Say "bat" and then "pat". To your ear they differ by one sound. To a microphone, the difference is a delay of a few hundredths of a second, and where your brain draws the line depends on the language you grew up with.
Both sounds start with your lips popping open. The question is when your vocal cords start buzzing afterwards. Linguists call that gap voice onset time, or VOT. In English, a "b" starts buzzing almost at once, around 12 milliseconds after the pop on average. A "p" waits much longer, around 60 ms, so a puff of breath escapes first.
Spanish uses the same two letters, but its sounds sit elsewhere. A Spanish "b" is typically pre-voiced: the buzz starts about 97 ms before the lips open. A Spanish "p" is a quick pop, about 14 ms. So a Spanish "p" is acoustically close to an English "b".
Drag the slider to change the delay. Then switch listener. The same sound can be a "b" to one brain and a "p" to another.
The delay is continuous, but what you perceive is not. Slide from 0 to 40 ms and, for an English listener, nothing seems to change gradually. It is "b", "b", "b", then suddenly "p". This is called categorical perception, and it was mapped across languages in classic experiments by Leigh Lisker and Arthur Abramson in the 1960s and 70s. For English the switch happens at around 20 ms. Spanish listeners flip near zero, so for them a sound that English hears as "b" is already a "p".
Notice what the slider does not do. It never produces a sound that is "half b, half p". Your brain throws away the in-between detail because, in your language, it rarely matters.
If the categories are learned, what do babies hear? In a famous 1984 study, Janet Werker and Richard Tees tested English-learning infants on consonant contrasts from Hindi, which English does not use. Infants aged 6 to 8 months could tell them apart. By around 10 months, English-learning infants could no longer do it. English-speaking adults also struggled, while adult Hindi speakers had no trouble.
This is called perceptual narrowing. In the first year, infants appear to start with broad sensitivity and then tune it toward the sounds that matter in the language around them. Distinctions that your language never uses are filed under "same", so you stop noticing them.
This is why it is hard to learn to hear, never mind say, a contrast your first language ignores. A Spanish speaker's "b" and "p" can land in the wrong English box. An English speaker may hear two different Hindi sounds as one. It is not a lack of effort: the sorting happens before you are aware of the sound.
It also explains something odd about your own speech. You produce "p" with a puff of air in "pin" but not in "spin", and you cannot hear the difference without cupping a hand to your mouth. Your categories hide it.
The good news is that categories are not welded shut. Adults can retrain with focused listening practice, and the boundary you carry is a habit of the brain, not a fixed limit of the ear.