Elision In speech, something often becomes somethin' — the final ŋ turns into n. It sounds almost the same, but it is a real spoken form, and the spelling already shows it.
Six pairs, compared by the final sound
Both cells of a pair are read by the same speaker. This page reports no milliseconds — swapping a velar nasal for an alveolar one does not change the length of the word. What changes is the last symbol.
Why you do not notice it
Every point below can be heard in this page's own audio.
Both are nasals, so the difference is minimal
ŋ and n are both nasal, with air going through the nose; they differ only in where the tongue blocks (back vs tip). Hearing somethin', the brain restores something — which is why a missing sound does not register.
This is the first pair. It is not a failing of your ears; this contrast is simply not carried by listening.
<em>-in'</em> is a fixed spoken variant, not laziness
It has its own spelling (an apostrophe for the g) and appears everywhere in lyrics, scripts and subtitles. It is a recorded form, not one person mumbling.
So the thing to learn is the form, not to start dropping your g's.
When to use it: speech and informal writing, almost nowhere else
Formal writing, speeches and interviews take the full -ing. Recognising it and using it are separate skills — keep them apart.
This page only covers recognition.
The specific difficulty for Mandarin speakers
Mandarin already distinguishes -n and -ng (fen1 vs feng1), so the sound itself is not hard for you. What is hard is noticing that it got swapped in English — because the meaning does not change, your auditory system never raises a flag.
So this page puts the swap on two lines of symbols where it cannot be missed.
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, and why the page is built this way.
Corpus: 12 English sentences synthesised by a TTS engine in 4 American voices (two male, two female), arranged as 6 pairs (full spelling / -in' spelling). Audio is 22.05 kHz mono at 64 kbps.
One speaker per pair: so the only difference between the two cells is that word ending.
No duration is reported: ŋ and n are both nasals, and replacing one with the other barely changes the word length (measured: something 383 ms against somethin' 381 ms, a 2 ms gap). Printing 2 ms would read as "no difference" — the difference is in the transcription.
Where the transcription comes from: on this page the phoneme sequence is the G2P reading of the spelling. The reason is that somethin' forces forced alignment to degrade (8 letters against 6 phonemes), and a degraded result contains no phoneme sequence at all — yet the n is exactly what this page is about.
The engine really does say -in': G2P reads somethin' as / s ʌ m θ ɪ n /, ending in n rather than ŋ; written as something it gives / s ʌ m θ ɪ ŋ /. The engine follows the spelling, so the two recordings genuinely differ.
Target words sit mid-sentence: when -in' degrades, the final word's end time is stretched to the end of the clip (measured: fishin' came out at 1252 ms when it sat last, which is nonsense). Mid-sentence, word boundaries come from the engine and are accurate.
-ing → -in' only: the other kind of elision is the apostrophe family ('cause / 'em / 'til); those are closer to reductions, and because → 'cause already has a page there.