Attention
Sound Width Task
Also called: interaural correlation discrimination, interaural coherence sensitivity, apparent source width task, binaural correlation threshold, interaural decorrelation detection, IACC sensitivity
Measure the smallest change in how well the noise in the two ears matches that listeners can hear: three noises through headphones, one wider or narrower than the other two, on two staircases.
8-10 minUntimed
Free to start: 10 free participants per account, then $0.99 per participant for academic research, or any yearly plan. Try it here first, with no account.
- Papers, 2021 to 2025
- 43
- Evidence grade
- D
- Registered clinical studies
- 0
What it measures
Measures how small a difference in the match between the two ears a listener can hear. Noise that is the same in both ears sounds compact and noise that differs between the ears sounds wide. Each trial plays three noises through headphones and participants say which one is different; two staircases find the thresholds. Needs headphones and normal hearing; results are research measures, not a hearing test.
Constructs
- Interaural correlation sensitivity
- Binaural hearing
- Apparent source width
- Spatial hearing
Research uses
- Spatial and binaural hearing research
- Auditory perception research
- Within-study group comparison
How the paradigm works
Sound width is how spread out a sound seems inside the head over headphones. The noise in the two ears carries it: the same noise in both ears sounds compact, like one spot in the middle of the head, and noise that is different in each ear sounds wide (Blauert and Lindemann, 1986). The measure is the interaural correlation, from 1 for the same noise to 0 for unrelated noises, the cue most tied to apparent source width (Whitmer, Seeber and Akeroyd, 2012). Listeners tell correlations apart far better near 1 than near 0. Pollack and Trittipoe (1959) found that the change in squared correlation needed for 75 percent correct fell from about 0.4 at a reference of 0 to about 0.04 at a reference of 1; Gabriel and Colburn (1981) found just noticeable differences of about 0.04 for wide bands at a reference of 1 and about 0.35 for bands wider than 115 Hz at a reference of 0; Culling, Colburn and Spurchise (2001) found the advantage near 1 in narrow bands from 250 to 1,500 Hz. This task measures the threshold from both references with two staircases.
More about the paradigm
Set a comfortable volume and complete the 6-trial headphone check. A failed check stops the task by default; researchers can choose to continue with a quality flag.
Each trial plays three noises of 500 ms, 400 ms apart, and the listener says which is different. The noise is Gaussian from 200 to 1,000 Hz, where the timing of the noise in each ear carries the correlation (Culling and colleagues found the sensitivity falling above 750 Hz), with 20 ms raised-cosine ramps that start and end together in both ears, new for every noise, 10 dB below the volume-setting tone and moved by -2 to +2 dB a noise so that no loudness difference gives the odd one away. The left ear's noise and a second noise made exactly uncorrelated with it are mixed to give the right ear its correlation, as Pollack and Trittipoe made their pairs from independent noises, so both ears are equally loud and the correlation of the two ears' samples is exact. Choose First, Second or Third with no time limit.
Two parts, in an order drawn from the session's seed. In the first, two noises are the same in both ears and one is wider because its ears match less; in the second, two noises are unrelated in the two ears and one is narrower because its ears match more. One 2-down 1-up staircase in each part (Levitt, 1971) tracks 70.7 percent right: two right answers in a row make the difference smaller and one error makes it larger. The difference is 1 minus the odd noise's correlation in the first part, from 0.316, and its correlation in the second, from 1, on a logarithmic grid from 1 to 0.001; steps are a factor of 0.56 until the third turnpoint and 0.79 after it. 40 trials in each part (30 to 60 as a setting): 80 trials, with no feedback, after 2 practice trials in each part at the largest difference.
The primary score is the decorrelation threshold of the first part: 1 minus the correlation at which the odd noise is heard among two noises that are the same in both ears, the geometric mean of the last eight turnpoints after the first three. The correlation threshold of the second part is the correlation at which the odd noise is heard among two unrelated noises. The share right in each part, the turnpoints, the answer times and the headphone checks come with them. Smaller is finer in both, and the second is expected to be much larger than the first.
Scores
- Decorrelation Thresholdscore
- How far below a correlation of 1 the odd noise must be to be heard among two noises that are the same in both ears (1 minus its correlation): the geometric mean of the last eight turnpoints after the first three.
- Smaller means a smaller drop in the match between the ears was heard; not a hearing level
- Correlation Thresholdscore
- The correlation at which the odd noise is heard among two noises that are unrelated in the two ears: the geometric mean of the last eight turnpoints after the first three.
- Smaller means a smaller gain in the match between the ears was heard; usually much larger than the decorrelation threshold
- Wider-Noise Accuracy%
- Share of trials in the first part, where the odd noise is wider, answered with the odd noise.
- Near 0.71 is expected on a staircase; near 0.33 means the noises were not heard to differ
- Narrower-Noise Accuracy%
- Share of trials in the second part, where the odd noise is narrower, answered with the odd noise.
- Near 0.71 is expected on a staircase; near 0.33 means the noises were not heard to differ
- Wider-Noise Turnpointscount
- Times the first part's staircase changed direction.
- A threshold needs at least seven
- Narrower-Noise Turnpointscount
- Times the second part's staircase changed direction.
- A threshold needs at least seven
- Median Answer Timems
- Median time from when the answers opened, after the three noises, to the answer.
- Informational; the answer has no time limit
- Anticipationscount
- Answers within 100 ms of the answers opening, which leave the staircase where it was.
- Half or more of the answers makes the session invalid
- Completed Trialstrials
- Trials answered in the scored run.
- Fewer than planned is flagged for review
- Planned Trialstrials
- Trials the run set out to play.
- Informational
- Headphone Check Passed0/1
- 1 when the headphone check passed (at least 5 of 6 right), 0 when it did not.
- 0 stops the test unless the study set it to continue
- Headphone Check Correctcount
- Headphone check trials answered right, of 6.
- 5 or 6 passes
- Reported Output Delayms
- Output delay the device reported to the browser; empty when the browser does not report it.
- Informational
Use and evidence
43 papers from 2021 to 2025 name this task in their title or abstract, up 43% compared with 2016 to 2020. That ranks it 493 of 711 test families in our census. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 9 |
| 2024 | 1 |
| 2023 | 8 |
| 2022 | 10 |
| 2021 | 15 |
| 2020 | 5 |
| 2019 | 6 |
| 2018 | 8 |
| 2017 | 6 |
| 2016 | 5 |
| 2015 | 10 |
| 2014 | 13 |
| 2013 | 13 |
| 2012 | 4 |
| 2011 | 5 |
| 2010 | 13 |
| 2009 | 13 |
| 2008 | 14 |
| 2007 | 14 |
| 2006 | 18 |
| 2005 | 8 |
| 2004 | 3 |
| 2003 | 5 |
| 2002 | 4 |
| 2001 | 8 |
| 2000 | 4 |
Use grade D
30 to 99 papers from 2021 to 2025.
Evidence grade D
Little published measurement evidence found.
- Papers on test-retest reliability
- 0
- Papers on reliability or validity
- 18
- Registered clinical studies using it as an outcome
- 0
Measured on ConductCognition
Pooled across the studies that feed our platform norms, by device group. A figure needs 50 or more people.
No pooled figure yet.
How the grades work
Use: A is 1,000 or more papers from 2021 to 2025, B 300 to 999, C 100 to 299, D 30 to 99 and E fewer than 30. Evidence: A means at least 25 papers report its test-retest reliability and at least 25 registered clinical studies use it as an outcome; B, at least 8 test-retest papers, and either 5 registered clinical studies or 150 reliability or validity papers; C, at least 2 test-retest papers, 30 reliability or validity papers or 3 registered clinical studies; D, little published measurement evidence found. The counts measure use. They include every paper that names the task in its title or abstract. Grades describe the published research on this task, not ConductCognition's version of it.
Source: ConductCognition test census 0.58.0, counted 14 November 2026 from OpenAlex and ClinicalTrials.gov.
Task design
The trial sequence as the test runs in ConductCognition.
- Sound setup
- The participant turns the device's volume until a steady 1 kHz tone is comfortable, then takes a 6-trial headphone check (Woods and colleagues, 2017). By default a failed check ends the test; its result is saved either way.
- Trial
- Press Play to hear three noises of 500 ms, 400 ms apart. Two match and one differs in how well the noise in the two ears matches. Choose First, Second or Third for the one that is different, with no time limit.
- Noise
- Noise from 200 to 1,000 Hz, new for every noise, with 20 ms raised-cosine ramps that start and end together in both ears, 10 dB below the volume-setting tone and moved by -2 to +2 dB a noise, so that no loudness difference gives the odd one away. Each ear's noise is mixed from independent noises, so both ears are equally loud and the correlation between the ears is exact.
- Parts
- Two, in an order drawn from the session's seed. In the first, two noises are the same in both ears (a correlation of 1, a compact sound) and one is wider. In the second, two are unrelated in the two ears (a correlation of 0, a wide sound) and one is narrower.
- Staircases
- One in each part. Two right answers in a row make the difference smaller and one error makes it larger, which keeps a listener near 71 percent right. The difference is 1 minus the odd noise's correlation in the first part and its correlation in the second, on a logarithmic grid from 1 to 0.001, moving by a factor of 0.56 until the third turnpoint and 0.79 after it.
- Trials
- 40 trials in each part (standard, 30 to 60): 80 trials, with no feedback, after 2 practice trials in each part at the largest difference, with feedback for 1,500 ms.
- Score
- Each threshold is the geometric mean of the last eight turnpoints after the first three. The main score is the decorrelation threshold of the first part, and the correlation threshold of the second part comes with it. Answers under 100 ms leave the staircase where it was.
The noise plays through the participant's own headphones at a volume they chose, and only the match between the ears adapts: every noise plays at the same level, moved by a small random amount. Listeners tell correlations apart far better near 1 than near 0, so the second threshold is much larger than the first. A staircase that stays at its easiest difference is flagged, as when sound reaches only one ear. The test needs normal hearing and headphones, wired if possible: wireless headphones can add delay and change the sound. It is a research measure, not a hearing test: it gives no hearing level and no pass or fail, and its scores are not a sign of hearing loss. No percentile is given.
Quality checks
Each session is marked valid, needs review or invalid, with the reason, before you see its scores.
Scored rows: sound_width_test from the selected run: three noises, then First, Second or Third for the one that was different, with no time limit, graded from the saved button and the place of the odd noise. The two 2-down 1-up staircases are replayed from the answers, and each threshold averages the last eight turnpoints after the first three. Chance is one third. Below chance invalidates and not above chance is flagged; one button on 90% of answers invalidates when accuracy is not above chance; answers under 100 ms leave the staircase where it was and invalidate when at least half. A headphone check that stopped the test invalidates; a failed check with the test run anyway, a staircase with too few turnpoints, a part with a quarter of its trials at the easiest difference and a run short of its planned trials are flagged for review.
Why these rules
Accuracy significantly below chance (one-sided exact test, p < .05) cannot come from guessing; it means the response mapping was reversed or the task was not followed (Slick, Sherman and Iverson 1999 treat below-chance forced-choice performance as definite response bias). Chance-level accuracy can be a real result in impaired participants, so it is only flagged. One option on at least 90% of answered trials, when the correct options are balanced and accuracy is not above chance, means the answers did not depend on the items (the task counterpart of long-string responding in surveys; Curran 2016). A staircase keeps a listener who hears the difference near 71 percent right (Levitt, 1971), so accuracy near one third means the answers did not follow the sounds. The answers open once all three noises have played, so an answer under 100 ms was pressed before they opened. Woods, Siegel, Traer and McDermott (2017) let only listeners who passed the headphone check go on, because a listener on loudspeakers hears the sounds differently; a session that ran after a failed check keeps its scores but not its percentile. Sound that reaches only one ear, or the same in both, cannot show a difference in the match between the ears, so a staircase that stays at its easiest difference is flagged.
No percentile. Scores and quality flags only.
Running this test
Settings you can change
| Setting | Standard | Allowed |
|---|---|---|
| Trials per part | 40 | 30 to 60 |
| When the headphone check fails | Stop the test | Stop the test, Continue the test |
Yearly plans can change settings within these limits. Free and pay-as-you-go studies use the standard settings.
Need a setting this test does not have?
Devices and languages
- Untimed test: runs on desktops, laptops, tablets and phones. Needs headphones. Participant devices
- Languages: English, Spanish, French, German, Italian, Portuguese (Brazil), Chinese (Simplified) and Japanese.
- Available on: Every plan, including the free plan.
- No percentile. Scores and quality flags only. How norms work
- Recruiting online? See the setup guides for Prolific, Sona Systems, CloudResearch Connect and Qualtrics.
Cite and describe
Participants completed the Sound Width Task in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Pollack and Trittipoe, 1959; Gabriel and Colburn, 1981; Culling, Colburn and Spurchise, 2001), with the standard settings (trials per part: 40; when the headphone check fails: Stop the test). The main score was decorrelation threshold (how far below a correlation of 1 the odd noise must be to be heard among two noises that are the same in both ears (1 minus its correlation): the geometric mean of the last eight turnpoints after the first three).
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- Pollack, I., & Trittipoe, W. J. (1959). Binaural listening and interaural noise cross correlation. The Journal of the Acoustical Society of America, 31(9), 1250-1252. https://doi.org/10.1121/1.1907852Identification of interaural correlations from references of 0 and 1, with pairs made from independent noises.
- Gabriel, K. J., & Colburn, H. S. (1981). Interaural correlation discrimination: I. Bandwidth and level dependence. The Journal of the Acoustical Society of America, 69(5), 1394-1401. https://doi.org/10.1121/1.385821Correlation thresholds from references of 0 and 1 in Gaussian noise.
- Culling, J. F., Colburn, H. S., & Spurchise, M. (2001). Interaural correlation sensitivity. The Journal of the Acoustical Society of America, 110(2), 1020-1029. https://doi.org/10.1121/1.1383296Sensitivity to correlation in narrow bands from 250 to 1,500 Hz, best near a correlation of 1.
- Blauert, J., & Lindemann, W. (1986). Spatial mapping of intracranial auditory events for various degrees of interaural coherence. The Journal of the Acoustical Society of America, 79(3), 806-813. https://doi.org/10.1121/1.393471How the spread of a sound inside the head changes with the match between the ears.
- Whitmer, W. M., Seeber, B. U., & Akeroyd, M. A. (2012). Apparent auditory source width insensitivity in older hearing-impaired individuals. The Journal of the Acoustical Society of America, 132(1), 369-379. https://doi.org/10.1121/1.4728200Interaural coherence as the cue most tied to apparent source width.
- Levitt, H. (1971). Transformed up-down methods in psychoacoustics. The Journal of the Acoustical Society of America, 49(2B), 467-477. https://doi.org/10.1121/1.1912375The 2-down 1-up rule, which tracks 70.7 percent right.
- Woods, K. J. P., Siegel, M. H., Traer, J., & McDermott, J. H. (2017). Headphone screening to facilitate web-based auditory experiments. Attention, Perception, & Psychophysics, 79(7), 2064-2072. https://doi.org/10.3758/s13414-017-1361-2The headphone check before the test.
Source papers
- Pollack and Trittipoe (1959): identification of interaural correlations from references of 0 and 1. Gabriel and Colburn (1981): correlation just noticeable differences from references of 0 and 1 in Gaussian noise. Culling, Colburn and Spurchise (2001): sensitivity to differences in correlation in narrow bands from 250 to 1,500 Hz.Paradigm source, Adults
Questions
How long does the Sound Width Task take?
About 8 to 10 minutes, as listed in the test library.
What device does a participant need?
Untimed test: runs on desktops, laptops, tablets and phones. Needs headphones. The browser must be Chrome, Edge, Firefox or Safari.
What scores does it produce?
Decorrelation Threshold, Correlation Threshold, Wider-Noise Accuracy, Narrower-Noise Accuracy, Wider-Noise Turnpoints, Narrower-Noise Turnpoints, Median Answer Time, Anticipations, Completed Trials, Planned Trials, Headphone Check Passed, Headphone Check Correct and Reported Output Delay. Each session is also marked valid, needs review or invalid, with the reason.
How widely is the Sound Width Task used in research?
43 papers from 2021 to 2025 name it in their title or abstract, up 43% compared with 2016 to 2020. That ranks it 493 of 711 test families in the ConductCognition test census (use grade D). Source: ConductCognition test census 0.58.0, counted 14 November 2026 from OpenAlex and ClinicalTrials.gov.
How much reliability evidence does the Sound Width Task have?
In the published research, 0 papers report its test-retest reliability and 0 registered clinical studies use it as an outcome (evidence grade D in the ConductCognition test census).
Does it report a percentile?
No. Scores and quality flags only.
Which settings can a researcher change?
Trials per part (30 to 60) and when the headphone check fails (Stop the test, Continue the test). Settings can be changed on yearly plans; free and pay-as-you-go studies use the standard settings.
Which languages can participants use?
Participants can take this test in English, Spanish, French, German, Italian, Portuguese (Brazil), Chinese (Simplified) and Japanese. English is on every plan, and the other languages come with yearly plans. Have a native speaker read the machine-translated text before launch.
Is this a clinical test?
No. ConductCognition is for research use only. It runs, scores and exports the test; clinical interpretation is outside the platform.
Every test: the full test list.
Run this test in your next study
Start free with 10 free participants on any test, then use participant credits or a yearly plan.