Masking Fundamentals

September 2016 -- Check out our new resource: mBook -- Making Masking Manageable

1) Air- and bone-conducted signals both can create the same traveling wave on the basilar membrane. In bone conduction, the distortion of the cochlea shell is one of the mechanisms that can produce the traveling wave.

Traveling wave from an air-conducted signal

Traveling wave from a bone-conducted signal

2) Since the signal must be present at the cochlea in order to be heard, then masking is about the signal and noise levels in the cochlea. In order not to mask the traveling wave from a pure tone, the envelope of the traveling wave from the noise signal must "cover it up".

3) The pure tone could be masked by white noise; however, the white noise would excite more neurons, and that would increase the loudness of the sound. Thus, narrow-band noise is used for masking pure tones.

4) Speech is comprised of separate sounds, with differing spectra, that occur at different times. Since vowel sounds tend to have greater intensity, the speech masking noise is wide band, and contains more low-frequency energy. Below the time and spectral waveforms for the word "ticket" are shown, and a rough illustration of the traveling waves for the sounds is shown. Speech noise needs to cover all possible speech frequencies. Because speech energy contains greater intensity low-frequency energy, speech masking noise will have greater low frequency energy.

5) Which of the following are true?
    a) Masking noise must be the same frequency, and at least as intense as the signal in order to mask the signal
    b) Masking noise must be lower in frequency due to the upward spread of masking in order to mask a signal
    c) Noise should be no wider in spectrum than needed to mask the signal to keep the signal intensity down, although pure tones shouldn't mask pure tones as that creates tone on tone interference and patient confusion.
    d) The important concept is having the signal and noise levels in the ear canal of appropriate intensities -- the noise must be more intense than the signal at the eardrum
    e) Masking occurs on the basilar membrane, so in the final analysis, we have to consider how loud the signals are in the cochlea

6) Although the masking occurs in the cochlea, the level in the cochlea cannot be exactly measured.

The reference intensity for masking noise is decibels of Effective Masking (EM), and it is measured as an air-conducted signal.

If a noise is X dB EM then it just masks an X dB HL sound, presented to the SAME ear. Thus, if you put a 40 dB HL pure tone into the right insert earphone, and with the second channel of your audiometer, put in a 40 dB EM signal into the right insert earphone, the 40 dB EM signal will no longer be heard (assuming the patient could hear the tone to begin with).

While the reference level for EM is defined when both signal and noise are sent to the same transducer, of course in clinical masking, the noise is put in one ear, and the signal goes to the other ear. The noise is meant to prevent the non-test ear cochlea from hearing the sound. An intense air-conducted sound can vibrate the skull, creating a compressional wave in the non-test ear.

So, in review, the air-conducted signal is heard at the non-test ear by:
a) air conduction
b) bone conduction

7) The occlusion effect is an enhancement of bone conducted sound that occurs when the conductive mechanism is no longer open. It enhances the signal level present at the non-test ear. The crossed-over bone conducted sound will also vibrate the walls of the middle ear and external auditory meatus. The bone vibration creates an air-conducted sound at the non-test ear. When the ears are open, much of this sound leaks out. However, with the non-test ear occluded, there is no escape path for the bone-conducted sound, and the energy sent to the cochlea of the non-test ear increases.

This is mainly an issue when testing with bone conduction. (Yes, you can vibrate the walls of the ear canal with the crossed over air conduction sound, but that doesn't add much energy.) The crossed over bone conducted sound is really where we see the occlusion effect occurring.

If the non-test ear has a significant conductive pathology, that pathology will likely either have already created the occlusion effect (e.g. if there were a foreign object in the non-test ear canal) or prevent the extra air-conducted sound from reaching the non-test ear cochlea (e.g. if there were fluid in the non-test middle ear).

Therefore, it is true that:
    A) The occlusion of the non-test ear always increases the sound at the non-test ear cochlea
    B) The occlusion of the non-test ear increases the sound at the non-test ear cochlea if there is conductive loss.
    C) The occlusion of the non-test ear increases the sound at the non-test ear cochlea if there is normal hearing.
    D) The occlusion of the non-test ear increases the sound at the non-test ear cochlea if there is cochlear loss.

8) Since the masking noise is presented to the non-test ear by air-conduction, and since the noise has to be present in the non-test ear cochlea to create masking, conductive loss decreases the effectiveness of the masking signal.

If there is a 25 dB air-bone gap in the non-test ear, then 70 dB EM in the non-test ear canal is only __ dB at the non-test cochlea.

9) Cross-over of the test ear signal to the non-test ear is a greater problem when using supra-aural earphones (e.g. TDH-49P in MX-41/AR cushions) than if using insert earphones, because there is more contact area of the sound with the skull.

Therefore when using supraural earphones, _________ masking will be required at the non-test ear. (More or less?)

10) Generally speaking, the occlusion effect is greater with supra-aural earphones too, assuming that the earphone is properly snug. (A stretched out earphone band doesn't permit full occlusion of the ear, so the low-frequency sound would leak out, minimizing the occlusion effect.)

An insert earphone in the non-test ear keeps a part of the non-test ear canal from vibrating, which minimizes the amount of the occlusion effect. The deeper the placement of the insert earphone in the non-test ear, the less of an occlusion effect occurs.

Dean & Martin (1991 - American Journal of Audiology) report the following occlusion effect size (for young, female college students, with mastoid oscillator placement):

	250 Hz	500 Hz	1k Hz
Supra-aural	17 dB	13 dB	3 dB
Shallow insert	16 dB	10 dB	6 dB
Deep insert	6 dB	6 dB	1 dB

Various texts recommend different occlusion effect values. You may read:

250 Hz	500 Hz	1000 Hz
15 dB	15 dB	10 dB
15 dB	10 dB	5 dB
20 dB	15 dB	10 dB

All these are reasonable values, but in each case, it is an approximation of the occlusion effect. Different patients will have different occlusion effect values, as the standard deviation of the occlusion effect is 4 to 7 dB. Of course, if the patient has conductive pathology in the non-test ear, the occlusion effect is reduced or eliminated.

11) It is possible to measure the patient's occlusion effect, and in tough masking cases, this may be a step worth taking. To do this, first obtain the patient's unmasked bone conduction thresholds. Then, put the masking earphone on. Before turning on the masking, re-establish the bone conduction threshold. The enhancement, if any, is the size of the occlusion effect.

What is the measured occlusion effect at 1k Hz? (Note, the dashed bone conduction thresholds in this figure are for convenience of the illustrator. They are not a standard, nor recommended, type of symbol.)

12) The degree to which sounds cross from test ear to non-test ear depends on the frequency of the sound (particularly for insert earphones) -- and the individual characteristics of the patient. In general, and especially for insert earphones, the crossover happens at the lower intensities in the mid frequencies.

Unlike the occlusion effect, there is no sure way of knowing whether sound has crossed to the non-test ear. You could ask your patient which ear or ears hears the sound, but that is not a reliable indicator. Therefore, we tend to err on the side of caution - masking the non-test ear if there is a possibility of crossover.

The minimum interaural attenuation for supra-aural earphones is often said to be 40 dB, and for insert earphones 50 dB.

Most patients will not have cross over at these levels -- the interaural attenuation values are on average about 70-90 dB for insert earphones and 55-65 for supra-aural earphones. However, you cannot assume that your patient has average or high interaural attenuation values: masking must be used if there is the possibility of crossover.

Is there any "harm" in assuming that the minimal interaural attenuation is 40 dB, even if using insert earphones?

13) Putting a masking sound in the non-test ear can increase the threshold of hearing in the TEST ear, as well as in the non-test ear. This is termed the central masking effect. Stimulating the non-test ear with a noise signal creates an afferent (ascending) stimulus, which ascends both ipsilateral and contralateral brain pathways. The presence of this neural activity can mask the nerve signals coming from the other (test) ear. Thus, central masking is a neural phenomenon coming from neural channels being "busy" with the additional duty of transmitting information about the non-test ear noise.

The newest ANSI calibration standards assume that bone conduction testing is conducted in the presence of contralateral masking. If you do not put in the contralateral masking, would your bone conduction thresholds tend to be better or worse? That is, would you tend to see air-bone gaps or bone-air gaps if you do not mask?

14) Next, we will examine a way of describing what happens during masking, using the "masking head" model. This masking explanation system is used in AudSim, the audiometer / masking simulator from AudStudent.com

Information about what is happening at the test ear is always on the left side of the figure; the non-test ear is always on the right.
The pink is information about the signal levels; green codes information about the masking noise levels.
Thresholds, if provided, are in the boxes towards the top of the screen. The air-bone gap size is noted in the square between/below the the signal levels for air and bone conduction.
The solid blue bar running up the center of the masking head illustrates the interaural attenuation.
The current version of the masking simulator generally assumes a 40 dB interaural attenuation value for air-conducted signals (TDH earphone use is default).

What is the level of the crossed-over signal at the non-test ear? Is it the signal level presented via air-conduction that crosses to the non-test ear, or is it the signal level after attenuation by the air-bone gap that crosses over to the non-test ear?

15)

How much masking noise needs to be presented (via air-conduction at the non-test ear) in order for there to be masking of the crossed-over signal?

16) Let's make sure we understand this point about central masking. The signal is 50 dB HL, the air-bone gap in the test ear is 45. There is 5 dB of signal at the test ear cochlea. Why doesn't the patient respond?

17) Let’s take a second “masking head case”. Let’s check your understanding of the masking head figure. What is the “true” test ear AC threshold and “true” BC threshold? What are the non-test ear (NTE) air and bone thresholds?

18) What is the interaural attenuation for the air conducted sound for this example, and what is the level of the crossed-over signal if the test ear is stimulated at 40 dB HL? Would the patient hear the sound?

19) If the signal level is raised to 45 dB HL, would it be heard? How about if it is 50?

20) Let’s keep track of these results graphically. Let’s record this threshold obtained without masking.

21) If we put in 5 dB EM masking to the NTE, would that change the hearing threshold or would the patient still hear the crossed over tone?

22) Let’s record that result as well.

23) As you probably predict, inputting 10 dB EM should alter threshold. How loud would the test ear signal be in the test ear in order for the sound to be heard (in either the test or non-test ear)?

24) Increase the masking level 5 dB to 15 dB EM. What would the threshold be? Mentally visualize the graph of hearing threshold versus masking intensity.

25) Increase the level of the masker to 20 dB HL, which will raise the patient threshold to 65 dB HL. Does the patient hear the tone in both the test and non-test ears now?

26) Increase the masking noise to 25 dB. What is the patient's threshold, and which ear or ears hear(s) the tone?

27) Will the patient respond when you increase the masking to 30 dB EM?

28) Increase the noise to 35, 40, and 45. Does threshold change, and at what point would the masking signal cross back to the test ear?

29) Clinically, a 15 dB "wide" plateau is considered indication that masking is appropriate - that the non-test ear is not participating. Let's continue with the scenario though and assume that increasing the masking to 50 dB causes a 5 dB central masking effect, which remains at this level as we put in 60 and 70 dB of contralateral masking. What happens to threshold with these masking levels?

30) How wide would the plateau be? When would threshold shift because of overmasking?