Frequency ear trainer

Because loudness is matched. A boost always makes the clip louder, and low bands add far more than high ones — on a music loop at +12 dB the spread between bands reaches 5-6 dB. We measure how much louder this particular material gets and compensate by exactly that. What remains is the difference in timbre, which is the entire point.

This skill behaves like a sport: short daily sets beat rare long ones. Ten minutes a day is a working routine. Watch the threshold in decibels rather than the score — it reports progress more honestly because it does not depend on how easy the questions happened to be.

It is the smallest boost you still recognise reliably. It is estimated as you play: difficulty follows your answers, and the final number accounts for the fact that some correct answers come from guessing. For scale: around 9 dB is a beginning, 4-6 dB is solid, and 2-3 dB is where people who do this professionally sit.

Yes — drop a file into the material block. The busiest excerpt of up to 20 seconds is taken and looped, so intros and fade-outs stay out of the training. Decoding and equalisation run in the browser through Web Audio, which is why switching between versions is instant.

A cut removes energy instead of adding it, and that reads much more quietly to the ear, especially on wide bands. Cuts are where most people plateau, so they are a separate switch you turn on once boosts come easily.

Any headphones work: what is being trained is relative discrimination between bands, not absolute reproduction accuracy. Phone and laptop speakers are fine too — they simply cannot deliver the extremes. That is what the audibility check at the start is for: it establishes what you actually hear on your device and only asks about that. You will not be marked wrong for something you could not have heard.