Velar nasal · ng The tongue back rises to the soft palate while the velum lowers and air leaves through the nose. The contrast with /n/ is not «nasal» but where the block is: tip for /n/, tongue back for /ŋ/.
Three steps to /ŋ/
Find the block first, then the nose.
Tongue back against the soft palate
This is the only posture: the tongue back, not the tip, rises to the soft palate. Put this drawing next to /n/ — the block is in a different place entirely.
Lower the velum
As with /n/ and /m/: the velum drops (blue dashes) and air goes through the nose. The three nasals differ only in where they block.
Self-check: no release at the end
Say sing. The ending stops cleanly, without a small /ɡ/ popping out. A hand on the throat should feel no burst. This is the most important check on the page.
Articulation (side view)
One speaker, one instant: tongue position and lip shape only. The amber circle marks what to look at; the caption says how it differs from the sound it is usually confused with.
Spelling: ng is /ŋ/, but suffixes complicate it
The rule is simple; what is hard is when -er / -ing bring the /ɡ/ back.
The most stable: final ng is /ŋ/ with no /ɡ/.
n before k becomes /ŋ/, so bank is /bæŋk/. ★ That is assimilation, not another phoneme.
★ Here the /ɡ/ is pronounced: finger /ˈfɪŋɡɚ/, longer /ˈlɔŋɡɚ/, angry /ˈæŋɡri/. Unlike sing /sɪŋ/.
★ The reverse of the row above: singer /ˈsɪŋɚ/ has no /ɡ/. The test is morphology: finger is one morpheme (ng belongs to the root), singer is sing + er.
Why learners replace /ŋ/
Reading /ŋ/ itself is easy. What fails is adding a /ɡ/ that is not there.
An extra /ɡ/ at the end
sing → /sɪŋɡ/, thing → /θɪŋɡ/. The g in the spelling is taken at face value.
★ The most frequent error here, and it changes the word: thin and thing differ by exactly this.
Replaced by /n/
sing → /sɪn/. Many Chinese dialects merge the n and ng codas, so the contrast is lost.
In the drawing /n/ uses the tip and /ŋ/ the tongue back. Both are nasal, so pinching the nose blocks both — look at the tongue.
The /ɡ/ dropped where it belongs
The reverse: finger → /ˈfɪŋɚ/, longer → /ˈlɔŋɚ/.
Morphology decides: ng inside the root takes /ɡ/ (finger, longer, angry); ng + suffix does not (singer, singing).
Word list
These 27 words are grouped by spelling environment. ★ /ŋ/ cannot be word-initial in English, so that dimension is empty. Note that the third group (finger, longer) pronounces the /ɡ/ while final -ng (sing) does not. All three groups use one speaker.
final -ng9
medial -nk9
medial -ng-9
★ These three groups are split by spelling environment, not position: /ŋ/ cannot be word-initial (English phonology forbids it), so the initial dimension is empty. The groups are: final -ng (sing), medial -nk (bank), medial -ng- + vowel (finger, where /ŋɡ/ is pronounced).
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 | ŋ duration | F0 | Pitch range | HNR |
|---|
Where the numbers come from
Every figure on this page can be recomputed from the description below.
Corpus: 12 English sentences synthesised with four American voices (two male, two female), each containing one target word, covering initial, medial and final positions. Audio is 22.05 kHz mono 64 kbps; each clip runs from 1.2 s before to 1.5 s after the target word.
Word boundaries: taken from the engine's word-level timestamps (WordBoundary) and corrected for the fixed offset between the engine timeline and the audio start (estimated per item). The target word is shaded in the waveform.
Phoneme boundaries: measured by forced alignment with the MMS_FA acoustic model. Each character's aligned frame range is kept, then a dynamic program maps the character sequence onto the phoneme sequence (th→θ, ng→ŋ, ck→k). Boundaries fall in the middle of the transition between characters; within-word coverage is 100%. Time precision is the model frame length, 20 ms, and boundaries are taken at midpoints, so effective resolution is about 10 ms.
Pitch: Praat's autocorrelation (to_pitch), 10 ms frame step, search range narrowed to the speaker's estimated range, octave errors removed. «Pitch range» is max minus min F0 within the word.
Why there are no paired differences here: this is a single-phoneme page. Eight other pages are contrast pages (θ/ð, l/r/n, vowel length) where the numbers come from paired words read by one speaker. A single-phoneme page has nothing to compare against — pairing it with a neighbour would measure that neighbour, which belongs to a different page.