Patent Grant
8165319
This invention relates to a method and system for reproducing an audio signal.
The entire contents of published international patent applications PCT/AU02/00852 (WO03/003790) and PCT/AU03/00301 (WO03/077236) attributed to the current inventor are incorporated herein by reference.
Earphones are contained within devices such as headsets, headphones, handsets, earbuds and inset earphones and have the potential to produce sound levels that can harm or cause discomfort to the listener of these devices. Harm such as the loss of hearing sensitivity can occur as a result of either excessive short-term exposure or long-term exposure to sound. Other hearing dysfunctions that may result from excessive exposure to sound include tinnitus, reduced speech understanding, hyperacusis and ear pain, the later two in particular have been observed to result from short-term exposure. Short-term exposure which is perceived by the listener to be loud and abrupt may result in symptoms affecting other parts of the body such as pain/ache within the head and/or neck. Injury resulting from short-term exposure to sound, which is perceived as being both loud and abrupt, has been described as an acoustic shock injury.
To reduce the occurrence and severity of injury to the listener methods of limiting the short-term and long-term sound exposure have been developed. These include the suppression of sounds known to cause injury, known as shriek rejection as well as broadband and frequency specific level control with a variety of response times. Methods of monitoring and recording the short-term and long-term exposure of a listener have also been developed. Devices have been developed to control the long-term sound exposure of a listener based on an estimate of the long-term exposure. One device predicts the future long-term sound exposure from past estimates of sound exposure combined with data on the anticipated use which it uses to control the current amplification of the signal.
Many of the injuries to users of earphones have resulted from short-term exposure and therefore long-term level control and recording offers nothing in the prevention of this injury or furthers the understanding of it.
In a first aspect the present invention provides a method of reproducing an audio signal by way of an audio system which includes an earphone, the method including the steps of: receiving an audio signal; estimating short-term characteristics of the acoustic exposure of a listener; recording the short-term characteristics; controlling the audio signal and; outputting the controlled audio signal for reproduction.
The short-term characteristics may be estimated based on characteristics of the audio system.
The short-term characteristics may be estimated based on characteristics of the controlled audio signal.
The short-term characteristics may be estimated based on characteristics of the received audio signal.
The short-term characteristics may include the short-term level.
The short-term characteristics may include the maximum of the short-term level within a specified time period.
The short-term characteristics may include characteristics which are frequency specific.
The short-term characteristics may include the time at which the maximum occurred.
The short-term characteristics may include the duration over which the short-term level exceeds a predetermined fraction of the maximum short-term level.
The short-term characteristics may include the abruptness of the maximum of the short-term level.
The abruptness may be determined by calculating the difference in the time between the time of the maximum and the preceding time in which the short-term level is below the maximum by a predetermined amount.
The short-term characteristics may include an identification code for the signal that produced the maximum short-term level.
The identification code may be determined to be a code associated with predefined characteristics.
The predefined characteristics may include the spectral content.
The predefined characteristics may include the temporal content.
In a second aspect the present invention provides a system for reproducing an audio signal produced by an audio system which includes an earphone, the system including: receiving means for receiving an audio signal; estimating means for estimating short-term characteristics of the acoustic exposure of a listener; recording means for recording the short-term characteristics; control means for controlling the audio signal and; outputting means for outputting the controlled audio signal for reproduction.
The system may further include identification means for producing an identification code representative of a particular type of received signal.
In a third aspect the present invention provides a computer software program providing instructions for controlling a computing system to carry out a method according to the first aspect of the invention.
In a fourth aspect the present invention provides a computer readable medium providing a computer software program according to the third aspect of the invention.
By recording details of the short-term characteristics it is possible to later analyse the cause or extent of an acoustic exposure incident. Further, this is achieved without the need to record the actual signal itself. By recording only characteristics of the signal the amount of data that needs to be recorded may be reduced.
Embodiments of the present invention will now be described, by way of example only, with reference to the accompanying drawings, in which:
Referring to
The system further includes estimating means for estimating characteristics of the acoustic exposure of a listener in the form of protected exposure analysis 24. The system further includes recording means for recording characteristics of the acoustic exposure in the form of storage 26. The system further includes control means for controlling the audio signal in the form of analysis and exposure control 22.
The system 10 or any part of it may be performed in the analog domain with the appropriate conversions between the domains, these conversions are not shown in
In operation, system 10 receives a signal at input 12. Analysis and exposure control 22 operates to control the level of the input signal 12 to produce a controlled output signal 14. This includes assessing the level of the input signal 12. This process further includes frequency weighting the input signal to produce an estimate of the level at the ear reference point ERP, the eardrum point, DRP or another point. The frequency weightings for a specific audio reproduction system including the earphone are obtained from measurements and are stored within system 10 as reproduction characteristics 28. Additional weighting for translation from the ear to the field and standard acoustic weightings such as A, B and C are optionally included in these characteristics. The short-term level is assessed on a broadband and frequency specific basis. The process of obtaining the short-term level includes squaring the signal and passing it through a filter with a low pass characteristic. The short-term level may then be compared to the exposure limits 30. If it exceeds these then the control process is such that it reduces the input signal by an amount at least equal to the amount by which the short-term level exceeds the exposure limits in dB or by the ratio of the short-term level to the exposure limits in linear terms to produce the output signal 14. Optionally, other processes may be applied to the input signal 12 within the analysis and exposure control process 22 such as fixed or adaptive filtering or gain control.
The protected exposure analysis 24 operates to analyse a signal to produce characteristics of the acoustic exposure of a listener in the form of exposure data 32. The signal to be analysed by the analyser 24 comes from the output signal, 14. The protected exposure analysis 24 receives the reproduction characteristics 28 to produce exposure data that is appropriate for the specific audio reproduction system 20 and earphone 18.
The reproduction characteristics 28 include frequency weightings for each specific audio reproduction system including the earphone are obtained from measurements and are stored within system 10 as reproduction characteristics 28. Additional weightings for translation from the ear to the field and standard acoustic weightings such as A, B and C are optionally included in these characteristics.
The protected exposure analysis 24 includes short-term level assessment on a broadband and frequency specific basis. The process of obtaining the short-term level includes squaring the signal and passing it through a filter with a low pass characteristic, such as a 1st order filter with a 125 millisecond time constant, RMS ‘Fast’ or ‘F’ as defined in the standard IEC 60651. The maximum of the short-term level is taken over a given analysis period. The protected exposure analysis 24 also records the time at which the maximum short-term level occurred within a given analysis period and includes this time within the exposure data 32 it produces.
In addition to the above the protected exposure analysis 24 analyses the characteristics of the signal at the time at which the maximum short-term level occurred and produces measures of the abruptness of the signal, the persistence of the signal and the character identification of the signal within a given analysis period and includes this within the exposure data 32 it produces.
The exposure data 32 produced by the protected exposure analysis 24 at the end of each given analysis period such as every half hour includes:
maximum short-term level (broadband & frequency specific)
time of the maximum short-term level (broadband & frequency specific)
abruptness of the signal that produced the maximum short-term level (broadband & frequency specific)
persistence of the signal that produced the maximum short-term level (broadband & frequency specific)
character identification code of the signal that produced the maximum short-term level (broadband & frequency specific)
also included in this data is the
maximum peak level
the time of the maximum peak level
the character identification code of the signal that produced the maximum peak level
broadband long-term exposure over the analysis period
analysis time
Storage 26 receives the exposure data 32 at the end of every analysis period which it stores in memory until the exposure data 32 is required. The storage process involves compressing the exposure data which includes conversion of linear power levels to decibels. The recorded data is available at the recorded data 16 output and is available for display 34 or storage in a database.
An example of the format of the recorded data is as follows:
An example of the data range and resolution in the above format is as follows:
Referring now to
The band signals 102 are squared by 103 to produce power signals 104 for each of the bands. The power signals 104 are weighted (multiplied) by frequency weights 105 to produce the frequency weighted power signals 106. The frequency weights represent the relationship between the digital signal level and the acoustic signal level produced by the earphone. These are the reproduction characteristics 28 and include any additional weights. The acoustic signal level is measured in a specific coupler or ear simulator such as those described in ITU-T Recommendation P57. Additional weights include A, B and C sound level weights and weights to translate the earphone measures at the ear (such as those taken at the eardrum reference point, DRP) to the field. Particular frequency weights are often associated with particular time weights in acoustic measurement. Many standards specify A weighting for long-term exposure in the field and C weighting for peak measures in the field. This embodiment applies the following three (N) additional weightings to the weights obtained for the digital to acoustic transfer function of the specific audio reproduction system including the earphone to produce a set of three frequency weighting functions:
1. A-weighting plus translation to the field for the broadband long-term average sound level,
2. C-weighting plus translation to the field for peak sound level,
3. A-weighting for broadband short-term average sound level and the narrow band (frequency specific) short-term average sound level.
The frequency weighted power signals 106 are summed by 107 to produce the broadband frequency weighted power signals 108. In an alternative embodiment of this method the broadband frequency weighted power signals 108 are obtained by squaring the output of filters applied to the input signal 100, these filters having the same magnitude response as the above frequency weighted digital to acoustic transfer functions.
The broadband frequency weighted power signals 108 are applied to time weighting operations 109 to produce the broadband long-term average 110 and the broadband short-term average 111 sound level estimations.
The broadband long-term average 110 is obtained by low pass filtering the appropriate frequency weighted broadband power signal 108. In this embodiment the low pass filter is a 1st order infinite impulse response filter with an exponential integration time constant in the order of many minutes. The broadband short-term average 111 is obtained by low pass filtering the appropriate frequency weighted broadband power signal 108. In this embodiment the low pass filter is a 1st order infinite impulse response filter with an exponential integration time constant of 125 milliseconds corresponding to the ‘Fast’ or ‘F’ integration time constant specified for sound level meters in the standard IEC 60651. Other filters and time constants may be employed.
The narrow band short-term averages 112 are obtained by applying low pass filtering to each of the appropriate frequency weighted power signals 106. The filtering is the same as that described for the broadband short-term average.
The real time clock 114 produces a time code 115. The time code is applied to timer 116 which produces an update command 117 at predefined time intervals, these being the analysis periods. In this embodiment the analysis period is 30 minutes however the period depends on the application. A trade off exists between the time resolution of the data and the amount of storage required to accommodate it.
The generation of detailed exposure data is performed by 118. The maximum levels 119, the times of maximum levels 120, the abruptness of maximum levels 121 the persistence of maximum levels 122, the analysis time 123, the broadband long-term exposure 124 and the identification code of the maximum level signals 125 are produced at the end of each predefined analysis period.
The maximum level (Peak) 119 is the maximum peak level of the appropriate frequency weighted broadband power signal 108 over the analysis period. The maximum level (BB) 119 is the maximum value of the broadband short-term average sound level 111 over the analysis period. The maximum levels (Band[k]) 119 are the maximum values of the narrow-band short-term sound levels 112 over the analysis period.
The time of maximum level (Peak) 120 is the sampled real-time clock value at the time at which the maximum of the appropriate frequency weighted broadband power signal 108 occurred during the analysis period. The time of the maximum level (BB) 120 is the sampled real-time clock value at the time at which the maximum of the broadband short-term average sound level 111 occurred. The times of the maximum level (Band[k]) 120 are the sampled real-time clock values at the times at which the maximum of the narrow-band short-term average sound levels 112 occurred.
The abruptness of the maximum levels 121 for the broadband and narrow-band short-term average sound levels are obtained as follows. The short-term average sound levels are sampled at periodic intervals and placed into circular buffers. In this embodiment this occur every 8 milliseconds. When a maximum of a short-term average sound level occurs its respective buffer contents is analysed in reverse order starting at the time of the maximum. The number of samples from the time of the maximum is counted until the short-term average sound level falls below the maximum level by a predefined factor. In this embodiment the factor is set to be 0.1. The number of samples counted multiplied by the sampling time interval is the abruptness rating in seconds.
The persistence of the maximum levels 122 for the broadband and narrow-band short-term average sound levels are obtained as follows. When a maximum short-term average sound level occurs the real-time clock value is sampled and saved as the start time. The short-term sound level is monitored and the real-time clock value is sampled again when the short-term sound level falls below the maximum by a predefined factor, this is the stop time. In this embodiment this factor is set to 0.1. The persistence is the difference in time between the start and the stop times.
The broadband long-term exposure 124 is obtained by accumulating the appropriate frequency weighted broadband power signal 108 over the analysis period and scaling it by the inverse of the product of the analysis period and the sampling rate.
Identification codes of the signals producing the maximum levels 125 for the peak, broadband and narrow-band short-term sound levels are obtained as follows. When a maximum sound level occurs a match request command 126 is issued to identification means in the form of analysis and characteristic matching process 127. The analysis and characteristic matching process contains a circular buffer which receives samples of the input signal 100. When a match request is received a predefined number of samples representing the signal over a predefined period prior to the match request being received are copied from the circular buffer into an analysis buffer. The analysis buffer then fills with a predefined number of samples received from the input following the match request. The contents of the analysis buffer is analysed and its characteristics are extracted. Those skilled in the art will be aware of many techniques available to analyse a signal and determine its character. This embodiment uses frequency analysis to obtain detailed spectral characteristics. The characteristics are compared with predefined reference characteristics 128 and the best match is determined. An identification code for the maximum level signal 125 corresponding to the reference characteristic which yields the best match is generated. In telecommunications there are many non speech signals that have known characteristics, such as service tones, DTMF tones, fax machine tones and so forth which may be identified and for which an identification code can be produced.
At the end of an analysis period as defined by timer 116 the time from the real time clock 114 output is sampled by exposure data generation 118 to produce the analysis time 123. The update command 117 issued at the end of analysis period provides a request to the storage 26 to store the current exposure data. The update command 117 then resets all the exposure data values within the exposure data generation 118 and the analysis and characteristic matching 127 to zero.
Referring to
The frequency analysis 301 splits the signal into a number (K) of frequency bands. In this embodiment the centre frequencies of the filters are linearly spaced and the bandwidths of the filters are constant. In another embodiment, the centre frequencies are logarithmically spaced and the bandwidths of the filters are third octave. In another embodiment, the filter centre frequencies and bandwidths are modelled on the human ear. Those skilled in the art will be aware of many techniques to achieve separation of the signal into a number of frequency bands including IIR filter banks, FIR filter banks, wavelets and discrete Fourier analysis.
The frequency synthesis 302 reconstructs the output signal from the (K) frequency bands of the output of the analysis and exposure control process 22. The method of reconstruction matches the method of frequency analysis performed by the frequency analysis 301.
Due to signals in this embodiment being in the frequency domain (other than the input signal and the output signal) frequency analysis is no longer required within the analysis and exposure control 22 and the protected exposure analysis 24. Referring again to detailed schematic of the protected exposure analysis 24 in
Referring to
Referring to
Referring to
Referring again to
The protective gains 604 are provided as control inputs to the exposure control 602. In this embodiment the exposure control 602 combines the peak, the broadband long-term average, and the broadband short-term average protective gains into a single broadband gain by taking the minimum of them. This single broadband gain is then combined with each of the narrow band short-term gains by taking the minimum of each narrow band short-term gain and the single broadband gain to produce a set of K multi band protective gains. In this embodiment the exposure control is multi band, the input signal is split into K frequency bands (frequency analysis) and modified (multiplied) by the K multi band protective gains and recombined (frequency synthesis) to produce the exposure controlled output signal 14. In another embodiment the minimum of all the gains, broad and narrow band is taken to produce a single broadband gain for a single band exposure control operation. The input signal 605 to exposure control 602 is a delayed version of the input signal 12 to the system, the delay is provided by 606. This delay is needed to compensate for the time delay introduced by the unprotected exposure analysis 401.
The protective exposure calculator 601 is similar to the protective exposure analysis 24 previously described. It differs in the following ways. It creates a set of protected sound levels by multiplying the sound levels 129 from the unprotected exposure analysis 401 by the corresponding set of protective gains 604. Referring now to
Referring to
Referring to
Referring to
In the foregoing description an earphone is intended to refer to any electro-acoustic transducer for converting electric signals into sounds which can be held over or inserted into the ear. An audio system is intended to refer to any sound reproduction system that reproduces sounds by way of an earphone such as telephone headsets or handsets, personal music players, mobile telephones, two way radios and the like.
The above described embodiments are meant to be illustrative and not limiting. It will be obvious to those skilled in the art that variations and modifications may be made without departing from the spirit and scope of the invention as defined by the appended claims.
| Number | Date | Country | Kind |
|---|---|---|---|
| 2005902653 | May 2005 | AU | national |
| Filing Document | Filing Date | Country | Kind | 371c Date |
|---|---|---|---|---|
| PCT/AU2006/000696 | 5/25/2006 | WO | 00 | 12/19/2007 |
| Publishing Document | Publishing Date | Country | Kind |
|---|---|---|---|
| WO2006/125265 | 11/30/2006 | WO | A |
| Number | Name | Date | Kind |
|---|---|---|---|
| 5355418 | Kelsey et al. | Oct 1994 | A |
| 5448620 | Gershkovich et al. | Sep 1995 | A |
| 20040234079 | Schneider et al. | Nov 2004 | A1 |
| 20050135631 | Yoshino | Jun 2005 | A1 |
| Number | Date | Country |
|---|---|---|
| WO-03003570 | Jan 2003 | WO |
| WO-03003790 | Jan 2003 | WO |
| WO-03077236 | Sep 2003 | WO |
| Number | Date | Country | |
|---|---|---|---|
| 20080199019 A1 | Aug 2008 | US |
