A small game worth keeping in your back pocket between reviews. The page plays you a sine tone, a sweep, or two tones in a row and the job is to put a number on what you heard. Five modes go from beginner identification up to the painful small-interval stuff. Stats stick around between visits and a heatmap shows you which band is currently your weakest.
Why bother? Once you can hear a 200 Hz boom or a 6 kHz sibilance peak the moment a track starts, the way you describe gear changes. "Sounds a bit off in the upper mids" stops being good enough. "There's a spike around 3 kHz, almost certainly the tweeter crossover" is what comes out instead. Two weeks of short daily rounds is usually enough to feel the shift.
Bigger bars = more practice. Bar colour = accuracy in that band (green good, red poor). At least 3 answers per band before a colour is assigned.
Score0/0Streak0
1 to 4 answer · Space replay · N next
Why train your hearing?
Mixers don't have magic ears
What they actually do is sit in front of a parametric EQ for years, sweeping a narrow bell up and down a track until they can do it without looking at the screen. After enough hours of that, certain numbers go from being numbers into something close to a sensation. You hear a frequency and your brain hands you the number before you have thought about it. No genetic shortcut, no mystery. Just reps with feedback, which is the part this page is here to provide.
The payoff lands the first time you put on a new pair of headphones and a 250 Hz dip jumps out on its own. Reviews stop hand-waving. Buying decisions stop being "I think this one sounds nicer." Words about audio start lining up with measurements other people publish about audio, which is the only common language honest comparison can happen in.
About the range
Healthy young hearing reaches roughly 20 Hz at the bottom and 20 kHz at the top. That ceiling drops with age and with cumulative loud exposure. Most people in their 40s have already lost the top octave, and by 60 the lid sits closer to 12 kHz. The Audiophile and Sweep modes here run all the way up to 22 kHz partly so you can find where your real ceiling sits, partly so you can put a teenager on the same chair and feel ancient about it.
Different parts of the spectrum do different work in a track. The lowest octave is closer to vibration than to tone unless your room can actually move air at those wavelengths. Voice and most melodic content live somewhere between 200 Hz and 5 kHz, and a small dip in there is what gives recordings the "distant" quality reviewers complain about. Up above 12 kHz is the air band. Lose it and a recording feels flat across the board rather than missing any one element.
The frequency band cheat sheet
Every review complaint - boomy, thin, shouty, honky, harsh, airy - sits in a specific band on the chart below. Once the band names live in your head, EQ graphs read like sentences, and a phrase like "3 dB peak around 3 kHz" lands on an actual ruler rather than a feeling.
Range
Band name
What lives there / what it sounds like
20-60 Hz
Sub-bass
Felt more than heard. Pipe organ pedals, kick drum body, EDM sub.
"S" and "T" consonants. Tuning here makes or breaks vocal listenability.
10-16 kHz
Upper treble / "air"
Cymbal shimmer, room ambience, breath noise.
16 kHz+
Extended treble
Most adults stop hearing here. Affects perceived "openness" more than detail.
FAQ
Golden Ear training FAQ.
How long it takes, what the five training modes are, and why it makes audio reviews more useful.
Can I actually train this?
Yes, and you can watch the curve go up on this page. The pattern-matching your brain runs on voices and instruments handles frequencies in the same way, once you start tagging things with names. Two weeks of short daily rounds is usually enough to get most listeners naming bands inside a third of an octave on cue. The heatmap below the mode picker shows you which band needs the most work first.
How long until I have a "Golden Ear"?
Two to six weeks gets you reliable third-octave identification on most material. Telling 3.5 kHz from 4 kHz on cue, which is roughly the resolution the Audiophile mode pushes toward, is closer to a six-month project for most people. Stay on one pair of headphones or monitors while you train so you are not also fighting changes in transducer or room. Plateauing twice on the way to the small intervals is normal and not a sign you have stopped improving.
What are the five modes?
Identify is the simplest: a sine tone, four answer options, pick the right one. Discriminate plays two tones back to back and you say which was higher. Octave throws a pair separated by a known musical interval (octave, fifth, tritone, or unison) and you name the gap. Audiophile is the painful one, where the gaps shrink down to roughly 10 cents and most ears tap out. Sweep plays a rising sine that stops somewhere on the spectrum, and you tell it where.
Why does this matter for audio reviews?
It is the difference between writing "a bit bright" and writing "6 kHz peak that pushes sibilance forward on female vocals." One sentence is information a reader can act on, A/B against their own setup, or argue with on the same terms. The other is a vibe report. Reviewers who skip frequency drills tend to keep writing the vibe-report version forever and never quite notice they are doing it.
