Ear Training — EQ Drills Online

This ear training drill teaches you to name the frequency band that a hidden peaking EQ has boosted or cut, using pink noise or a synth pad as the reference signal. It runs entirely in your browser on the Web Audio API, so the audio is synthesized locally and nothing is uploaded.

How to use it

1. Pick the difficulty ladder. L1 Broad hides the change among five widely spaced bands: 63 Hz, 250 Hz, 1 kHz, 4 kHz and 12 kHz. L2 10-Band ISO opens up to 31 Hz, 63 Hz, 125 Hz, 250 Hz, 500 Hz, 1 kHz, 2 kHz, 4 kHz, 8 kHz and 16 kHz. L3 14-Band Pro adds the in-between centers: 400 Hz, 630 Hz, 1.6 kHz, 2.5 kHz, 6.3 kHz and 12.5 kHz.
2. Choose a source. Pink Noise carries equal energy per octave, so the hidden filter is the only tonal change you hear. Synth Pad plays a sustained Cmaj9 voicing (C3 130.8 Hz, E3 164.8 Hz, G3 196.0 Hz, B3 246.9 Hz, D4 293.7 Hz, G4 392.0 Hz) whose partials can either mask or expose the same filter.
3. Compare filtered against flat. Listen Filtered (Mystery EQ) routes the loop through a peaking biquad set to the secret band; Listen Flat (Bypass Reference) drops that filter to 0 dB gain without restarting playback. The swap ramps over 50 ms, and the mode chip shows the current gain.
4. Name the band. Click the equalizer center-frequency button you believe was altered. The target is highlighted green on a correct answer, your wrong pick is highlighted red, and the feedback line names both your choice and the real band.
5. Watch the dashboard. Level, accuracy, score, current streak and best streak update after every trial. Five correct answers in a row promote you one level and reset the streak to zero, so streaks reward consistency as much as raw accuracy.
6. Reset when you change method. Reset Progress clears the locally stored stats and returns the ladder to Level 1.

What the engine does under the hood

The drill is one biquad peaking filter with Q = 1.4, which works out to a bandwidth of roughly one octave: BW = (2 / ln 2) * asinh(1 / (2Q)) = 1.01 octaves. That width is what makes band naming a practical skill rather than a fine-pitch exercise, because you are hearing a broad tilt in the spectrum rather than a single frequency.

Ladder level	Bands	Boost gain	Cut gain	Typical use
L1 Broad	63 Hz, 250 Hz, 1 kHz, 4 kHz, 12 kHz	+9.0 dB	-12.0 dB	Learn the five macro zones
L2 10-Band ISO	31 Hz to 16 kHz, octave steps	+6.0 dB	-9.0 dB	Standard EQ recognition
L3 14-Band Pro	31 Hz to 16 kHz, 14 centers	+4.0 dB	-6.0 dB	Fine-grained mix decisions

The signal loop is 2.5 seconds long and repeats without a gap. Pink noise is generated with Paul Kellet’s recursive filter, and the pad sums three harmonics per note. A 40 ms fade in and out avoids clicks, and the engine selects a boost 65% of the time and a cut 35% of the time, so you cannot win by always guessing the same direction.

Frequency bands, intervals and cents

Naming an EQ band and naming a musical interval are different listening tasks: the first is spectral, judged from timbre and brightness, and the second is sequential, judged from the distance between two pitches. This page trains the spectral skill, and the table below is the reference for the on-site pitch tools.

Interval	Cents	Frequency ratio	Second note above A4 440 Hz
Unison	0	1.0000	440.00 Hz
Minor 2nd	100	1.0595	466.16 Hz
Major 2nd	200	1.1225	493.88 Hz
Minor 3rd	300	1.1892	523.25 Hz
Major 3rd	400	1.2599	554.37 Hz
Perfect 4th	500	1.3348	587.33 Hz
Tritone	600	1.4142	622.25 Hz
Perfect 5th	700	1.4983	659.26 Hz
Major 6th	900	1.6818	739.99 Hz
Octave	1200	2.0000	880.00 Hz

There are 100 cents in a semitone and 1200 cents in an octave by definition. The octave is the one interval that maps directly onto the EQ ladder, because both are built on frequency doubling: ISO 266 sets the preferred octave centers at 31.5 Hz, 63 Hz, 125 Hz, 250 Hz, 500 Hz, 1 kHz, 2 kHz, 4 kHz, 8 kHz and 16 kHz, which is exactly the spacing the 10-band level drills.

Practice methodology: ten focused minutes beats an hour

Hearing memory consolidates between sessions rather than during them, and high-frequency sensitivity is among the first things to blur when you push past 20 or 30 minutes. That is why the training effect comes from short, repeated visits instead of one long grind.

• Session shape. Two minutes of warm-up on a band you already know, six to eight minutes of scored trials at 8 to 12 answers per level, then two minutes comparing a familiar track through the Flat and Filtered buttons. Stop while your accuracy is still rising.
• Daily beats weekend. Ten to fifteen minutes a day for two weeks moves your score further than a single 90-minute session, and the local streak counter makes consistency visible.
• Benchmark, do not guess. Set a baseline with the Blind Frequency ABX test for Hz discrimination and the Blind Pitch ABX test for cents, then re-test every two weeks. Score progress by statistically significant sessions, not one hot run.
• Log your confusions. The feedback line names the real band after every miss. Pairs such as 250 Hz against 400 Hz, or 8 kHz against 12.5 kHz, deserve targeted reps.
• Keep the level reasonable. Low frequencies need more acoustic energy to sound as loud, so do not fix a missed 63 Hz trial by turning the volume up; loud listening shifts the perceived spectral balance and invites fatigue.

Accuracy, limits and troubleshooting

• 31 Hz is often unplayable on laptops and phones. Small speakers roll off steeply below 100 Hz, so the lowest L2 and L3 centers may not exist acoustically no matter how well you hear. Use headphones or full-range monitors, and treat a device limit as a device limit, not an ear limit.
• 16 kHz is a device and age filter. Tweeter roll-off, earbud mesh and codec low-pass filtering all attenuate the top band. If you keep missing 16 kHz, verify with a known-good pair before drawing conclusions about your hearing.
• Room and headphone colorations are part of the test. Standing waves boost some bands and cancel others, adding a bias that scoring cannot remove. Learn your system’s peaks on the Spectrum Analyzer first.
• Disable system audio processing. Windows Audio Enhancements, macOS Voice Isolation, spatial audio and loudness equalization reshape pink noise, which changes the cue you are trying to hear.
• Stats are browser-local. Progress lives in localStorage for this origin only, so clearing site data or switching browsers resets it.
• This is not a hearing test. The drill measures band recognition, not hearing thresholds, and it cannot diagnose anything. If you are concerned about your hearing, see an audiologist.

Related tools

• Audio Spectrum Analyzer — watch the exact band the hidden filter is altering.
• Blind Frequency ABX Test — measure plain Hz discrimination without a musical frame.
• Perfect Pitch Test — train note naming when you want to add the sequential skill.
• Tone Generator — build your own reference tones for the bands you keep confusing.