Reductions going to becomes gonna, want to becomes wanna, because becomes 'cause. This is not laziness — it is how English is normally spoken. Textbooks print the full form, which is why you cannot hear it.
Six pairs, heard side by side
Both cells of a pair are read by the same speaker. The number is the whole span from the first word's onset to the last word's offset — how much shorter the phrase got, not how long a single sound is.
Why you cannot hear it
Every point below can be heard in this page's own audio.
Textbooks teach the full form; the recording says the reduced one
You memorised going to; the recording says gonna. The two differ far more than you would guess: /ɡoʊ ɪ ŋ t ə/ against /ɡɑ n ə/ — apart from the opening /ɡ/ they share almost nothing.
This is the first pair on the page. Play it twice and count the syllables in gonna.
A reduction is fixed, not just fast speech
wanna is the established form of want to: fixed phonemes /wɑ n ə/, fixed two syllables. It is not slurred because you spoke quickly — it is said this way slowly too.
So the way to practise is to learn the form itself, not to say the full form faster.
What is dropped is usually the predictable part
because loses /bɪ/; in kind of the of turns into a weak /tə/-like syllable. What goes missing tends to be what context already tells you — which is exactly why native listeners do not notice a gap.
The because → 'cause pair has the largest difference on this page; listen to it side by side.
You understand it but do not dare use it
Reductions are normal in speech and almost absent from formal writing. Recognising them and using them appropriately are two separate skills — do not practise them as one.
This page only solves recognition. When to use them is a different question.
Corpus and acoustic measurements
Click any word in a sentence to play from that word. The pale band over the waveform marks the measured boundary of the target word.
Per-token data
Every number below is measured directly from the audio, not hand-annotated.
| Token | Accent | Sex | Target | Word length | target word length | F0 | Pitch range | HNR |
|---|
Where the numbers come from
This section documents how each number is computed, so it can be checked.
Corpus: 12 English sentences synthesised by a TTS engine in 4 American voices (two male, two female), arranged as 6 pairs (full form / reduced form). Audio is 22.05 kHz mono at 64 kbps.
One speaker per pair: so the difference between the two sides comes from the reduction alone, not from a change of voice — a build-time assertion enforces it.
Phrase duration: measured from the start of the first word to the end of the last. going to is two words and gonna is one; comparing single-word durations would invert the result.
Word boundaries: MMS_FA forced alignment followed by character-anchored correction; timing precision is the 20 ms model frame.
No phoneme-level figures on this page: a reduction changes syllable structure (three syllables become two), so phoneme boundaries are not comparable. Showing phoneme highlights would imply we are comparing which sound is longer, when what is actually measured is how much shorter the whole phrase got.
The audio is synthesised, not conversational speech. The engine is faithful to spelling (write gonna and it says gonna), so the reduced forms are genuine; but real conversation has further variants that this page does not cover.