconds. While the 83 picosecond interchannel deviation here is hundreds of times smaller than the ITD threshold, it is a full-bandwidth systematic offset. On the transient components at very high frequencies, through the precise analysis of the auditory center, it is entirely possible that it could be perceived, in a system of extremely high resolution, as a “reduced tightness of the high-frequency image” or a “relaxation of transient sharpness.”

III. Shielding

In the Power Cable Edition, I have already discussed in detail how RF interference injects itself into equipment via cables acting as antennas. For analog signal cables, the severity of this problem escalates by an order of magnitude.

The reason: the voltage levels carried by analog signal cables are extremely weak. Signals upstream of a phono preamplifier are merely hundreds of microvolts to a few millivolts; line-level signals are no more than about 2V RMS. This is the lowest-level, most vulnerable link in the entire system. And what it connects to is often the extremely high-gain input stage of an amplifier. Ambient RF fields picked up by a poorly shielded analog cable, after being amplified by the subsequent amplifier’s 60 dB or even higher gain, will become clearly audible at the speaker output as a “hiss,” crosstalk from radio broadcasts, or the “dit-dit-dit” pulse noise of a GSM mobile phone.

The shielding of an excellent analog signal cable follows the “Faraday cage” principle: the conductor is completely enveloped by a layer of conductive material. Typically, a dual covering of braided copper mesh plus aluminum foil is used; the braided layer covers low-frequency magnetic field shielding, while the aluminum foil provides 100% coverage against high-frequency electric fields. Some high-end constructions even employ dual-layer independent shielding, separately addressing interference in different frequency bands.

Furthermore, the grounding scheme of the analog signal cable’s shielding structure has given rise to the long-debated issue of directionality within the audiophile community. In the traditional structure, the shield is grounded at both ends, forming a complete loop that provides the best suppression of high-frequency interference. However, if a ground potential difference (ground loop) exists between the upstream and downstream equipment, power-frequency current will flow through the shield, and the magnetic field it generates may couple into the signal core wire, causing audible hum.

Another design approach employs a scheme where the shield is grounded only at the source end and left floating at the load end, in order to break the ground loop. In this case, the shield acts purely as an electrostatic shield and does not constitute a loop current path. This renders the cable’s impedance characteristics asymmetric in the two directions, which is one of the reasons why some cables indicate a signal transmission direction. The superiority of one approach over the other is not absolute but depends on the specific system grounding environment—this is one of the objective physical bases for the “matching theory” that objectively exists for analog cables.

IV. Conductor Material and Grain Boundaries

Once the above macro-structural factors—capacitance, inductance, shielding—have all been optimized through engineering, it is the conductor itself that determines the final fine sonic texture of an analog cable.

In DC power transmission, the sole significance of a conductor is its resistance. But in wideband analog music signals spanning from a few hertz to a megahertz, the microscopic physical processes within the conductor begin to reveal their acoustic consequences.

Metal conductors are not ideally continuous media; they are composed of countless microcrystals called “grains,” and between these grains exist grain boundaries where the atomic arrangement is discontinuous. When driven by an electric field, electrons crossing grain boundaries undergo inelastic scattering, causing minute energy dissipation. More importantly, the impedance behavior of grain boundaries at different frequencies exhibits nonlinear components, causing the extremely subtle undulations in the signal waveform—micro-dynamics—to undergo selective loss.

The significance of higher-purity metal conductor technology lies precisely here. Through directional solidification processes, the metal interior is drawn into extremely long, extremely few grains, drastically reducing the number of grain boundaries along the signal path. Theoretically, this reduces the probability of electron scattering, allowing extremely low-level, weak signal components—those microvolt-level fluctuations that constitute spatial reverberation and the phase information of instrumental overtones—to be transmitted more completely.

V. Physical Equalization

As for the practice, often recommended by veteran audiophiles, of using longer analog cables, this is essentially employing the cable as a physical equalizer (EQ). In headphone systems, the connection between preamp and power amp often requires only 0.5 meters or even less. In such cases, the analog cable’s capacitance and inductance are small, the transient response is fast, and theoretically, this is the most faithful to the original signal. However, influenced by contemporary popular listening tastes, headphones tend to be manufactured with ever higher resolution, which can sometimes sound overly sharp, rushed, or even “piercing.” The subtle “filtering effect” of a long cable can precisely soften this sharpness, making the sound appear more “settled,” “relaxed,” or, as veteran enthusiasts say, “richer in flavor.” Therefore, when the system’s sound leans toward brightness or hardness, swapping in a pair of high-quality long signal cables, with their minute “softening effect,” can aptly balance the overall listening experience.

Let me conclude with a summary. As the final connecting cables in the entire audio system, analog cables will magnify exponentially the various small problems accumulated in the preceding stages. And the transmission of analog signals is far more complex than that of digital signals. In fact, the “cables are useless” argument debated among enthusiasts for many years is, more often than not, a manifestation of “empiricism.” When we truly delve into the world of audio, to understand the physics and acoustics underpinning each signal’s transmission, we quickly shed that sense of taking things for granted, and thereby come to a truly clear recognition of every stirring moment in the HIFI world. HIFI is, in its essence, an exceedingly rigorous science. The science guides in my series are merely offered as a modest spur to induce deeper contributions. I hope everyone can correctly understand the significance of cables, and I also remind everyone to act within their means and not blindly worship cable brands. If the opportunity arises in the future, I may also write a science guide on cable brands.