Assessment sets
Ten batteries. They are not interchangeable, and that is the point — each one buys depth in an area by giving up coverage somewhere else. Every set lists what it assesses and what it cannot, so you can pick on evidence rather than on the name.
Start with the Core Diagnostic. It is the set the band boundaries were calibrated against, so it is the only one whose result stands on its own. Use a focused set afterwards to go deeper on whatever the core named as your first problem — and use Form B for a re-test, so an improvement is not just familiarity with a passage you have read before.
Sign in to choose a set. Without an account you get the Core Diagnostic, which is the right starting point anyway. An account is what makes the other nine useful: they are designed to be read against a previous result, and history needs somewhere to live. Create an account →
Core Diagnostic — Form A
CalibratedThe full battery. Start here.
5 tasks · about 6 minutes
The complete instrument: every item the product can measure, in about twelve minutes. This is the one the band boundaries were set against, so it is the one to use for a baseline and for the re-test you compare against it.
Assesses 16 of 16 itemsdetails
Measured by this set
- Phonemic Accuracy — needs a pronunciation scorer
- Word Stress — needs a pronunciation scorer
- Articulation — needs a pronunciation scorer
- Connected Speech — needs a pronunciation scorer
- Stress-Timing
- Pitch Variation
- Terminal Contours
- Uptalk
- Focus Words — needs a pronunciation scorer
- Contrastive Emphasis
- Speech Rate
- Disfluency Rate — needs a transcriber
- Pause Placement — needs a pronunciation scorer
- Vocal Stability
- Projection
- Register Shift
Items marked “needs a pronunciation scorer” or “needs a transcriber” depend on a provider being configured. If one is not, the report says so per item rather than guessing.
Tasks: The Mirror · The Exchange · The Emphasis Drill · Off the Page · Two Rooms
Trade-off. Nothing is left out, which means nothing is examined in unusual depth either. If you already know your problem is vowels, a focused set will tell you more about vowels than this will.
Core Diagnostic — Form B
Parallel formSame instrument, different words. Use for your re-test.
5 tasks · about 6 minutes
A parallel form of Form A: the same five tasks, the same coverage, matched difficulty, entirely different material. Use it for a 90-day re-test so an improvement is not just familiarity with a paragraph you have now read twice.
Assesses 16 of 16 itemsdetails
Measured by this set
- Phonemic Accuracy — needs a pronunciation scorer
- Word Stress — needs a pronunciation scorer
- Articulation — needs a pronunciation scorer
- Connected Speech — needs a pronunciation scorer
- Stress-Timing
- Pitch Variation
- Terminal Contours
- Uptalk
- Focus Words — needs a pronunciation scorer
- Contrastive Emphasis
- Speech Rate
- Disfluency Rate — needs a transcriber
- Pause Placement — needs a pronunciation scorer
- Vocal Stability
- Projection
- Register Shift
Items marked “needs a pronunciation scorer” or “needs a transcriber” depend on a provider being configured. If one is not, the report says so per item rather than guessing.
Tasks: The Witness · The Follow-Up · The Emphasis Drill · Second Thoughts · Two Rooms Again
Trade-off. Matched by construction, not yet by measurement. Until the reliability study compares the two forms on the same speakers, treat a small difference between Form A and Form B as noise rather than progress.
Vowel Contrasts
FocusedThe vowel pairs that change which word a listener hears.
2 tasks · about 3 minutes
Dense in the five vowel contrasts that carry the highest functional load in English — ship/sheep, bad/bed, luck/lock, full/fool, caught/coat. If the core diagnostic named vowels as your top finding, this is where you find out which ones and in which words.
Assesses 7 of 16 itemsdetails
Measured by this set
- Phonemic Accuracy — needs a pronunciation scorer
- Word Stress — needs a pronunciation scorer
- Articulation — needs a pronunciation scorer
- Connected Speech — needs a pronunciation scorer
- Stress-Timing
- Focus Words — needs a pronunciation scorer
- Contrastive Emphasis
Not measured by this set
- Pitch Variation — this set has no dialogue
- Terminal Contours — this set has no dialogue
- Uptalk — this set has no dialogue
- Speech Rate — this set has no unscripted speech
- Disfluency Rate — this set has no unscripted speech
- Pause Placement — this set has no unscripted speech
- Vocal Stability — this set has no unscripted speech
- Projection — this set has no register shift
- Register Shift — this set has no register shift
Items marked “needs a pronunciation scorer” or “needs a transcriber” depend on a provider being configured. If one is not, the report says so per item rather than guessing.
Tasks: Minimal Pairs · The Emphasis Drill
Trade-off. No dialogue and no unscripted speech, so nothing about intonation, hesitation or pacing. It is a magnifying glass, not a second opinion — run it alongside a core result, not instead of one.
Consonants & Endings
FocusedFinal sounds, clusters, and the endings that carry grammar.
2 tasks · about 3 minutes
Word-final consonants and clusters, plus the -ed and -s endings that mark tense and number. Dropping these is the most common single cause of a listener asking for a repeat, because the loss changes grammar rather than just accent.
Assesses 7 of 16 itemsdetails
Measured by this set
- Phonemic Accuracy — needs a pronunciation scorer
- Word Stress — needs a pronunciation scorer
- Articulation — needs a pronunciation scorer
- Connected Speech — needs a pronunciation scorer
- Stress-Timing
- Focus Words — needs a pronunciation scorer
- Contrastive Emphasis
Not measured by this set
- Pitch Variation — this set has no dialogue
- Terminal Contours — this set has no dialogue
- Uptalk — this set has no dialogue
- Speech Rate — this set has no unscripted speech
- Disfluency Rate — this set has no unscripted speech
- Pause Placement — this set has no unscripted speech
- Vocal Stability — this set has no unscripted speech
- Projection — this set has no register shift
- Register Shift — this set has no register shift
Items marked “needs a pronunciation scorer” or “needs a transcriber” depend on a provider being configured. If one is not, the report says so per item rather than guessing.
Tasks: The Endings · The Emphasis Drill
Trade-off. No dialogue and no unscripted speech. Deliberately unnatural in places — real sentences do not stack six final clusters in a row, but real sentences also do not test them.
Intonation & Questions
FocusedWhether you sound like you are asking or telling.
2 tasks · about 3 minutes
A long dialogue with statements, yes/no questions, tag questions and echo questions in close alternation — the pattern English uses to mark telling from asking. This is the set for anyone told they sound uncertain, or whose statements are heard as questions.
Assesses 5 of 16 itemsdetails
Measured by this set
- Pitch Variation
- Terminal Contours
- Uptalk
- Focus Words — needs a pronunciation scorer
- Contrastive Emphasis
Not measured by this set
- Phonemic Accuracy — this set has no read-aloud passage
- Word Stress — this set has no read-aloud passage
- Articulation — this set has no read-aloud passage
- Connected Speech — this set has no read-aloud passage
- Stress-Timing — this set has no read-aloud passage
- Speech Rate — this set has no unscripted speech
- Disfluency Rate — this set has no unscripted speech
- Pause Placement — this set has no unscripted speech
- Vocal Stability — this set has no unscripted speech
- Projection — this set has no register shift
- Register Shift — this set has no register shift
Items marked “needs a pronunciation scorer” or “needs a transcriber” depend on a provider being configured. If one is not, the report says so per item rather than guessing.
Tasks: The Long Exchange · The Emphasis Drill
Trade-off. No read-aloud passage and no unscripted speech, so nothing about individual sounds, hesitation, or pacing. It answers one question thoroughly and the others not at all.
Fluency Under Load
FocusedHesitation and pacing when the topic is genuinely hard.
2 tasks · about 5 minutes
Two and a half minutes of unscripted speech on a prompt chosen to be cognitively demanding, plus a register shift. More speech means a far better estimate of hesitation — a rate measured over sixty seconds is noisy, and this triples the sample.
Assesses 6 of 16 itemsdetails
Measured by this set
- Speech Rate
- Disfluency Rate — needs a transcriber
- Pause Placement — needs a pronunciation scorer
- Vocal Stability
- Projection
- Register Shift
Not measured by this set
- Phonemic Accuracy — this set has no read-aloud passage
- Word Stress — this set has no read-aloud passage
- Articulation — this set has no read-aloud passage
- Connected Speech — this set has no read-aloud passage
- Stress-Timing — this set has no read-aloud passage
- Pitch Variation — this set has no dialogue
- Terminal Contours — this set has no dialogue
- Uptalk — this set has no dialogue
- Focus Words — this set has no emphasis drill
- Contrastive Emphasis — this set has no emphasis drill
Items marked “needs a pronunciation scorer” or “needs a transcriber” depend on a provider being configured. If one is not, the report says so per item rather than guessing.
Tasks: Under Load · Same Point, Two Rooms
Trade-off. No read-aloud passage and no dialogue, so no individual-sound scoring and no intonation. Also the hardest set to sit: an abstract prompt with nothing to read is uncomfortable, which is the point.
Rhythm & Timing
FocusedThe English beat: long stressed syllables, squeezed unstressed ones.
2 tasks · about 3 minutes
Built on long words with strong alternating stress and heavy unstressed reduction. If your speech is described as flat, robotic, or evenly paced — common for speakers of syllable-timed first languages such as Spanish, French, Turkish or Japanese — this is the set that shows it.
Assesses 7 of 16 itemsdetails
Measured by this set
- Phonemic Accuracy — needs a pronunciation scorer
- Word Stress — needs a pronunciation scorer
- Articulation — needs a pronunciation scorer
- Connected Speech — needs a pronunciation scorer
- Stress-Timing
- Focus Words — needs a pronunciation scorer
- Contrastive Emphasis
Not measured by this set
- Pitch Variation — this set has no dialogue
- Terminal Contours — this set has no dialogue
- Uptalk — this set has no dialogue
- Speech Rate — this set has no unscripted speech
- Disfluency Rate — this set has no unscripted speech
- Pause Placement — this set has no unscripted speech
- Vocal Stability — this set has no unscripted speech
- Projection — this set has no register shift
- Register Shift — this set has no register shift
Items marked “needs a pronunciation scorer” or “needs a transcriber” depend on a provider being configured. If one is not, the report says so per item rather than guessing.
Tasks: The Beat · The Emphasis Drill
Trade-off. No dialogue and no unscripted speech. The passage is deliberately word-heavy and reads like a tongue-twister in places, because natural prose does not stack five-syllable words often enough to test this.
Register & Delivery
FocusedDo you sound different at work than you do with friends?
2 tasks · about 4 minutes
Three deliveries of the same news to three different audiences, plus an unscripted answer. Register can only ever be measured as a difference, so holding the content constant and changing the listener is the only way to see it.
Assesses 6 of 16 itemsdetails
Measured by this set
- Speech Rate
- Disfluency Rate — needs a transcriber
- Pause Placement — needs a pronunciation scorer
- Vocal Stability
- Projection
- Register Shift
Not measured by this set
- Phonemic Accuracy — this set has no read-aloud passage
- Word Stress — this set has no read-aloud passage
- Articulation — this set has no read-aloud passage
- Connected Speech — this set has no read-aloud passage
- Stress-Timing — this set has no read-aloud passage
- Pitch Variation — this set has no dialogue
- Terminal Contours — this set has no dialogue
- Uptalk — this set has no dialogue
- Focus Words — this set has no emphasis drill
- Contrastive Emphasis — this set has no emphasis drill
Items marked “needs a pronunciation scorer” or “needs a transcriber” depend on a provider being configured. If one is not, the report says so per item rather than guessing.
Tasks: The Case · Three Rooms
Trade-off. No read-aloud passage and no dialogue. Register and Projection are reported as observations rather than bands even here — microphone and room affect them more than your speech does, and a number would overstate what we know.
Interview Simulation
FocusedThe specific speech that gets judged in interviews.
3 tasks · about 5 minutes
An interview dialogue, an unscripted answer to a question you cannot rehearse, and a closing pitch. Interviews compress the things this product measures — hesitation, uptalk, and trailing off at the ends of sentences — into fifteen high-stakes minutes.
Assesses 9 of 16 itemsdetails
Measured by this set
- Pitch Variation
- Terminal Contours
- Uptalk
- Speech Rate
- Disfluency Rate — needs a transcriber
- Pause Placement — needs a pronunciation scorer
- Vocal Stability
- Projection
- Register Shift
Not measured by this set
- Phonemic Accuracy — this set has no read-aloud passage
- Word Stress — this set has no read-aloud passage
- Articulation — this set has no read-aloud passage
- Connected Speech — this set has no read-aloud passage
- Stress-Timing — this set has no read-aloud passage
- Focus Words — this set has no emphasis drill
- Contrastive Emphasis — this set has no emphasis drill
Items marked “needs a pronunciation scorer” or “needs a transcriber” depend on a provider being configured. If one is not, the report says so per item rather than guessing.
Tasks: The Room · The Hard Question · Pitch and Truth
Trade-off. No read-aloud passage, so no phoneme-level scoring. This set tells you how you come across under pressure, not which sounds you produce incorrectly.
Quick Screen
ScreenFour minutes. Is this worth twelve?
2 tasks · about 2 minutes
A short passage and half a minute of unscripted speech. Enough to see whether the obvious problem is sounds, rhythm or fluency, and enough to check your microphone is usable before committing to the full battery.
Assesses 9 of 16 itemsdetails
Measured by this set
- Phonemic Accuracy — needs a pronunciation scorer
- Word Stress — needs a pronunciation scorer
- Articulation — needs a pronunciation scorer
- Connected Speech — needs a pronunciation scorer
- Stress-Timing
- Speech Rate
- Disfluency Rate — needs a transcriber
- Pause Placement — needs a pronunciation scorer
- Vocal Stability
Not measured by this set
- Pitch Variation — this set has no dialogue
- Terminal Contours — this set has no dialogue
- Uptalk — this set has no dialogue
- Focus Words — this set has no emphasis drill
- Contrastive Emphasis — this set has no emphasis drill
- Projection — this set has no register shift
- Register Shift — this set has no register shift
Items marked “needs a pronunciation scorer” or “needs a transcriber” depend on a provider being configured. If one is not, the report says so per item rather than guessing.
Tasks: The Short Read · Thirty Seconds
Trade-off. Every band from this set is low-confidence and the report says so. Thirty seconds of speech cannot support a hesitation rate you should act on, and the passage does not cover the full phoneme inventory. Treat the result as a direction, not a diagnosis.
Why the results are not interchangeable
A band means “this much control, measured this way”. Change the material and you change the measurement — so a higher band on a different set is not evidence that anything improved. It may only mean the second test was easier for you.
Your progress view therefore compares like with like: it draws a line between two runs only when they are the same set, or deliberate parallel forms of it (Core Form A and Form B). Everything else is listed but never trended. Two sets are declared parallel by construction — same tasks, same coverage, matched difficulty — and that claim will be checked by measurement in the reliability study, not before.