Revision 7T · Full Tutorial Edition · Updated 1 September 2026
PWR-076 · SELF GUIDED full tutorial
Separate a sound scene into sources, time courses and distinguishing cues
The brain can learn a specific sound pattern without becoming a universal noise filter. Auditory scene analysis means organising mixed sound into streams. The learner hears a synthetic scene, lists sources and when they change, names the cue—pitch, timing, spatial position or timbre—then checks isolated stems. The task never uses covert recordings or surveillance. A stream-continuity probe temporarily removes one source segment; the learner predicts whether the remaining sound will split, merge or reveal a previously masked event.
What you will produceOn coded eight-second mixtures A and B, the learner inventories the three generated sources, marks each supplied onset or offset to the nearest one-second grid line, names the pitch, timing or timbre cue used, and tests one written unmasking prediction against the companion file.
Full step-by-step individual tutorial · TLU-PWR-076
On coded eight-second mixtures A and B, the learner inventories the three generated sources, marks each supplied onset or offset to the nearest one-second grid line, names the pitch, timing or timbre cue used, and tests one written unmasking prediction against the companion file.
Inventory sources in coded Scene A on one listen, mark their time boxes on one replay and name pitch, timing or timbre evidence before stems are revealed.
Score source misses, false sources and one-second timing errors, then write and test one A/B probe unmasking prediction.
Use synthetic or consented learner-owned audio, usual hearing support and an accessible timeline response at fixed comfortable level.
Qualified help is required for
Assessing a new/persistent hearing or auditory-processing complaint and selecting clinical testing or hearing-device treatment.
Using scene separation for alarms, surveillance, traffic, work monitoring, mobility or another safety/privacy-sensitive activity.
Changing hearing-device programmes or clinical listening therapy because a synthetic mixture changed.
Never do this from the page alone
Covertly record people/private places, study isolated stems before the inventory or loop the mix until every burst is memorised.
Raise level or claim one confirmed probe creates a universal noise filter, alarm skill or professional monitoring ability.
Continue through ear pain, louder tinnitus, vertigo, headache, distorted sound or marked listening burden.
2 · Get ready
Gather what you need and check the starting conditions
What you need
A helper-built mono Scene A, eight seconds long: pink noise runs from 0.0–8.0 seconds at −18 dBFS; a 120-Hz sine runs from 0.0–8.0 at −24 dBFS; and an 880-Hz sine makes 100-ms bursts beginning at 2.0, 5.0 and 7.0 seconds at −12 dBFS. Give every generated track a 10-ms fade at its outer edges, mix the tracks, label the learner file only A and keep the three isolated tracks hidden until scoring.
A helper-built mono Scene B, eight seconds long: a 180-Hz sine runs from 0.0–8.0 seconds at −24 dBFS; 600-Hz, 150-ms bursts begin at 1.0, 3.0 and 6.0 seconds at −15 dBFS; and a 1,200-Hz sine starts at 4.0 seconds and ends at 8.0 seconds at −21 dBFS. Use 10-ms outer fades, mix, code the learner file B and retain isolated tracks as the key.
Two companion files made without changing the mix level: A-probe mutes only the 120-Hz track from 4.0–6.0 seconds; B-probe mutes only the 180-Hz track from 2.0–4.0 seconds. The learner receives a probe only after committing the original analysis.
The same headphones or speakers at a comfortable locked setting, plus a sheet with an eight-box one-second timeline and columns for source description, start, stop/change, cue, confidence, miss and false source.
Before you start
Use synthetic or consented learner-owned audio only.
Lock level, device and hearing-device setting.
Set one replay rule and a maximum short block.
Do not use the method for covert listening, professional monitoring or safety alarms.
3 · The method
Follow these steps in order
Learn the cue vocabulary
Review pitch, timing, spatial position and timbre—the sound quality that distinguishes sources.
Why: Cue labels make stream separation explicit.
Check: The learner can match each cue to an example.
Lock the scene setup
Record coded file A or B, device, headphones/speakers, level and the fixed rule of one inventory listen plus one timeline replay.
Why: Scene composition and playback define the task.
Check: The exact scene version is named.
First pass: list sources
Listen once to coded mixture A or B and list each perceived source without checking the stems or timeline key.
Why: A first pass captures spontaneous segregation.
Check: Sources and uncertain guesses are both recorded.
Second pass: map time
Replay exactly once and shade the one-second boxes in which each source starts, stops or produces a burst; use a question mark when the boundary is uncertain.
Why: Time course helps separate overlapping streams.
Check: Each listed source has a timeline mark.
Name the separating cue
For each source, state whether pitch, timing, position or timbre made it distinct.
Why: A source label without the cue does not teach the method.
Check: Every confident source has at least one cue.
Check against stems
After committing the scene analysis, uncover the source recipe and play the three isolated tracks. Mark a missing real track as a miss, an invented track as a false source and a boundary more than one grid box away as a timing error.
Why: Stems provide specific feedback without erasing the mixture response.
Check: Original scene answers remain visible.
Retest a novel arrangement
At least 48 hours later, analyse the other coded scene using the same two-listen rule before viewing its recipe.
Why: A novel scene tests the analysis method rather than memorised content.
Check: The retest arrangement is unfamiliar.
Use a continuity probe
After scoring the original mixture, write which burst or onset should be easier to hear when the named low-frequency track is removed during the marked interval. Then play A-probe or B-probe once and record whether the predicted event became clearer, unchanged or less clear.
Why: Removing one source tests the proposed grouping cue and reveals masking that an isolated-stem review cannot show.
Check: The prediction names the affected source and interval, and the companion result is recorded without changing the original scene score.
4 · Worked example
See the whole method used once
Scenario
Mika receives coded Scene A without its recipe and uses the eight-box timeline.
Walkthrough
Mika writes device and level, hears A once and lists a hiss, a low hum and two high beeps; the last item is marked uncertain.
On the one permitted replay, Mika shades the hiss and hum across boxes 0–8 and puts beep marks at 2, 5 and 7 seconds.
Mika labels timbre as the hiss-versus-hum cue and timing plus higher pitch as the beep cue.
The hidden recipe confirms all three tracks and all three burst times. Because the inventory said only two beeps but the timeline has three, Mika records one inconsistent count rather than erasing either answer.
Before hearing A-probe, Mika predicts that removing the 120-Hz hum from 4–6 seconds will make the 5-second burst clearer.
The probe makes that burst clearer, so the prediction is confirmed for this file; Mika does not infer that all background noise can be filtered.
Result
Mika improves source and timing identification in synthetic scenes. This does not create a universal noise filter or professional monitoring skill. The probe connects a missed bell to masking by the steady fan rather than to a general listening deficit.
5 · Right and wrong
Compare correct or safer execution with the common wrong version
Right and wrong comparison
Moment
Right / safer
Wrong / riskier
Why it matters
Audio source
Use synthetic or consented audio.
Record people or environments covertly.
Privacy and legality control the material.
First pass
Commit a source inventory before stems.
Study isolated stems first.
Pre-exposure turns the task into recognition memory.
Cue
Name pitch, timing, position or timbre evidence.
Say only “I heard it”.
Cue naming makes segregation teachable.
Replay
Use the same limited replay rule.
Loop the scene until all stems are memorised.
Unlimited replay changes burden and evidence.
Score
Count missed and false sources plus timing errors.
Use a vague listening score.
Specific errors show what changed.
Testing stream interaction
Mute one declared synthetic stem interval and predict the change.
Turn every stem on and off repeatedly until the answer is obvious.
One controlled probe tests masking; repeated dismantling becomes stem memorisation.
6 · Common mistakes
Spot the error and apply the correction
Common mistakes and corrections
Mistake
Fix
Record people or environments covertly.
Use synthetic or consented audio.
Study isolated stems first.
Commit a source inventory before stems.
Say only “I heard it”.
Name pitch, timing, position or timbre evidence.
Loop the scene until all stems are memorised.
Use the same limited replay rule.
Use a vague listening score.
Count missed and false sources plus timing errors.
The learner uses spatial position as the cue even though the scene is presented in mono.
Check the file/channel metadata and restrict cue labels to pitch, timing or timbre when spatial information is absent.
7 · Practice
Turn the steps into a usable skill
First session
A helper generates Scene A and its three stems from the stated recipe, codes the mixture A and keeps the recipe, stems and A-probe out of view.
The learner draws eight one-second boxes, records the playback chain and completes the first-listen inventory without pausing.
On exactly one replay, the learner draws each steady span and each short-event mark, then assigns pitch, timing or timbre to every confident row.
The helper reveals the A recipe and stems; the learner separately totals source hits out of three, misses, false sources and burst marks within one second of 2, 5 and 7.
The learner writes the 5-second-burst prediction, hears A-probe once and records clearer, unchanged or less clear.
Scene B stays sealed for a transfer session at least 48 hours later.
Repeat plan
After 48 hours, analyse Scene B with the identical inventory/replay limits and predict what B-probe will do to the 3-second burst. On later weeks, a helper may move one burst time or one steady-track onset while preserving three tracks and eight seconds; change only that overlap feature, never playback level.
Progress when
Target-source accuracy rises on novel scenes.
False sources and timing errors decline.
Listening burden remains acceptable at the same level.
Do not progress when
Only covert or non-consented audio is available.
Listening fatigue or ear symptoms rise.
The learner changes level or unlimited replay to maintain score.
8 · Check the result
Measure what changed
Target-source accuracy in novel multi-source scenes with fatigue
How: Record correct, missed and false sources, onset/offset error, named cue, replay, burden and exact scene/audio configuration. Record whether the one-stem continuity probe confirmed the predicted masking or grouping relation.
Good result: On Scene B or a newly timed three-track file, source hits or one-second boundary marks improve without extra listens, more false sources or greater burden.
This does not prove: Predeclare number/type of sources, cue dimensions, target, masker, accuracy and new-scene transfer; neural responses remain mechanism measures.
Self-check
Can you demonstrate “Lock the scene setup”? The exact scene version is named.
Can you demonstrate “Second pass: map time”? Each listed source has a timeline mark.
Can you demonstrate “Retest a novel arrangement”? The retest arrangement is unfamiliar.
Can you predict which source and time interval will change when one synthetic stem is briefly muted?
9 · Stop, adapt or get help
Keep the safety boundary practical
Stop and get help
Stop for ear pain, tinnitus increase, dizziness, headache or marked fatigue.
A new one-sided hearing drop, spinning sensation with inability to stand or other abrupt ear/neurological change is not a scene-analysis problem; stop and seek urgent assessment.
Stop and review consent/privacy if real people or private recordings enter the task.
Accessibility and adaptations
Use visual timeline markers, tactile tokens or a scribe for source/time responses.
Reduce source count as a declared level while keeping level comfortable.
Preserve usual hearing devices and caption only the instructions, not the scene content.
10 · Evidence and limits
Why these instructions are here
primary research
Primary research shows perceptual learning can alter responses during a defined vowel-segregation task.
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.
01
Learn the cue vocabulary
Review pitch, timing, spatial position and timbre—the sound quality that distinguishes sources.
Why this step exists
Cue labels make stream separation explicit.
Success check
The learner can match each cue to an example.
I’m stuck on this step
Reset: Re-read this authored instruction — “Review pitch, timing, spatial position and timbre—the sound quality that distinguishes sources.” — and its success check, then attempt only this step.
Possible snag: Use a vague listening score.
Correction: Count missed and false sources plus timing errors.
Possible snag: The learner uses spatial position as the cue even though the scene is presented in mono.
Correction: Check the file/channel metadata and restrict cue labels to pitch, timing or timbre when spatial information is absent.
Stop / get help: Stop for ear pain, tinnitus increase, dizziness, headache or marked fatigue.
02
Lock the scene setup
Record coded file A or B, device, headphones/speakers, level and the fixed rule of one inventory listen plus one timeline replay.
Why this step exists
Scene composition and playback define the task.
Success check
The exact scene version is named.
I’m stuck on this step
Reset: Re-read this authored instruction — “Record coded file A or B, device, headphones/speakers, level and the fixed rule of one inventory listen plus one timeline replay.” — and its success check, then attempt only this step.
Possible snag: Record people or environments covertly.
Correction: Use synthetic or consented audio.
Possible snag: Loop the scene until all stems are memorised.
Correction: Use the same limited replay rule.
Stop / get help: Stop for ear pain, tinnitus increase, dizziness, headache or marked fatigue.
03
First pass: list sources
Listen once to coded mixture A or B and list each perceived source without checking the stems or timeline key.
Why this step exists
A first pass captures spontaneous segregation.
Success check
Sources and uncertain guesses are both recorded.
I’m stuck on this step
Reset: Re-read this authored instruction — “Listen once to coded mixture A or B and list each perceived source without checking the stems or timeline key.” — and its success check, then attempt only this step.
Possible snag: Study isolated stems first.
Correction: Commit a source inventory before stems.
Stop / get help: Stop for ear pain, tinnitus increase, dizziness, headache or marked fatigue.
04
Second pass: map time
Replay exactly once and shade the one-second boxes in which each source starts, stops or produces a burst; use a question mark when the boundary is uncertain.
Why this step exists
Time course helps separate overlapping streams.
Success check
Each listed source has a timeline mark.
I’m stuck on this step
Reset: Re-read this authored instruction — “Replay exactly once and shade the one-second boxes in which each source starts, stops or produces a burst; use a question mark when the boundary is uncertain.” — and its success check, then attempt only this step.
Possible snag: The result from “Replay exactly once and shade the one-second boxes in which each source starts, stops or produces a burst; use a question mark when the boundary is uncertain.” does not yet meet this declared check: Each listed source has a timeline mark.
Correction: Return to the start of “Second pass: map time”, reduce complexity or pace, and repeat only the part needed to satisfy: “Each listed source has a timeline mark.”
Stop / get help: Stop for ear pain, tinnitus increase, dizziness, headache or marked fatigue.
05
Name the separating cue
For each source, state whether pitch, timing, position or timbre made it distinct.
Why this step exists
A source label without the cue does not teach the method.
Success check
Every confident source has at least one cue.
I’m stuck on this step
Reset: Re-read this authored instruction — “For each source, state whether pitch, timing, position or timbre made it distinct.” — and its success check, then attempt only this step.
Possible snag: Say only “I heard it”.
Correction: Name pitch, timing, position or timbre evidence.
Stop / get help: Stop for ear pain, tinnitus increase, dizziness, headache or marked fatigue.
06
Check against stems
After committing the scene analysis, uncover the source recipe and play the three isolated tracks. Mark a missing real track as a miss, an invented track as a false source and a boundary more than one grid box away as a timing error.
Why this step exists
Stems provide specific feedback without erasing the mixture response.
Success check
Original scene answers remain visible.
I’m stuck on this step
Reset: Re-read this authored instruction — “After committing the scene analysis, uncover the source recipe and play the three isolated tracks. Mark a missing real track as a miss, an invented track as a false source and a boundary more than one grid box away as a timing error.” — and its success check, then attempt only this step.
Possible snag: The result from “After committing the scene analysis, uncover the source recipe and play the three isolated tracks. Mark a missing real track as a miss, an invented track as a false source and a boundary more than one grid box away as a timing error.” does not yet meet this declared check: Original scene answers remain visible.
Correction: Return to the start of “Check against stems”, reduce complexity or pace, and repeat only the part needed to satisfy: “Original scene answers remain visible.”
Stop / get help: Stop for ear pain, tinnitus increase, dizziness, headache or marked fatigue.
07
Retest a novel arrangement
At least 48 hours later, analyse the other coded scene using the same two-listen rule before viewing its recipe.
Why this step exists
A novel scene tests the analysis method rather than memorised content.
Success check
The retest arrangement is unfamiliar.
I’m stuck on this step
Reset: Re-read this authored instruction — “At least 48 hours later, analyse the other coded scene using the same two-listen rule before viewing its recipe.” — and its success check, then attempt only this step.
Possible snag: The result from “At least 48 hours later, analyse the other coded scene using the same two-listen rule before viewing its recipe.” does not yet meet this declared check: The retest arrangement is unfamiliar.
Correction: Return to the start of “Retest a novel arrangement”, reduce complexity or pace, and repeat only the part needed to satisfy: “The retest arrangement is unfamiliar.”
Stop / get help: Stop for ear pain, tinnitus increase, dizziness, headache or marked fatigue.
08
Use a continuity probe
After scoring the original mixture, write which burst or onset should be easier to hear when the named low-frequency track is removed during the marked interval. Then play A-probe or B-probe once and record whether the predicted event became clearer, unchanged or less clear.
Why this step exists
Removing one source tests the proposed grouping cue and reveals masking that an isolated-stem review cannot show.
Success check
The prediction names the affected source and interval, and the companion result is recorded without changing the original scene score.
I’m stuck on this step
Reset: Re-read this authored instruction — “After scoring the original mixture, write which burst or onset should be easier to hear when the named low-frequency track is removed during the marked interval. Then play A-probe or B-probe once and record whether the predicted event became clearer, unchanged or less clear.” — and its success check, then attempt only this step.
Possible snag: The result from “After scoring the original mixture, write which burst or onset should be easier to hear when the named low-frequency track is removed during the marked interval. Then play A-probe or B-probe once and record whether the predicted event became clearer, unchanged or less clear.” does not yet meet this declared check: The prediction names the affected source and interval, and the companion result is recorded without changing the original scene score.
Correction: Return to the start of “Use a continuity probe”, reduce complexity or pace, and repeat only the part needed to satisfy: “The prediction names the affected source and interval, and the companion result is recorded without changing the original scene score.”
Stop / get help: Stop for ear pain, tinnitus increase, dizziness, headache or marked fatigue.
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.
Audio source — Privacy and legality control the material.
Correct / safer
Use synthetic or consented audio.
Wrong / riskier
Record people or environments covertly.
First pass — Pre-exposure turns the task into recognition memory.
Correct / safer
Commit a source inventory before stems.
Wrong / riskier
Study isolated stems first.
Cue — Cue naming makes segregation teachable.
Correct / safer
Name pitch, timing, position or timbre evidence.
Wrong / riskier
Say only “I heard it”.
Replay — Unlimited replay changes burden and evidence.
Correct / safer
Use the same limited replay rule.
Wrong / riskier
Loop the scene until all stems are memorised.
Score — Specific errors show what changed.
Correct / safer
Count missed and false sources plus timing errors.
Mute one declared synthetic stem interval and predict the change.
Wrong / riskier
Turn every stem on and off repeatedly until the answer is obvious.
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.