This page helps you evaluate headphones or earbuds online, and determine which one offers the best performance when comparing different pairs. Do not forget to bookmark this page — or download the sound files to your portable audio player if you are a patron — when going out shopping for a new pair of headphones or earphones.
All our tests should be performed with the headphones on your ears.
|10 Hz >> 200 Hz
+ Voice Over
The first file tests your headphones' bass extension. Play back the file until you start hearing the underlying sweeping tone as it rises. The voiceover tells you the frequency you have reached: this number represents the lowest boundary of your headphones or earbuds frequency extension. Good headphones will go as low as 20 Hz, the lowest limit of our hearing.
|22 kHz >> 8 kHz
+ Voice Over
The second file tests your headphones’ treble extension. To measure your headphones’ highest frequency, play back the second file until you start hearing the underlying (high pitched) sweep tone as it descends. Good headphones will reproduce frequencies up to 20 kHz, the upper limit of human hearing range.
Beware though: 1/ this limit decreases as we age. If all of your headphones suddenly exhibit the same upper frequency limit, and this frequency is lower than what is specified by the manufacturer, suspect your hearing, not the quality of your headphones. 2/ If you hear strange ups and downs or anything else that does not sound like a continuously descending sweep, suspect your sound card and aliasing. Have a check → here ← !
Ideally, you want your headphones to reproduce all frequencies consistently between the lower and upper limits, without a dip or peak in a particular frequency range. Absolute flatness is not exactly what you are looking for, though. Your hearing may not be perfect, and if you think about it, the best headphones are those that compensate for your particular hearing curve. It doesn't matter if your headphones are not flat, as long as they sound flat through your ears.
Spectral Flatness Test
Frequency responses are measured using sine sweeps and special test equipment. With a prominent sensitivity bump around 1–3 kHz, humans are not good at judging flatness: frequencies in the upper medium range will always sound louder than what they are because of the increased hearing sensitivity. Our test tone is not a regular sweep. It embeds an inverted hearing sensitivity curve that turns it perceptually flat. Our compensation only works at hearing threshold levels: turn your computer level down so that the test tone plays as quietly as possible.
If the sweep keeps playing at your hearing thresholds, your headphones are matched to your ears. If your hearing is perfect, this also means that the headphones have a flat frequency response. If your hearing is imperfect, this then means that your headphones are simply matched to your ears. If the test tone sometimes disappears, or gets too loud, you know you have a problem, either with your hearing or your headphones.
This test is of particular importance for earbuds. Depending on how deep you insert them into your ear canal, their frequency response will change drastically. Use this test to find which insertion depth gives you the flattest frequency response.
Dynamic range represents the ratio between the loudest signal you can hear and the quietest. Dynamic range is not part of any headphone specification, but will help you when benchmarking the isolation offered by your headphone in a noisy environment.
+ Voice Over
The file starts by playing noise at a full scale level. Adjust the level in your headphone so that this noise plays loudly, without being uncomfortably loud. Right after the noise, a voice is played back at a specified level, expressed in dBFS (decibels below full scale). Noise references and voiceovers alternate with each other, with the voice being played at decreasing levels. Play the file until you can't hear the voiceover anymore. The dynamic range that has been reached is given by the level the voice message was playing at when it was still (barely) audible.
The higher the dynamic range reached, the better the isolation offered by your headphones.
In general, "closed" headphones and "in-ear" earphones provide more isolation than the "open" type of headsets.
Poorly built or extensively worn headphones may start to rattle whenever loud or deep bass content is played. The next file scans bass frequencies and will literally shake your drivers when turning the level up. Adjust the volume in your headphone so that the test is made at a high level: the sweeping tone should remain pure and clear at all frequencies, without any parasitic buzz or rattle appearing in one earpiece or the other.
Better headphones have tighter tolerances in the variation of their drivers' frequency responses. To reproduce a faithful stereo image, the left and the right drivers must respond equally to every frequency in the audible spectrum. When this condition is fulfilled, the drivers are said to be "matched."
|Full Range Sweep|
Our driver matching test sweeps all these frequencies (up to 10 kHz) and sends exactly the same levels to both earpieces. Play the test tone at a moderate level and listen: the sweeping tone should keep a perfect central position across all frequencies, playing right in the middle of your head, without any deviation.
A panning that departs from its central position for a particular frequency will highlight poorly matched drivers... or mismatched ears.
To ensure that your headphone’s drivers cause the problem, not your ears, perform the test again, with the earpieces swapped: the left channel now feeding your right ear and vice versa. The panning should now deviate in the opposite direction.
Headphones that are properly wired will route the left channel to the left earpiece, and the right channel to the right (this makes sense). More important, relative polarity between drivers must be preserved: when presented with the same input signal, both drivers should move in the same direction, not opposite of each other.
To test polarity, use the audio files labeled "Center" and "Twisted." When polarity is preserved, the "Center" test signal will play either from a well defined spot between your ears or in front of you (depending upon how your brain interprets our test signal). The "Twisted" file should be harder to locate in space, sounding like it is being twisted inside your brain. If it feels like the opposite, suspect faulty cabling in your headphones or earbuds.
That would be the last straw if your headphone's manufacturer made wrong connections, but it sometimes happens as reported with very cheap brands, or counterfeit replicas. It happens more frequently though, when your headphone has been serviced and the four drivers’ wires have been messed-up during the replacement of a worn cable or broken jack.
Binaural recordings are made by placing microphones directly in one's ear, capturing the sound that reaches the pinnae. By playing back the recording through headphones, one ensures that each of the listener's ears get the exact same signal as originally captured. The result is immersive and truly surprising, but only works with headphones.
Our test signal consists of a binaural recording of someone knocking on wooden doors. When comparing headphones, judge how realistic the recording sounds: do you feel as if the doors are right next to you? Does it sound like real wooden doors?
|Custom Musical Track|
Last but not least, the ultimate test is to listen to your favorite music, and tracks that you've listened to so many times so you know them by heart. We can't help you with this one, as only you best know, but we can provide you with a short musical test, designed to cover the whole spectrum in a short amount of time.
From time to time, I receive emails asking for advice on new headphones. As this page seems to be the starting point of most inquiries, I will post my answer below.
As an engineer and professional sound designer, I am looking for a flat frequency response. In my opinion, headphones should sound transparent. If they depart for the ideal flat response, I can tolerate a little emphasis on both the bass and treble extremes: this means that I can listen to my source material at lower levels, and have the headphones simulate the sound I would be hearing if the levels were higher (the human hearing sensitivity at both spectrum ends drops drastically when the sound becomes quieter). Lower listening levels are the secret for sustained hours of work. As I am often wearing headphones several hours a day, comfort and durability are key factors for me. I am lucky to be working in a quiet environment: so, my headphones do not need to block external noises. Rather, I want to understand someone talking to me occasionally, without removing the headphones. An open design represents the best choice in such a case. Compared to closed headphones, open headphones offer a flatter frequency response too, and are much less fatiguing during long working sessions. The price-quality ratio matters to me. I wish I had an unlimited budget...
I've been trying different headphones for 20 years. For a couple of years now, I have set my mind - and ears - on the Sennheiser HD-600. They offer indisputable sound and build quality around USD 500. This may represent a lot of money, but it is a wise investment. Quality headphones will become your best friends over the years - or decades when used with care - which contrasts with the portable electronic equipment they are often connected to. They will also offer you the greatest listening experience that no loudspeaker will ever be able to achieve for the same price. Treat yourself; your ears are worth it! ;-)
Other alternatives worth considering are:
Is AudioCheck free? Not for me. Your support keeps this site running and growing without ads. Every donation gets rewarded with an access to • uncompressed .wav files downloads, for every test • sample rates up to 192 kHz, in the tone generator section • a suggestion box, on every page • no ads, never.
If you already are a patron, please log in.
Ads will disappear with 3 visitors contributing today. Still missing 2...