Hearing correction system

Abstract

The novel hearing correction system presented comprises a device and method for a hearing test to assess hearing loss, a software to process the measurement data and provide settings to the digital signal processing elements, and a digital signal processing topology to perform dynamic compensation. This novel hearing test is based on equal loudness impressions and measures hearing ability at several sound pressure levels, not just hearing thresholds as performed by standard hearing tests.

Description

[Background technology]

[0001] By the end of World War II, Raymond Carhart and Norton Canfield started hearing hospitals for soldiers who were experiencing hearing loss. In these hospitals, they developed electronic audiometers used to test hearing. The results of an audiometer test were called audiograms. Although technology and technical devices have improved significantly since the end of World War II, the basic function of the audiometer has remained the same to this day.

[0002] The most common hearing test today for assessing hearing loss and hearing aid fitting is called pure-tone audiometry, which measures the hearing thresholds of both ears separately at eight standard frequencies ranging from 250 Hz to 8000 Hz. The results of the test are displayed on an audiogram, which shows the hearing thresholds at each of the frequencies tested.

[0003] From interviews with hearing aid users and people with hearing loss who have tried to use hearing aids, a pattern emerges. Many express dissatisfaction with the amplification of sounds from their hearing aids. Most regular hearing aid users say they are helped with dialogue intelligibility in environments with low to moderate sound pressure, but generally consider that at slightly higher sound pressure levels, the sounds from their hearing aids are too loud, brittle, and metallic, especially at higher frequencies. People who have tried hearing aids but choose not to use them consider the loud, brittle, metallic sounds to be intolerable, and any improvements are outweighed by this significant drawback. It appears that a user usually needs to have a significant degree of hearing loss to accept the universal drawback of loud, brittle sounds at higher frequencies.

[0004] The correlation between typical audiometric tests and audiograms relies on the assumption that the human hearing mechanism has an essentially linear response to input. A hearing threshold raised by, say, 20 dB at a certain frequency is assumed to give a proportional attenuation of 20 dB across the entire audible dynamic range at that frequency. This is obviously not entirely true, since hearing aids are not usually programmed to fully compensate for the hearing loss indicated by a hearing threshold measurement in an audiogram.

[0005] There is a considerable amount of scientific literature and published philosophy covering the subject of compensation levels in relation to audiogram data. A very rudimentary summary provides the conclusion that approximately half to two-thirds of the measured impairment should be compensated depending on the situation.

[0006] Considering the apparently weak relationship between measured elevated hearing thresholds and optimally applied compensation, together with the non-ideal user experience feedback, there seems to be reason to suspect that currently used measurements and compensations for hearing loss in hearing aids are not optimal, and new methodologies are needed.

[0007] The novel hearing correction system presented comprises a hearing test to assess hearing loss, software to process the measurement data and provide settings to a digital signal processing element, and a digital signal processing topology to perform dynamic compensation. The hearing test measures hearing ability at several sound pressure levels, as well as hearing thresholds, and is based on the impression of equal loudness. The digital signal processing topology utilizes dynamic elements that provide amplification that varies dynamically based on the measured hearing loss and the instantaneous sound pressure level. Summary of the Invention

[0008] The present invention is directed to a device for hearing assessment and / or correction, said device comprising: - administering a hearing test to a subject at different frequencies and different sound levels, said hearing test involving the use of reference sounds and test sounds and adjusting the test sound levels until the subject perceives the test sound levels as more or less equal to the loudness of the reference sounds, obtaining data on test sound levels and reference sounds for different frequencies at different sound levels, Here, one or more dynamic filters are involved to compensate the data.

[0009] In connection with the above, it is noted that the device according to the invention may be any type of device suitable for including a software unit, such as a mobile phone, a PC or other computer unit, a tablet, etc. Furthermore, personal sound units, such as personal sound amplifier products, are also embodied as such by the invention. [Brief description of the drawings]

[0010] Detailed description of the hearing correction system A typical pure tone hearing test produces an audiogram with hearing threshold data for each ear. The validity of the test to assess hearing loss is based on the assumption that an elevated hearing threshold translates into a perceived decrease in level corresponding to the measured increase. In other words, the assumption is that if the hearing threshold is elevated by 20 dB at a particular frequency, there is a corresponding perceived attenuation of sound by 20 dB across the entire audible dynamic range at that frequency. Figure 1 shows hearing threshold data from an audiogram of a person with significant hearing loss at eight standard frequencies (250 Hz, 500 Hz, 1 kHz, 2 kHz, 3 kHz, 4 kHz, 6 kHz, and 8 kHz) starting from 250 Hz shown on the left side of the figure and ending at 8 kHz on the right side.

[0011] The new hearing correction system uses a novel hearing test that evaluates hearing loss over a wide dynamic range, significantly different from pure tone hearing tests, which only provide information about hearing thresholds. The test is based on the perceived equal loudness of a reference tone compared to a test tone. The test tone level is adjusted until the tester perceives it as being equally loud as the reference tone. Typically, the reference tone and the test tone are played alternately with an alternating frequency of 0.5 Hz. It is clear that the alternating frequency can be changed to higher or lower frequencies. A feasible range is from 0.1 Hz to 2.5 Hz. The alternation between the two compared tones can be automatic or under manual control.

[0012] The equal loudness levels used are based on the ISO 226 standard levels at different frequencies and sound pressures. The test and reference tones may be pure sine wave tones, as in hearing tests, or other types of sounds, bandwidth limited to the desired test frequency band. Such sounds may consist, by way of example, of warbles, noises, or any type of complex signal containing multiple frequencies. Any type of signal may be used that has an appropriate bandwidth to fit the test frequency band.

[0013] In a standard hearing test, the frequency range is divided into eight well-known standard test frequency bands. The equal loudness test results shown in the figure use 81 frequencies ranging from 100 Hz to 10 kHz. This number of test frequencies is, of course, not critical, and the number of frequencies tested can be anywhere between 1 and the 81 presented, or more. Preferably, the number of test frequencies should be in the range of 5 to 35. Fewer test frequencies result in quicker tests that are easier for the test subject to perform, but also provide less information, and a good balance is needed. The individual test frequencies can be either pure sine wave tones, or complex signals with most of their frequency components contained within the test frequency band between adjacent test frequencies. Complex test signals with a wider frequency spectrum than pure sine waves can provide more information about hearing ability within the test frequency band, and are particularly useful in tests with fewer test frequencies. The equal loudness test frequency range can extend to all frequencies covered by the ISO 226 standard, i.e., from below 100 Hz to 20 Hz, and from above 10 kHz to 12.5 kHz. If equal loudness data are available and it is desired to test a wider frequency range than 20 Hz to 12.5 kHz, the test frequency range can of course be further extended. Under normal conditions, the range of 100 Hz to 12.5 kHz is preferred, and in many cases can be reduced to the range of 100 Hz to 10 kHz. It may also be expedient to further reduce the hearing test frequency range, for example, to 250 Hz to 8 kHz.

[0014] Equal loudness testing is typically performed at sound pressure level intervals of 10 dB, with the minimum test level being just above the subject's hearing threshold at the reference frequency. Any other sound pressure level granularity can be used, whether more closely spaced, for example 5 dB or 3 dB, or more widely spaced, such as 15 dB or 20 dB. Usually, if the subject has no or low hearing loss at the reference frequency, the minimum test level is set at 10 phon. If the subject experiences hearing loss at the reference frequency, the minimum level can be increased. Equal loudness testing is usually limited to a maximum sound pressure level of 70 phon to avoid subjecting the testee to sound pressures that could cause hearing damage. Of course, there are no other restrictions limiting the maximum sound pressure during the test, but the 70 phon level appears to be the highest level that most subjects feel comfortable receiving and is within safety limits.

[0015] Equal loudness tests should be performed at the reference frequency and at a minimum one higher sound pressure level, with a test level just above the hearing threshold of the subject, i.e. from 0 to 10 dB above, preferably less than 5 dB. More sound pressure test levels provide a better assessment of the hearing ability of the test subject. At least two sound pressure test levels are required, and usually four levels with a sound pressure level interval of 10 dB give good results. Testing at more levels increases the burden on the subject to perform the test, so it is desirable to perform the test at fewer sound pressure levels. Three test levels with a sound pressure level interval of 10 dB often constitute a reasonable compromise that gives reasonable results. The number of levels required depends on the sound pressure level interval between tests. The smaller the interval, the more tests are needed to produce good data, and the larger the interval, the fewer tests are needed. The number of tests can, of course, be any number from two onwards.

[0016] First, it is necessary to ensure that the subject can hear the reference tone at the lowest level used in the test. For this purpose, a normal hearing test is used, which establishes the hearing threshold at a selected reference frequency. If the subject is found to have severe hearing loss at the reference frequency, it may be beneficial to select another reference frequency where the hearing loss is less severe. The reference frequency should be located approximately in the middle of the frequency range being tested, because it makes equal loudness comparison easier for the subject. However, if the hearing loss in the middle of the frequency range is severe, moving the reference to another frequency can provide a useful compromise. Another technique to facilitate comparison is the step approach, which utilizes multiple reference frequencies. Multiple reference frequencies reduce the frequency distance between the test tone and the reference tone, making the comparison easier. The basic one reference tone test aims to place the reference tone in the middle of the frequency range, while the step approach uses at least two reference tones, but there can be any number, and the reference tones are spread approximately evenly across the frequency range being tested. In the stepwise approach, first the equal loudness levels of the reference tones are tested so that all reference tones are perceived to have the same loudness starting with the reference tone located at the frequency where the hearing loss is as small as possible. Then the test tone is compared to one or more reference tones to complete the test.

[0017] Figure 2 shows the equal loudness test data for a normal hearing person at levels of 10 phon, 20 phon, and 30 phon. The test was performed with a reference tone of 880 Hz, and the test tone was varied in 81 steps from 100 Hz to 10 kHz. The reference tone and the test tone alternate automatically at 0.5 Hz. At frequencies above about 1.8 kHz, some wiggles are present in the graph at all levels. These smaller wiggles do not signify hearing loss or hearing deviations. They are a combination of normal variations in headphone response at higher frequencies experienced by individuals, and subject specific deviations from the ISO226 average loudness data. The variations in headphone response are normal and arise due to the interaction of the headphones with the outer ear and ear canal of the test subject. The very slight deviation indicating the smallest hearing threshold increase is the small peak in the 10 phon graph just below 3 kHz. The perceived level differences between 10 phon, 20 phon, and 30 phon are all 10 dB, as indicated by the even spacing of 10 dB between the graphs at all frequencies. Clearly, the assumption and basis of audiometer testing, that the human hearing mechanism exhibits an essentially linear response to changes in input level, applies to individuals with normal hearing. However, as will be shown later, this is not true for people who suffer from hearing loss.

[0018] Figure 3 shows audiogram data for eight standard frequencies from a person with significant hearing loss (blue graph (2)) compared to the equal loudness test data from the same person (red graph (1)) at a minimum equal loudness test level just above the hearing threshold. Apart from small local variations, the equal loudness data and audiogram data correlate well, validating the validity of the equal loudness test.

[0019] Figure 4 reveals equal loudness test data at levels 10 phon, 20 phon, 30 phon, and 40 phon for a person with significant high frequency hearing loss. The bottom trace shows the lowest 10 phon level, which is just above or at the value of the subject's hearing threshold at a reference frequency of 880 Hz. The graph shows, to the right, a significant rise in the hearing threshold starting just above 1 kHz. At lower frequencies, to the left of the figure, the graphs are spaced apart by the expected 10 dB. At higher frequencies, starting from about 2 kHz and above, the spacing is significantly less than 10 dB. At 300 Hz, the 10 to 40 phon traces are separated by nearly 40 dB, which corresponds to an increase in sound pressure level. The slight rise in the 10 phon trace between 100 Hz and 1 kHz produces a slightly smaller separation for the 20 phon trace than between the other traces. This is caused by a slight rise in the hearing threshold affecting the lowest 10 phon level. Above 20 phon the rise disappears and the spacing is exactly 10 dB apart with some small local deviations. At higher frequencies above 2kHz, where there is a significant increase in the hearing threshold, the 10 to 40 phon traces are separated by only a few decibels. This makes it clear that at, say, 3kHz, the sound pressure level needs to increase by a few decibels to be perceived by the subject as having increased by 40dB. Very obviously, when the sound pressure level exceeds the hearing threshold, the brain begins to radically compensate for the increased hearing threshold, and the ear / brain no longer behaves as a linear device, and in fact is very far from it in this case.

[0020] This obviously creates a big problem when the elevated hearing thresholds are compensated for by hearing aids fitted using current standard methods that rely on the assumption of linear hearing behavior assessed only by looking at hearing thresholds: with standard compensation of about half to two-thirds of the measured hearing threshold elevation, sounds are under-amplified at lower levels and at frequencies above about 2 kHz, and are over-amplified at higher levels.

[0021] The subject in Figure 4 obtained professionally fitted hearing aids but did not use them because they were perceived as not helping speech intelligibility at lower levels and, at somewhat higher levels, were perceived as amplifying the higher frequencies too much. The hearing test data shown in Figure 4 clearly supports the subjective analysis from the test subject and provides an objective explanation for the subjective evaluations and problems reported.

[0022] Typically, the dynamic range compression at frequencies where the hearing threshold is elevated is less than in the case shown in Figure 4, but all subjects show similar dynamic compression at frequencies where the hearing threshold is elevated. Looking at all the measurement data we collected, it is clear that the brain automatically compensates for the elevated hearing threshold at higher sound pressure levels when sounds exceed the hearing threshold and are therefore heard. The brain compensation can be very dramatic, as in Figure 4, or smaller, but always present to a significant degree. Figures 5, 6, and 7 show equal loudness test data from three more subjects with similar compression behavior. The compression measured at the higher frequencies where there is hearing loss is still very significant, but not as dramatic as in Figure 4.

[0023] The equal loudness test data shown in FIG. 7 was obtained from a subject who was professionally fitted with hearing aids and wore them daily. As can be seen, the person has significant hearing loss at higher frequencies and there is normal compression of the dynamic range. The equal loudness test data also reveals some abnormal hearing loss at higher levels of 50 and 60 phons in the low frequency range of about 150 Hz to 800 Hz. Standard audiometer testing and audiograms do not reveal this loss and do not provide information about the degree of dynamic compression at higher frequencies. The audiogram data shown in FIGS. 1 to 3 was obtained from the same subject as the lowest level equal loudness test data shown in FIG. 3. The complete equal loudness test data is shown in FIG. 7.

[0024] The equal loudness test data shown in Figure 8 is from the same subject as Figure 7, but the equal loudness test was performed while the subject was wearing professionally fitted hearing aids. A comparison between the traces in Figures 7 to 8 reveals that the hearing aids amplify increasingly higher frequencies, starting at about 1 kHz. The 20-fold trace in Figure 8 at 4 to 5 kHz when compared to the same trace in Figure 7 suggests approximately 25 dB of amplification at these frequencies.

[0025] The amplification of the hearing aid is too low at 20 phon, as seen by the gradual increase in levels for frequencies above 1 kHz; in this case, the compensation is about half the increase in the measured hearing threshold displayed in the audiogram of Figure 3. The amplification at the highest sound pressure level in the equal loudness test, 60 phon, is too high, as shown by the gradual downward slope of the trace at higher frequencies at that level. At sound pressure levels higher than 60 phon, a moderate sound pressure level experienced in real life, the downward slope increases even further and the high frequency amplification becomes very high.

[0026] If the hearing aid had restored the subject's hearing to normal, the equal loudness test data would look like the data from a person with normal hearing shown in Figure 2. Clearly, the equal loudness test data in Figures 2 and 8 are very different, and the hearing aid is not able to adequately compensate for the subject's experience of hearing loss. The subject gives the same feedback to his hearing aid experience as many other hearing aid users, namely, too little amplification of high frequencies at low levels, and too much amplification of high frequencies at high sound pressure levels. Hearing aids are still used routinely by people with significant hearing loss who find them helpful in restoring speech intelligibility, but with too much amplification at the higher frequencies of louder sounds. The equal loudness test data objectively describes the subjective experience.

[0027] FIG. 9 is again the equal loudness test data from the same subject as FIG. 7-8, but the hearing loss has been dynamically compensated for by the DSP portion of the novel hearing correction system using input data derived from the equal loudness test shown in FIG. 7. Comparing FIG. 9 to the equal loudness test data from the normal hearing individual shown in FIG. 2, it is clear that the correction is very good and the subject's hearing ability has been restored to essentially normal hearing. This is undoubtedly a major improvement over the results from the regular hearing aid shown in FIG. 8.

[0028] Figure 9 shows that abnormal hearing loss at higher levels, 50 and 60 phons, in the low frequency range between about 150 Hz and 800 Hz has been almost completely eliminated by the novel hearing correction system, leaving only a slight trace. High frequency compensation above 1 kHz is now linear, with the lowest 20 phon levels being appropriately amplified above 1 kHz, while at the same time the amplification at higher levels is reduced, resulting in a flat response at all levels equal to the results of a normal hearing individual. The traces are also evenly spaced by about 10 dB, eliminating dynamic compression.

[0029] Subjective responses of subjects to the experience of using the novel hearing correction system have been exceptionally positive, reported to be very natural sounding with no adverse effects from amplification. Speech intelligibility is significantly improved without drawbacks, and the sound experience has been likened to wearing glasses, everything is significantly clearer with more audible detail, and overall much better with no perceptible drawbacks.

[0030] The compensation used to obtain the equal loudness test data shown in Figure 9 uses six dynamic filters ranging from 200 Hz to 12.8 kHz. The center frequencies of the filters are 200 Hz, 800 Hz, 1.6 kHz, 3.2 kHz, 6.4 kHz, and 12.8 kHz. The number of filters required to compensate for hearing loss varies from individual to individual, and any number of filters from only one to several hundred is possible, but in practical applications, a range of one to twenty filters is preferred. Most normal compensation can be solved using one to eight filters. The center frequencies of the filters can of course also be changed to fit the compensation required. The center frequencies mentioned above are all multiples of the two missing 400 Hz, which was not required in this particular case. An equal spread of center frequencies is preferred, although not necessary, since the filters usually overlap to some extent, and the interaction between the amplifications of the filters needs to be taken into account when calculating the composite amplification from all the filters. An equal spread makes the calculation easier. The multiplication factor of two between the center frequencies is just one example of a useful and suitable number that can be changed to any other value.

[0031] The software algorithm in the hearing correction system uses the equal loudness test data shown in FIG. 7 to generate inputs to the DSP section, which performs compensation to achieve the equal loudness test data shown in FIG. 9. The maximum amplification level at the center frequency of each filter is determined, taking into account the total amplification from adjacent filter bands. The maximum amplification is also limited by the algorithm to avoid acoustic feedback if the target device is a hearing aid, where acoustic feedback may occur. In pure playback applications, where there is no microphone feedback path and therefore no risk of feedback, this is not necessary. The table generated by the software algorithm for each filter frequency is used by the DSP section to determine the dynamically changing amplification. FIG. 10 shows the filter table data output from the algorithm for each filter used to achieve the compensation of FIG.

[0032] In FIG. 10, the input sound pressure level is displayed in phons on the X-axis and the amplification in dB on the Y-axis. Looking at the graph showing the table data for 1600 Hz, it can be seen that the maximum amplification at 0 phons is slightly less than 20 dB. The amplification gradually decreases and reaches 0 dB at about 70 phons, i.e. there is no amplification. The input sound pressure level is measured individually in each filter band, i.e. if there is a pure tone at 1600 Hz, the amplification at 1600 Hz is mainly affected. At 0 phon, the table data for 3200 Hz is a maximum of 35 dB, which in this case is set to avoid acoustic feedback. It goes without saying that the 35 dB limit can be higher or lower depending on the feedback characteristics of the hearing aid used. This limit can also vary with frequency, but in this example it is set to 35 dB for all frequencies. The value of 35 dB is a reasonable value that can be achieved by an average modern hearing aid that can cancel feedback fairly. The 3200Hz amplification table data is at 0 dB at a slightly higher level than the 1600Hz table, just below 80 phons. The 6400Hz and 12800Hz table data are similar to 1600Hz and 3200Hz, despite showing higher amplification for the input sound pressure level. The 200Hz and 800Hz table data differ in that instead of gradually decreasing amplitude with increasing input sound pressure level, they increase amplification with increasing input sound pressure level. This is necessary to compensate for the hearing loss at the lower frequencies at the higher sound pressure levels shown in Figure 7.

[0033] A typical hearing aid implementation also requires a limiter applied to the DSP on the output to reduce the maximum sound pressure. The limiter keeps the reproduced sound pressure within safe limits, limiting levels to stay within the dynamic reproduction range of the output device, thus preventing the risk of hearing damage.

[0034] The table data for each frequency is calculated from the equal loudness test data shown in Figure 7. The input sound pressure is amplified so that the experienced level is fully compensated if applicable within the limits set by the acoustic feedback. Essentially, the table data is collected by drawing a vertical line at a particular center frequency in Figure 7 and creating the table data from the deviation between the experienced level and the actual level as shown along the vertical line for each input sound pressure level. Since the upper level of the measured equal loudness test data is limited, the table data for sound pressures that exceed the limits of the test must be extrapolated. One possible way to extrapolate the data is by polynomial fitting to the available test data, with first order polynomial fits being used in actual testing with good results. Obviously, there are numerous ways to extrapolate the extended data and the polynomial fit is just one example that has proven to be valid.

[0035] However, more accurate methods involving multiple steps are preferred. The acquired hearing data are somewhat irregular, fluctuating up and down, which can be seen by the wiggle movements on the traces in Figures 4 to 6, as an example. The wiggle movements are partly due to frequency deviations in the measurement setup, but are also related to human error and uncertainty in the test. The dynamic compression variations seen on the right at the higher frequencies of these figures show irregular compressions that vary up and down from a frequency- and level-dependent trend. If the test is repeated many times, these fluctuations will even out and disappear, but in practice this is not possible. This test procedure is mentally challenging for the tester, and after a while mental fatigue sets in, resulting in a decrease in accuracy. To regain accuracy, the tester needs significant breaks, and many repetitions of the test run the risk of extending the test period over several days, which is clearly undesirable and rather suboptimal.

[0036] One method that is useful in hearing restoration products is to maximize accuracy and make the best use of available data, and is described below. The method uses five steps. The first three steps can be understood by looking at the traces in Figure 14. Trace 1 connects four raw measured data points, one at each 10 dB increment starting at 40 dB and ending at 70 dB. The first step is to expand the number of sample points well beyond the four measured points to even out the variation in the measured data from a general underlying trend. There are many possible ways to do this mathematically, and two possible solutions are to use cubic spline data interpolation or linear first order polynomial interpolation between the raw data points, followed by multiple order linear phase FIR average filtering of the new interpolated data. The interpolation can produce the desired number of data points over a dynamic range of 40 dB to 70 dB, in this case 0.5 dB spacing between data points is used. Trace 2 in Figure 14 shows the expanded and filtered data using linear interpolation and linear phase FIR filtering.

[0037] The second step is to extend the dynamic range below the lowest measurement level, 40 dB. For this purpose, the first portion of the interpolated and filtered data, represented by trace 2, is used. A linear derivative fit is performed on the first derivative of the interpolated and filtered data. The straight line can then be used to extend the data below 40 dB. Trace 3 shows the straight line. Similarly, in the third step, the last portion of the interpolated and filtered data, represented by trace 2, is used to perform a linear derivative fit on the end derivative of the interpolated and filtered data. The resulting straight line, which can be used to extend the data points above 70 dB, is shown by trace 4.

[0038] FIG. 15 shows the overall dynamic range from 0 dB to 110 dB. As in FIG. 14, trace 1 is the measured data points, trace 2 is the expanded and filtered data using linear interpolation and linear phase FIR filtering, trace 3 is the fitted straight line for the lower level range, and trace 4 is the fitted upper level range. Looking at FIG. 15, it can be seen how trace 3 is used to expand the level range below 40 dB down to 0 dB, and how trace 4 is used to expand the range above 70 dB. It can also be seen that a simple first order polynomial fit to the measured data, i.e. a straight line, would not have produced accurate results in this case. The measured data is collected from a person suffering from what is commonly referred to as cookie bite hearing loss, measured at 400 Hz. It is somewhat unusual in that it requires less amplification at the lower levels, not just at the higher ones as is usually the case. The data is verified to be accurate over several tests, so the measurements, although surprising, are correct.

[0039] Figure 16 shows another example of results from the first three steps of the method. Data is measured at 2400 Hz between 40 dB and 70 dB at four levels. Again, trace 1 is the measured data points, trace 2 is the expanded and filtered data using linear interpolation and linear phase FIR filtering, trace 3 is the fitted lower level range line, and trace 4 is the fitted upper level range. Figure 17 shows the expanded dynamic range of this data from 0 dB to 110 dB.

[0040] FIG. 18 shows the fourth step of the method. The gain required to restore the hearing experience is measured in phon levels, which relate to human loudness experience and not to technical sound pressure levels measured in dBspl. Sound pressure levels are referenced to 20 μPa at all frequencies, i.e., a sound pressure of 20 μPa is equal to 0 dBspl at all frequencies. 1 dB phon is defined as equivalent to 1 dBspl at 1 kHz, where the equivalent sound pressure level varies greatly depending on frequency and sound pressure level. As an example, 20 dB phon at 100 Hz is equal to approximately 28 dBspl. To create gain table data useful in the DSP, the phon-related data must first be mathematically transposed to spl-related data. Trace 1 shows the transposed gain table data over the entire dynamic range from 0 dB to 110 dB based on the measurement and extraction information to which the extraction data below 40 dB and above 70 dB have been added as described above. The table data set is limited to a minimum gain of 0 dB and a maximum allowable gain. The maximum gain is normally set to an appropriate level to avoid acoustic feedback, but in this case no upper limit had to be applied since the maximum required amplification amounted to only about 34 dB.

[0041] After applying the gain limit, filtering of the table data is performed. The filtered table data is displayed by trace 2. The filter is a multi-order linear phase FIR average filter. The filter step is used to avoid abrupt gain changes that may cause audible artifacts such as clicks.

[0042] In the fifth and final step of the method, adjacent filter gain contributions are managed. As will be described later, the bandwidth of the filters cannot be very narrow, resulting in gain leakage between the filters. The stringent time domain requirements imposed on the filters demand that the filter bandwidths be wide enough to produce gains significantly greater than zero dB at adjacent frequencies. For example, consider filters at two adjacent frequencies, 1 kHz and 2 kHz. At each of these frequencies, 6 dB of amplification is required to restore the measured hearing loss. The 1 kHz filter is set to add a 1 kHz gain of +6 dB, and the other filter at 2 kHz is set to add a 2 kHz gain of +6 dB. Unfortunately, the 1 kHz filter produces a gain of +2 dB at 2 kHz, and the 2 kHz filter has a gain of +2 dB at 1 kHz. Simply setting the gain to +6 dB at each frequency would produce a total gain of +8 dB at both 1 kHz and 2 kHz, due to gain leakage from one filter to the other adjacent frequency. If the filters are normal second order parametric equalizers with a Q of 0.96 and center frequencies of 1 kHz and 2 kHz respectively, which provide +2 dB gain at adjacent frequencies as just described, then a gain setting of about 4.5 dB will produce the desired +6 dB gain at 1 kHz and 2 kHz.

[0043] Figure 19 shows a 3D graph with frequency on the left axis, sound pressure level on the right axis, and the amplification required on the height axis. The height axis displays the amplification required to restore the measured hearing loss at each frequency and sound pressure level. The dynamic gain behavior of the hearing restoration system with respect to sound pressure level and frequency can be studied in the diagram. There are seven separate boost frequencies in the diagram. Each boost is provided by the filter block of Figure 11 and the dynamic filter within the filter block. The filters that generate the gain at each of the frequencies in the diagram of Figure 19 must be wideband filters that span adjacent boost frequency bands, as mentioned above. Therefore, adjacent frequency filter boosts must be taken into account, otherwise the total gain will be significantly too large.

[0044] Figure 20 shows the sum of the gains from adjacent filters with realistic and reasonable bandwidths, without taking into account the total gain. Comparing the levels in Figure 20 with the desired levels in Figure 19, it becomes very clear that the total amplification produces much too much gain, over +90 dB in some areas that should have been around +35 dB. Clearly, it is very important to take into account the adjacent filter gain contributions when generating the gain table.

[0045] The table amplification data and filter center frequencies are imported from the software algorithm into the DSP section. The DSP section includes one or more filter blocks that provide dynamically varying amplification, each processing an individual frequency range. Figure 11 shows an example filter topology that contains, from the left, a bandpass filter that limits the frequency range before the level detector, a level detector that detects the instantaneous input sound pressure level, and a gain table that uses the table data from the algorithm and converts the current input level into a gain setting in the dynamic filter. The dynamic filter amplifies sounds within the frequency band of the filter block according to the gain table data input.

[0046] Filter blocks can, of course, be constructed using topologies other than that shown in FIG. 11, the example being just one of many possible implementations.

[0047] It is clear that means other than the gain table described above can be used to determine the sound pressure level dependent gain of the dynamic filter. It is easy to understand that a polynomial or some other type of mathematical expression can be used instead, and the gain table serves only as an example of a possible implementation approach. The gain table shown in FIG. 11 is replaced in such a case by the applied alternative method used to obtain the dynamic filter gain.

[0048] At each frequency that needs to be amplified in the hearing restoration process, a dynamic filter is needed. The dynamic filter is operating with a gain that changes dynamically depending on the input signal level that changes dynamically at the filter frequency. There are many basic filter topologies possible for dynamic filters. For example, IIR topology, FIR topology, etc., or a combination of both. Regardless of the type of filter, there is always a direct relationship between the filter bandwidth and the time domain response. A filter with a narrower bandwidth will have a poorer time domain response compared to a filter with a wider bandwidth. As the filter bandwidth decreases, there will be more and more ringing present at the filter output. Also, the output response after a transient input signal to the filter will be more sluggish and delayed by a narrower bandwidth filter than a wider bandwidth filter. The filter order also has a similar effect on the time domain response, with higher order filters resulting in better stopband attenuation or passband gain, but exhibiting poorer time domain behavior.

[0049] Human hearing is sensitive to time domain anomalies, and filters with large ringing and slow response times will clearly cause audible sound degradation. In products and systems aimed at improving hearing, sound degradation is clearly unacceptable, and therefore filters with sufficiently wide bandwidth are required. If the 1 kHz and 2 kHz filters mentioned above are narrow enough that each of them only gives 0.1 dB gain at adjacent frequencies, then they must have a Q of about 5.3 at 6 dB gain. This is a high Q that will cause large ringing on the filter output for more than 10 ms after an input signal step, which will clearly degrade the sound audibly. Therefore, filters with much lower Q values ​​are needed; filters with a Q below 1 behave much better, with the only significant drawback being that adjacent filter gain contributions must be compensated for.

[0050] As with the dynamic filter, the time domain behavior of the bandpass filter used before the level detector is important. The bandpass filter is needed to measure the signal level at the dynamic filter frequency and suppress audio signals present at other frequencies. A high Q filter output will not track the input signal well and the detected level will not be accurate due to filter output ringing and slow response. Therefore the bandpass filter must be low order, preferably second order, with a Q less than 1. Functionally, there is also an advantage if the bandpass filter and the dynamic filter have matching bandwidths.

[0051] It is not easy to build dynamic filters in the digital domain using standard building blocks while maintaining signal fidelity. The ubiquitously used infinite impulse response, IIR, filters and finite impulse response, FIR, filters do not behave predictably well with dynamically changing filter coefficients. Neither do any combinations of these filters. IIR filters operate through nested feedback paths with different delay lengths and coefficients used throughout for multiplication, regardless of the selected DSP implementation topology. For example, the common IIR biquadratic building block comprises five coefficients and four delay elements that hold data from the previous sample cycle. Any non-zero input signal to an IIR filter produces an impulse response on the output that decreases progressively with time. The tail time is by definition infinite, but in practice the output response will eventually fall into the noise floor, either through rounding of errors in the bit-depth limited precision of the DSP calculation, or through the actual noise floor present in the input signal. The length of the tailing signal depends on the selected filter order and Q, with higher order and higher Q filters having longer tails. FIR filters have similar delay elements that hold data from previous samples and a number of coefficients in their filter structure. They do not rely on nested feedback loops, but use a finite number of delay elements and coefficients defined by the filter length. The finite length of the filter, the number of samples, also limits the tailing response to samples of the same finite length. Although the finite tailing response seems beneficial, a significant drawback of FIR filters compared to IIRs is the latency they introduce. FIR filters always delay the output signal by a finite number of samples, which becomes very troublesome in real-time applications such as hearing restoration systems when the filters are long. Humans are sensitive to latency, and even a few tenths of a millisecond are clearly perceptible, and sounds that do not occur at the correct time in relation to the visual impression will result in a totally strange and confusing sensation that is unacceptable for the application.

[0052] The tail at the output of any filter after the input signal goes to zero is a measure of the energy stored in the delay elements within the filter. The filter models and formulas used to calculate the filter coefficients from the parameters assume that the filter structure is initially at a steady state of zero, which is reached only after processing perhaps hundreds of samples due to nested feedback paths with different delay lengths or filter lengths in the case of FIR filters after the input signal goes to zero. Whether it is an IIR or FIR filter, the energy stored within the filter is similar and the number of samples required to reach a steady state of zero within the filter is more or less the same.

[0053] In real-time hearing restoration applications, the filter never reaches a steady-state zero within itself, since there is always an input signal present. The update of the filter parameters occurs all the time while energy is present within the filter structure. If the filter parameters were updated every now and then, say once every 10 seconds, this would not be a big problem, but dynamic filters must be updated more frequently, potentially every sample. Updating the coefficients at a rate that is not far from or approaching the sample rate would result in significant distortion. The updated coefficients are multiplied with the old data stored in the filter, and as a result, the output from the filter will be erroneous for as long as it takes for the stored data to propagate through the filter structure. If the coefficients are updated every sample, the distortion would be very large.

[0054] Filters built on the digital integrator cascade technique developed by Hal Chamberlin are an alternative to IIR and FIR filters in dynamic applications. These filters are more suitable for dynamic updates with much less distortion, but when resonances are introduced, the filter response is less accurate. The basic integrator form of these filters can only produce a limited set of filters, and second-order filters with a Q value, i.e., resonances, are not as accurate. Thus, although they are commonly used in real-time dynamic applications such as gaming software, their characteristics make them less ideal for hearing restoration systems.

[0055] FIG. 21 shows a possible implementation of the dynamic filter block present in FIG. 11 using an alternative structure that does not have the drawbacks caused by the dynamic update of the filter coefficients mentioned above. This structure uses a fixed IIR filter, a fixed IIR boost filter, that adds gain at the desired boost frequency. This filter can obviously be replaced with another type of filter or combination of filters, the IIR filter being only a suitable example. The coefficients of the filter are statically set so that the maximum required boost is always provided by the filter. As an example, the gain table of FIG. 18 shows that a maximum gain of about 34 dB is required, and therefore in this case the filter gain is set to 34 dB. On the left side of FIG. 21 are the input signals to the dynamic filter structure, the audio signal input and the gain table input. The audio signal input is obviously the audio input signal, and the gain table input is the gain control input to the dynamic filter. The gain control input signal varies between 0 and 1 depending on the dynamically changing gain required. When the signal is 1, the dynamic filter produces a maximum gain at the output, and when it is 0, the dynamic filter passes the audio signal without any gain being applied. The gain control signal is fed to multiplier 1 which multiplies the audio input signal with the gain control signal. The gain control signal is also fed to multiplier 2 after being subtracted from a constant value of 1. The gain control signal after subtraction varies between 1 and 0, i.e. when the gain control signal to multiplier 1 is 1 the signal to multiplier 2 is 0 and when the gain control signal to multiplier 1 is 0 the signal to multiplier 2 is 1. The audio signal output from multiplier 1 is fed to a fixed IIR boost filter which applies a maximum gain and the output from the fixed IIR boost filter is fed to an adder, Add, which sums the signal with the output from multiplier 2. The sum of the audio signal output from multiplier 1 and multiplier 2 will always be the same as the input audio signal but the ratio between what comes out of multiplier 1 and multiplier 2 varies depending on the gain control input signal to the dynamic filter. By changing the gain control input signal to the dynamic filter a different mix of the unmodified but gain controlled output from multiplier 2 is added together with the boosted output from multiplier 1.By varying the signal ratio via the gain control input signal to the dynamic filter, the boost provided by the dynamic filter can be varied from zero to maximum boost. This is achieved through a simple change in the mix of the two audio signals, without adding any distortion whatsoever.

[0056] Often, DSP sections use more than one filter block, and there are two obvious ways to combine the filter blocks: serial or parallel topology. Figure 12 shows an example of a serial topology, and Figure 13 reveals a parallel topology. Of course, many combined filter block configurations can be constructed using topologies other than those shown in Figures 12-13.

[0057] In the example of the hearing correction system described, where the amplification increases with level, there are two frequencies, 200Hz and 800Hz, but in most cases the amplification decreases as the level increases. This dynamic reduction of amplification at higher input levels provides a clear advantage in hearing aids and similar applications, where there is always a risk of acoustic feedback. If there is feedback and the level at the feedback frequency increases, the amplification is automatically reduced, thus reducing the feedback and limiting it to a low level. In contrast, a fixed gain system would only increase the feedback until it reaches the maximum output from the system, which is obviously very unpleasant and needs to be actively prevented and avoided.

[0058] All modern hearing aids use various types of feedback reduction technology. Aggressive feedback reduction systems run the risk of introducing auditory irregularities and degrading the sound quality of the hearing aid. The more aggressive, the greater the risk. The automatic reduction of gain at higher levels in the new hearing correction system puts less strain on the feedback reduction system, which does not need to be as aggressive as required in systems with constant gain.

[0059] The described hearing correction system is not only applicable to hearing aids, but can be used with any sound reproduction system. Such a system can use any type of speaker or headphones. The hearing correction system will function properly if the output level at the listener's ear is known, i.e. if a certain sound reproduction system output produces a known sound pressure level experienced by the listener. Most applications of hearing correction systems utilize some form of real-time digital signal processing for real-time applications. However, it is entirely possible to pre-process audio material with the compensation of the hearing correction system, thereby creating a library of personally compensated audio.

[0060] Digital signal processing can be implemented in many ways, from pure hardware implementations to pure software / firmware or a mix of the two. The DSP functions in the described examples use code written for a digital signal processor. The described hearing tests and algorithms that generate the input data to the DSP section are implemented in software running on a personal computer. The software can of course be implemented to run on any computing system, such as a phone, tablet, or other device. It can also be implemented on a dedicated target system similar to an audiometer for the new hearing correction system.

[0061] The digital signal processing part of the system is typically physically implemented in a hearing aid or headphones in real-time applications. It can also be implemented in a phone, tablet, TV set, headphone amplifier, or computer to improve call quality and sound reproduction, generally using any regular headphones. It can also be implemented in sound reproduction systems in cars, where the relationship between the system output and the sound pressure at the listener's ears is predictable due to the generally fixed physical position of the people in the car.

[0062] All these configurations are provided only as examples; obviously many other possible implementation scenarios can be envisaged.

[0063] EMBODIMENTS OF THE PRESENTINVENTION Several embodiments of the present invention are provided and further described below.

[0064] According to one embodiment, the one or more dynamic filters are dynamically changed based on the input signal and at least one parameter, preferably the one or more dynamic filters change amplification in response to changes in sound pressure.

[0065] According to one embodiment, the input sound pressure is amplified such that the experienced sound level is fully compensated.

[0066] Moreover, according to yet another embodiment, the input sound pressure is amplified such that the experienced sound level is fully compensated within the limits set by the acoustic feedback.

[0067] Furthermore, according to one embodiment, the method also comprises a step of processing and compensating the data to provide more or less equal loudness at each used frequency at different sound levels.

[0068] According to yet another embodiment, multiple dynamic filters are involved in compensating the data, and more preferably, between 2 and 20 dynamic filters are involved in compensating the data.

[0069] As will be appreciated from the above, suitably the method includes digital signal processing. Furthermore, according to yet another embodiment, the method involves determining a dynamically varying amplification for each filter frequency. This step may involve increasing and / or decreasing the amplification depending on the measured data and the input sound pressure level.

[0070] In accordance with yet another embodiment of the present invention, the method involves determining a dynamically varying amplification for each filter frequency.

[0071] Further, according to one embodiment, the table data is collected by drawing a vertical line at a particular center frequency, preferably using amplification of the input signal at a particular frequency with different input sound pressures to create the table data from the deviation between the experienced level and the actual level.

[0072] Furthermore, according to one embodiment, the method includes digital signal processing of the amplified data and filter center frequencies, preferably by involving one or more filter blocks.

[0073] Tests according to the invention may be used to identify frequencies / frequency ranges that need to be adjusted, and the filter is then used to adjust these. The center frequency may be fixed or may be adjusted based on measured data.

[0074] According to yet another embodiment of the present invention, the digital signal processing includes one or more filter blocks, preferably each filter block processing an individual frequency.

[0075] Moreover, according to another embodiment, each dynamic filter involved has a center frequency in the range of 200 Hz to 12.8 kHz.

[0076] Furthermore, according to one embodiment, the one or more dynamic filters are low order, preferably second order, wideband filters, where the method involves compensation of adjacent frequency filter boost to ensure control of the total gain obtained.

[0077] Further, according to another embodiment, a dynamic filter is provided for each frequency that needs to be amplified in the hearing restoration process.Further, according to one embodiment, the one or more dynamic filters are operating with a dynamically varying gain depending on the dynamically varying input signal level at a filter frequency.

[0078] According to another embodiment of the present invention, a maximum amplification is determined for one or more frequencies, preferably for the center frequency of each dynamic filter, and more preferably also taking into account the total amplification from adjacent filter bands.

[0079] Each filter has a certain bandwidth, and the bandwidths of different filters may overlap. For example, if filter 1 has a center frequency of 1 kHz and filter 2 has a center frequency of 2 kHz, filter 1 can contribute some amplification to the frequency at 2 kHz, even though this is a much lower amplification than at 1 kHz. In this regard, according to the invention, a total amplification may be required.

[0080] Moreover, in accordance with yet another embodiment of the present invention, the maximum amplification is limited to avoid acoustic feedback.

[0081] The filter block may comprise different components according to the invention. According to one embodiment, said one or more filter blocks comprise a bandpass filter, a sound pressure detector and a dynamic filter. According to one embodiment, the bandpass filter is configured to filter out and the detector is configured to measure the signal level at the dynamic filter frequency and to suppress sound signals present at other frequencies.

[0082] Furthermore, according to another embodiment, the bandpass filter has a low order, preferably second order, and more preferably a Q less than one.

[0083] Furthermore, the one or more filter blocks may comprise a gain table. A gain table may be implemented as one possible way to convert from sound pressure to gain. However, it should be noted that other alternatives are possible according to the invention. Furthermore, according to another embodiment, multiple filter blocks are used. These may be arranged in series or parallel or a combination thereof.

[0084] Furthermore, according to another embodiment, the method comprises transposing phong-related data into spl-related data.

[0085] Furthermore, as illustrated above in relation to Figs. 14-18, the method according to the invention may also involve a numerical method for processing smart data outside the measurement area. Along these lines, according to one embodiment, the method involves extending the model beyond the obtained measured test data points by providing a curve fit between at least a plurality of measured test data points, then providing an interpolation of the obtained curve fit, then providing a derivative curve fit of the interpolation, preferably a first derivative curve fit at a relatively low sound pressure level of the interpolation and a second derivative curve fit at a relatively high sound pressure level of the interpolation. Furthermore, according to another embodiment, the method involves a filtering step for removing data points of the obtained measured test data points that are outside the relationship of the obtained measured test data points, in order to enable the provision of a curve fit between at least a plurality of measured test data points.

[0086] According to yet another embodiment, the present invention relates to a hearing correction system, comprising: - a hearing test configured to perform a hearing test on a subject at different frequencies and sound pressure levels, said hearing test involving the use of reference sounds and test sounds to obtain data on test sound levels and reference sounds for different frequencies and sound levels, and adjusting the test sound levels until the subject perceives the test sound levels as more or less equal to the loudness of the reference sounds, a data processing unit configured to process the acquired data and provide settings to a digital signal processing unit; a digital signal processing unit configured to perform dynamic compensation, Here, one or more dynamic filters are involved for dynamic compensation of the data.

[0087] It should be noted that the hearing correction system according to the present invention may be implemented as a software unit in any type of suitable hardware device, such as a mobile phone, a PC or other computer unit, a tablet, etc. Furthermore, personal sound units, such as personal sound amplifier products, are also so embodied by the present invention.

[0088] Furthermore, the digital signal processing unit may be based on software, hardware or a combination thereof.

[0089] Furthermore, according to yet another embodiment of the present invention, there is provided a system comprising the hearing correction system according to the present invention and a hearing aid unit, headphones or a sound reproduction system, an example of which is a loudspeaker.

[0090] Furthermore, the present invention is directed to a method for hearing assessment and / or correction, said method comprising the steps of: - administering a hearing test to a subject at different frequencies and different sound levels, said hearing test involving the use of reference sounds and test sounds and adjusting the test sound levels until the subject perceives the test sound levels as more or less equal in loudness to the reference sounds; obtaining data on test sound levels and reference sounds for different frequencies at different sound levels, Here, one or more dynamic filters are involved to compensate the data.

[0091] Furthermore, all embodiments presented above with respect to the device according to the invention should also be considered as possible embodiments with respect to the method for hearing assessment and / or correction according to the invention.

Claims

1. A device for hearing assessment and / or correction, comprising: The device comprises a software unit configured to execute a method, the method comprising: - administering a hearing test to a subject at different frequencies and different sound levels, said hearing test involving the use of reference sounds and test sounds and adjusting the test sound levels until the subject perceives the test sound levels as more or less equal to the loudness of the reference sounds, - obtaining data for said test sound levels and said reference sounds for different frequencies at different sound levels, The device with one or more dynamic filters to compensate for the data.

2. 2. The device of claim 1, wherein the one or more dynamic filters are dynamically changed based on an input signal and at least one parameter, preferably the one or more dynamic filters change amplification in response to changes in sound pressure.

3. 3. A device as claimed in claim 1 or 2, wherein the input sound pressure is amplified such that the experienced sound level is fully compensated.

4. 4. The device of claim 3, wherein the input sound pressure is amplified such that the experienced sound level is fully compensated within the limits set by the acoustic feedback.

5. The method comprises: A device according to any one of claims 1 to 4, further comprising a step of processing and compensating the data in order to provide a more or less equal loudness at each used frequency at different sound levels.

6. 6. The device according to claim 1, wherein a plurality of dynamic filters are involved to compensate the data, preferably between 2 and 20 dynamic filters are involved to compensate the data.

7. A device according to claim 1 , wherein the method comprises digital signal processing.

8. 8. The device of claim 6 or 7, wherein the method involves determining a dynamically varying amplification for each filter frequency.

9. A device as claimed in any one of claims 6 to 8, wherein the table data is collected by drawing a vertical line at a particular centre frequency, preferably using amplification of the input signal at a particular frequency at different input sound pressures to create the table data from the deviation between the experienced level and the actual level.

10. 10. A device according to any of the preceding claims, wherein the method comprises digital signal processing of the amplified data and filter centre frequencies, preferably by involving one or more filter blocks.

11. 11. A device according to any one of claims 7 to 10, wherein the digital signal processing comprises one or more filter blocks, preferably each filter block processing an individual frequency.

12. 12. A device according to any preceding claim, wherein each associated dynamic filter has a centre frequency in the range 200Hz to 12.8kHz.

13. 13. A device according to any one of claims 6 to 12, wherein a maximum amplification is determined for one or more frequencies, preferably for a centre frequency of each dynamic filter, more preferably also taking into account the total amplification from adjacent filter bands.

14. 14. A device according to any one of claims 1 to 13, wherein maximum amplification is limited to avoid acoustic feedback.

15. 15. A device according to any one of claims 1 to 14, wherein the one or more dynamic filters are low order, preferably second order, wideband filters, and the method involves compensation of adjacent frequency filter boost to ensure control of the total gain obtained.

16. 16. A device according to any one of claims 1 to 15, wherein a dynamic filter is provided for each frequency that needs to be amplified within the hearing restoration process.

17. 17. The device of claim 1, wherein the one or more dynamic filters operate with dynamically varying gain dependent on dynamically varying input signal levels at particular filter frequencies.

18. 18. The device of claim 10, wherein the one or more filter blocks comprise a bandpass filter, a sound pressure detector, and a dynamic filter.

19. 20. The device of claim 18, wherein the bandpass filter is configured to filter out and the detector is configured to measure the signal level at a dynamic filter frequency and suppress sound signals present at other frequencies.

20. 20. A device according to claim 18 or 19, wherein the bandpass filter has a low order, preferably second order, and more preferably has a Q less than 1.

21. The device of claim 20 , wherein the one or more filter blocks further comprise a gain table.

22. 22. The device of claim 1, wherein the method comprises replacing phone-related data with spl-related data.

23. 23. A device according to any one of claims 10 to 22, in which multiple filter blocks are used.

24. 24. A device according to any one of claims 1 to 23, wherein the method involves extending a model beyond the obtained measured test data points by providing a fitted curve between at least a plurality of measured test data points, then providing an interpolation of the obtained fitted curve, then providing a derivative fitted curve of the interpolation, preferably a first derivative fitted curve at a relatively low sound pressure level of the interpolation and a second derivative fitted curve at a relatively high sound pressure level of the interpolation.

25. 25. The device of claim 24, wherein the method involves a filtering step to remove data points of the acquired measured test data points that are outside a relationship of the acquired measured test data points to enable provision of a curve fit between at least a plurality of the measured test data points.

26. 1. A hearing correction system, comprising: a hearing test configured to perform a hearing test on a subject at different frequencies and sound pressure levels, said hearing test comprising using said reference sounds and test sounds to obtain data on test sound levels and reference sounds for said different frequencies and sound levels, and adjusting said test sound levels until said subject perceives the test sound levels as more or less equal to the loudness of said reference sounds, a data processing unit configured to process the obtained data and provide settings to a digital signal processing unit; said digital signal processing unit configured to perform dynamic compensation, wherein one or more dynamic filters are involved in said dynamic compensation of said data.

27. 27. The system of claim 26, wherein the digital signal processing unit comprises software, hardware, or a combination thereof.

28. A system comprising a hearing correction system according to claim 26 or 27 and a hearing aid unit, a headphone or a sound reproduction system.

29. 1. A method for hearing assessment and / or correction, comprising: - administering a hearing test to a subject at different frequencies and different sound levels, said hearing test involving the use of reference sounds and test sounds and adjusting the test sound levels until the subject perceives the test sound levels as more or less equal to the loudness of the reference sounds, - obtaining data for said test sound levels and said reference sounds for different frequencies at different sound levels, A method in which one or more dynamic filters are involved to compensate the data.

Images