Section 258 of 440
Complete canonical tutorial. This reader section contains the same teaching body as PWR-075 · Speech-in-noise understanding. Open the Power dossier.
PWR-075 · SELF GUIDED full tutorial
Write exactly what you heard, score key words, and keep replay and noise fixed
Listening in noise can improve on selected tasks; an app score is not the same as easier real conversation. Speech-in-noise understanding is tested with a fixed synthetic sentence/noise mix at comfortable level. The learner transcribes before checking, scores key words rather than confidence, records replay and listening burden, and retests a new sentence. It is not a hearing treatment or a substitute for conversation outcomes. The scoring sheet distinguishes acoustic confusions from language reconstruction by marking the exact substituted consonant, vowel or content word and whether the sentence remained plausible.
1 · Permission and limits
Know exactly what you may do
2 · Get ready
Gather what you need and check the starting conditions
What you need
- An exact 20-sentence corpus with five keyed words each. Set A: A1 ‘The red kite crossed the quiet lake’ [red,kite,crossed,quiet,lake]; A2 ‘Mina placed seven blue cups beside the window’ [Mina,seven,blue,cups,window]; A3 ‘A small dog waited under the wooden bridge’ [small,dog,waited,wooden,bridge]; A4 ‘The baker carried warm bread to the market’ [baker,carried,warm,bread,market]; A5 ‘Three green buses left before the heavy rain’ [three,green,buses,left,rain]; A6 ‘Jon found the silver key behind the sofa’ [Jon,silver,key,behind,sofa]; A7 ‘The nurse packed clean towels in the yellow bag’ [nurse,clean,towels,yellow,bag]; A8 ‘Two white boats passed the old lighthouse’ [two,white,boats,old,lighthouse]; A9 ‘Sara sent the final letter on Monday morning’ [Sara,final,letter,Monday,morning]; A10 ‘The tall tree shaded the stone garden path’ [tall,tree,shaded,stone,path]. Set B: B1 ‘The black cat watched three birds near the fence’ [black,cat,three,birds,fence]; B2 ‘Omar moved the round table into the kitchen’ [Omar,round,table,into,kitchen]; B3 ‘Four red trains crossed the narrow valley’ [four,red,trains,narrow,valley]; B4 ‘Lena stored the glass jar above the sink’ [Lena,glass,jar,above,sink]; B5 ‘The young farmer closed the gate before sunset’ [young,farmer,closed,gate,sunset]; B6 ‘Maya chose a purple scarf for the concert’ [Maya,purple,scarf,for,concert]; B7 ‘Six grey stones lined the river bank’ [six,grey,stones,river,bank]; B8 ‘The driver left the empty box near reception’ [driver,empty,box,near,reception]; B9 ‘Noah wrote the short answer with a pencil’ [Noah,short,answer,with,pencil]; B10 ‘The bright moon lit the snowy field’ [bright,moon,lit,snowy,field].
- A helper creates the clips in one audio editor using the same synthetic voice, speaking rate and mono pink-noise track. Normalise all voice clips to the same RMS level, mix the continuous pink noise 6 dB below that voice level, add 250 ms noise before and after each sentence, export under neutral A1–A10/B1–B10 codes and fold the transcript/key. If the editor cannot preserve the same voice/noise ratio, label the new mix a different configuration rather than comparing it.
- Same playback device and headphones/speakers at locked comfortable level.
- Transcript keys with premarked key words.
- Writing/AAC method and trial record.
Before you start
- Choose a quiet room and one comfortable level; lock voice, noise and level.
- Use one replay rule for every sentence.
- Keep hearing aids or implants in the ordinary declared programme.
- Do not raise level through discomfort or use the task as a hearing-aid fitting check.
3 · The method
Follow these steps in order
- Lock voice, noise and level
Record speaker, noise type, signal-to-noise mix, device, output and hearing-device setting.
Why: Each element changes intelligibility.
Check: The audio configuration is complete.
- Learn the scoring rule
Mark only exact key words as correct; keep substitutions and omissions visible.
Why: A fixed word rule is more reliable than “understood the gist”.
Check: A clear practice sentence is scored correctly.
- Listen under one replay rule
Play each sentence at the fixed level once, or twice only if that rule applies to all trials.
Why: Selective replay gives difficult items extra evidence.
Check: Play count is recorded.
- Write before checking
Transcribe exactly what was heard, including uncertain words, before seeing options or the key.
Why: Free report reveals real substitutions rather than recognition from choices.
Check: A full response is locked before feedback.
- Score words and errors
Count correct key words, substitutions, omissions and added words separately.
Why: Different errors can yield the same percent correct.
Check: Raw word-level errors remain available.
- Record burden
Rate listening effort, fatigue and discomfort after each short block.
Why: Accuracy gained through unsustainable effort may not help conversation.
Check: Burden is separate from word score.
- Retest a new mix
After delay, use unfamiliar sentences and, only after stable baseline, change one noise feature.
Why: New language prevents sentence memorisation; one change tests transfer.
Check: No sentence repeats and the changed feature is declared.
- Classify each word substitution
For every wrong key word, write the heard word beside the target and mark sound-neighbour, omitted, inserted or meaning-based guess. Do this before replaying any isolated word.
Why: Two people with the same key-word score may have different problems: one misses speech sounds while another fills gaps with a plausible sentence.
Check: Every wrong key word has one preserved heard form and one error category, or is explicitly marked unintelligible.
4 · Worked example
See the whole method used once
Scenario
Dara uses Set A's ten synthetic sentences mixed with the specified continuous mono pink noise at one locked level.
Walkthrough
- Dara records the synthetic voice, pink-noise track at 6 dB below the voice RMS, headphones and hearing-aid programme.
- They listen once and type exactly what they heard before checking.
- The key scores five words per sentence; substitutions and omissions remain visible.
- After five sentences Dara records effort and rests, without raising level.
- A week later new sentences use the same mix before any different-noise transfer set.
- On A1 Dara writes “coat” for keyed word “kite” and marks the preserved heard form as a sound-neighbour candidate; an invented but plausible location word is marked a meaning-based guess.
Result
Dara improves key-word accuracy on novel sentences at the same mix with stable effort. This does not prove easier real conversation or clinical hearing improvement. The error profile identifies what changed in the listening task instead of treating all wrong sentences as equivalent.
5 · Right and wrong
Compare correct or safer execution with the common wrong version
| Moment | Right / safer | Wrong / riskier | Why it matters |
|---|---|---|---|
| Level | Keep the comfortable level fixed. | Raise volume on difficult sentences. | That changes audibility and exposure. |
| Response | Write free recall before the key. | Choose from options after seeing likely words. | Options provide language cues not present in listening. |
| Replay | Apply the same replay rule to all items. | Replay only misses. | Selective replay biases comparison. |
| Score | Count key words and error types. | Score confidence or gist as accuracy. | Confidence and gist can hide wrong words. |
| Transfer | Use new sentences and measure conversation separately. | Call app improvement hearing treatment. | Task gains may not transfer to daily participation. |
| Wrong key word | Preserve the exact heard word and classify the substitution. | Replace every error with a blank or mark only sentence wrong. | The heard form shows whether noise obscured a sound or the listener reconstructed meaning. |
6 · Common mistakes
Spot the error and apply the correction
| Mistake | Fix |
|---|---|
| Raise volume on difficult sentences. | Keep the comfortable level fixed. |
| Choose from options after seeing likely words. | Write free recall before the key. |
| Replay only misses. | Apply the same replay rule to all items. |
| Score confidence or gist as accuracy. | Count key words and error types. |
| Call app improvement hearing treatment. | Use new sentences and measure conversation separately. |
| The transcript is corrected for grammar before scoring, hiding what the learner actually perceived. | Freeze the verbatim first transcript, score it, and place any grammatical interpretation in a separate note. |
7 · Practice
Turn the steps into a usable skill
First session
- A helper creates and conceals the A and B mixes from the exact corpus, same synthetic voice and pink-noise rule, then folds the five-word keys.
- Lock the audio chain, hearing-device setting, comfortable level and one-play rule; practise scoring one clear unscored sentence.
- Run A1–A5, freeze each verbatim transcript before feedback and count the five keyed words.
- Rest and record burden, then run A6–A10 under the same rule.
- Classify every wrong keyed word before reviewing total accuracy, and after at least 48 hours run novel B1–B10 at the unchanged mix.
Repeat plan
Use 5–10 novel sentences once or twice weekly with breaks. Progress by a single noise/SNR change only after key-word accuracy is stable and burden acceptable.
Progress when
- Novel-sentence key-word accuracy improves.
- Substitutions/omissions decline without more replay.
- Listening effort stays stable or falls.
Do not progress when
- Volume or hearing-device programme changes.
- Tinnitus, pain, fatigue or distress rises.
- The score replaces clinical or real-conversation evaluation.
8 · Check the result
Measure what changed
Exact keyed-word accuracy and preserved substitution types on the supplied sentence-in-pink-noise corpus
How: Record correct key words, substitutions, omissions, added words, replay, effort, fatigue, voice/noise/SNR and device configuration. Add the preserved heard word and substitution category for every incorrect key word.
Good result: Novel-sentence accuracy improves without increased level, replay or burden and a separate real-conversation goal is not worse.
This does not prove: Declare speech corpus, language, masker, spatial layout, SNR, device state and fatigue; separate trained-material accuracy from new-speaker and real-conversation outcomes.
Self-check
- Can you demonstrate “Learn the scoring rule”? A clear practice sentence is scored correctly.
- Can you demonstrate “Write before checking”? A full response is locked before feedback.
- Can you demonstrate “Retest a new mix”? No sentence repeats and the changed feature is declared.
- Can you show one acoustic substitution and one meaning-based guess in the transcript without changing the learner’s original wording?
9 · Stop, adapt or get help
Keep the safety boundary practical
Stop and get help
- End the sentence block if the speech/noise mix distorts or provokes ear pain, new fullness, louder tinnitus, dizziness or headache.
- Seek urgent care for sudden hearing loss or severe vertigo.
- See an audiologist when speech understanding changes, differs by ear or causes meaningful participation difficulty.
Accessibility and adaptations
- Use typing, handwriting, speech-to-text correction, AAC or partner transcription while preserving the learner’s words.
- Provide visual instructions and non-auditory break cues.
- Keep usual hearing technology and communication strategies as declared supports.
10 · Evidence and limits
Why these instructions are here
- primary research
A randomised trial found benefits from a defined phoneme-discrimination programme in older adults with mild hearing loss.
Benefits of phoneme discrimination training in a randomized controlled trial of 50- to 74-year-olds with mild hearing loss - official guidance
WHO guidance supports controlling sound level and cumulative duration during headphone tasks.
Deafness and hearing loss: Safe listening
Limits
- Declare speech corpus, language, masker, spatial layout, SNR, device state and fatigue; separate trained-material accuracy from new-speaker and real-conversation outcomes.
- Unsupported claim: Hear any conversation anywhere.
- Unsupported claim: Train away hearing loss.
- Unsupported claim: Stop using hearing aids or captions.
- Synthetic sentence gains do not diagnose hearing, fit a device or guarantee easier conversation in new speakers and rooms.
Open the complete canonical research register
- Primary empirical supportLimiting / contraryBenefits of phoneme discrimination training in a randomized controlled trial of 50- to 74-year-olds with mild hearing loss
M. A. Ferguson; H. Henshaw; D. P. Clark; D. R. Moore · 2014 · Primary research
- Primary empirical supportLimiting / contraryClinical Effectiveness of an At-Home Auditory Training Program: A Randomized Controlled Trial
L. E. Humes; K. G. Skinner; D. L. Kinney; S. E. Rogers; A. K. Main; T. M. Quigley · 2019 · Primary research
- Primary empirical supportLimiting / contraryEffect of Serious Gaming on Speech-in-Noise Intelligibility in Adult Cochlear Implantees: A Randomized Controlled Study
P. Reynard; V. Attina; S. Idriss; R. Hermann; C. Barilly; E. Veuillet; C. A. Joly; H. Thai-Van · 2022 · Primary research
- Limiting / contraryOfficial boundary contextDeafness and hearing loss: Safe listening
World Health Organization · 2026 · Official guidance
Read the complete evidence interpretation on the Power dossier.
Tutorial delivery controls
Learn, adapt, troubleshoot and resume
Progress is saved only in this browser on this device.
Step-by-step learner mode
Each activity includes its success check, a nearby accessible alternative and an “I’m stuck” correction path. Alternatives preserve the target where possible; when they change the task, Titan labels them as related rather than equivalent.
Lock voice, noise and level
Record speaker, noise type, signal-to-noise mix, device, output and hearing-device setting.
Each element changes intelligibility.
The audio configuration is complete.
I’m stuck on this step
Reset: Re-read this authored instruction — “Record speaker, noise type, signal-to-noise mix, device, output and hearing-device setting.” — and its success check, then attempt only this step.
Possible snag: Call app improvement hearing treatment.
Correction: Use new sentences and measure conversation separately.
Stop / get help: End the sentence block if the speech/noise mix distorts or provokes ear pain, new fullness, louder tinnitus, dizziness or headache.
Learn the scoring rule
Mark only exact key words as correct; keep substitutions and omissions visible.
A fixed word rule is more reliable than “understood the gist”.
A clear practice sentence is scored correctly.
I’m stuck on this step
Reset: Re-read this authored instruction — “Mark only exact key words as correct; keep substitutions and omissions visible.” — and its success check, then attempt only this step.
Possible snag: The result from “Mark only exact key words as correct; keep substitutions and omissions visible.” does not yet meet this declared check: A clear practice sentence is scored correctly.
Correction: Return to the start of “Learn the scoring rule”, reduce complexity or pace, and repeat only the part needed to satisfy: “A clear practice sentence is scored correctly.”
Stop / get help: End the sentence block if the speech/noise mix distorts or provokes ear pain, new fullness, louder tinnitus, dizziness or headache.
Listen under one replay rule
Play each sentence at the fixed level once, or twice only if that rule applies to all trials.
Selective replay gives difficult items extra evidence.
Play count is recorded.
I’m stuck on this step
Reset: Re-read this authored instruction — “Play each sentence at the fixed level once, or twice only if that rule applies to all trials.” — and its success check, then attempt only this step.
Possible snag: Raise volume on difficult sentences.
Correction: Keep the comfortable level fixed.
Possible snag: Replay only misses.
Correction: Apply the same replay rule to all items.
Stop / get help: End the sentence block if the speech/noise mix distorts or provokes ear pain, new fullness, louder tinnitus, dizziness or headache.
Write before checking
Transcribe exactly what was heard, including uncertain words, before seeing options or the key.
Free report reveals real substitutions rather than recognition from choices.
A full response is locked before feedback.
I’m stuck on this step
Reset: Re-read this authored instruction — “Transcribe exactly what was heard, including uncertain words, before seeing options or the key.” — and its success check, then attempt only this step.
Possible snag: Choose from options after seeing likely words.
Correction: Write free recall before the key.
Possible snag: The transcript is corrected for grammar before scoring, hiding what the learner actually perceived.
Correction: Freeze the verbatim first transcript, score it, and place any grammatical interpretation in a separate note.
Stop / get help: End the sentence block if the speech/noise mix distorts or provokes ear pain, new fullness, louder tinnitus, dizziness or headache.
Score words and errors
Count correct key words, substitutions, omissions and added words separately.
Different errors can yield the same percent correct.
Raw word-level errors remain available.
I’m stuck on this step
Reset: Re-read this authored instruction — “Count correct key words, substitutions, omissions and added words separately.” — and its success check, then attempt only this step.
Possible snag: Score confidence or gist as accuracy.
Correction: Count key words and error types.
Stop / get help: End the sentence block if the speech/noise mix distorts or provokes ear pain, new fullness, louder tinnitus, dizziness or headache.
Record burden
Rate listening effort, fatigue and discomfort after each short block.
Accuracy gained through unsustainable effort may not help conversation.
Burden is separate from word score.
I’m stuck on this step
Reset: Re-read this authored instruction — “Rate listening effort, fatigue and discomfort after each short block.” — and its success check, then attempt only this step.
Possible snag: The result from “Rate listening effort, fatigue and discomfort after each short block.” does not yet meet this declared check: Burden is separate from word score.
Correction: Return to the start of “Record burden”, reduce complexity or pace, and repeat only the part needed to satisfy: “Burden is separate from word score.”
Stop / get help: End the sentence block if the speech/noise mix distorts or provokes ear pain, new fullness, louder tinnitus, dizziness or headache.
Retest a new mix
After delay, use unfamiliar sentences and, only after stable baseline, change one noise feature.
New language prevents sentence memorisation; one change tests transfer.
No sentence repeats and the changed feature is declared.
I’m stuck on this step
Reset: Re-read this authored instruction — “After delay, use unfamiliar sentences and, only after stable baseline, change one noise feature.” — and its success check, then attempt only this step.
Possible snag: The result from “After delay, use unfamiliar sentences and, only after stable baseline, change one noise feature.” does not yet meet this declared check: No sentence repeats and the changed feature is declared.
Correction: Return to the start of “Retest a new mix”, reduce complexity or pace, and repeat only the part needed to satisfy: “No sentence repeats and the changed feature is declared.”
Stop / get help: End the sentence block if the speech/noise mix distorts or provokes ear pain, new fullness, louder tinnitus, dizziness or headache.
Classify each word substitution
For every wrong key word, write the heard word beside the target and mark sound-neighbour, omitted, inserted or meaning-based guess. Do this before replaying any isolated word.
Two people with the same key-word score may have different problems: one misses speech sounds while another fills gaps with a plausible sentence.
Every wrong key word has one preserved heard form and one error category, or is explicitly marked unintelligible.
I’m stuck on this step
Reset: Re-read this authored instruction — “For every wrong key word, write the heard word beside the target and mark sound-neighbour, omitted, inserted or meaning-based guess. Do this before replaying any isolated word.” — and its success check, then attempt only this step.
Possible snag: The result from “For every wrong key word, write the heard word beside the target and mark sound-neighbour, omitted, inserted or meaning-based guess. Do this before replaying any isolated word.” does not yet meet this declared check: Every wrong key word has one preserved heard form and one error category, or is explicitly marked unintelligible.
Correction: Return to the start of “Classify each word substitution”, reduce complexity or pace, and repeat only the part needed to satisfy: “Every wrong key word has one preserved heard form and one error category, or is explicitly marked unintelligible.”
Stop / get help: End the sentence block if the speech/noise mix distorts or provokes ear pain, new fullness, louder tinnitus, dizziness or headache.
Correct versus incorrect execution
These accessible process diagrams are built from the tutorial’s own right/wrong teaching. They are not anatomical illustrations and do not add technique beyond the canonical tutorial.
Keep the comfortable level fixed.
Raise volume on difficult sentences.
Write free recall before the key.
Choose from options after seeing likely words.
Apply the same replay rule to all items.
Replay only misses.
Count key words and error types.
Score confidence or gist as accuracy.
Use new sentences and measure conversation separately.
Call app improvement hearing treatment.
Preserve the exact heard word and classify the substitution.
Replace every error with a blank or mark only sentence wrong.
Method-structure checklist
10 of 10 structural checks present
- Ordered, Power-specific instructions — present
- Every activity has a success check — present
- Materials or supplied records are declared — present
- Measurement or assessment rule is present — present
- Tutorial-specific troubleshooting is present — present
- Stopping or escalation boundary is present — present
- Every activity has an adjacent alternative — present
- Correct-versus-incorrect comparison is present — present
- Evidence context is bound to the Power record — present
- Planning metadata is present — present
The method-readiness band and presence checklist assess tutorial presentation and are separate from evidence quality for the underlying Power. They are automated editorial aids, not human approval.
Manual editorial sign-off: Pending. This tutorial must not display a human-approved state until an identified editor signs the exact content hash.