İşitmeAtölyesiClinical Masking Calculator
Enter thresholds; the tool works out, for each measurement, whether masking is needed, the noise level to start from if it is, and the level not to go up to. It draws the workable range in between and warns you when that range closes: the masking dilemma. Every number is traced back, step by step, to the threshold it came from.
For educational use · not a clinical record or a diagnostic tool · the tool can be wrong too
- Air and bonea separate calculation for each
- Plateau and dilemmaits width and its closing
- Step by stepwhere each number came from
- 1 Enter the thresholds Click the chart or start from a ready-made example. The calculation uses both ears' thresholds together.
- 2 See which measurement needs it The grid beside the chart states, for each measurement, whether masking is needed. Click a cell to select that measurement.
- 3 Read the level and the plateau For the measurement you select, the initial masking level, the overmasking limit, the workable range between them and the steps of the calculation all open up.
Abbreviations and units
- TE
- test ear: the ear you present the sound to and whose threshold you are looking for
- NTE
- non-test ear: the opposite ear you deliver the noise to
- IA
- interaural attenuation: the level the sound loses while crossing the skull to the opposite ear
- ABG
- air-bone gap: the difference between air and bone conduction thresholds in the same ear
- dB HL
- hearing level: the unit both of the threshold you read on the audiogram and of the masking level you read on the audiometer's noise channel. Some sources write the masking level as dB EM (effective masking level); this page uses dB HL
- YY · M
- in the table yy means no response, and m means the measurement was made with masking
Why and when is masking needed?
The sound you present to the ear under the earphone does not stay in that ear. Once it is loud enough it sets the skull vibrating and reaches the opposite cochlea as well: cross-hearing. The sound is attenuated along the way, and the level it loses is called interaural attenuation. The problem is this: when the patient presses the button, you cannot tell which ear heard the sound. Raise the tone in the diagram below and see for yourself.
Is masking needed? Step by step
After every unmasked measurement one question is asked, and the difference between two thresholds answers it. The flow runs down separate branches for air conduction and bone conduction; at each step, beside the criterion, it says why that criterion is the one used. Rather than memorising the rule, it helps to follow which sound reaches which ear.
You measured an unmasked threshold. That threshold belongs to which ear?
The tone was delivered through the test ear's earphone.
Is the test ear's air conduction threshold higher than the opposite ear's bone conduction threshold by at least the interaural attenuation?
TE air conduction − NTE bone conduction ≥ IA
The sound delivered to the earphone also travels across the skull to the opposite cochlea, losing IA on the way. If what arrives there exceeds that cochlea's bone conduction threshold, the response comes from that ear. What you measured may then belong not to the test ear but to the opposite ear: a shadow threshold. The comparison is therefore made against the opposite ear's bone conduction threshold, not its air conduction.
What if bone conduction was not measured?
If the opposite ear's bone conduction was not measured, air conduction is used as a proxy. That is a cautious assumption, so it calls for masking earlier than strictly necessary.
The tone was delivered through the bone vibrator, on the mastoid or the forehead.
Is the air-bone gap in the test ear greater than 10 dB?
TE air conduction − TE bone conduction > 10 dB
In bone conduction the interaural attenuation is small enough to be treated as zero: wherever the vibrator is placed, the sound reaches both cochleae at roughly the same level. An unmasked bone conduction threshold therefore measures the better of the two cochleae and says nothing about which side of the head it came from. If there is a gap, you cannot tell which cochlea the better one is.
Why is it not needed when there is no gap?
With no gap, both cochleae already have similar thresholds, so which one you measured does not change the result; masking is then unnecessary.
No masking is needed. Record the threshold as it stands.
Unnecessary masking costs more than time: central masking lifts the threshold by a few decibels, so it makes your measurement worse for no reason.
Comparing air conduction with the opposite ear's air conduction. The criterion is the opposite ear's bone conduction threshold. In bilateral conductive loss the two air conduction thresholds look close to each other and the rule is quietly skipped.
Skipping masking in bone conduction. The moment the gap exceeds 10 dB, an unmasked bone conduction threshold measures the better of the two cochleae. This is the most common cause of a wrong type of loss being reported.
Let these audiogram patterns bring masking to mind
Whether the rule calls for masking and whether a shadow threshold is actually being read in that measurement are two separate questions; the tool below shows them separately.
Enter the thresholds and the earphone
You can enter them by clicking the chart or by typing into the table; each updates the other at once. The masking calculation looks at both ears' thresholds together: the test ear's air and bone conduction, and the non-test ear's air-bone gap. That is why you have to enter the opposite ear's thresholds as well. To start from a ready-made example, pick one of the buttons below; if this is your first visit, the guided tour will walk you through the page.
Which measurements need it?
The two rules from the Why section are applied here to every pair of thresholds you enter. Each box is one measurement. Click a box: that measurement is selected and why the decision came out that way is written out with the numbers below the grid. The calculation, the plateau band and the demo below all follow that measurement.
The grid fills in once thresholds are entered.
Type them into the table
| Hz | Right air | Right bone | Left air | Left bone |
|---|
For a measurement with no response type yy into the cell. To mark a masked measurement add m after the number.
Advanced settings
The defaults are the set textbooks usually assume: supra-aural earphones, the supra-aural occlusion effect table, a flat 55 dB interaural attenuation with inserts, and central masking left out of the calculation. Change these only to match your own clinic's practice.
Raise the noise and watch both sides at once
The one thing you cannot see in a real measurement is which ear the response came from. This demo shows it. On the left, the audiometer display: as you raise the noise in the opposite ear you see how the threshold measured in the test ear changes and which ear is responding at that moment. On the right, the same thing is plotted on the masking function; the undermasking, plateau and overmasking regions emerge on their own. The measurement you selected above is used. Try a few steps here before moving on to the formulas: once the plateau makes sense to the eye, the calculation below is far easier to follow.
The rules and numbers you need to know
Everything on one page before you move on to the procedure: when masking is needed, how the initial masking level is calculated, in what situation the occlusion effect enters the formula, where overmasking begins and when the plateau counts as found. The boxes below are kept short enough to memorise; the detail behind each one sits in its own section.
air conduction · TE air conduction − NTE bone conduction ≥ IA
bone conduction · TE air conduction − TE bone conduction > 10 dB
speech · speech level − NTE best bone conduction ≥ IA
If the NTE bone conduction threshold has not been measured, the NTE air conduction threshold stands in; that is a cautious assumption and leads to masking earlier than strictly necessary.
air conduction · NTE air conduction + 15 dB
bone conduction · NTE air conduction + occlusion effect + 15 dB
Of the 15 dB, 5 dB starts the masking and 10 dB is the safety margin. The noise is delivered at this level and then raised while looking for the plateau.
- Only in bone conduction masking.
- Only while an earphone is on the opposite ear.
- Taken as zero at 2000 Hz and above.
- In the opposite ear it does not appear when there is conductive pathology; the middle ear is already holding back the low frequency energy.
overmasking · noise ≥ IA + TE bone conduction + 5 dB
dilemma · initial level ≥ overmasking limit
If the initial level reaches the limit, no workable range is left; the first thing to try is switching to insert earphones.
- 5 dB steps and the threshold does not change over 3-4 measurements.
- 10 dB steps and the threshold does not change over 2-3 measurements.
- In short, if you raise the masking level by 15-20 dB and the threshold does not change.
Use 5 dB steps when the workable range is narrow; a 10 dB step can skip past a shortened plateau and lead to overmasking.
- The value read on the plateau is what gets written, with the masked symbol.
- For central masking no correction is made.
- If no plateau can be found the measurement is not written as masked; it is stated that it was left undetermined.
The values used in the calculation
Occlusion effect
When an earphone is placed on the opposite ear, that ear's bone conduction threshold improves at low frequencies; this is why the occlusion effect is added to the initial level in bone conduction masking. The effect is taken as zero at 2000 Hz and above. The reason for it and how it is measured are explained in the next section.
| Hz | Supra-aural | Insert · deep | Insert · shallow |
|---|
The literature disagrees on insert earphone values. Shallow insertion closes the canal but adds no volume, so it gives a larger effect than a supra-aural earphone; studies that measure it directly find a markedly smaller value with deep insertion. The two columns are therefore kept apart; pick the one that matches your insertion depth.
Interaural attenuation
Sources agree on supra-aural earphones: the lowest value found in the studies, 40 dB, is used. For insert earphones the sources diverge; all three sets below are in current use and you can pick one above.
| Source | ≤1 kHz | >1 kHz |
|---|
In the clinic, interaural attenuation for bone conduction is taken as 0 dB. With mastoid placement it has been reported to vary between 0 and 15 dB at 2000 and 4000 Hz; with forehead placement no value is given.
Central masking
When noise is delivered to the opposite ear, the test ear's threshold rises by a few decibels independently of cross-hearing. The effect arises at the central level of the auditory pathways and is typically around 5 dB. It is a small number, but a common reason why three measurements do not come out identical while you are looking for the plateau: if you see a rise on the masking function that is clearly shallower than 1 dB per decibel, the first thing to think of is central masking, not cross-hearing.
The tool does not include central masking in the plateau calculation; the formulas do not contain it. It is added only to the threshold measured on the masking function and is reported separately in the results section.
How you carry out masking, step by step
The order followed in the clinic is this: an initial masking level is calculated, the noise is delivered at that level, and it is then raised step by step until the range where the threshold settles, the plateau, is found. Two numbers come out of this: where to start and which level not to reach. The steps below are written with the selected measurement's own numbers.
In order, for the selected measurement
If you pick another measurement from the strip at the top or from the grid in the threshold section, the numbers here change with it.
Why does the occlusion effect enter the bone conduction formula?
In bone conduction testing you deliver the noise to the opposite ear through an air conduction earphone. The moment the earphone closes the ear canal, the low frequency energy trapped in that canal cannot escape and that ear's bone conduction threshold improves. In other words the opposite ear now hears better than it did before the earphone went on. The same noise is no longer enough to silence it; the occlusion effect has to be added to the initial level.
air conduction · initial = NTE air conduction + 15 dB
bone conduction · initial = NTE air conduction + occlusion effect + 15 dB
How do you establish the occlusion effect?
First way · a published table. Values fixed in advance by frequency and earphone type are used. The calculation on this page follows this way; the table it uses can be chosen from Advanced settings in the Calculation section. The effect is taken as zero at 2000 Hz and above.
Second way · the audiometric Bing test. The occlusion effect varies from person to person and does not appear at all when there is conductive pathology, so measuring it is more reliable. Once the unmasked bone conduction thresholds have been found, the earphone is placed on the opposite ear and, with no noise delivered at all, the bone conduction threshold is measured again. If the threshold has improved, an occlusion effect has occurred; the difference between the two measurements is that person's occlusion effect, and this value is used in the initial level calculation.
In conductive loss the middle ear is already holding back some of the low frequency energy, so closing the canal adds nothing. The Bing test shows this directly; a published table cannot.
In what steps do you raise the noise?
The noise is delivered at the initial level and the threshold is measured. The noise is then raised and the threshold measured again. If the measured threshold rises along with the noise, the sound is still being heard by the opposite ear: undermasking. If the threshold does not change as you raise the noise, the opposite ear has been silenced; that is the plateau, and the threshold you are looking for is there.
3-4 kez without the threshold changing, that level counts as the true threshold. Use this step when the workable range is narrow.
2-3 kez without the threshold changing, that level counts as the true threshold. This is the basic step Hood described, but it risks overmasking when the range is narrow.
Put another way: if the measured threshold does not change when you raise the masking level by 15-20 dB, that level counts as the plateau. The masking level that produces the plateau is called the the effective masking level, and the width of the plateau the the effective masking range.
If the threshold starts rising again as you keep raising the noise, the noise is now crossing the skull into the test ear: overmasking. The level at which overmasking begins is found with IA + TE bone conduction + 5 dB; do not go up to that level.
Which value do you write on the audiogram?
The threshold you read on the plateau is what gets written, and the symbol marks the measurement as masked. Because of central masking this value is expected to come out a few decibels above the underlying cochlear threshold; no correction is made for that margin in clinical masking. If no plateau can be found, that is, if overmasking already occurs at the initial level, do not write the measurement as masked: see the masking dilemma section.
Need for masking and masking levels
For every measurement the table shows three things together: whether masking is needed by the rule, whether cross-hearing is actually present in that measurement, and if masking is needed, which noise level to start from and which level not to exceed. Click a row and that measurement is selected, with the steps of the calculation opening beneath it.
Click a symbol: that measurement is selected, and the table and the calculation below open up for it. The same selection can also be made by clicking a row in the table.
Measurement flow: where are you in the selected measurement?
The flow in the Why section asks whether masking is needed; this one follows the measurement itself from start to finish. The dark boxes show the path the selected measurement takes.
| Measurement | Hz | Threshold | Clinicalby the rule | Cross-hearingactually present | Initial leveldB HL | Overmasking limitdB HL | Workable rangedB |
|---|
The table fills in once thresholds are entered.
Show me the criterion once more
The same two rules appear in the flow in the Why section and on the first card in the Rules section; here they are written with the interaural attenuation of the earphone you selected.
Air conduction
If the test ear's air conduction threshold is higher than the opposite ear's bone conduction threshold by the interaural attenuation (IA) or more, the measurement may be a shadow threshold. If the opposite ear's bone conduction has not been entered, the tool uses air conduction as a proxy; that is a pessimistic assumption, so it calls for masking earlier than strictly necessary.
TE air conduction − NTE bone conduction ≥ IA
Bone conduction
If the air-bone gap in the same ear exceeds 10 dB, you cannot tell which cochlea an unmasked bone conduction measurement belongs to, because interaural attenuation in bone conduction is small enough to be treated as zero.
TE air conduction − TE bone conduction > 10 dB
Masking in speech testing
Cross-hearing is not peculiar to pure tones: the speech signal also crosses the skull to the opposite ear, so the same question arises in speech reception threshold and word recognition measurements. Is the response you are measuring coming from the test ear?
When it is needed
Masking is needed when the difference between the speech level presented to the test ear and the opposite ear's best bone conduction threshold reaches the interaural attenuation. Because speech is broadband, the lowest IA value is taken as the basis in practice: with supra-aural earphones, 40 dB.
speech level − NTE best bone conduction ≥ IA
In word recognition testing the presentation level is well above threshold, so the need for masking arises more often than with pure tones.
Which noise, which level
Narrow band noise is used with pure tones; with speech, speech noise or white noise is used, because the signal to be masked is broadband.
The same logic applies to the level: high enough to silence the opposite ear, low enough not to cross to the test ear. A common starting point is to add a safety margin above the opposite ear's speech reception threshold; the upper limit is again set by the interaural attenuation and the test ear's bone conduction threshold.
The calculations on this page are for pure-tone testing. The values you use in speech testing depend on your clinic's protocol; this section is here to tie the concept together.
The plateau band
While the noise is not yet enough to silence the opposite ear, what you are measuring is still that ear's threshold, and it rises along with the noise: undermasking. Once the noise starts crossing into the test ear the threshold rises again: overmasking. The band in between is the region where the threshold does not change even as you raise the noise; the true threshold is there. The band below belongs to the selected measurement and is drawn from the masking function; the initial masking level and the overmasking limit are chosen so that they fall inside this band.
The band is drawn once thresholds are entered.
If you raise the masking level by 15 to 20 dB without the measured threshold changing, that level counts as the plateau. At frequencies where the workable range comes out narrower than this, raising the noise in 5 dB steps instead of 10 dB helps you not to miss a shortened plateau.
Common mistakes
The masking dilemma
The masking dilemma is the situation where overmasking already begins at the very first noise level. Undermasking and overmasking are present at the same time: by the time you deliver enough noise to silence the opposite ear, the noise has already crossed to the test ear, and if you deliver less, what you are measuring is still a shadow threshold. No workable range is left in between. The typical picture is bilateral mild to moderate conductive hearing loss with a wide air-bone gap in both ears: the test ear's gap makes it easier for the noise to cross into the test ear, while the non-test ear's gap raises the noise needed to silence that ear. The two effects squeeze the plateau band from both ends until it closes.
Is there a dilemma in the thresholds you entered?
For each measurement the tool checks whether the plateau band has closed.
What do the numbers say for the selected measurement?
The classic textbook presentation: put the initial masking level next to the overmasking limit. If the initial level equals or exceeds the limit, there is a dilemma.
The criterion
The plateau narrows as the air-bone gaps are subtracted from the interaural attenuation. In air conduction both ears' gaps act together; in bone conduction only the non-test ear's gap does.
air conduction · TE gap + NTE gap ≥ 2·IA − 15
bone conduction · NTE gap ≥ IA − 15
With supra-aural earphones this limit works out at 65 dB for air conduction and 25 dB for bone conduction. With insert earphones the same limits rise to 95 and 40 dB; these two numbers explain why the dilemma becomes rarer with insert earphones.
What to do
- Switch to insert earphones. This raises the interaural attenuation from 40 dB to somewhere between 55 and 75 dB. It does two things at once: the need for masking itself decreases, and when masking is needed the plateau band widens from both ends, so the dilemma becomes less likely. The gain is most marked at low frequencies. This is the most direct solution to the dilemma.
- Treat the middle ear first. If effusion or blockage is creating a temporary conductive component, the gap closes when the measurement is repeated after treatment and the dilemma disappears on its own.
- Consider the SAL test. It is a method for deriving the bone conduction threshold from the threshold shift produced by noise delivered through bone; it is explained below.
- Move to objective tests. Acoustic reflexes, otoacoustic emissions and auditory brainstem response give information that either needs no masking or follows different masking rules.
- Report the unmasked measurement as it is. State plainly that you could not determine which ear the threshold belongs to. An invented masked threshold is more harmful than a measurement left undetermined.
The SAL test sensorineural acuity level
It finds the bone conduction threshold indirectly: air conduction thresholds are measured again while a strong noise is delivered from a vibrator placed on the forehead, and what is read off is how far the threshold has shifted.
How is the test done, and what is the logic?
First the air conduction thresholds are measured in the usual way, unmasked. Then, while noise is delivered at maximum level from a vibrator placed on the forehead, the air conduction thresholds are repeated. The noise reaches the cochlea by bone conduction and masks the tone coming through air; the threshold shifts upwards.
How much shift the same noise produces in normal-hearing listeners is established beforehand (typically with a group of 10 to 15 people). Subtracting the patient's shift from that normal value gives the bone conduction threshold.
bone conduction threshold = normal shift − the patient's shift
The logic is this: in conductive loss the cochlea is intact, the noise arriving through bone is as effective as in a normal listener, and the shift comes out the same as normal. If there is a cochlear problem the noise is also perceived weakly and the shift is smaller. The shortfall in the shift gives the cochlear loss directly.
In current practice it is done with narrow band noise and insert earphones. SAL does not replace a masked bone conduction measurement; it is a route taken when the dilemma makes masked measurement impossible, and its result is interpreted accordingly.
Where the calculation comes from
The page's main calculation is Martin's (1980) set: the initial masking level and the overmasking limit come from it. Lidén and colleagues approach the same measurement with a different question, and their set sits in a collapsed box under the calculation steps for comparison. The criteria for when masking is needed follow the British Society of Audiology's recommended procedure for pure-tone audiometry. The masking function the tool draws is built independently of these formulas: the measured threshold is taken as the lower of the test-ear path and the cross-hearing path. This model's plateau limits coincide exactly with the formulas, and the engine's test suite checks that on every build.
Interaural attenuation for insert earphones defaults to a flat 55 dB, the minimum given by the British Society of Audiology's recommended procedure for a correctly placed insert earphone. Sources diverge: CSHBC 2023 and the textbook tradition give 75 dB at 1 kHz and below and 50 dB above it, and a flat 60 dB is sometimes used as a more cautious single value. All three can be selected from Advanced settings. For supra-aural earphones the established lowest value of 40 dB is used.
Test yourself
Five short questions. Each set is generated from a random case, so it is different every time you refresh. When you pick an option, the reasoning and the rule the calculation came from open up. A wrong answer is not a problem; read the reasoning.
Let us do a measurement together
Instead of asking questions it has you do the job itself: a case opens, you make the decisions, and at each step you are told whether you were right and why.
A five-question check
The questions are generated from a random case; they are different every time you refresh.
How is a masked measurement entered on the audiogram?
A masked measurement replaces the unmasked one; the two do not sit side by side on the same audiogram. Which threshold was masked has to be clear from the symbol, because it can stay uncertain which ear an unmasked threshold belongs to.
In this tool a masked air conduction measurement is shown by drawing a square around the unmasked symbol. In the ANSI S3.21 and ASHA 1990 set the same measurement is marked with a triangle for the right ear and a square for the left ear. Every tool on this site uses the same set; you need to know this difference, because an audiogram from another clinic may use a triangle.
What goes in the report
The threshold alone is not enough; the following are written down as well, so that whoever reads it later can judge the measurement.
- Earphone type. Whether you used supra-aural or insert earphones changes the interaural attenuation you assume, and with it the masking decisions.
- The noise level used. Recording which masking level you used at which frequency lets a later reader judge how likely overmasking was.
- If no plateau was found, say so. An invented masked threshold is more harmful than a measurement left undetermined. If there is a masking dilemma, name it.
- Noise calibration. If your audiometer's narrow band noise is not calibrated in effective masking level, write down which scale your values refer to.
Terms used on this page
Masking is described with a small set of terms that recur throughout this page. They are collected here so you can look one up without leaving the section you are reading.
| Term | What it means |
|---|---|
| clinical masking | Quieting the non-test ear with noise so that the measurement belongs to the ear being tested. |
| initial masking level | The noise level masking starts at. On this page it is calculated with Martin's (1980) set and written in dB HL. |
| effective masking level | The noise level calibrated by how many decibels it shifts the threshold at that frequency in a normally hearing ear. Some sources write this level as dB EM. |
| interaural attenuation · IA | The level a sound delivered to one ear loses while crossing the skull to the opposite cochlea. At least 40 dB with supra-aural earphones, and markedly higher with inserts. |
| cross-hearing | A sound delivered to the test ear being heard in the opposite ear. |
| shadow threshold · shadow curve | A threshold measured because of cross-hearing that in fact belongs to the opposite ear. On the audiogram it follows the opposite ear's curve like a shadow. |
| plateau method | Raising the noise step by step and looking for the band over which the threshold does not change. |
| undermasking | The noise is not enough to quiet the opposite ear; the measured threshold is still a shadow threshold. |
| overmasking | The noise crosses the skull to the test ear and lifts its threshold too. |
| central masking | The small rise seen in the test ear's threshold when noise is delivered to the opposite ear, arising at the central level of the auditory pathways. |
| occlusion effect | The improvement in the bone conduction threshold at low frequencies when the ear canal is closed. |
| masking dilemma | The plateau band closing when both ears have a conductive component; masking is required but cannot be achieved. |
| narrow band noise · NBN | Noise limited to a narrow band around the test frequency. The type of noise used when masking for pure-tone thresholds. |
| test ear · TE | The ear whose threshold you are measuring. |
| non-test ear · NTE | The ear you deliver the noise to. |