Patent Grant
8838259
The present document relates to media players, such as portable electronic devices, vehicle audio systems, home stereo systems, etc. For example, it relates to the management of the sound pressure level generated by portable electronic devices.
Mobile media players have emerged as one preferred platform for listening to music. Music playback has become a feature of most mobile phones as well. While the exposure to occupational noise has decreased in recent years due in part to workplace legislation, the exposure to so called “social noise”—including music—has increased drastically. Music listening can become a health risk if a user chooses to listen to music for longer periods of time at high audio volume levels, which studies suggest may lead to hearing impairments like loss of hearing sensitivity, disability to separate different sounds or tinnitus.
The present disclosure is explained below in an exemplary manner with reference to the accompanying drawings, wherein
a illustrates exemplary graphs of the sound pressure level sensitivity for human listeners, also referred to as the equal-loudness contour;
b illustrates exemplary perceptual weighting curves;
According to an aspect, a method for controlling the consumed audio dose of a user of a media player is described. The media player may e.g. be an audio player (such as a personal music player), a video player (such as a portable DVD player) or other portable electronic devices. The audio dose may be given by the sound pressure level which a user has been exposed to during a given time interval. An audio dose is assumed to be “consumed” by a user when the audio dose is output by the media player and the user could be exposed to the audio dose. For purposes of the method, an audio dose is deemed to be “consumed” even if the user is not actually exposed to the audio dose. In other words, the method is not dependent upon any action or inaction by the user.
The method relates to a first frequency range from the total frequency range relevant for the human ear. The first range is typically a sub-range of the total frequency range. In particular, it may relate to the frequency range within which the human ear is most sensitive. Alternatively, the first frequency range may relate to a low band or a high band frequency range so as to selectively focus on low or high frequencies. In addition, the first frequency range may be determined by splitting the total frequency range into N sub-ranges. N is typically greater than one. One of the N sub-ranges may be selected as the first frequency range. The N sub-ranges may correspond to the Bark scale or an octave scale. Furthermore, the N sub-ranges may be associated with the modifiable frequency bands of an equalizer of the media player, thereby linking the frequency range in which the audio is determined to the hardware constraints of the media player.
The method may comprise the step of determining the audio dose already consumed by the user within the first frequency range. This may comprise determining the audio dose consumed in the first frequency range within a pre-determined time interval prior to the time instance of playing back a particular media track. The consumed audio dose in the first frequency range may be directly determined as the physically produced sound pressure level at the headphones and/or speakers of the media player. The audio dose of a media track within the first frequency range may also be determined from a digital representation of the audio track, e.g. the digital samples of the media track. A scaling factor may be applied to take into account the rendering characteristics of the media player, i.e. notably the volume settings and/or the equalizer settings of the media player and/or the sensitivity of the headphones. Notably in view of the frequency dependent settings of an equalizer, the scaling factor may depend on the frequency range. As such, the consumed audio dose in the first frequency range may be determined from the digital representation of the media track and a scaling factor representing the rendering characteristics of the media player in the first frequency range.
The step of determining the consumed audio dose within the first frequency range may comprise weighting the consumed audio dose with a weight associated with the time instance at which the audio dose was consumed. The weight may decrease with increasing anteriority of the consumed audio dose, thereby reflecting the physiological memory of the human ear.
The method may further comprise the step of evaluating the audio dose of a media track within the first frequency range and the already consumed audio dose of the user within the first frequency range. In other words, a media track may be considered for playback on the media player. The audio dose of the considered media track in the first frequency range is determined and evaluated jointly with the already consumed audio dose of the user within the first frequency range.
The step of determining the audio dose of a media track in the first frequency range may comprise determining spectral components of the media track and/or weighting the spectral components using weights associated with human auditory perception and/or determining the audio dose of the media track based on the weighted spectral components. In other words, the audio dose of a media track may take into account the human auditory perception, e.g. through weighting with an A-curve. These steps may be performed on the digital representation of the audio track. The determined value of the audio dose may need to be multiplied with the scaling factor representing the rendering characteristics of the media player, in order to obtain an audio dose value which corresponds to the perceived sound pressure level of the user of the media player.
The step of determining the audio dose of a media track in the first frequency range may comprise the steps of extracting a plurality of segments of the media track using a window function and/or of determining the audio doses within the first frequency range for the plurality of segments of the media track and/or of determining the audio dose of the media track as the sum of the audio doses of the plurality of segments of the media track. Such windowing may be beneficial in order to isolate quasi-stationary segments of a media track. As a result, the spectral components of a media track may be determined on such quasi-stationary segments of the media track for determining the audio dose of the segment of the media track within the first frequency range.
It may be beneficial to determine an average audio dose of the plurality of segments of the media track. Such average audio dose may also be referred to as an audio dose contribution. The total audio dose of the media track within the first frequency range may then be determined by multiplying the average audio dose within the first frequency range with a factor related to the length of the media track and the length of the window function.
The method may further comprise the step of controlling the audio dose generated by the media player when playing back the media track based on the evaluating step. This controlling step may comprise selecting the media track for play back on the media player. The media tracks may e.g. comprise audio tracks, music tracks or video tracks with an associated audio track.
As already outlined above, the media player may comprise an equalizer. Such equalizer may have a first gain associated with the first frequency range. Furthermore, the equalizer may comprise other gain values which are associated with other frequency ranges outside the first frequency range. In such cases, the controlling step may comprise setting of the first gain and changing the audio dose of the media track within the first frequency range using the first gain. The step of changing the audio dose may comprise amplifying or attenuating the volume of the played back media track within the first frequency range by the first gain. Consequently, if it is determined that the consumed audio dose in the first frequency range exceeds a pre-determined value, the audio dose generated by the media player in the first frequency range may be attenuated, i.e. the playback volume of the media track may be reduced in the first frequency range, while the volume remains unchanged in the other frequency ranges outside the first frequency range.
The method may further comprise the steps of selecting a second frequency range from the total frequency range relevant for the human ear; of determining the audio dose already consumed by the user within the second frequency range; and of evaluating the audio dose of a media track and the already consumed audio dose of the user within the second frequency range. This evaluating step is typically performed separately from the first evaluating step, i.e. the evaluation is performed separately in each frequency range.
The method may comprise the further steps of weighting the already consumed audio dose in the first and second frequency range by a first weight and/or weighting the audio dose in the first and second frequency range of a media track by a second weight and/or determining a weighted sum of the consumed audio dose and the audio dose of the media track in the first and second frequency range. The determination of the weighted sum is performed separately for the first and second frequency range, thereby yielding a first weighted sum and a second weighted sum. The second weight may depend on the duration of the media track. The first and second weight may add up to 1. The second weight may decrease with an increased duration of the media track. The first and second weighted sum typically yields the value of the consumed audio dose after play back of the media track in the first and second frequency range, respectively. The weights may be used to model the physiological memory characteristics of the human ear.
The audio dose consumed by the user may be updated, wherein the updating may be based on a leaky integration of the previously consumed audio dose and the audio dose of the selected media track. The leaky integration is performed separately for the first and the second frequency range. Such leaky integration may e.g. be implemented by weighting of the previously consumed audio dose and the audio dose of the selected media track.
The method may further comprise the step of determining the audio dose within the first and second frequency range of a set of media tracks that are available on the media player; and of determining a playlist for playing back media tracks on the media player by selecting a plurality of media tracks from the set of media tracks based on the separate evaluating steps in the first and second frequency range.
The method may further comprise the step of determining the audio dose of a plurality of media tracks that are available on the media player. The audio dose is determined separately for the first and the second frequency range. As a consequence, the individual audio dose of the media tracks may be used for selecting a particular media track for play back. The media track with the lowest determined audio dose in the first and/or the second frequency range may be selected from the plurality of media tracks for play back on the media player.
The audio dose values may also be used to determine a playlist of media tracks. A playlist typically comprises a plurality of media tracks which are played back in a random or predetermined order. Such a playlist for playing back media tracks on the media player may be determined by selecting media tracks from the plurality of media tracks based on the individual audio doses of the media tracks and the already consumed audio dose of the user. The selection of the media tracks may be performed such that the requirements with regards to a maximum cumulated consumed audio dose are met within the first and/or the second frequency range.
A playlist of media tracks may be generated by the steps of determining the first and the second weighted sum for a plurality of media tracks and by selecting a media track with a smallest first and/or second weighted sum amongst the plurality of media tracks or a first and/or second weighted sum smaller than a pre-determined value (a value that is determined before the playlist generation begins). In other words, the potentially consumed audio dose in the first and/or second frequency range for a plurality of media tracks may be calculated in advance. This may be done under consideration of the previously consumed audio dose in the first and/or second frequency range. Subsequently, the plurality of media tracks may be selected for play back in a playlist, which provides the smallest calculated potentially consumed audio dose in the first and/or second frequency range or which provides a calculated potentially consumed audio dose in the first and/or second frequency range which does not exceed a predefined value, e.g. a maximum allowed audio dose. The predefined value may be defined separately for each frequency range.
The method may further comprise the steps of selecting a media category including a plurality of media tracks that are available for playback on the media player, wherein the selection of a media track is restricted to media tracks from the selected category. In other words, a playlist may be generated under consideration of the audio dose of the media tracks and in addition under consideration of user preferences, such a media categories, genres, interprets, etc.
According to an aspect, an electronic device is described. The electronic device may comprise an audio rendering component configured to generate an audio dose to a user. Typically the audio rendering component is associated with a scaling factor representing its rendering characteristics, e.g. the volume settings, the equalizer settings and the headphone sensitivity. The device may further comprise a memory configured to store a plurality of media tracks. The device may also comprise a processor configured to execute the method steps outlined in the present patent document. In particular, the processor may be configured to select a first frequency range from the total frequency range relevant for the human ear; to determine the audio dose already consumed by the user within the first frequency range; to determine the audio dose within the first frequency range of at least one of a plurality of media tracks; to evaluate the audio dose of the at least one of the plurality of media track within the first frequency range and the already consumed audio dose of the user within the first frequency range; and to control the audio dose generated by the media player based on the evaluating step.
According to an aspect, a storage medium is described. The storage medium comprises a software program adapted for execution on a processor and for performing any of the method steps outlined in the present document when carried out on a computing device.
According to an aspect, a computer program product is described. The computer program product represents a tangible storage item (including but not limited to an optical disk or magnetic storage medium) that includes executable instructions that can cause a processor to perform any of the method steps outlined in the present document when carried out on a machine such as a computer, dedicated media player, mobile telephone or smartphone.
It should be noted that the methods and systems including its preferred embodiments as outlined in the present patent application may be used stand-alone or in combination with the other methods and systems disclosed in this document. Furthermore, all aspects of the methods and systems outlined in the present patent application may be arbitrarily combined. In particular, the features of the claims may be combined with one another in an arbitrary manner.
Mobile media players, such as mobile audio players, have become an important source of “social noise,” which may present a hearing impairment risk to users of the media players. In order to reduce this risk, national governments as well as the European Community (EC) want to follow the scientific advice by limiting the audio dose to sound pressure levels that are less likely to cause hearing impairments over the years. For the work place, the EC has limited the sound pressure level (SPL), weighted by the human frequency sensitivity curve (A-curve) to 80 dB(A) for an eight hour working day (40 hours per week). An equivalent audio dose would be double the sound pressure energy (83 dB(A)) for 20 hours accumulated exposure per week or four times the SPL energy (86 dB(A)) for 10 hours accumulated exposure per week. The unit “dB(A)” refers to the actual sound pressure levels (measured in dB), weighted by the respective A-curve.
Table 1 shows the examples of equivalent time-intensity pressure levels, also referred to as action levels, specified by the European Community directive 2003/10/EC for Noise at Work.
The sound pressure levels (SPL) for typical sounds are shown below in Table 2.
Furthermore, the human frequency sensitivity A-curve is illustrated in
Furthermore, the sensitivity of the ear also depends on the sound level itself. At a sound level of 40 phon, the A-curve 180 drops steeper with increasing frequency than the A-curve 181 at a higher sound level of 80 phon. A “phon” is a unit which describes the perceived loudness level for pure tones, i.e. the phon scale aims to compensate for the effect of frequency on the perceived loudness of tones. By definition, 1 phon is equal to 1 dB sound pressure level at a frequency of 1 kHz. This can be seen in
b illustrates exemplary weighting curves, whereas the curve 190 corresponds to one of the human frequency sensitivity curves illustrated in
With the emergence of personal music players (PMP), notably MP3-based music players, the use of such devices has significantly increased. In 2007, between 40 and 50 million portable audio devices were sold in the countries of the European Union. These devices, which users may control to increase the volume of the sound output, may expose their users on a regular basis to sound pressure levels that may range from 60 dB(A) to 120 dB(A). It has been assumed by the EC that approximately 10% of the users are at risk of developing a permanent hearing impairment due to an excessive exposure to sound pressure levels above 85 dB(A).
Consequently, a significant percentage of the daily audio dose of a PMP user may originate from the PMPs by listening to music via headphones or the built-in speaker(s). Headphones can reach SPLs of 115 dB(A) and even more if they are tightly coupled to the ear drum (e.g. in-ear headphones). As such, they may significantly exceed the sound pressure levels considered to be harmful. Such high sound pressure levels may be experienced without harm for a short period of time, but it is strongly suggested that the accumulated sound pressure level over a given period of time is kept below a certain limit. This is also reflected in the equivalent sound pressure levels listed in Table 1.
It is therefore desirable to provide media players with an ability to limit the overall sound pressure level generated by the media player. In particular, it may be beneficial to provide media players which keep the audio dose that is generated over a certain period of time below a predefined or allowed limit. This target should preferably be achieved for fixed volume settings. That is to say, while the cumulated audio dose is kept below a predefined or predetermined limit (such as, but not limited to, a limit set by a regulatory agency or standards body), the user experience should be enhanced to a degree preferred by the user (for example, enabling a user to choose to listen to audio at a fixed—and perhaps generally high—volume). In other words, unless the user adjusts the volume manually, the volume settings of the media player are generally kept unchanged during a predefined period of time. Such a predefined period of time may be given e.g. by a predefined time interval or by a predefined set of audio tracks.
Furthermore, the sound pressure level generated by a media player may be monitored within specific frequency ranges. As already outlined in the context of
The total acoustic frequency range which is relevant for the human ear may be sub-divided into a plurality of frequency ranges. Such sub-division may follow psychoacoustic scales such as the Bark scale. The Bark scale provides a sub-division of the total frequency range into 24 ranges with the frequency boundaries being at 20 Hz, 100 Hz, 200 Hz, 300 Hz, 400 Hz, 510 Hz, 630 Hz, 770 Hz, 920 Hz, 1080 Hz, 1270 Hz, 1480 Hz, 1720 Hz, 2000 Hz, 2320 Hz, 2700 Hz, 3150 Hz, 3700 Hz, 4400 Hz, 5300 Hz, 6400 Hz, 7700 Hz, 9500 Hz, 12000 Hz, 15500 Hz. Other scales could be the basis for defining a plurality of frequency ranges, e.g. a sub-division wherein each frequency range corresponds to an octave starting from a base frequency. In such cases, the higher frequency boundary of a frequency range would be two times the lower frequency boundary.
The sub-division of the total frequency range may also be associated with the capabilities of the media player. In particular, the media player may comprise an equalizer with frequency dependent equalizer settings. Such equalizer settings may enable a user to amplify or attenuate a certain number of frequency bands of an audio track independently. This may be implemented by assigning a different equalizer weight or gain to each of the number of frequency bands. These weights may be changed by the user. The number of frequency bands which can be modified separately may vary from media player to media player. In an embodiment the sub-division of the total acoustic frequency range may correspond to the number of frequency bands provided by the equalizer of the particular media player.
In view of the above, it may be beneficial to provide a media player with means to evaluate the sound pressure level generated over a predefined amount of time within a plurality of different frequency ranges. The media player should be enabled to ensure that the cumulated sound pressure level within a given frequency range remains below a frequency dependent threshold value. Preferably, this should be ensured for all frequency ranges from the plurality of frequency ranges. In an embodiment, this should be achieved for fixed equalizer settings of the media player.
According to an aspect, a playlist of media tracks is suggested to the user so that the accumulated sound pressure dose within a frequency range of the proposed playlist on top of the listening exposure of the past is below a predefined limit. In general, a media track is a recorded sound or sounds, generally having a beginning, an ending and a playback duration. The recorded sounds may be accompanied by media information other than audio information, such as video information. Because the techniques discussed herein are generally applicable to the audio portion of a multi-media track, the terms “media track” and “audio track” are used herein synonymously. The predefined limit may be set differently for a plurality of different frequency ranges. The playlist of audio tracks should be generated such that the accumulated sound pressure dose, including the listening exposure of the past, stays below the predefined limits for all relevant frequency ranges.
The playlist typically comprises one or more audio tracks which are played back on the media player in a predetermined or arbitrary manner. In order to enhance the overall user experience, the audio volume setting and the equalizer settings should remain unchanged during playback of the playlist (unless the user adjusts any of the settings manually to the user's own preferred settings). Instead, the audio content may be changed to meet the cumulated audio dose target, while keeping the volume level of the media player constant. In other words, one or more audio tracks are selected that can be played at the fixed volume settings and at the fixed equalizer settings, while maintaining the cumulated audio dose below or at the predefined limit for all frequency ranges.
A playlist is typically specified by a set of media tracks, e.g. audio tracks and/or video tracks. The length of the playlist may be defined as the number of media tracks which it comprises and/or as the cumulated duration of the playback of the set of media tracks. The set of media tracks which is comprised in a playlist is typically selected from a larger collection of media tracks, e.g. from a media track database that is stored on the user's media player and/or from appropriate web sites. The selection of the set of media tracks may be based on, for example, the author of an audio track, the genre of the media track, and/or other preferences of the user. The set of media tracks of a playlist may be played back in a predefined order or randomly. In other words, the generation of a playlist may be submitted to constraints. As outlined above, such constraints may be related to the audio dose contribution of the selected media tracks within the different frequency ranges. Furthermore, such constraints may be related to user preferences, such as genre, etc.
According to a further aspect, a plurality of average SPL values, weighted by the A-curve, may be computed for a media track. Each average SPL value is related to the sound pressure value generated by the media track within a particular frequency range. As discussed below, various signal processing techniques can be employed to determine SPL values. Typically the plurality of average SPL values covers the total acoustic frequency range relevant for the human ear. It is also possible to determine average SPL values for partial audio tracks, e.g. for blocks of a given duration of an audio track. Consequently, each audio or music track i, i=1, . . . , N, is modeled by a set of average SPL value Si,n, wherein n=1, . . . , N indicates the respective frequency range. These SPL values may be pre-computed and they may reflect the complete audio dose of the audio track or the audio dose of a predetermined time segment of the audio track. In the latter case, the complete audio dose may be determined by cumulating the sectional audio dose values over the length of the audio track.
In an embodiment, the set of SPL values for a music track i can be computed by taking the short-time Fourier spectrum of a suite of windowed signal segments (a suite of windowed signal segments being a set of short-duration pieces of the audio track), by applying the A-weighting curves 180, 181 or 190 shown in
In order to reduce computational complexity, it may be beneficial to down-sample the number of windows of a music track, since the sounds are typically stationary for a short period of time.
In the above example, the SPL value Si,n corresponds to the average SPL value of the audio track i in the frequency range n within a certain window. Given the duration or length Tw of the window and the duration or length Ti of the audio track i, the total SPL value of the audio track i within the frequency range n may be given by
Ai,n may also be referred to as the audio dose of the audio track i within the frequency range n. It should be noted that the length Tw of the window typically depends on the form/progression of the window function. For a rectangular window Tw may be the actual length of the window, whereas for a Gaussian window Tw may depend on the underlying variance of the Gaussian window.
The process of audio dose computation for a music or audio track is illustrated in
The frequency subband coefficients are subsequently weighted using weights which are associated with human auditory perception. This is performed in the weighting unit 203 and yields the weighted subband coefficients Xi′(k,w). The weights may be derived from the A-curves of
Based on the weighted subband coefficients Xi′(k,w) the perceived sound pressure level in the frequency ranges n=1, . . . , N, e.g. the sound pressure level measured in dB(A), is determined in the SPL determination unit 203. This yields the set of perceived SPL estimates Si,n(w) for the windowed section of the audio track xi(n). The SPL determination unit 203 may comprise an inverse transform, converting the frequency subband coefficients of a particular frequency range n into the time domain, thereby yielding a weighted subsection xi,n′(n,w) of the frequency range n of the audio track. This weighted subsection xi,n′(n,w) is transformed into sound pressure within the frequency range n by the audio rendering means of the respective media player, e.g. a D/A converter and an amplifier in combination with a speaker or a headphone. The specification of the audio rendering means and/or volume settings and/or the equalizer settings influence the actually generated sound pressure level within the particular frequency range n. However, a normalized SPL value may be determined for the audio track xi(n) within this particular frequency range n. This normalized SPL value may be multiplied by a scaling factor to determine the actual perceived sound pressure level during playback. The scaling factor will typically depend on the specification of the audio rendering means, its actual volume settings and the weight or gain of the equalizer in the respective frequency range n. The normalized SPL value Si,n(w) for the frequency range n may be determined as the root mean squared value of the samples of the weighted subsection xi,n′(n,w) of the audio track. Furthermore, the determination of the normalized SPL value Si,n(w) may involve normalization by a reference sound pressure and/or determination of a logarithmic value of the sound pressure.
It should be noted that the transformation into the frequency domain may be done such that the number of subbands corresponds to the number of frequency ranges N. In other words, the number of points used for the transformation, e.g. the FFT or DFT, may correspond to the number of frequency ranges N. In such cases, the subband coefficient Xi(k,w) can be directly associated with a particular frequency range and the transformation of the corresponding weighted subband coefficient Xi′(k,w) into the time domain can be directly used for the determination of the perceived audio dose of the audio track xi(n) in the particular frequency range.
Eventually, the normalized audio dose of the audio track xi(n) in the frequency range n is determined in the audio dose computation unit 205. This may be performed for all frequency ranges n=1, . . . , N. The average SPL value Si,n of the audio track xi(n) in the frequency range n may be determined as the average SPL value Si,n(w) across the complete set of windows. In such cases, the SPL value represents the average audio dose of the audio track xi(n) within a predefined window of length Tw. The complete audio dose Ai,n is obtained by integrating the Si,n values over the length Ti of the audio track xi(n). In other words, the audio dose Ai,n of audio track i is obtained by multiplying the average Si,n value with the length Ti of the audio track i. Furthermore, the length Tw of the window may have to be taken into consideration. As such, the audio dose Ai,n of audio track i may be obtained by multiplying the average Si,n value with the length Ti of the audio track divided by the length Tw of the window.
The audio dose computation may be triggered every time new music tracks are detected. A music watcher application is started in step 301. This music watcher application scans particular web sites for new audio or music tracks in the interest of the user. If a new music track is available, it is downloaded to the device, e.g. via USB or via a wireless communication network (step 302). The device checks the availability of new audio tracks (step 303) and if such tracks are available, a set of audio dose values is calculated for the new audio tracks (step 304).
By using the above methods and systems, media tracks i may be associated with a set of audio dose values Ai,n and/or a set of average SPL values or audio dose contributions Si,n. This may be done for the complete set of media tracks stored in the database of a media player and/or for the media tracks available at particular web sites. It should be noted that audio dose values Ai,n and/or average SPL values Si,n may be normalized, i.e. they may be independent from the actual rendering characteristics of the particular media player. These rendering characteristics, e.g. the volume settings, the equalizer settings, the speaker sensitivity and/or the headphone sensitivity, may be reflected by a scaling factor F associated with the actual rendering characteristics. Such a scaling factor F may be different for different frequency ranges n. This may be due to the frequency response of the amplifier and/or the frequency dependent equalizer settings. In an embodiment a set of scaling factors Fn, n=1, . . . , N may be defined for the set of frequency ranges n=1, . . . , N. Consequently, the actual audio dose in the frequency range n may be determined by multiplying the normalized audio dose value in that frequency range with the scaling factor Fn of that frequency range. In other words, the computation is done in the digital domain. The resulting sound pressure levels after digital-to-analog (D/A) conversion, amplification and conversion into acoustic energy via the speakers or headphones of a media player can be pre-computed for a particular media player configuration, if the design parameters of the media player and of the speakers/headphones are known. If these parameters are not known, then the sound pressure levels may be estimated e.g. by using a worst-case scenario. By way of example, the use of very sensitive headphones may be assumed in a worst-case scenario. Using such assumptions, a set of scaling factors Fn can be determined.
In the following, it is assumed without loss of generality, that the set of audio dose values Ai,n and/or the set of average SPL values Si,n correspond to the actually rendered audio dose values and/or SPL values.
Typically, a user has an audio listening history, i.e., what the user has been exposed to (and/or has actually heard) in the past until a certain time (t=0). From the audio listening history can be determined a cumulated audio dose An(0) in the frequency range n. This audio dose may be referred to as the already consumed audio dose in the frequency range n.
At the starting time (t=0) the system proposes or adapts a playlist by inserting music (or other audio) tracks so that the accumulated audio dose in the frequency range n, which is composed of the already consumed audio dose An(0) and the individual playlist contributions Si,n remains below the maximum allowed audio dose for that particular frequency range n. This condition should be preferably met at all times. Furthermore, this condition should be met for all frequency ranges n=1, . . . , N.
If at any time, the accumulated audio dose exceeds the pre-determined level in any for the frequency ranges n=1, . . . , N, the playlist may be adjusted such that eventually the accumulated audio dose in that particular frequency range drops below the allowed limit for that particular frequency range. If (for example) the starting value An(0) is above the limit for the frequency range n, the playlist may be assembled (e.g., by selecting or by declining to select tracks as a function of the tracks' own audio doses) to aim at reducing the audio dose in the frequency range n over time so that the final value is below the maximum limit for the frequency range n.
It may be assumed that the volume level and the equalizer settings remain constant for the selection process of the playlist. If the user changes the volume level and/or the equalizer settings, an equivalent correction factor or scaling factor may be applied to the SPL contributions of each music track in the playlist. In other words, the above mentioned scaling factor Fn for the respective frequency range may be increased or decreased in accordance to the changes in volume and/or equalizer settings.
As already outlined above, the overall audio dose for a user should take into account the listening history of the device or user and the potential audio dose contributions of the music tracks played in the future. This may be done in different manners, whereby apart from the accumulation of the audio doses in the different frequency ranges, also the time aspect should be taken into consideration. In particular, it should be taken into account that longer pieces of music would have a higher impact than shorter pieces of music. Furthermore, the impact of previously heard music tracks on the cumulated audio dose should decrease over time to model physiological memory effects of the human ear (which are discussed below).
As such, the accumulation process of audio doses of the different frequency ranges may be modeled as a leaky integrator. Mathematically speaking the audio dose An(t) in the frequency range n which has been consumed by a user at time t may be represented by a recursive filter
where a music track i with a duration Ti and a set of audio dose contributions Ai,n is played next after time instance t. If only a partial audio track i is played, then the set of audio doses of the partial audio track may be obtained from the set of average SPL values Si,n of the audio track i. For this purpose the set of average SPL values Si,n, typically normalized by the length Tw of the window which was used to determine the set of SPL value Si,n, is multiplied by the duration Tp during which the audio track i was played back. This will provide the partial audio dose Ai,n,p of the audio track i. In such cases, the values Ai,n,p and Tp replace the values Ai and Ti in the above equation.
The constant c determines a time constant of the audio dose integration. It may be used to model the auditory “memory” of the human ear, i.e. it may be used to reflect the physiological fact that typically the impact of a consumed audio dose on the ear decreases over time. As such, the constant c models a decay which is typically in the order of a few days.
Based on the evaluation of the user's cumulated audio dose An(t) in the set of frequency ranges n=1, . . . , N, a playlist may be selected. In other words, a set of audio tracks may be selected for playback from a reservoir of audio tracks, e.g. a database on the media player or a web site. The set of audio tracks may be selected such that the cumulated audio dose An(t) stays below a predefined value An,max, i.e. An(t)≦An,max. This condition may need to be met at all time, i.e. ∀t. This conditions should also be met for all frequency ranges n=1, . . . , N. If, at a point of time, the cumulated audio dose An(t) exceeds An,max in a frequency range n, the set of audio tracks may be selected such that the time to reduce the cumulated audio dose An(t) below the predefined value An,max is minimized.
A further aspect to be considered in the selection process of the audio tracks for the playlist is the length of the playlist, i.e. including but not limited to the number of tracks which are included in the playlist. Typically, the available degrees of freedom for meeting the target of keeping the cumulated audio dose below a predefined value increase with the number of audio tracks in the playlist. If the number of audio tracks is large, a mixture of tracks with relatively high average SPL values Si,n for particular frequency ranges and tracks with relatively low average SPL values Si,n for particular frequency ranges may be selected. By way of example, audio tracks having predominant low frequency contribution and audio tracks having predominant high frequency contribution may be selected. Using the above recursive formula for the cumulated audio dose An(t) in the different frequency ranges, an order of playback of the playlist could be determined which meets the condition An(t)≦An,max. By way of example, audio tracks having a large high frequency contribution could follow audio tracks having a large low frequency contribution. If, on the other hand, the number of tracks within the playlist is small, the selected audio tracks will typically have medium average SPL values Si,n, such that each individual audio track in the playlist approximately meets the condition that its average SPL value Si,n does not exceed a predefined maximum SPL value Sn,max.
In other words, when selecting a given number of audio tracks from a database or website to form the playlist, the set of audio doses Ai,n and/or the set of average SPL values Si,n of the audio tracks are taken into consideration. Furthermore, other criteria, e.g. the similarity of a certain music track i to a desired category of music and/or the genre and/or the author of the audio track, may be taken into account when selecting music tracks for the playlist.
Apart from selecting a set of audio tracks for a playlist, other factors, such as the order of the playlist, the skipping of certain audio tracks, the partial playback of certain audio tracks, etc., may influence the user's cumulated audio dose An(t) in the different frequency ranges. By way of example, the audio tracks in a playlist may be played back randomly, while the cumulated audio dose An(t) is monitored for each of the different frequency ranges n=1, . . . , N. If, at a point of time, the cumulated audio dose exceeds the maximum allowed audio dose An,max within at least one of the frequency ranges, audio tracks with low average SPL values Si,n in the respective frequency range may be selected from the playlist, and played back until the cumulated audio dose in the respective frequency range has dropped to a threshold value, which is typically lower than An, max in order to provide an audio dose buffer. Once the latter condition is met, the random playback of audio tracks of the playlist may be resumed. In this context, different pieces of music may be sorted according to their SPL values or relative audio dose contribution Si,n in the different frequency ranges. If at a particular point of time, the cumulated audio dose An(t) exceeds the allowed limit within a particular frequency range, audio tracks with low Si,n values in this particular frequency range may be easily inserted in order to reduce the cumulated audio dose.
In an embodiment, the equalizer settings may be modified when the cumulated audio dose An(t) in a particular frequency range exceeds the allowed limit An, max. In particular, the equalizer gain which is associated with the particular frequency range may be reduced until the cumulated audio dose in the particular frequency range has dropped to the pre-defined threshold value. The equalizer gain will typically be selected such that the pre-defined threshold value is reached within a minimum time interval, while still maintaining an acceptable acoustic quality.
In step 401, the user initiates a playback mode of his media player. First, the system determines the set of audio doses An(0) which has already been consumed by the user. Furthermore, the current volume settings and equalizer settings and possibly the specification of the audio rendering means, e.g. the speakers or the headphones, are determined (step 402), thereby providing a set of scaling factor Fn. The set of already consumed audio doses may be stored in and retrieved from a memory of the media player. Alternatively or in addition, the set of audio doses which has already been consumed by the user on other devices may be taken into account. By way of example, the current device may retrieve the set of already consumed audio doses from a central network server, where such data is collected and stored for a plurality of media players. The set of already consumed audio doses may also be transferred from one media player to a next using short range communication means such as Bluetooth™.
In step 403, the media player generates a playlist according to the methods outlined in the present document. This playlist takes into account the set of already consumed audio doses, the current volume and equalizer settings and/or the specification of the audio rendering means, and aims at maintaining the cumulated consumed audio doses in the different frequency ranges below a predetermined limit. This condition should be achieved for all frequency ranges n=1, . . . , N. The playlist may be determined in different manners. Depending on the length of the playlist, a certain number of audio tracks may be selected from a database or website. This selection process should take into account the relative audio contribution values Si,n of the audio tracks, such that a mix of audio tracks is available in the playlist which jointly can meet the requirements with regards to the cumulated audio doses in the different frequency ranges. Furthermore, musical preferences and similarities or genres or interprets may be considered, when selecting audio tracks for a playlist. In addition to selecting the audio tracks for the playlist, an order of the playlist may be determined, such that the conditions with respect to the cumulated audio doses in the different frequency ranges are met. Furthermore, selective measures may be taken, if at a point of time, the cumulated audio dose exceeds a predefined value within a particular frequency band. By way of examples, audio tracks with an excessive audio dose in the particular frequency band may be skipped and/or audio tracks with a low audio dose contribution in the particular frequency band may be inserted.
In an embodiment, a plurality of predefined levels of cumulated audio dose is considered when generating the playlist, i.e. when selecting the audio tracks of the playlist and when determining their order of playback. Such a plurality of predefined levels may be used to define different sets of rules for the generation of the playlist. By way of example, if a first level of cumulated audio dose is reached in a particular frequency range, only audio tracks which significantly exceed the targeted audio dose level in the particular frequency range are excluded from the playlist. With increasing level of cumulated audio dose further audio tracks may be excluded, until eventually only audio tracks with a low audio dose contribution may be played back, in order to meet the overall cumulated audio dose target in the different frequency ranges. It may also be contemplated to completely block the playback of audio tracks or to completely block the playback of particular frequency ranges, if a certain level of cumulated audio dose has been reached.
A playlist may be generated by determining in advance the cumulated audio dose in the different frequency ranges of the set of audio tracks using the methods outlined above. By way of example, a first set of audio tracks may be selected and the cumulated audio dose in the different frequency ranges may be determined in advance using the above formula. If the cumulated audio dose exceeds the predetermined level in a particular frequency range, the audio tracks which provide the highest audio dose contribution in the particular frequency range may be replaced with audio tracks which contribute a reduced audio dose in the particular frequency range. By performing such an iterative process, a playlist may be generated which comprises audio tracks that meet the desired audio dose target for all the relevant frequency ranges. Such a generation scheme for a playlist which takes into account a plurality of future audio tracks may be referred to as a predictive generation of a playlist. A predictive generation scheme is opposed to an ad hoc generation scheme of a playlist, where at any time only the immediately next audio track in the playlist is selected.
Different schemes for the computation of the cumulated audio dose may be used. The set of audio dose of the currently played audio track may be added to the set of previously consumed audio dose, e.g. using the formula provided above. The accumulation may be performed smoothly, such that continuously a fraction of the set of audio doses of the audio track is added to the set of cumulated audio doses when the audio track is played back. This has the advantage that when the playback of an audio track is interrupted, the set of cumulated audio doses is accurate. Alternatively, the set of audio doses of an audio track may be added to the set of cumulated audio doses, once the complete audio track has been played back. If the set of audio tracks is interrupted, only a respective fraction of the set of audio doses is added to the set of cumulated audio doses.
If no user input is performed, the audio tracks of the determined playlist are played back on the media player (step 404). However, if it is determined that the user has changed the volume settings and/or the equalizer settings of the device or that the user has modified the playlist (step 405), the system returns to steps 402 and 403, in order to determine an updated playlist, e.g. an updated set of audio tracks and/or an updated order of playback of the set of audio tracks, which takes into account the modifications made by the user. It should be noted that if the user has interrupted an audio track which was currently on playback, only a fractional part of the set of audio doses of that audio track should be added to the set of cumulated audio doses. This could be done by only considering the fraction of the set of audio doses which corresponds to the already played time of the audio track.
In an embodiment, the equalizer settings may be modified by the user as outlined above. It may be contemplated to provide forced limits of equalizer gain values in particular frequency ranges which are consumed excessively by a user. As such, the user may be prevented from setting an equalizer gain which exceeds the forced limit in the particular frequency range.
According to an aspect, a media player may be used by a plurality of users. In such cases, it is desirable that the set of consumed audio doses is monitored for the different users separately. For this purpose, a plurality of user accounts associated with the plurality of users could be managed on the media player. At the beginning of a session, a particular user would be prompted for a user identification and possibly a password. In addition, the user may be requested to provide the media player with information related to the already consumed audio dose in the different frequency ranges. By using the user identification, the media player could execute the above methods for each user separately and thereby monitor and possibly limit the consumed audio dose in the different frequency ranges.
It may be contemplated to allow a plurality of users to register with the media player at the same time. This may be beneficial when monitoring the audio dose or sound pressure level exposure consumed by a plurality of users using the same media player. By way of example, a plurality of headphones may be connected to the same media player. In a further example, a set of speakers may be used, thereby exposing a plurality of users to the audio dose. By allowing a plurality of users to be registered on the media player in parallel, the consumed audio dose per frequency range could be monitored for each individual user in parallel. Each user could be given the possibility to inform the media player of the set of already consumed audio doses, when registering on the media player. It should be noted that as a result of different users entering different initial set of consumed audio dose values, conflicts between the separate monitoring processes for the different users may arise. By way of example, a user having entered a set of high initial consumed audio dose value may reach the maximum allowed audio dose in a particular frequency range, while others are still within the allowed range. To resolve such conflicts, the generation of the playlist may be performed according to the above methods, such that the maximum allowed audio dose in the different frequency ranges is not exceeded for any one of the registered users.
Upon interruption of a session and/or upon leaving the media player, a user of the media player may de-register from the media player, e.g. by entering a user identification and possibly a password. Upon de-registration the media player may inform the user about the set of cumulated consumed audio doses, such that the user may provide this information to a subsequent media player. In view of the fact that the media player monitors each active user on the media player separately, such de-registration will typically not impact the monitoring for the other users registered with the media player.
The above examples are not intended to be an exclusive list of techniques whereby an audio dose may be controlled based upon the evaluation of the audio dose of one or more media tracks and the already consumed audio dose of the user within one or more frequency ranges. In some instances, variations or combinations of the above techniques may be employed.
Referring to
In addition to the microprocessor 128, other parts of the wireless device 100 are shown schematically. These include: a communications subsystem 170; a short-range communications subsystem 102; the keyboard 114 and the display 126, along with other input/output devices including a set of LEDs 104, a set of auxiliary I/O devices 106, a serial port 108, a speaker 111 and a microphone 112; as well as memory devices including a flash memory 116 and a Random Access Memory (RAM) 118; and various other device subsystems 120. The wireless device 100 may have a battery 121 to power the active elements of the wireless device 100. The wireless device 100 is in some embodiments a two-way radio frequency (RF) communication device having voice and data communication capabilities. In addition, the wireless device 100 in some embodiments has the capability to communicate with other computer systems via the Internet.
Operating system software executed by the microprocessor 128 is in some embodiments stored in a persistent store, such as the flash memory 116, but may be stored in other types of memory devices, such as a read only memory (ROM) or similar storage element. In addition, system software, specific device applications, or parts thereof, may be temporarily loaded into a volatile store, such as the RAM 118. Communication signals received by the wireless device 100 may also be stored to the RAM 118.
Further, one or more storage elements may have loaded thereon executable instructions that can cause a processor, such as microprocessor 128, to perform any of the method outlined in the present document.
The microprocessor 128, in addition to its operating system functions, enables execution of software applications on the wireless device 100. A predetermined set of software applications that control basic device operations, such as a voice communications module 130A and a data communications module 130B, may be installed on the wireless device 100 during manufacture. In addition, a personal information manager (PIM) application module 130C may also be installed on the wireless device 100 during manufacture. As well, additional software modules, illustrated as another software module 130N, may be installed during manufacture. Such additional software module may also comprise an audio and/or video player application according to the present disclosure.
Communication functions, including data and voice communications, are performed through the communication subsystem 170, and possibly through the short-range communications subsystem 102. The communication subsystem 170 includes a receiver 150, a transmitter 152 and one or more antennas, illustrated as a receive antenna 154 and a transmit antenna 156. In addition, the communication subsystem 170 also includes a processing module, such as a digital signal processor (DSP) 158, and local oscillators (LOs) 160. The communication subsystem 170 having the transmitter 152 and the receiver 150 includes functionality for implementing one or more of the embodiments described above in detail. The specific design and implementation of the communication subsystem 170 is dependent upon the communication network in which the wireless device 100 is intended to operate.
In a data communication mode, a received signal, such as a text message or web page download of a video/audio track, is processed by the communication subsystem 170 and is input to the microprocessor 128. The received signal is then further processed by the microprocessor 128 for an output to the display 126, the speaker 111 or alternatively to some other auxiliary I/O devices 106, e.g. a set of headphones or other audio rendering means. A device user may also compose data items, such as e-mail messages, using the keyboard 114 and/or some other auxiliary I/O device 106, such as a touchpad, a rocker switch, a thumb-wheel, or some other type of input device. The composed data items may then be transmitted over the communication network 110 via the communication subsystem 170.
In a voice communication mode, overall operation of the device is substantially similar to the data communication mode, except that received signals are output to a speaker 111, and signals for transmission are generated by a microphone 112. The short-range communications subsystem 102 enables communication between the wireless device 100 and other proximate systems or devices, which need not necessarily be similar devices. For example, the short range communications subsystem may include an infrared device and associated circuits and components, or a Bluetooth™ communication module to provide for communication with similarly-enabled systems and devices.
In a particular embodiment, one or more of the above-described methods for audio track download are implemented by the communications subsystem 170, the microprocessor 128, the RAM 118, and the data communications module 130B, collectively appropriately configured to implement one of the methods described herein. Furthermore, one or more of the above-described methods for video and/or audio playback are implemented by a software module 130N, the RAM 118, the microprocessor 128, the display 126, and an auxiliary I/O 106 such as a set of headphone and/or the speaker(s) 111.
In the present document methods and systems have been described which may be used to protect a user of media players or mobile telephones against hearing impairments caused by an excessive exposure to high sound pressure levels. It is proposed to perform an automatic music selection or more generally an automatic audio selection which meets pre-defined audio dose requirements and which at the same time enhances the overall user experience. Such audio dose requirements are specified and monitored separately for a plurality of frequency ranges. This can be achieved by taking into account the listening history of the particular user or device. The proposed methods can be implemented with low computational complexity and are therefore well adapted for the use in portable electronic devices. Further, the techniques described herein offer the potential advantage of adaptation to the listening habits of different users.
The methods and systems described in the present document may be implemented as software, firmware and/or hardware. Certain components may e.g. be implemented as software running on a digital signal processor or microprocessor, e.g. the microprocessor 128 of the mobile device 100. Other components may e.g. be implemented as hardware or as application specific integrated circuits. The signals encountered in the described methods and systems may be stored on media such as random access memory or optical storage media. They may be transferred via networks, such as radio networks, satellite networks or wireless networks. Typical devices making use of the method and system described in the present document are dedicated media players (including, but not limited to, dedicated audio players), mobile telephones or smartphones.
| Number | Name | Date | Kind |
|---|---|---|---|
| 5172358 | Kimura | Dec 1992 | A |
| 6177623 | Ooseki | Jan 2001 | B1 |
| 7003123 | Kanevsky | Feb 2006 | B2 |
| 20050147004 | Kumagai et al. | Jul 2005 | A1 |
| 20060147049 | Bayley et al. | Jul 2006 | A1 |
| 20080212787 | Goldstein | Sep 2008 | A1 |
| Number | Date | Country |
|---|---|---|
| 0529158 | Mar 1993 | EP |
| 2430319 | Mar 2007 | GB |
| 2007114348 | Oct 2007 | WO |
| WO2008095013 | Aug 2008 | WO |
| Entry |
|---|
| Scenihr; European Commission, Directorate-General for Health and Consumers; Potential health risks of exposure to noise from personal music players and mobile phones including a music playing function; Sep. 23, 2008; http://ec.europa.eu/consumers/safety/projects/docs/noise-workgroup-report.pdf. |
| Scenihr; European Commission, Directorate-General for Health and Consumers; Potential health risks of exposure to noise from personal music players and mobile phones including a music playing function; Sep. 23, 2008; http://ec.europa.eu/health/ph risk/committees/04 scenihr/docs/scenihr o 018.pdf. |
| Office Action dated Feb. 11, 2011; European App. No. 09176389.6. |
| Examination Report dated May 20, 2011. In corresponding application No. 09176389.6. |
| F. Duck: “Acoustic Dose and Acoustic Dose-Rate”; Science Direct Ultrasound in Medicine & Biology vol. 35, No. 10, Aug. 3, 2009 pp. 1679-1685, XP002567397 USA ; Retrieved on Feb. 5, 2010. |
| EESR, European Patent App. No. 09176389.6; mailed Feb. 26, 2010. |
| Summons to attend oral proceedings pursuant to Rule 115(1) EPC mailed Apr. 13, 2012, in corresponding European patent application No. 09176389.6. |
| David Havelock, Sonoko Kuwano, Michael Vorlander: “Handbook of Signal Processing in Acoustics”, Jan. 1, 2008, Springer ISBN:0387776982, vol. 1, pp. 486-487. |
| Communication under Rule 71(3)EPC mailed Oct. 10, 2012, in corresponding European patent application No. 09176389.6. |
| Druckexemplar accompanying communication under Rule 71(3)EPC mailed Oct. 10, 2012, in corresponding European patent application No. 09176389.6. |
| Office Action mailed Oct. 4, 2012; in corresponding Canadian patent application No. 2,721,599. |
| Extended European Search Report mailed Aug. 7, 2013; in corresponding European patent application No. 13154856.2. |
| English abstract for WO2007114348; published on Oct. 11, 2011; and retrieved on Mar. 26, 2014. |
| Number | Date | Country | |
|---|---|---|---|
| 20110118859 A1 | May 2011 | US |
