When we talk about insertion lossthat is, the actual loss of SPL due to the physical presence of an earmold or hearing aid in the eardo we really understand the full problem? Many of us assume that REUR is lost once a hearing aid is inserted in the ear canal. But what actually happens acoustically when a hearing aid is inserted into the ear and turned off? This article looks at insertion loss, how we think of it and deal with it clinically, and how we might be wiser to revise some of our thoughts relative to this concept. In particular, the author suggests that a wide range is possiblefrom little or no attenuation to excessive attenuation within real-world fittings.
While using the probe microphone system to obtain real ear measurements for the purpose of hearing aid fitting, four terms are widely used representing the real-ear sound pressure level (SPL) as a function of frequency1:
- Real-Ear Unaided Response (REUR): The real-ear SPL measured at or near the eardrum with the external ear canal unoccluded;
- Real-Ear Occluded Response (REOR): The real-ear SPL measured at or near the eardrum with an hearing aid inserted in place and turned off;
- Real-Ear Aided Response (REAR): The real-ear SPL measured with the hearing aid in place and the hearing aid turned on;
- Real-Ear Insertion Response (REIR): The difference between the REAR and REUR curves.
A variety of prescriptive formulae (eg, NAL-NL1, DSL[i/o], etc) exist, and fitting hearing aids via a prescriptive method is one of the most common approaches used in clinical settings. Certain correction factors are usually necessary when transforming the prescribed target gain and the 2-cc coupler gain of a hearing aid (For a review of transform functions, see Revits article in the November 1997 HR, pgs 35-38). For example, the CORFIG (Coupler Response for Flat Insertion Loss) correction factor2-5 represents the average difference between the 2-cc coupler gain and the real-ear insertion gain of a hearing aid. This correction factor can be used to convert the coupler gain to real-ear gain (or vice versa) to achieve the prescribed target gain. The CORFIG consists of three factors5-7:
- Correction for RECD (Real Ear Coupler Difference);
- Correction for REUR, and
- Correction for the hearing aid microphone location for the purpose of a customized hearing aid fitting.
This article will not discuss how the correction factors are implemented or how the desired target gain is accurately achieved, but instead ask why the correction for REUR was once recommended. One might think that this topic is not particularly timely and shouldnt merit interest; after all, the procedures of using these correction factors are unnecessary if todays highly flexible programmable aids and software are used. However, this article suggests that more attention to the recommended correction for REUR may, indeed, be worthwhile.
Boiled down to its essentials, this article seeks to provide a better, more thorough answer to the simple, frequently-asked question: What happens when a hearing aid is inserted in the ear and turned off? How do we teach graduate students, CFYs, and new clinicians about insertion loss? Can we answer this question as we might have learned in the classroom: Namely, that the REUR is lost? Understanding the assumptions behind the recommended correction factors for REUR might lead to a better realization of insertion loss, and this could lead to better hearing aid fittings.
The Underlying Assumption
An individuals REURwhich is usually determined by the mass/stiffness of the outer/middle earnormally shows selective amplification of the input signal at the high frequency regions and is viewed clinically as being critical to hearing sensitivity and speech intelligibility. It is generally agreed thatbecause the REUR is lost when a hearing aid is inserted into the external ear canalthe electronic circuits of a hearing aid should provide, in addition to the prescribed target gain, additional gain to compensate for the occluding earmold.
Along with the correction for REUR, various proposals were once suggested. Because individual REUR may vary tremendously among patients,8,9 using the difference between the average REUR and individual REUR was once recommended.10-14
In the early 90s, Mueller & Bryant13 and Mueller15 pointed out some exceptional situations in which the individualized correction for REUR (ie, by using the difference between the individual-vs-average REURs) is unnecessary and inappropriate. One such situation occurs when a patient shows normal hearing up to 2 kHz with hearing loss starting at or above 3 kHz, while the subjects REUR has a greater-than-average value in the 2 kHz region. Minimizing the amount of amplification in the 2 kHz region due to subjects normal hearing sensitivity is necessary in this case, and the authors wondered if it is wise to prescribe additional 2-cc coupler gain to compensate for this greater-than-average REUR.13,15 A similar situation occurs when a patients hearing loss starts at or above 3-4 kHz but the REUR dips within the 4 kHz region. Mueller & Bryant13 and Mueller15 again wondered whether it is wise to prescribe less-than-average 2-cc coupler gain for the 4 kHz region to compensate for the less-than-average REUR. For these two types of situations, they suggested ignoring and making no correction for the unusual REUR. Fikret-Pasa & Revit6 suggested using the individuals REUR directly for compensating the REUR in an individualized manner.
From the above-mentioned examples, it appears that the use of individual REUR is practical and reasonable. The underlying assumption for the real-ear procedures is that the entire REUR is lost when the earmold is inserted in the ear. This assumption is at the heart of the proposals for using either the difference between the individual-average REURs or the individuals REUR directly. That is, since the individuals entire REUR is assumed to be lost upon insertion of the hearing aid, there is a need to compensate for the amounts of individual REUR. This assumption also holds true for the two exceptional situations mentioned above in which no correction for the REUR was suggested.13,15
The Overlooked Assumption?
The Real Ear Occluded Response (REOR) is a real-ear measurement indicating the SPL in the ear canal when the hearing aid is inserted in the ear but is turned off. Insertion loss is, therefore, the difference between the SPL measured at or near the eardrum of the unoccluded ear (REUR) and the REORor the actual attenuation of the individuals natural selective amplification system. If this assumption were true, the REOR would then be a flat response across frequencies at the input level, indicating that the entire REUR is lost at all frequencies. Stated differently, the amount of the ears natural amplification is completely attenuated. Although this area has not been completely studied, data suggest that this assumption does not reflect reality.
Figure 1. Relationship between the REUR (black dotted line on top) versus four hypothetical REOR curves. The orange REOR essentially mimics the REUR, suggesting a case of no attenuation of the natural unaided system. In contrast, the purple REOR demonstrates complete attenuation in which exact REURs are attenuated at all frequencies. The green and dark-blue REORs are shown as examples of other possible responses. In this case, the green REOR stays above the input level and below the REUR, indicating that the original natural hearing condition at the high-frequency region is partly (but not entirely) attenuated relative to REUR (partial attenuation). In the middle frequencies, the green REOR stays even with or above the REUR, indicating possible resonance over the amount of REUR. The dark-blue curve stays below the input level across frequencies, indicating that the amount of attenuation exceeds the original amount of amplification at all frequencies (excessive attenuation).
Possible Insertion Losses: For discussion purposes, a clearer definition of the possible relationships between the REOR and REUR is called for. Figure 1 displays hypothetical relationships between patients REUR(black dotted line that overlaps the orange line) and REORs (Curves 1-4). Only Curve 3 (the purple flat line at the input level) shows the complete attenuation in which the exact REURs are completely attenuated at all frequencies. All other curves denote some possible theoretical patterns of insertion loss due to the presence of the hearing aid. Curve 1 (in orange) basically overlaps with the REUR at all frequencies, suggesting practically no insertion lossa case of no attenuation of the individual natural amplification. The high-frequency portion of Curve 2 (in green) stays above the input level and below the REUR. This indicates that the original natural amplification at the high-frequency region is only partly, but not entirely, attenuated relative to the REURthis could be termed partial attenuation. The mid-frequency portion of this curve stays even with or above the REUR. This suggests that the insertion of hearing aid causes not only no attenuation but also resonance over the amount of REURa unique and possible pattern in the REUR-REOR relationship. Curve 4 (in blue) stays below the input level across frequencies. This means that the amount of attenuation exceeds the original amount of natural amplification at all frequencies relative to the REURand could be termed excessive attenuation.
In fact, all five of these theoretical insertion loss patterns have been documented in several excellent studies related to this topic. Mueller16 and Mueller & Bryant13 displayed mean REORs relative to mean REURs obtained from 38 hearing-impaired subjects fitted with monaural in-the-ear (ITE) and in-the-canal (ITC) hearing aids. The resulting mean REOR curves, caused by the vented aids, clearly showed no attenuation up to 2 kHz and then only partial attenuation above 2 kHz. The more occluded (ie, less vented) ITE and ITC aids resulted in a mean REOR that exhibited partial attenuation at the lower frequencies and excessive attenuation at the higher frequencies. Since the mean REUR and mean REOR were obtained using different subjects fitted with a single device, and the ITE aid used in the study was fitted relatively loosely on subjects, the mean data cannot be viewed as representative of the relationship between individual REUR and individual REOR. However, the data does illustrate that excessive attenuation, partial attenuation, and no attenuation are all possible cases that can occur in addition to the assumed complete attenuation.
Sullivan17 described a classification system to qualitatively categorize hearing aid fittings into four basic fitting classes. In this system, the degree of acoustic coupling of the earmold (and the ITE instruments) and the resulting insertion loss was used as the indices to classify the fitting. For the Class I fittingthe least occluding coupling (eg, tube only or non-occluding earmold)the direct acoustic input (REUR in todays parlance) was essentially unimpeded so that there was essentially no insertion loss and only slight modification of the external ear canal resonance. For the Class II fittinga medium occluded coupling (eg, acoustic modifier mold or large Select-A-Vent)the REUR was reduced with slight insertion loss shown in the high frequency region. For the Class III fittinga closed moldthe REUR was minimized and significant insertion loss occurred at the high and middle frequency regions. Lastly, for the Class IV fittinga closed, deep mold with no ventthe REUR was absent and insertion loss expanded from the high to the low frequency regions. From the data presented in the Sullivan article,17 it can be seen that the amount of insertion loss in the least-occluding condition of the fitting resulted in partial attenuation relative to the mean REUR. Likewise, in the highly occluded condition, excessive attenuation occurred even at the lower frequency ranges. Because the main purpose of Sullivans article was to classify hearing aid fittings into distinct groups, the individual relationship between the REUR and REOR was not dealt with per se. However, it was actually suggested in the article that the measure of REOR may be optional or omitted for the Class III and Class IV fittings. Further, the article noted that, in a real-life clinical setting, it is possible for an individual not to fall cleanly into one of the four classifications.17
Kopun et al.18 investigated the magnitude of attenuation on the external ear canal resonance resulting from the insertion of several devices (ie, a light-weight headphone and different styles of earmolds on both adults and children). The study measured the degree of occlusion produced by the earmolds, and then demonstrated that the magnitude of insertion loss was related to the degree of occlusion. Their data showed that, generally, no attenuation was found up to 1 kHz, then attenuation increased in the high frequency region. The researchers concluded that, on average, only a portion of the external ear canal resonance was lost with the CROS earmold, and for many subjects the attenuation from the snap-ring earmold was actually greater than the external ear canal resonance (as much as 35 dB SPL for some subjects). The study was set up to show how FM systems can be fit on patients with normal hearing sensitivity. The data presented were concerned with attenuation, and individual REUR and REOR were not presented or discussed. But the authors did conclude that partial attenuation was more common in their subjects, and excessive attenuation occurred on some of their subjects.18
Further Indirect Evidence: Several studies have examined the REUR and REAR, but there are relatively few studies dealing with the REOR and the individual actual attenuation.1,19-23 The following is a look at some excellent studies in which the individual REORs were shown.
One clinical application of the individual REOR, proposed by Mueller,9 was to trouble-shoot a vent-related resonance peak for the aided response. The REIR and REOR of a hearing aid fitting for a subject with normal hearing through 1 kHz were displayed. There was a large peak of gain shown on both the REIR and REOR at the 0.75 kHz region. Because this peak of gain is not necessary for normal low-frequency hearing sensitivity, a portion of the vent was filled in. Once the vent size was reduced, the low-frequency peak on the REOR, and consequently on the REIR, disappeared. Thus, by comparing the REOR and REIR, good trouble-shooting can be achieved for this non-circuitry vent-related resonant peak problem. Vent-related resonance was also reported in the mean REOR data in an earlier Mueller study16 in which the mean REOR was actually greater than the mean REUR at the 1.5 kHz region. Apparently, this shows that the relationship between the REOR and REUR may carry the unique pattern, namely a resonance peak, at the low and middle-frequency regionsin addition to the none, partial, complete, and excessive attenuation. Another noticeable feature of the figures provided by Mueller9,16 is that excessive attenuation might have occurred at and above 1 kHz.
Another suggested use of the REOR is related to reducing the occlusion effect: the enhanced low frequency response to a subjects own voice when a hearing aid or earmold is placed in the ear. With the loudspeaker of the probe microphone system disabled, open ear canal response and occluded ear canal response (a hearing aid in the ear but turned off) can be measured while the subject is vocalizing a vowel sound at a constant level. The difference between the open and occluded responses represents the occlusion effect, which usually induces the complaints of hollow, echoing, and in-the-barrel voice or a fullness sensation.
One approach of reducing the occlusion effect is to use a deep-canal fitting, once termed minimal contact technology, with the earmolds or hearing aids. The concept is to design the canal portion of the device for limited contact with the cartilaginous ear canal, moving the earmolds seal to the bony portion of the external ear canal. Mueller & Hawkins24 showed the open and occluded ear canal responses to subjects vocalization of /ee/ for both the standard earmolds and long-canal earmolds. The study demonstrated that the occlusion effect was essentially eliminated with the long canal earmold, on which the occluded low-frequency responses to self-vocalization approached the open ear canal response. Similarly, Bryant, Mueller & Northern25 reported similar measurements of the occlusion effect from long-canal ITE and standard ITE hearing aids. Their study indicated that the occlusion effect was largely reduced for the long-canal ITE aid as compared to the standard ITE aid.
In general, as the REOR falls below the REUR (indicating greater attenuation), the occlusion effect increases relative to ones own voice. It was suggested, therefore, that the REOR may be used as an indirect measure of the occlusion effect. For example, Mueller9 displayed the REOR and REUR of one of the subjects from Bryant, Mueller & Northern study25 in which the REOR of a standard ITE did indeed fall substantially below the REOR of the long-canal ITE aidverifying the usefulness of REOR as indirect indicator of the occlusion effect. Although the purpose of the studies was related to the occlusion effect, both the partial and excessive attenuationrather than the assumed complete attenuationwere noticed from the REOR and REUR curves for both long/standard canal sizes. Additionally, it is also observable from the direct measurements of the occlusion effect in these two studies9,25 that partial attenuationnot complete attenuationappears to be fairly common.
Sweetow26 studied the use of more open, non-occluding coupling techniques designed to reduce the occlusion effect in low frequencies while producing less insertion loss in high frequencies. His preliminary REOR data from various non-occluding earmolds and ITE aids suggested that, as the degree of occlusion resulting from the earmolds or hearing aids increased, the amount of insertion loss increased as well. It was pointed out that the insertion loss with various devices did not occur only between 2.5 kHz and 4 kHz, but actually began at 0.6 kHz in the more closed earmolds and ITE aids. The study showed that only the tube fitting produced no insertion loss, and only this fitting could be truly viewed as a non-occluding coupling technique. Although low in occlusion, the helix-style ITE aid (the least occluding ITE) still resulted in some (albeit minimal) insertion loss. It can be seen from the curves in the Sweetow article that patterns of partial and excessive attenuationnot the complete attenuationappear to be the prevalent case.
The studies detailed above indicate that several clinical uses of the REOR have already been suggested. However, none of the studies directly addressed the assumption of insertion loss in hearing aid fittings. The relationship between individual REUR and individual REORincluding the resonance peak, no attenuation, partial attenuation, complete attenuation, and excessive attenuationwas indirectly observable or inferred, but it appears that it has yet to be fully addressed.
Individual Insertion Loss: The Actual Attenuation
Theoretical Considerations: Shaw27 studied the basic acoustic characteristics of REUR and revealed that the REUR curve could be attributed primarily to the sound diffraction or resonance resulting from the external ear canal and concha. Sullivan17 pointed out that the amount of modification on the resonance/diffraction of the pinna, concha, or canal via the placement of a hearing aid was related to the insertion loss.
The earmold or hearing aid itself is conceivably a significant contributing factor in the mass or stiffness of the device. The type, shape. size, material density, etc, of the hearing aid and earmold, internal electronic components, internal volume, faceplate, battery compartment and battery, length of the canal portion of the earmold and hearing aid, and the size and type of the vent could also contribute to the resonance or diffraction of sound. Additionally, when an hearing aid or earmold is placed in the ear, other influencing factors could include: impedance of the residual volume between the eardrum and the receiver, the location of acoustic seal, the impedance change over the pinna/concha/canal areas, the impedance change caused by the presence of vent, and the overall degree of occlusion. It is possible these could all play a role in reflecting and/or passing sound waves and modifying the resonance/diffraction features of the earthus, influencing the REUR-REOR relationship.
Clearly, the physiological impedance features, related to the anatomical structures of the outer and middle ear, are conducive to the formation of the natural selective amplificationthe REUR. The artificial impedance features, created by the combination of the earmold or hearing aid shell and the impedance changes over the outer and middle ear, are obstructive to sound transmission. Only when this artificial obstructive impedance is exactly the same as the conducive physiological impedance would the natural selective amplification be completely removed at all frequencies and result in complete attenuation.
This seems unlikely in most cases. Logically, since the mass and stiffness of each hearing aid or earmold fitting varies dramatically and involve as numerous variables as described above, the chance that the net resulting impedance of the fitting is exactly the same as the individuals physiological impedance appears to be rare. This means that the real individual insertion loss may rarely equal complete attenuation as often assumed. On the contrary, it appears far more likely that the artificial impedance of the fitting would be different from the physiological impedance and induce various response patterns for the REUR-REOR relationship other than complete attenuation.
Examples of Clinical Use: Clinical experience also indicates that complete attenuation is rare. Figures 2-5 reveal examples of the individual REOR and REUR when the hearing aid with battery is worn in place but the aid is turned off. These data were collected only to illustrate the frequently seen patterns of response between the individual REOR and REUR.
Figure 2. A relationship between the individual REUR and REOR, produced by a BTE hearing aid with a shell earmold.
Figure 3. A relationship between the individual REUR and REOR, produced by a CIC hearing aid.
Figure 4. A relationship between the individual REUR and REOR, also produced by a CIC hearing aid.
Figure 5. A relationship between the individual REUR and REOR, produced by a vented ITE hearing aid.
It can be seen that the four possibilitiesresonant peak, no attenuation, partial attenuation, or excessive attenuation relative to the REUR is more prominent than complete attenuation. In fact, all these patterns of individual insertion loss have been repeatedly observed with the insertion of hearing aids, earmolds, and other coupling devices (including trimmed/untrimmed ear impressions and various earplugs). Partial and excessive attenuation are frequently observed. Further, in the rare cases when complete attenuation does appear, it usually does so as data point(s) within limited frequencies, rather than over a wide frequency range (ie, with the REOR flat at all frequencies at the input level).
The intention here is not to display a standard or typical REUR-REOR response pattern because there is simply no so-called typical response pattern or typical actual attenuation to report. The actual REOR in a patients ear may exhibit any form of frequency responses shown in Figures 1-5 with large variations in dB SPLs across frequency. The response patterns for the REUR-REOR relationship contain not only the complete attenuation (rarely seen), but also all of the other four possibilities. Additionally, the excessive and partial attenuation is observed more often at high frequencies, while the resonance peak is observed more often at the lower and middle frequencies.
Evidence suggests that the assumption behind the recommended corrections for REUR are oversimplified. The individual insertion loss in hearing aid fittings (ie, the actual attenuation versus a patients natural selective amplification) is rarely equal to the individuals measured REUR. The assumption that the REUR is gone when a hearing aid is inserted in an ear does not adequately describe the situation.
Admittedly, the concept of actual attenuation and the five attenuation patterns discussed above may not be surprising to some clinicians; however, these possibilities have not been explicitly discussed enough in the field and within the literature. The acknowledgement and appreciation of the actual REOR-REUR relationship should allow better confidence in teaching and answering the simple but frequently-asked question: What happens when a hearing aid is inserted in the ear and turned off? Of course, highly flexible modern hearing aids can be programmed in-situ to deliver the desired target gain without the practitioner knowing the underlying assumption and phenomena. While this may be true, a good background in fitting fundamentals often comes to ones aid when thinking out thorny fitting problems.
One clinical implication is perhaps toward what the phrase, individualization of hearing aid fittings, means. This discussion has shown that the proposed correction for REUR by using the individual REUR values is oversimplified. A logical alternative could be compensating for the individuals actual insertion loss by utilizing the magnitude of actual attenuation, which is induced by the individuals hearing aid or earmold. As discussed, the key is looking into the individual REOR relative to individual REUR. After all, the correction for the actual insertion lossinstead of the correction for REURis supposed to be the original goal. Clinically, it takes only a few more seconds to measure REOR during the fitting or verification procedure. This represents one more meaningful application of the REOR in the individualized fitting with modern programmable hearing aids. It appears that this line of consideration requires more future research.
|Bailey K. Wang, PhD, is an associate professor and audiologist in the Department of Communication Sciences and Disorders of the College of Health Sciences and Human Services at the University of Texas-Pan American, Edinburg, Tex.|
Correspondence can be addressed to Bailey K. Wang, PhD, University of Texas-Pan American1201 W. University Drive, Edinburg, TX 78541; email: [email protected].
1. Mueller HG, Hawkins DB, Northern JL. Probe Microphone Measurements: Hearing Aid Selection and Assessment. San Diego, CA: Singular Publishing Group, Inc; 1992.
2. Killion M, Monser E. CORFIG: Coupler response for flat insertion gain. In: Studebaker G, Hochberg I, eds. Acoustic Factors Affecting Hearing Aid Performance. Baltimore, MD: University Park Press; 1980: 149-168.
3. Killion MC, Berger EH, Nuss RA. Diffuse field response of the ear. J Acoust Soc Am. 1987; 81(Suppl. 1), S75.
4. Bentler RA, Pavlovic CV. Transfer functions and correction factors used in hearing aid evaluation and research. Ear Hear. 1989; 10(1), 58-63.
5. Killion MC, Revit LJ. CORFIG and GIFROC: Real-ear to coupler and back. In: Studebaker GA, Hochberg I, eds. Acoustical Factors Affecting Hearing Aid Performance (2nd ed.). Boston, MA: Allyn and Bacon; 1993: 65-85.
6. Fikret-Pasa S, Revit LJ. Individualized correction factors in the preselection of hearing aids. J Sp Hear Res 1992; 35(2): 384-400.
7. Hawkins DB. Corrections and Transformations relevant to hearing aid selection. In: Mueller HG, Hawkins DB, Northern JL, eds. Probe Microphone Measurements: Hearing Aid Selection and Assessment. San Diego: Singular Publishing Group Inc; 1992: 251-268.
8. Upfold G, Byrne D. Variability of ear canal resonance and its implications for the design of hearing aids and earplugs. Austral Jour Audiol. 1988; 10:97-102.
9. Mueller HG. Terminology and procedures. In: Mueller HG, Hawkins DB, Northern JL, eds. Probe Microphone Measurements: Hearing Aid Selection and Assessment. San Diego: Singular Publishing Group Inc; 1992: 48-56.
10. Mueller HG. Individualizing the ordering of custom hearing instruments. Hear Instrum. 1989; 40(2):18-22.
11. Valente M, Valente M, Vass W. Selecting an appropriate matrix for ITE/ITC hearing instruments. Hear Instrum. 1990; 41: 20-24.
12. Bratt G., Sammeth C. Clinical implications of prescriptive formulas for hearing aid selection. In: Studebaker G, Bess F, Beck L, eds. The Vanderbilt Hearing Aid Report II. Parkton, MD: York Press; 1991: 23-35.
13. Mueller HG, Bryant MP. Some commonly overlooked uses of probe microphone measures. Sem in Hear. 1991; 12, 1, 73-91.
14. Valente M, Valente M, Vass W. Use of real-ear measures to select the gain and output of hearing aids. Sem Hear 1991; 12:53-61.
15. Mueller HG. Individualizing the ordering of custom hearing aids. In: Mueller HG, Hawkins DB, Northern JL, eds. Probe Microphone Measurements: Hearing Aid Selection and Assessment. San Diego: Singular Publishing Group Inc; 1992: 183-200.
16. Mueller HG. Probe tube microphone measures: Some opinions on terminology and procedures. Hear Jour. 1990; 42(1):1-5.
17. Sullivan R. An acoustic coupling-based classification system for hearing aid fittings. Hear Instrum. 1985; 36(9) Pt I: 25-28, 36; 12 (Part II & III): 16-22.
18. Kopun J G, Stelmachowicz PG, Carney E, Schulte L. Coupling of FM systems to individuals with unilateral hearing loss. J Sp Hear Res. 1992; 35, 1, 201-207.
19. Byrne D, Upfold G. Implications of ear canal resonance for hearing aid fitting. Sem in Hear. 1991; 12(1):34-41.
20. Mueller HG. Probe microphone measurements: Yesterday, today, and tomorrow. Hear Jour. 1998; 51, 4, 17-22.
21. Palmer CV, Lindley IV GA, Mormer EA. Selection and fitting of conventional hearing aids. In: Valente M, Hosford-Dunn H, Roeser RJ, eds. Audiology Treatment. New York, NY: Thieme Medical Publishers Inc; 2000: 397-431.
22. Revit LJ. Real-ear measures. In: Valente M, Hosford-Dunn H, Roeser RJ, eds. Audiology Treatment. New York, NY: Thieme Medical Publishers Inc; 2000: 105-145.
23. Staab WJ. Hearing aid selection: An overview. In: Sandlin RE, ed. Textbook of Hearing Aid Amplification: Technical and Clinical Considerations (2nd ed). San Diego: Singular Publishing Group; 2000: 55-135.
24. Mueller HG, Hawkins DB. Assessment of fitting arrangements, special circuitry, and features. In: Mueller HG, Hawkins DB, Northern JL, eds. Probe Microphone Measurements: Hearing Aid Selection and Assessment. San Diego, CA: Singular Publishing Group, Inc; 1992: 221-224.
25. Bryant M, Mueller H, Northern J. Minimal contact long canal ITE hearing instruments. Hear Instrum. 1991; 42:12-15.
26. Sweetow R. The truth behind "non-occluding" earmolds. Hear Instrum. 1991; 42(6):25.
27. Shaw E. Transformation of sound pressure from the free field to the eardrum in the horizontal plane. J Acoust Soc Amer.1974; 56: 1848-1861.