The IELTS Speaking band descriptor for Pronunciation is one of four equally-weighted criteria. It's also the criterion candidates obsess over least — usually because they don't know what examiners are actually listening for.
This post fixes that. Here's what the descriptors mean in IPA terms, the specific sounds that separate band 6 from band 8, and how to drill them efficiently.
What "pronunciation" means on the IELTS rubric
The publicly available band descriptors describe pronunciation in vague terms — "individual sounds," "features of connected speech," "intelligibility." Translated to phonetics, examiners are listening for:
- Phoneme accuracy — are you producing distinct /ʃ/ vs /s/, /θ/ vs /s/, long /iː/ vs short /ɪ/?
- Word stress — do you stress photograph on the first syllable, photographer on the second?
- Sentence stress and rhythm — do content words get prominence, function words stay short?
- Connected speech features — linking, assimilation, weak forms.
- Intonation — do questions rise, statements fall, lists pattern correctly?
If you only fix individual sounds, you cap at band 6. To reach band 7+, you need rhythm and connected speech.
Band 6 vs band 8 — what changes
Band 6 pronunciation:
- Mostly intelligible.
- Individual sounds attempted but often substituted (think → "sink").
- Limited use of stress and intonation features.
- Some accent influence that occasionally affects clarity.
Band 8 pronunciation:
- Wide range of features used flexibly and naturally.
- Sustained accurate phoneme production.
- Effective use of stress for meaning.
- Connected speech features (weak forms, linking, elision) used naturally.
- L1 accent present but not affecting clarity.
The single fastest band jump comes from fixing the 5-7 most common phoneme substitutions for your L1, and then drilling rhythm.
The phonemes English learners substitute most
| Target IPA | Common substitution | Example | Fix | |---|---|---|---| | /θ/ | /s/, /t/, /f/ | think /θɪŋk/ → "sink"/"tink"/"fink" | Tongue between teeth, light friction. | | /ð/ | /d/, /z/ | this /ðɪs/ → "dis"/"zis" | Same position as /θ/ but voiced. | | /ʃ/ vs /s/ | merged | ship vs sip | /ʃ/ has rounded lips and tongue body raised. | | /v/ vs /w/ | merged (Slavic, Indian, German speakers) | vest vs west | /v/ touches lips to teeth; /w/ doesn't. | | /æ/ | /e/ or /a/ | cat /kæt/ → "ket" or "cot" | Wide open mouth, tongue front and low. | | /ɪ/ vs /iː/ | merged | ship /ʃɪp/ vs sheep /ʃiːp/ | Duration AND quality differ. | | /ʌ/ | /a/ or /ɔ/ | cut /kʌt/ → "cot"/"caught" | Central, mid, unrounded. |
Examiners hear these substitutions instantly. They're the difference between "clearly L1-influenced" and "L1 influence noticeable but doesn't affect clarity."
Word stress patterns examiners notice
English is a stress-timed language. Stressed syllables are louder, longer, and higher in pitch. Misplaced stress is one of the biggest comprehension blockers.
Patterns to drill:
- Two-syllable nouns and adjectives: stress usually first syllable. PHOto, HAPpy, BEAUtiful.
- Two-syllable verbs: stress usually second syllable. to forGET, to deCIDE.
- Words ending in -tion, -sion, -ic, -ity: stress the syllable before the suffix. eduCAtion, sciENtific, posSIbility.
- Compound nouns: stress the first word. BUS stop, BLACK board, GREENhouse.
A stressed syllable in IPA is marked with /ˈ/ before it: photograph /ˈfəʊtəɡrɑːf/, photography /fəˈtɒɡrəfi/.
If you're not sure where the stress goes, look up the word in IPAtics and the /ˈ/ marker tells you instantly.
Sentence rhythm and weak forms
This is the band 7+ unlock. In natural English, function words (the, of, and, can, to, for, was, were, has, have, do, does, must, should) are reduced to weak forms in connected speech.
| Function word | Strong form | Weak form | |---|---|---| | the | /ðiː/ | /ðə/ | | of | /ɒv/ | /əv/ | | and | /ænd/ | /ənd/, /n̩/ | | can | /kæn/ | /kən/ | | to | /tuː/ | /tə/ | | for | /fɔː/ | /fə/ | | was | /wɒz/ | /wəz/ | | have | /hæv/ | /həv/, /əv/ |
A sentence like I have to go to the shop in natural speech becomes /aɪ həv tə ɡəʊ tə ðə ʃɒp/ — most syllables reduced to schwa. Pronouncing every function word fully marks you as a learner.
Intonation patterns
Four patterns examiners listen for:
- Falling intonation — wh-questions, statements, commands. Where do you LIVE? (falls on LIVE)
- Rising intonation — yes/no questions. Do you live HERE? (rises on HERE)
- Rising then falling — emphatic statements. That was AMAzing!
- List intonation — rise on each item, fall on the last. I bought APples, BAnanas, and PEARS.
Robotic intonation (flat or wrong pattern) immediately caps you at band 6.
The drill that works
A 20-minute daily routine:
- Phoneme audit (5 min). Pick the 3 phonemes you substitute most. Record minimal pair drills: ship/sheep, think/sink, vest/west. Compare to native audio.
- Word stress audit (5 min). Read aloud 10 multi-syllable words from a topic you'll likely discuss in the exam (education, technology, environment). Mark the stress in IPA before recording. Compare to native audio.
- Sentence rhythm (5 min). Read a short paragraph aloud, then mark all function words. Re-read it with weak forms. Record both versions and compare.
- Spontaneous practice (5 min). Answer a Part 2 cue card while applying everything above. Listen back. Note 2-3 specific phoneme/stress errors.
Loop daily for 4 weeks. Examiners will hear the difference.
IPAtics' Speech Analyzer automates steps 1 and 4 — it records you, scores each phoneme, and tells you exactly which articulator (jaw, lips, tongue) was off. You stop guessing what your accent issue is.
The honest summary
IELTS pronunciation isn't about losing your accent. It's about producing distinct phonemes, placing stress correctly, and using rhythm naturally. A heavy accent at band 8 is fine — incomprehensible substitutions at band 6 are not.
Most of the work is muscle memory. Most of the muscle memory is built on accurate input. IPA is the accurate input.
Download IPAtics free for macOS and Windows, or transcribe in your browser without installing. Either way, stop guessing.
Related reading: How I built a pronunciation Anki deck in 10 minutes · Minimal pairs in English · From PDF to fluent: the 2-second pronunciation workflow