"It is what happens every time an engineering discipline solves the noise sources it can see, and only then discovers the one it could not."
For most of the analog era, timing instability had a name, a measurement standard, and a number on the spec sheet. Wow and flutter told you, in a single weighted percentage, how faithfully a tape transport or a turntable held its speed. When digital audio arrived, that number quietly disappeared from spec sheets, and with it went the assumption that timing was still something worth worrying about. It was not a correct assumption. The problem did not go away. It changed shape, changed timescale, and became harder to explain to anyone who had not gone looking for it. Jitter is that problem, and it deserves the same seriousness wow and flutter once received.
There Is Always Something to Battle
Wow and flutter was never the only adversary. Rumble, the low-frequency noise transmitted into the signal from a turntable's motor and bearings, got its own weighted measurement standard and its own engineering response: better bearing tolerances, mass-loaded platters, suspended sub-chassis designs to decouple the cartridge from structure-borne vibration. Hum, energy coupled in at the mains frequency and its harmonics through poor grounding or inadequate transformer shielding, got its own discipline of star grounding, humbucking coil geometry, and shielded transformers. Hiss, the broadband noise floor inherent to magnetic tape and early electronic gain stages, got Dolby and dbx noise reduction, lower-noise tape formulations, and better-designed preamplifier stages. Each of these had a name, a standard, a measurement, and a body of engineering practice built specifically to reduce it, because each was audible, and because reducing it was understood to be real engineering rather than marketing.
Digital audio was, correctly, credited with eliminating several of these intrinsic limitations. A well-designed digital storage and transmission chain eliminates mechanical rumble and tape hiss as fundamental constraints. Any remaining hum or noise originates in the associated analog electronics rather than the digital encoding itself. What got lost in that genuine progress was the recognition that a new adversary had taken their place, one operating at a timescale small enough, and in a domain unfamiliar enough, that it took the industry the better part of two decades to characterize it rigorously and understand its practical significance. That adversary is jitter, and the pattern is not a coincidence. It is what happens every time an engineering discipline solves the noise sources it can see and only then discovers the one it could not.
What Wow and Flutter Actually Measured
Wow and flutter describe speed variation in a mechanical transport, a capstan that does not turn at a perfectly constant rate, a platter that is not perfectly isolated from motor cogging, a belt that stretches unevenly under load. Wow refers to the slow component of that variation, on the order of a few hertz, audible as a slow pitch wander. Flutter refers to the faster component, from roughly 6 Hz into the tens of hertz, audible as a coarser, more granular instability. Both were standardized, weighted, and measured as a percentage of nominal speed, with published audibility thresholds down around 0.05 to 0.1 percent for well-trained listeners on sustained tones.
The engineering response to wow and flutter was mechanical: heavier platters and flywheels for rotational inertia, tighter motor speed regulation, better bearing tolerances, quartz-locked direct-drive systems that measured actual shaft speed and corrected it in real time. None of this was cosmetic. It was a direct, first-principles attack on a measurable, audible source of timing error, and it defined a large share of what separated a mediocre transport from a reference one.
Why Digital Seemed to Close the Question
Digital audio replaces a continuously variable physical process with a discrete sampling process, and it is easy to assume that discreteness itself solves the timing problem. A sample is either read or it is not. There is no belt to stretch, no platter to wobble. For a while, this was treated as settled. Wow and flutter, as a category, essentially vanished from consumer digital equipment because the mechanism that produced it, continuous mechanical speed variation, mostly did too.
What that reasoning missed is that digital audio replaces one timing dependency with another. Every conversion event, on the way in at the ADC and on the way out at the DAC, depends on a clock telling the converter precisely when to sample or reconstruct the waveform. If that clock's edges do not arrive at perfectly even intervals, the resulting timing uncertainty ultimately appears in the reconstructed waveform as distortion correlated with the audio signal. The mechanism moved from the mechanical domain into the electrical one. It did not disappear.
At the instant a DAC reconstructs a waveform, it assumes each sample arrives precisely when the reconstruction filter expects it. If the timing of those sampling instants shifts, even though the numerical sample values remain correct, the converter is forced to reconstruct the waveform from slightly misplaced points in time. On rapidly changing signals, where the waveform slope is steep, even extremely small timing errors translate into measurable voltage errors. Timing uncertainty therefore becomes amplitude error, which is ultimately what reaches the loudspeaker.
Jitter, Defined Properly
Jitter is the deviation of a clock's actual transitions from their ideal, perfectly periodic positions, typically specified in picoseconds or nanoseconds RMS. Where wow and flutter modulate an entire signal's pitch as the transport speed drifts, clock jitter operates sample by sample at the moment of conversion and produces a different signature: modulation sidebands spaced around each frequency component of the audio signal, with amplitude that scales with signal level, jitter magnitude, and signal frequency, since a faster-changing waveform has a steeper slope at each sampling instant, and with audibility that correspondingly increases as signal frequency rises. This is not a cosmetic distinction. It means jitter and flutter are not literally the same artifact wearing different clothes. They are, however, the same category of problem: uncorrected timing uncertainty at the point of signal capture or reconstruction, degrading the thing every other part of the system is trying to preserve.
The audibility of clock jitter has been documented in peer-reviewed engineering literature since the early 1990s, much of it presented through the Audio Engineering Society, with published studies reporting audibility thresholds ranging from tens of nanoseconds for random jitter to considerably lower effective thresholds for certain forms of deterministic or signal-correlated jitter. The exact threshold depends strongly on jitter spectrum, signal content, converter architecture, listener training, and test methodology. The existence of the effect and its underlying mechanism are not seriously disputed.
Where Jitter Actually Comes From
Jitter is not a single defect with a single fix, any more than wow and flutter was solved by one component. It accumulates from several independent sources, and a well-engineered digital chain has to address all of them.
Clock generation itself is the first source: the phase noise and short-term instability inherent to the oscillator generating the sample clock, whether a simple crystal, a temperature-compensated crystal oscillator, or a full oven-controlled design. Power supply noise is the second: any ripple or switching noise reaching the clock circuitry can modulate its output timing, which is why isolated, well-filtered supplies matter as much for a clock stage as for an analog gain stage. Ground and shield contamination is the third: common-mode noise coupled in from elsewhere in the system, or from the mains, that finds its way into timing-critical circuitry, a mechanism treated in detail in Ground Noise Coupling. Clock recovery through interface receivers and phase-locked loops can introduce additional timing uncertainty depending on interface quality and PLL design. Signal path reflections are another source, arising directly from the transmission-line behavior of any cable carrying a clock signal at nanosecond-scale edge speeds. An impedance discontinuity anywhere along that path reflects a portion of the signal back toward the source, potentially perturbing the threshold-crossing time of subsequent edges and thereby introducing additional timing uncertainty, even when the originating clock itself is exceptionally stable.
That last source is the one most directly within a cable manufacturer's control, and it is worth being precise about it rather than vague. A clock or word clock reference typically expects a 50- or 75-ohm environment from source to receiver, depending on the interface standard. A mismatch of that specification produces a calculable reflection coefficient and, under the right conditions, measurable timing perturbations. This is not a subjective claim about cable sound. It is standard transmission-line theory applied to a timing-critical signal, and we go into the specifics in Word Clock Impedance.
Why This Deserves the Same Rigor Wow and Flutter Got
The analog industry did not treat wow and flutter as a minor footnote. It built dedicated measurement standards, dedicated mechanical engineering disciplines, and an entire tier of premium transport design around minimizing it, because listeners could hear the difference between a transport that held speed and one that did not. Digital audio deserves the same discipline applied to its own timing dependency, rather than the assumption, inherited from digital's early marketing, that a bitstream that is numerically correct is automatically also timed correctly at the point of conversion.
Modern well-engineered DACs already manage jitter far better than early digital equipment did, through internal reclocking, isolation stages, asynchronous architectures where appropriate, and increasingly sophisticated PLL design. That is real, measurable progress, and it means the residual jitter in a good modern DAC is typically well below the more conservative published audibility thresholds. It does not mean the problem has been closed out the way wow and flutter effectively was in the direct-drive turntable era. It means the remaining errors are smaller, more subtle, and more dependent on getting every downstream detail right: the oscillator, the power supply, the ground scheme, the interface implementation, and yes, the cable, in order to realize the full performance the converter is capable of. We cover the system-level case for addressing this directly, rather than assuming it away, in External Master Clocks and The Digital Hierarchy.
The Bottom Line
Rumble, hum, hiss, wow, and flutter each forced their era to take a specific, nameable source of degradation seriously because listeners could hear what happened when it was not addressed. Jitter asks the same question of the digital era, at a vastly smaller timescale and in a domain far easier to wave away as already solved. It is not already solved. It has simply moved into the electronics, the power supplies, the grounding, the clock architecture, and the cabling, where the same first-principles engineering that once tamed platters, capstans, transformers, and tape formulations is now needed to tame it.
There is always something to battle. Treating jitter with the seriousness its predecessors earned is not nostalgia. It is the same engineering mindset, applied to where the problem actually lives today.