You press Start and hear two tones, one after another, about a second apart. You decide whether the second one is higher or lower than the first — that's the whole task. On a computer you tap p for higher and l for lower; on a phone you tap the top half of the screen for higher, the bottom half for lower. There's no countdown, no music theory required, no chord names. You just listen to two notes and say which way the second one moved.
Before you even press Start, the grid on the page already shows you the whole ladder — ten cells, each labeled with its own cents value, from 20 cents down to 4. Nothing is hidden: you can see exactly how hard question ten is going to be before you've answered question one. The first pair is genuinely easy — the kind of gap you'd hear even with a fan running in the next room. By the time you're staring down a cell labeled "4 cents," you already know you're walking into the hardest comparison in the set, and that anticipation is part of the experience. Most people's confidence collapses somewhere around question seven or eight — that's normal, and it's the point. The section isn't trying to see if you can nail all ten; it's trying to find out where your discrimination actually runs out.
This isn't a test of musical knowledge. It doesn't care if you can name an interval or read a key signature. It's closer to a hearing-science measurement: how small a change in frequency can you reliably detect? That's a different skill than "having a good ear" in the way musicians usually mean it, and it's worth separating the two.
Working musicians lean on pitch discrimination constantly without thinking about it — tuning a string against a reference note, hearing that a horn section is a hair sharp, noticing a vocalist drifting flat over the course of a phrase, matching pitch when singing harmony. None of that requires absolute pitch — the rare ability to name a note with no reference at all. It requires good relative pitch: the ability to compare two sounds and reliably tell which is higher, even when the gap is small. That's exactly what this section isolates, stripped of everything else — no rhythm, no melody, no memory, just two tones and a direction.
I've watched experienced players get surprised by this section both ways. Some assume decades on their instrument guarantees an elite score and land somewhere ordinary, because an instrument gives constant pitch feedback — frets, keys, valves — that a raw psychoacoustic test doesn't. Others who call themselves "not really an ear person" turn out to have a genuinely sharp threshold. Years on an instrument and raw frequency discrimination are correlated, but loosely — part of why I built the section this way instead of just asking people to self-rate.
This is the part most sites like this one gloss over, so let me be specific, because I wrote the code and there's no reason to be vague about it.
You get ten pairs of tones. Each pair starts from a fresh, randomly chosen base pitch somewhere between roughly 150 and 800 Hz, so you can't lock onto a fixed reference note and pattern-match — the starting pitch moves around from trial to trial, low one time, higher the next. What matters isn't the starting pitch; it's the gap to the second tone.
That gap is measured in cents, the standard musician's unit for pitch distance: 100 cents is exactly one semitone, so a cent is a hundredth of the smallest interval on a piano. The ten trials use a fixed, pre-set ladder of gaps — 20, 15, 12, 10, 8, 6, 6, 5, 4, then 4 cents again. Question one is a fifth of a semitone apart, which most people catch without much effort. Question ten is 4 cents apart — one twenty-fifth of a semitone — a genuinely difficult discrimination even for trained ears. Whether the second tone goes up or down is decided randomly each time, fifty-fifty, so you can't guess from a pattern there either.
One thing worth being precise about: this ladder is fixed, not adaptive. It doesn't loosen up if you're missing every question, and it doesn't tighten faster if you're acing it — everyone gets the same ten gaps in the same order. There's also no "listen again" in this section, unlike Count the Voices — you hear each pair exactly once, and the app locks out your answer until the second tone has actually started playing, so you can't jump the gun on a fast guess.
Each correct answer is worth two points, flat, regardless of how easy or hard that particular trial was — question one and question ten are worth the same. Ten trials at two points apiece caps the section at 20 points. A wrong answer scores zero for that trial but doesn't end the section or cost you anything beyond the missed points; you keep going through all ten regardless of how you're doing. There's no time pressure and no penalty for a wrong guess beyond simply not banking the points.
That 20-point ceiling matters in context: MusIQ's final score is out of 100, built from five sections worth 20 points each, added together. Pitch sensitivity isn't upweighted or downweighted relative to tempo, voice counting, tonal memory, or rhythm recall — it's one-fifth of the total, same as the others.
Pitch discrimination is a resolution problem, and resolution is exactly what gets lost first when audio is compressed or reproduced on cheap hardware. A pair of tones four cents apart is a genuinely small signal, and anything that adds noise, coloration, or distortion between the speaker and your eardrum eats into the margin you have to work with.
Laptop speakers are the worst offenders — small drivers have resonance peaks and dips across the frequency range, so they don't reproduce every frequency at the same relative loudness. That coloration can make a real pitch difference sound smaller than it is, or introduce artifacts that weren't in the original signal at all. Bluetooth headphones and earbuds add a second problem on top of that: most consumer Bluetooth codecs are lossy, meaning they're deliberately discarding audio information the codec's designers judged you wouldn't miss. Wired headphones skip both problems — the signal reaches your ears essentially intact, with nothing decoded, recompressed, or run through a compromised driver in between. If you're testing your actual ear rather than your laptop speaker's ear, plug in.
A rough pitch score means your raw frequency-discrimination threshold was on the coarser end, on this particular day, on this equipment. It does not mean you're "unmusical," and it doesn't mean you can't sing, play, or hear phrasing and expression — those draw on different skills than the narrow psychoacoustic measurement this section takes.
It's also worth saying plainly: this is a casual, browser-based test, not a diagnostic tool. If you're wondering whether you might have amusia — a real, recognized difficulty with pitch processing, sometimes called tone deafness — MusIQ can't tell you that, and a ten-minute web test shouldn't be treated as a medical or clinical assessment. That's a conversation for a specialist, not a website. I go into what amusia actually is, and isn't, on the tone-deafness page, along with the Wikipedia entry on amusia if you want the clinical background.
What actually moves the needle on this specific skill: focused interval practice, not passive listening. Play two notes close together on an instrument or a tuner app and try to name the direction before you look; sing along with a sustained drone and try to lock onto it exactly, noticing when you drift sharp or flat; slow everything down before you speed it up. Pitch discrimination responds to deliberate, attentive practice more than to just logging hours with music on in the background. I've laid out a fuller practice routine on the ear training guide, and it's worth pairing this section with how your brain actually processes pitch if you want the mechanism behind the exercise.
It varies a lot by individual and by training. Research consistently finds that trained musicians tend to detect much smaller gaps than untrained listeners, and this section's ladder — from 20 cents down to 4 — is built to span exactly that range, so most people land somewhere in the middle of the ten questions rather than acing or failing the whole thing.
No. Perfect pitch (absolute pitch) is the ability to identify or produce a specific note with no reference tone at all — a rare trait, not something this section measures or requires. This is a relative pitch test: you're only ever comparing two tones to each other. See the perfect pitch page for more on what absolute pitch actually is and whether it can be learned.
The ten gaps are arranged from easiest to hardest on a fixed schedule that's identical for every player — there's no way to skip questions, and there's no adaptive difficulty reacting to your answers in real time. Everyone takes the same ladder.
No. A miss just scores zero for that question; the section runs through all ten trials no matter what you answer.
No — the section doesn't ask for music-theory knowledge, interval names, or instrument experience. It's a raw hearing measurement. Training tends to help because it comes bundled with a lot of focused listening practice, but it isn't a prerequisite.