๐Ÿ“ˆFrequency test

โœ“ Tested & verified Updated: How the test works

Frequency test

The sweep climbs from 20 Hz to 20 kHz at one constant level, so everything you hear change on the way is your ears and your speakers, not the signal. Mark the two moments: when the tone appears, and when it disappears.

Audible band of this setup

โ€”

Both ends of this band belong to the whole chain, not to your ears: the bottom is where the speaker stops moving air and the top is usually where the speaker gives up as well. Headphones will give a different answer, and both answers are true.

  • Octaves coveredโ€”
  • Playing nowsilent
  • Sweep lengthโ€”
  • Output ceilingโ€”

One gear check a month

Monthly

One thing a month worth measuring before it fails properly โ€” a mic that reset itself, a stick that drifts, a switch sending two clicks. Two sentences and the test.

Free, no account. One click unsubscribes and deletes your address.

Try also

What frequency range can I actually hear?

The frequency test sweeps continuously from the bottom of the range to the top and reports the band you could hear โ€” both ends, from a single pass. A sweep catches what stepped tones step over: the narrow band where a panel buzzes, the dip where a crossover hands over badly, the one note the room swallows at your chair.

As everywhere in this group, the result describes a chain rather than a person. The low end is set by your speakers nearly every time, the top end by your ears and your drivers together, and the middle by the room you are sitting in. Textbook hearing runs 20 Hz to 20 kHz; real results on real equipment are narrower at both ends, and that is the ordinary outcome rather than a fault.

Run it once on headphones and once on speakers. The two answers disagree, and the disagreement is the useful part.

The interesting part of a sweep is rarely the two ends. Note where anything changes and the table below will usually name it.

What you hear during the sweepWhat it usually isWhere to go next
Nothing until well above 100 HzThe speakers cannot reach lower, which is normal for laptops and phonesBass test puts a figure on the low limit
A buzz or rattle over a narrow bandA resonating object, not the driverHold the panel or move the speaker while the sweep runs
A dip in the 2โ€“4 kHz regionA crossover handover, or something covering the driverMove the speaker off the desk or take the grille off
A note that vanishes at one seat onlyA null caused by reflections in the roomMove your head; if it follows you, it was the room
A sudden stop somewhere above 8 kHzA Bluetooth voice profile, not your hearingClose whatever opened the microphone; check the audio test
A gradual fade between 12 and 17 kHzThe ordinary top end of ears and drivers togetherHearing test reports that limit as a number

The other pages in this group each take one slice of what a sweep shows. The hearing test reports only the top end, the bass test only the bottom, and the tone generator holds one frequency still so you can walk around the room with it. Routing is a different question again: when a whole channel is missing rather than a band, use the speaker test, the headphone test or, for more than two channels, the surround sound test. The full group sits under tones and hearing.

Frequently asked questions

I searched for a Hz test and meant my monitor. Where is that?

One word, two measurements, and this page is the audio one. A screen’s refresh rate is also quoted in hertz โ€” 60, 144, 240 โ€” and it counts frames per second rather than sound waves per second. That test lives on the refresh rate test, which measures the frames your browser was actually handed and converts them to a frame time in milliseconds. Nothing on this page will tell you anything about a display.

Is this a spectrum analyzer?

No, and the difference is the direction. A spectrum analyzer listens and shows you what frequencies are present in a sound; this page sends a known frequency and asks what you heard. Analyzing requires a microphone, and any measurement of your speakers through a laptop microphone would describe the microphone’s own response as much as theirs. Sending is the half a browser can do exactly, which is why it is the half we built.

What frequency is a human voice?

The fundamental sits between roughly 85 and 180 Hz for most adult men and 165 to 255 Hz for most adult women, but that is not where intelligibility lives. Consonants โ€” the difference between s, f and t โ€” sit up between 2 and 4 kHz, which is why a system that reproduces the fundamentals can still be hard to follow. A voice that arrives boomy is heavy in the low band; one that arrives clear but thin has the fundamentals missing and the consonants intact.

Part of the sweep goes quiet and comes back. What is that?

A narrow gap usually belongs to the room rather than the speakers. Sound reflecting off a wall arrives slightly after the direct sound, and at frequencies where the two are out of step they cancel at that spot, which produces a null you can hear at one seat and not half a metre away. Move your head while the sweep runs: a null moves with you and a genuine speaker fault does not. Dips that stay put in the 2 to 4 kHz region are more often a crossover or something covering the driver.

Does the sweep speed change the result?

It changes what you can notice, not what is played. A fast sweep passes a narrow resonance in a fraction of a second and your ear may register nothing at all, while a slow sweep sits in the problem band long enough for a buzz to become obvious. For finding the ends of your range, faster is fine; for hunting a rattle, slow the sweep down and be ready to stop, because the point of a sweep is knowing exactly where it was when you heard the fault.

How the test works

The sweep is one oscillator whose frequency is ramped exponentially rather than linearly, which means it spends equal time in each octave. That matters more than it sounds: a linear sweep from 20 Hz to 20 kHz would spend half its life above 10 kHz and rush the entire bass range in a moment, making the bottom end almost impossible to judge. The level is held constant in digital terms throughout, so any change in loudness across the sweep is being introduced by the output device, the room or the ear rather than by the signal โ€” human sensitivity varies by tens of decibels across the range, which is why a constant-level sweep sounds anything but constant. Start and end are ramped over a few milliseconds to avoid a click. The page reports the band you marked; it cannot hear the room, so this is not a frequency-response measurement and would need a calibrated microphone to become one. Everything is generated locally, nothing is downloaded, no microphone is opened and nothing leaves the page.

Worked example

With bottom end marked 45 Hz, top end marked 15 000 Hz, this page works out 45 Hz to 15 kHz, 8.4 octaves. This setup answered from 45 Hz to 15 kHz, which is 8.4 octaves. Ears, speakers, volume and room are all inside this number and cannot be separated by a web page, so treat it as a measurement of the setup, not of your hearing.

  • Bottom end45 Hz
  • Top end15 kHz
  • Octaves between them8.4

Sources:

Why a sweep and not more tones

Resolution. A stepped test asks a handful of questions and a sweep asks every one in between, which is the only way to find a fault that occupies a band a few hertz wide.

Order. Because the frequency rises with time, the moment you hear something is the frequency of the problem โ€” no hunting through a list to reproduce it afterwards.

Both ends at once. The bottom of your range and the top come out of the same pass, so the result is a band rather than a single number, and the band is what actually describes a system.