Total Harmonic Distortion (THD) Explained

A DAC or amplifier's box now routinely advertises a THD+N figure with more zeros after the decimal point than the model it replaced, and the number reads as a self-evident ranking: fewer zeros before the first real digit, cleaner sound, better product. What the box almost never states is the level, frequency, load or bandwidth that figure was measured at, and without those four conditions attached, a THD or THD+N percentage is not simply incomplete — it is a number that cannot be compared to any other manufacturer's number, and sometimes not even to the same manufacturer's number for the same product measured a different way.

THD and THD+N measure two different things

Total harmonic distortion, THD, is a ratio: the combined strength of the harmonic frequencies a device adds to a pure input tone, divided by the strength of that tone itself. Test-equipment maker Audio Precision's glossary explainer on the two measurements states plainly that "noise is not included in a THD measurement" — a modern analyzer isolates it by taking an FFT of the output and summing only the specific frequency bins where harmonics of the input tone fall, a result the same page describes as "not influenced by DUT noise or spurious interfering signals."

THD+N is the more commonly published of the two, and it works differently. Rather than picking out harmonic bins, a notch filter removes the fundamental tone from the output and everything left over — harmonics, hiss, hum, anything uncorrelated to the input — is summed and reported as one number. AES17, the Audio Engineering Society's own standard for measuring digital audio equipment — read here in a copy mirrored by l7audiolab.com — defines it in exactly that spirit: "harmonic distortion and noise is the ratio of the output noise and distortion level to the output signal level. Both levels shall include all harmonic, inharmonic and noise components." The same clause of AES17 immediately hedges what the number is good for: "the results are indicative of anomalies in device behavior but may not be indicative of audible performance." A single THD+N percentage is a health check for the device, not a prediction of what a listener will hear.

The percentage is meaningless without level, frequency, load and bandwidth

AES17 does not let a device be characterized by one THD+N reading. Its procedure for the measurement calls for testing "at –1,0 dB FS and repeated with a sine wave at –20 dB FS," then separately "at each level from 0 dB FS to –80 dB FS in steps no larger than 10 dB," and separately again "at each octave frequency from 20 Hz to one-half the upper band-edge frequency" — because the same circuit's distortion genuinely changes with level and frequency, a single spot-check would misrepresent it either way. Audio Precision's own comparison of a complete and an incomplete THD+N specification makes the same point about what a spec sheet owes the reader: a usable figure reads "THD+N less than 0.01%, 1 Vrms, 20 Hz - 20 kHz, unity gain, 20 kHz BW," while "THD less than 0.01%" alone tells a buyer almost nothing.

Bandwidth matters for a reason that is easy to miss: it can silently exclude the very harmonics being measured. An independent write-up on audio test methodology at nihtila.com works through the arithmetic — "if the input bandwidth is, say 22 kHz, the third harmonic of the signal will only be included up to fundamental frequencies of about 7 kHz," so a fixed measurement bandwidth quietly narrows which harmonics of a high-frequency test tone actually get counted, and a manufacturer choosing that bandwidth is also choosing, deliberately or not, how much of the real distortion gets reported.

Level and load show up plainly in real spec sheets rather than as theory. Yamaha's own technical data sheet for the PX3 power amplifier — mirrored by the dealer saleswl.com — lists "Total Harmonic Distortion 0.1% (1kHz, 10W), 0.3% (1kHz, Half power)": the same amplifier, the same frequency, and a threefold jump in the published figure purely from raising the output level toward its rated maximum. The sheet rates the PX3 into 8 Ω, 4 Ω and 2 Ω loads at different power figures, but does not state which of those loads either THD reading was taken into. A different manufacturer's convention looks nothing like that shape: Crown's own product page for the XLC 21300 gives a single swept-band figure, "THD: <0.5%, 20 Hz–20 kHz," with no power level or load stated at all. Neither figure is dishonest, and neither can be lined up against the other — one is a worst-case ceiling across the whole audio band, the other is two best-case snapshots at a single frequency, and nothing on either page says which loudspeaker load either amplifier was driving when the number was taken.

Why published figures don't line up across manufacturers

Nothing obliges a manufacturer to publish a THD or THD+N figure the way AES17 measures one. The standard itself only asks that a compliant report say so: it recommends a note stating results were "measured in accordance with AES17," a statement most consumer spec sheets carry no version of, which is why nothing guarantees they followed the same level sweep, frequency sweep or bandwidth the standard requires. For amplifiers specifically, the usual reference is a different standard: AES17 itself lists IEC 60268-3, the sound-system-equipment standard covering amplifiers, among its own normative references, which is as much as this article can responsibly say about it. That paywalled standard is named here only because AES17 names it — IEC 60268-3 itself sits behind a paywall this article did not access, so what it actually requires is left out rather than guessed at. The practical upshot is the same either way: a THD figure with no stated level, frequency, load or bandwidth is not evidence of anything relative to a competitor's figure, only evidence that the two conditions might, or might not, have been the same.

A single number hides which harmonics a listener would actually notice

THD and THD+N both collapse every harmonic into one summed percentage, but the ear does not treat all harmonics equally. Dolby engineer Louis Fielder's AES convention paper on perceptual assessment of headphone distortion proposes scoring distortion against "a critical-band spectral comparison of the distortion and noise to an appropriate masked threshold," building separate masking curves for different frequencies rather than one blanket threshold — because a harmonic's audibility depends on where it lands in frequency relative to the fundamental and how strongly nearby energy masks it, not just on how large it is as a fraction of the signal. Two devices with an identical summed THD, one dominated by a masked high harmonic and one by an unmasked low one, would not sound equally clean, and a single percentage cannot tell them apart. That model, by its own abstract, was built for a sine-wave stimulus reproduced through headphones, with only a brief mention of extending the approach to loudspeakers — it is not a claim about music playback, which this article does not attempt to source.

That difficulty is also the honest answer on audibility thresholds. There is no single, sourceable percentage below which distortion "becomes inaudible" across harmonic orders and programme material — the reason Fielder's paper builds a multi-band masking model instead of proposing one number is that a fixed threshold does not survive contact with how differently a 2nd harmonic and a 9th harmonic, or a sine tone and a piece of music, get masked. Where this article cannot cite a study read in full for a specific number, it says so rather than repeating one of the many uncredited percentages that circulate in buying guides and forum threads.

Distortion is outside what this site's tools measure — there is no THD meter here, and nothing about a browser's audio output lets one be built reliably. What a difficult headphone load does to level and frequency response, rather than distortion, is covered in headphone impedance and why it matters; how cleanly a digital signal chain stores a waveform in the first place, a separate question from how a real amplifier reproduces it, is covered in sample rate and bit depth explained.