Audio signal equalizer and method for vehicles using data communication bus
The vehicle audio signal equalizer adapts to diverse noise sources using a CAN bus-based system, combining speed and functional filters to maintain spectral balance efficiently and cost-effectively, addressing the limitations of existing methods.
Patent Information
- Application Number
- JP2024065747
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2023-04-14
- Filing Date
- 2024-04-15
- Publication Date
- 2024-11-06
AI Technical Summary
Existing vehicle audio signal equalization methods face challenges in handling diverse and unpredictable noise sources, with SDVC and SDEC techniques being either ineffective or costly and complex, while NDEC requires significant computational resources.
A vehicle audio signal equalizer using a data communication bus, such as a CAN bus, processes noise data to apply adaptive equalization filters based on vehicle speed, engine rotation, and HVAC parameters, without additional hardware, using a combination of speed and functional filters to maintain spectral balance.
Achieves effective audio signal equalization across varying noise conditions at a reasonable cost, utilizing existing vehicle architecture and reducing computational overhead.
Smart Images

Figure 2024156621000001_ABST
Abstract
Description
[Technical field]
[0001] The present invention relates to the field of vehicle audio signal processing, and more particularly to the equalization of such audio signals in potentially noisy environments, especially when the noises in question are prone to change over time. [Background technology]
[0002] Masking a first audio signal with a second audio signal is the process by which the hearing threshold of the first audio signal is raised by the presence of the second audio signal, i.e., spectral masking occurs in a given frequency band when the presence of the second signal makes it impossible to detect the first signal at a lower amplitude and in the same frequency band.
[0003] In a car, this effect is commonly caused by the aerodynamic noise associated with the movement of the car and by the engine noise, in the presence of which some frequencies are masked and can impair the perception of the spectral balance of the music being emitted within the car's cabin.
[0004] The perceived tonal balance depends on the difference between the outgoing sound level and the masking threshold: since the dynamic range of a musical signal (the difference between its maximum and minimum amplitude) is constant, when its average level, given in dBSPL (for "Sound Pressure Level"), is close to the threshold, some components of the signal will be perceived while others will be masked.
[0005] To avoid masking phenomena and maintain the perceived tonal balance, some frequencies of the transmitted speech signal must be above the masking threshold. Due to the lack of sufficient acoustic isolation, signal processing techniques have been developed to obtain masking filters that compensate for the masking phenomena.
[0006] In the prior art, two types of techniques are commonly used. The first technique, called SDVC (for "Speed Dependent Volume Control"), consists in bringing the volume of the audio signal (i.e. the entire audio signal) above the masking threshold by adding a gain calculated from a vehicle speed table. The second technique, called SDEC (for "Speed Dependent Equalisation Control"), consists in applying a low-shelf filter with parameters that depend on the vehicle speed and the attenuation of the entire system.
[0007] These two techniques have the advantage that they do not require much computing power, which is extremely important in the automotive environment, both from the perspective of cost and ease of integration.
[0008] However, in a vehicle such as a car, there are multiple sources of background noise, as in the following example: -The structure of the vehicle itself, related to acoustic isolation and aerodynamics, tyre type, engine etc; -Some vehicle functions that may change the background noise, such as the HVAC system, starting / stopping the heat engine in a hybrid vehicle; -Road noise and aerodynamic noise increase with vehicle speed; - Road surface characteristics: asphalt roads and coarse-grained roads (which will be noisier than smooth asphalt roads); -Environment: Open country roads produce less noise than tunnels. Rain and wind also increase background noise for vehicles.
[0009] In general, background noise can be described as a broadband noise with a drop of 6 dB per octave in the high frequency range. However, this definition may be insufficient to describe the masking phenomenon encountered in practice, depending on the noise source mentioned above. For example, in the presence of rain, high frequencies may be masked. Also, depending on the nature of the vehicle and its speed, the frequency bands actually masked may change over time.
[0010] For such a variable masking effect, the SDVC function would be expected to increase the overall level of the signal when the vehicle speed increases. However, the perception of the emitted audio frequency signal is non-linear around the masking threshold (e.g., higher frequencies may be more easily perceived than lower frequencies). That is, while the SVDC function increases the overall level of the signal when the vehicle speed increases, this is not necessarily the case for all frequencies.
[0011] The SDEC function, on the other hand, introduces certain spectral processing to avoid this effect. By amplifying the signal with a low-shelf filter, the SDEC function ensures that the low frequencies of a typical nasal noise profile (i.e. a profile with a drop of 6 dB per octave at higher frequencies) are above the masking threshold and maintains the perceived spectral balance. Again, it is assumed that the background noise profile depends only on the vehicle speed. However, as mentioned above, such a profile can change dramatically and unpredictably, e.g. depending on the road surface and the environment.
[0012] For this reason, the Applicant has developed an equalization by masking called NDEC (for "Noise Dependent EQ Control"). It consists in performing an equalization adjustment depending on the noise collected by an acoustic sensor in the vehicle. The process is based on an algorithm that analyses the noise collected by a microphone, estimates therefrom its spectrum in real time and applies a compensation filter to the audio signal. This has the great advantage of being very adaptable to different noises. However, it also has a number of disadvantages, such as the need to use a microphone, a very high consumption of computational and memory resources, and the need for audio-specific processing. Summary of the Invention [Problem to be solved by the invention]
[0013] Thus, in the current situation, one must choose between equalization by masking, which is simple but not very effective due to the diversity of noise sources, and equalization, which is effective but very costly and complicated to implement. [Means for solving the problem]
[0014] The present invention improves this situation. For this purpose, it provides an audio signal equalization device for a vehicle using a data communication bus, comprising: a memory unit arranged to store a speed filter data set, which consists of three items relating a speed value to a gain value associated with a coefficient defining a speed masking filter, and a functional filter data set, which consists of three items relating an engine speed value or a ventilation value to a gain value associated with a coefficient defining an engine masking filter or a ventilation masking filter, a collection unit arranged to receive at least two parameters available on said communication bus defining the speed of said vehicle and at least one functional parameter selected from the group consisting of an engine speed and a ventilation data set of said vehicle; a calculation unit arranged to receive as input a noise margin value and said at least two parameters available on said communication bus, to derive therefrom at least one velocity filter data set and at least one functional filter data set, and to combine them as a masking equalization filter to be applied to said audio signal; The present invention proposes an apparatus comprising:
[0015] The present apparatus is highly advantageous since it is able to apply effective equalization for each major noise source, while keeping implementation costs reasonable, without additional hardware, and using architectures present in all vehicles.
[0016] In various embodiments, the present invention may include at least one of the following features: the calculation unit is arranged to determine a velocity filter data set and a functional filter data set in the storage unit from the at least two parameters determined by the collection unit, to calculate a velocity gain value from at least the gain value and the noise margin value of the determined velocity filter data set and to derive a velocity filter data set therefrom, and to determine a functional gain value from at least the gain value, the velocity gain value and the noise margin value of the determined functional filter data set and to derive a functional filter data set therefrom; the calculation unit is arranged to calculate the velocity gain value by selecting a minimum value of the gain value and the noise margin value of the determined velocity filter data set, and to calculate the functional gain value by selecting a minimum value of the gain value and the difference between the noise margin value and the velocity gain value of the determined functional filter data set; the calculation unit is arranged to determine, in the presence of an engine speed functional parameter and a ventilation data set functional parameter, a primary functional filter data set and a second secondary functional filter data set from the engine speed functional parameter and the ventilation data set functional parameter such that a gain value of the primary functional filter data set is greater than a gain value of the secondary functional filter data set; determine a first functional gain value from at least the gain values, the speed gain value and the noise margin value of the primary functional filter data set and derive a first functional filter data set therefrom; determine a second functional gain value from at least the gain values, the first functional gain value, the speed gain value and the noise margin value of the secondary functional filter data set and derive a second functional filter data set therefrom; calculate a functional gain value and a functional filter data set from the first functional gain value, the second functional gain value, the first functional filter data set and the second functional filter data set; the calculation unit is arranged to determine a functional filter data set and a velocity filter data set in the storage unit from the at least two parameters determined by the collection unit if the velocity parameter is below a predefined threshold, to determine a functional gain value from at least the gain value and the noise margin value of the determined functional filter data set and to derive a functional filter data set therefrom from the functional gain value, and to determine a velocity gain value from at least the gain value, the functional gain value and the noise margin value of the determined velocity filter data set and to derive a velocity filter data set therefrom from the velocity gain value; the computing unit is configured to calculate the functional gain value by selecting a minimum value among the gain value and the noise margin value of the determined functional filter data set, and to calculate the velocity gain value by selecting a minimum value among the gain value of the determined velocity filter data set and a difference between the noise margin value and the functional gain value; the computing unit is arranged to determine, in presence of an engine speed functional parameter and a ventilation data set functional parameter, a primary functional filter data set and a second secondary functional filter data set from the engine speed functional parameter and the ventilation data set functional parameter such that a gain value of the primary functional filter data set is greater than a gain value of the secondary functional filter data set; determine a first functional gain value from at least the gain value and the noise margin value of the primary functional filter data set and derive a first functional filter data set therefrom; determine a second functional gain value from at least the gain value, the first functional gain value and the noise margin value of the secondary functional filter data set and derive a second functional filter data set therefrom; and calculate a functional gain value and a functional filter data set from the first functional gain value, the second functional gain value, the first functional filter data set and the second functional filter data set; The computing unit is arranged to apply a high pass filter to the masking equalisation filter in response to the difference between the noise margin, the velocity gain value and the function gain value.
[0017] The present invention further provides a method for equalizing an audio signal in a vehicle using a data communication bus, comprising the steps of: a) storing a speed filter data set, which consists of three items relating a speed value to a gain value associated with a coefficient defining a speed masking filter, and a functional filter data set, which consists of three items relating an engine RPM value or ventilation value to a gain value associated with a coefficient defining an engine masking filter or ventilation masking filter; b) receiving at least two parameters available on the communication bus defining the vehicle's speed and at least one functional parameter selected from the group consisting of an engine speed of the vehicle and a ventilation data set, and a noise margin value; c) deriving at least one velocity filter data set and at least one functional filter data set from the parameters of step b) and said noise margin value and combining them as a masking equalization filter to be applied to the speech signal; The present invention relates to a method for equalizing an audio signal, comprising:
[0018] In various embodiments, the present invention may include at least one of the following features: - step c) comprises the sub-steps c1) of determining a velocity filter data set and a functional filter data set from the at least two parameters determined by the collection unit, c2) of calculating at least a velocity gain value from the gain value and the noise margin value of the determined velocity filter data set and deriving a velocity filter data set therefrom, and c3) of determining at least a functional gain value from the gain value, the velocity gain value and the noise margin value of the determined functional filter data set and deriving a functional filter data set therefrom; the sub-step c2) comprises calculating the velocity gain value by selecting a minimum value among the gain values and the noise margin value of the determined velocity filter data set, and the sub-step c3) comprises calculating the functional gain value by selecting a minimum value among the gain values and the difference between the noise margin value and the velocity gain value of the determined functional filter data set; said sub-step c3) comprises, if step b) returns engine speed and ventilation data set function parameters, c3a) determining a primary functional filter data set and a second secondary functional filter data set from said engine speed and ventilation data set function parameters such that a gain value of said primary functional filter data set is greater than a gain value of said secondary functional filter data set, c3b) determining a first functional gain value from at least the gain value, the speed gain value and the noise margin value of said primary functional filter data set and deriving therefrom a first functional filter data set, c3c) determining a second functional gain value from at least the gain value, the first functional gain value, the speed gain value and the noise margin value of said secondary functional filter data set and deriving therefrom a second functional filter data set, and c3d) calculating a functional gain value and a functional filter data set from said first functional gain value, said second functional gain value, said first functional filter data set and said second functional filter data set; the step c) comprises, if the speed parameter of the step b) is less than a predefined threshold, a sub-step c1) of determining a functional filter data set and a speed filter data set in the storage unit from the at least two parameters determined by the collecting unit, c2) a sub-step of determining a functional gain value from the gain value and the noise margin value of the determined functional filter data set and deriving a functional filter data set from the functional gain value therefrom, and c3) a sub-step of determining a speed gain value from the gain value, the functional gain value and the noise margin value of the determined speed filter data set and deriving a speed filter data set from the speed gain value therefrom; said sub-step c1) comprising calculating said feature gain value by selecting a minimum value among said gain values and said noise margin value of said determined feature filter data set, and said sub-step c2) comprising calculating said velocity gain value by selecting a minimum value among said gain values and a difference between said noise margin value and said feature gain value of said determined velocity filter data set; the sub-step c2) comprises, if step b) returns engine speed and ventilation dataset function parameters, c2a) determining a primary functional filter dataset and a secondary functional filter dataset from the engine speed and ventilation dataset function parameters such that a gain value of the primary functional filter dataset is greater than a gain value of the secondary functional filter dataset, c2b) determining a first functional gain value from the gain value and the noise margin value of the primary functional filter dataset and deriving a first functional filter dataset therefrom, c2c) determining a second functional gain value from the gain value, the first functional gain value and the noise margin value of the secondary functional filter dataset and deriving a second functional filter dataset therefrom, and c2d) calculating a functional gain value and a functional filter dataset from the first functional gain value, the second functional gain value, the first functional filter dataset and the second functional filter dataset; The method further comprises a step d) of applying a high pass filter to the masking equalization function depending on the noise margin and the difference between the speed gain value and the function gain value.
[0019] The invention further relates to a computer program comprising instructions for carrying out the method according to the invention, a data storage medium having such a computer program recorded thereon, and a computer system comprising a processor connected to a memory having such a computer program recorded thereon.
[0020] Other configurations and advantages of the invention will appear more clearly on reading the following description, given by way of illustrative and non-limiting examples, with reference to the drawings, in which: [Brief description of the drawings]
[0021] [Figure 1] FIG. 1 shows a system for transmitting audio signals in a vehicle embodying the present invention. [Diagram 2] FIG. 2 is a diagram showing an example of the structure of an apparatus for applying equalization by masking to the system of FIG. 1 embodying the present invention. [Diagram 3] FIG. 3 shows a first embodiment of equalization by the device of FIG. 2; [Figure 4] FIG. 4 shows a first embodiment of the equalization of FIG. 3. [Diagram 5] FIG. 4 shows a second embodiment of the equalization of FIG. [Figure 6] FIG. 4 shows a third embodiment of the equalization of FIG. [Figure 7] FIG. 3 illustrates a second embodiment of equalization by the device of FIG. 2; DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0022] The drawings and the following description essentially contain constructions of a certain nature, which means that they are not only for the purpose of facilitating understanding of the invention, but also contribute appropriately to the definition of the invention.
[0023] This description contains material which, by its nature, may be protected by copyright. The holder of such rights has no objection to anyone reproducing this patent document or any of the descriptions therein in the official patent file. All other rights are reserved.
[0024] 1 shows a system 110 for transmitting audio signals implemented with the present invention in a transmitting environment in the form of a vehicle 100. In this figure, the vehicle is depicted in the form of a car, but the present invention has application to any type of motorized vehicle. As will become apparent, the present invention is particularly based on the vehicle 100 using a CAN bus as its data communication bus.
[0025] The emission system 110 comprises a number of loudspeakers 112 and an equalization device 114 according to the invention, configured to perform equalization by masking according to any of the embodiments described in Figure 3 or Figure 7. Figure 2 shows an overview of an example use of the device 114.
[0026] In various embodiments, the equalizer 114 is not integrated into the originating system 110, but rather is connected to the originating system 110 via a wired (e.g., USB connection or equivalent) or wireless (e.g., Bluetooth, Wi-Fi, or equivalent) connection to exchange data, such as the originating audio signal and the equalized audio signal. The originating system 110 may have only one loudspeaker 112.
[0027] 2 shows an overview of an example of use of the equalizer 114. In the example shown in the figure, the equalizer 114 includes a storage unit 200, a collection unit 210, and a calculation unit 220.
[0028] The storage unit 200 may be any type of data storage capable of receiving digital data, such as a hard disk, a hard disk with flash memory, any form of flash memory, random access memory, magnetic disk, local or cloud distributed storage, etc. However, for the automotive application of the present invention, the storage unit 200 is likely to be a flash memory or a hard disk, which can be accessed by the system 110, the collection unit 210 and the calculation unit 220.
[0029] In the example described herein, the storage unit 200 receives all data related to the device 114, i.e., the programs or software instantiating the collection unit 210 and the calculation unit 220, their parameters, data received as input (if any), intermediate filter gain and coefficient values, buffered data, and masking filter data output. Data calculated by the device may be stored in any type of storage unit similar to the storage unit 200 or in the storage unit 200. These data may be erased or retained after the device has performed its task.
[0030] The storage unit 200 receives a speed filter data set, which consists of three entries relating a speed value, a gain value and a speed masking filter coefficient, and a functional filter data set, which consists of three entries relating an engine speed value or ventilation value, a gain value and an engine masking filter coefficient or a ventilation masking filter coefficient. These coefficient sets allow a fast and cost-effective implementation of various filters. The coefficient sets may be generic or vehicle-specific, the latter being preferred in the context of the present invention. In the example shown in the figure, the masking filter used is a biquadratic filter. In a variant, the filter may be of any kind implemented in the time domain, such as a recursive filter, a finite impulse response filter, etc. The filter may also be implemented in the frequency domain.
[0031] The storage unit 200 receives the parameters as input data and at least one noise margin data.
[0032] Said parameters serve to characterize the masking noise phenomenon that is to be filtered by the device 114. Said parameters are obtained by the collection unit 210 from a communication bus, such as a CAN bus, of the vehicle, which is highly advantageous since it carries a large number of information items that allow the quantification of the vehicle's state, such as speed, engine speed, and other functional information, such as parameters of the HVAC system (such as, but not limited to, operational indicators, ventilation speed values, ventilation power values, etc.). Hereafter, speed will represent the speed parameter, and engine speed and parameters of the HVAC system will represent the functional parameters. The reason for this distinction is that speed is a well-known noise source due to the road surface and the aerodynamics and movement of the vehicle, and the associated masking filters are also well-known, whereas the filtering of the masking of the engine and HVAC system noise is not well controlled, since it is desired to calculate it at a reasonable cost. This is why the latter are not treated, for example, in the SDEC method.
[0033] The applicant has surprisingly found that, contrary to all expectations, the CAN can provide all the necessary data to define functional noises such as engine noise, HVAC noise, etc., to augment the SDEC filtering. As will be seen, this differentiation of various noise sources and their characteristics allows a differentiation process that matches the performance of NDEC, but at a much lower cost, is extremely versatile independent of the vehicle platform, and does not require microcontrollers. Moreover, the above simplicity of using the existing CAN (or other communication bus) makes the integration of the device 114 into the vehicle easy.
[0034] That is, to save computational costs, the speed range, the engine speed and the HVAC system data are separated and each is associated with a filter gain and a filter coefficient stored in the memory unit 200. Thus, the three items of the speed value (engine speed or HVAC system data), the gain and the filter coefficient set define a speed filter data set (engine data set or ventilation data set). Such filter data sets are operated by the computation unit 220 to calculate the masking equalization filter. Preferably, these filter data sets are individualized for each vehicle type and are adjusted before the device 114 is used.
[0035] The noise margin data can be of two natures: the amount of attenuation required for the volume of the audio signal, and the maximum gain required for the device 114. Hereinafter, the noise margin refers to the smaller of these two values. As a variant, only one of these two values may be taken into account, in which case the noise margin is equal to the required attenuation or the maximum required gain, depending on the particular case.
[0036] The collecting unit 210 and the computing unit 220 have direct or indirect access to the storage unit 200. The collecting unit 210 and the computing unit 220 may be formed in the form of suitable computer code executed on one or more processors. By processor, any processor suitable for the calculations described below is to be understood. Such a processor may be formed in any known manner, in any form capable of providing the necessary computing power for the completion of the processes described below, such as a microprocessor for a personal computer or laptop or tablet or smartphone, a dedicated chip of the FPGA or SoC type, a computing resource on a grid or cloud, a cluster of graphical processors (GPU), a microcontroller, etc. At least one of these may be formed in the form of a specialized electronic circuit, such as an ASIC. Also, a combination of a processor and an electronic circuit may be considered. Since the invention is in the automotive context, it is desirable for the processor to be as simple as possible in order to optimize the cost of the device 114.
[0037] It is also clear that the only function of the collection unit 210 is to retrieve said parameters and said noise margin to drive the calculation unit 220. As a variant, the collection unit 210 and the calculation unit 220 may be integrated, or the calculation unit 220 may be divided into smaller functional units, provided that the operation of the device 114 remains functionally unchanged.
[0038] Fig. 3 shows a first embodiment of the masking equalization implemented by the device of Fig. 2. In this embodiment, it is assumed that the speed is the most significant noise source. For this reason, its processing is prioritized. Afterwards, the noise on the feature parameters is processed, which, as will be seen, depends on the processing of the speed noise. In the example shown in Fig. 3, it is assumed that there is only one feature parameter. As a variant, it is also possible to process several feature parameters, as will be explained for example in the second masking equalization implementation.
[0039] In the process 300, the collection unit 210 retrieves from the CAN bus the data required for processing by the calculation unit 220, namely the speed of the vehicle V (speed parameter), the engine speed rpm or ventilation data HVAC[], and the noise margin HR. The speed V, the engine speed rpm and the ventilation data HVAC[] can be obtained directly or indirectly from any data traveling on the CAN (or other communication bus of the vehicle).
[0040] Next, in steps 310, 320, and 330, the speed noise is processed. In step 310, the calculation unit 220 calls the storage unit 200 to retrieve a speed filter data set that matches the speed V using the function Filt1() that receives the speed V as an argument. If the value of the speed V does not exist in the speed filter data sets present in the storage unit 200, the function Filt1() may return a speed filter data set with a speed closest to the speed V, or a speed filter data set derived from an extrapolation of at least one speed filter data set. Next, in step 320, the gain LSG is determined using the function Min(). The role of the function Min() is to take into account that the noise margin (i.e., "headroom") is finite in the processing by the calculation unit 220. It is not possible to simply apply correction to the audio signal according to its volume. Moreover, if the audio signal to be reproduced is extremely strong enough to exceed the risk of saturation of the loudspeaker 112, it is considered that there is no need to amplify the signal and filter the masking. Thus, in the example shown in the figure, the function Min() takes the minimum of the noise margin and the gain value of the velocity filter data set of step 310. Finally, if step 320 generates a gain LSG different from the gain of step 310, then in step 330 a function Filt2() is executed in the memory unit 200 to determine a velocity filter data set of a gain GV that best matches the gain LSG. Again, the function Filt2() may return the velocity filter data set of a gain GV that is closest to the gain LSG or a velocity filter data set derived from an extrapolation of at least one velocity filter data set.
[0041] After steps 310-330, the calculation unit 220 performs steps 340-360 to process the noise related to the functional parameters. As described above, in the embodiment of FIG. 3, only one of the functional parameters, engine speed and ventilation data, is processed. Step 340 is similar to step 310, and determines the functional filter data set in the memory 200 that is closest to the functional parameter by performing the function Filt1() using the functional parameter of interest. Next, in step 350, the gain P is calculated taking into account the noise margin, as in step 320. However, in order to take into account the processing of steps 310-340, the gain of the functional filter data set of step 340 is compared not with the noise margin but with the noise margin after cutting the speed gain GV of step 330. Finally, in step 360, the functional filter data set of step 340 is reprocessed according to the functional gain P determined in step 350, and a functional filter data set FP[] is returned.
[0042] Finally, in step 370, the calculation unit 220 executes a function Comb() that combines the velocity gain GV and velocity filter data set FV[] from step 330 with the functional gain GP and functional filter data set FP[] from step 360 to calculate a masking equalization filter MF.
[0043] In a preferred embodiment, optional step 380 modifies the masking equalization filter MF with a high-pass filter to protect the loudspeaker 112 from excessive gain in the low frequency range.
[0044] It will be appreciated that steps 310 and 340 can be performed in parallel, while the subsequent step 350 depends on steps 320 and 330.
[0045] It is clear that the masking equalization filter is constructed by successive keys, taking into account the available noise margin. To further reduce the cost of the invention, the calculation of the filter coefficients may be performed by storing an all-pass filter data set, which is modified according to the gain determined at each stage. For example, the filter data set of step 360 may be determined by the following calculations: b0=0.5*((1 + gainLin)+b0_ap*(1-gainLin)); b1=0.5*((1 + gainLin)*a1_ap+b1_ap*(1-gainLin)); b2=0.5*((1+gainLin)*a2_ap+(1-gainLin)); a1=a1_ap; a2=a2_ap; (where gainLin is the conversion of the function gain value (dB) to a linear gain according to the formula: gainLin=10^(P / 20), ai and bi are the coefficients of the function filter data set FP[], and ai_ap and bi_ap are the coefficients of the all-pass filter data set.)
[0046] FIG. 4 shows an example of a masking equalization filter that is generated when the noise margin is greater than the sum of the velocity gain and the function gain.
[0047] This example is based on the following data: speed: 70 kn / hr (corresponding to a speed filter data set with a gain of 3.3 dB), noise margin: 13 dB, and fundamental frequency of engine noise: 120 Hz (determined from the product of engine speed (rpm) and number of cylinders divided by 120 (another formula may be used) rpm value: 3600; corresponding to a speed filter data set with a gain of 5 dB).
[0048] The Min() function of step 320 returns a velocity gain of 3.3 dB, i.e. the velocity filter data set of step 310 is left unchanged. The functional gain is then determined by comparing the 5 dB gain with 13-3.3, or 9.7 dB, i.e. again the functional filter data set of step 340 is left unchanged. Due to the nature of the engine noise, the coefficients of the engine filter data set are related to harmonics of the fundamental frequency of the engine noise, which is the aforementioned frequency of 120 Hz.
[0049] Finally, the final gain is calculated by subtracting the velocity gain and the function gain from the noise margin, i.e. 13-3.3-5=4.7 dB. This is used as the reference value by the high pass filter in step 380.
[0050] FIG. 5 shows an example of a masking equalization filter that is generated when the noise margin is less than the gain of the velocity filter data set for the vehicle speed.
[0051] This example is based on the following data: speed: 105 kn / hr (corresponding to a speed filter data set with a gain of 4 dB), noise margin: 3.1 dB, and a fundamental frequency value of the engine noise: 120 Hz (corresponding to a speed filter data set with a gain of 5 dB).
[0052] The function Min() in step 320 returns a velocity gain of 3.1 dB. This means that the velocity filter data set in step 310 needs to be modified in step 330. Unfortunately, step 330 returns a velocity filter data set with a velocity gain of 3.1 dB. As a result, step 350 necessarily returns 0 dB (3.1-3.1), making the calculation of the function gain unnecessary. This is not the case for a high-pass filter. In this case, the application of the function filter is eliminated (a zeroth order filter is applied).
[0053] FIG. 6 shows a third example where the noise margin is greater than the gain of the velocity filter data set for the vehicle speed, but less than this speed gain plus the gain of the feature filter data set for the feature parameters.
[0054] This example is based on the following data: speed: 105 kn / hr (corresponding to a speed filter data set with a gain of 4 dB), noise margin: 3 dB, and a fundamental frequency value of the engine noise: 120 Hz (corresponding to a speed filter data set with a gain of 5 dB).
[0055] The function Min() in step 320 returns a velocity gain of 3 dB. This means that the velocity filter data set in step 310 needs to be modified in step 330. In this example, step 330 returns a velocity filter data set with a velocity gain of 2 dB. As a result, the function gain is calculated as Min(5;3-2)=1 dB. In step 360, the function filter is determined accordingly. Since the noise margin of 3 dB is equal to the sum of the velocity gain (2 dB) and the function gain (1 dB), a zero-order high-pass filter is applied.
[0056] The above is a suitable example, but it is possible to understand how taking into account the noise margin can significantly change the definition of the masking equalization filter.
[0057] FIG. 7 shows an embodiment illustrating the process when the function speed parameter and the ventilation data function parameter are present, as an alternative to allow for subtlety.
[0058] In this embodiment, the same processes as those in Fig. 3 have the same last two digits as those in the reference numerals in Fig. 3. Only the differences will be described.
[0059] The main difference in this embodiment is that the speed threshold is taken into account to determine the order of processing of the speed noise relative to the functional noise. As with the noise margin in Fig. 3, each processing process processes the residual noise starting from the residual margin.
[0060] At high speeds, aerodynamic noise is very loud and needs to be prioritized over functional noise, whereas at low speeds, e.g. below 50 km / h, it may be advantageous to reverse this process.
[0061] Thus, in step 705, the velocity V is compared to a threshold value Vm. If this comparison indicates a high velocity, then in step 709 the processing described in FIG. 3 is performed (with the addition of processing of two functional parameters, described below).
[0062] Otherwise, the functional noise is processed preferentially. In the embodiment shown in the figure, two different types of functional noise are processed: engine noise and ventilation noise. To determine which one is to be processed preferentially, in step 710, similar to step 310, a function Filt3() is executed to extract the engine filter data set that matches the rpm functional parameter and the ventilation filter data set that matches the ventilation data functional parameter HVAC[]. These two sets are separated based on the gains associated with them. In steps 720 and 722, a first functional gain PG1 and a set of coefficients thereof are determined similarly to step 320, and in steps 724 and 726, a second functional gain PG2 and a set of coefficients thereof are determined taking the first functional gain PG1 into account. Finally, in step 730, the functional gain coefficients PG1 and PG2 and their corresponding sets of coefficients are combined. In this way, a hierarchical process is iteratively performed for functional parameters of various natures while decreasing the noise margin.
[0063] Next, in steps 740, 750 and 760, the velocity gain and velocity filter data sets are determined in a manner similar to steps 340-360, except that the roles of "function" and "velocity" are swapped. A masking equalization filter is calculated in step 770, and a high pass filter is applied in optional step 780.
[0064] Above, the filter data set is formed by a triplet associating the index values (speed, engine speed and ventilation data) with the gain values associated with the coefficients that jointly define the masking filter. However, the gain value / coefficient pairs may also be represented by the coefficients alone, with the gain value being implicitly included in the norm of the coefficients.
[0065] In other words, as mentioned above, the applicant of the present application has developed a process that takes into account various types of noise, extracts information about the noise from a direct source available (the communication bus), and is able to calculate a masking equalization filter at a very low computational cost. The fact that such a result can be obtained starting from SDEC was unexpected, as can be seen from the applicant's application for a patent on NDEC (FR2211921).
Claims
1. 1. An audio signal equalization apparatus for a vehicle using a data communication bus, comprising: a storage unit (200) arranged to store a speed filter data set, which consists of three items relating a speed value to a gain value associated with a coefficient defining a speed masking filter, and a functional filter data set, which consists of three items relating an engine speed value or ventilation value to a gain value associated with a coefficient defining an engine masking filter or ventilation masking filter, a collection unit (210) arranged to receive at least two parameters available on said communication bus defining the speed of said vehicle and at least one functional parameter selected from the group consisting of engine speed and ventilation data sets of said vehicle; a calculation unit (220) arranged to receive as input a noise margin value and said at least two parameters available on said communication bus, to derive therefrom at least one velocity filter data set and at least one functional filter data set and to combine them as a masking equalization filter to be applied to said audio signal; An apparatus comprising:
2. 2. The apparatus according to claim 1, wherein the computing unit (220) comprises: - determining a velocity filter data set and a functional filter data set in said storage unit (200) from said at least two parameters determined by said collection unit (210); - calculating a velocity gain value from at least the determined gain value and the noise margin value of the velocity filter data set and deriving a velocity filter data set therefrom; determining functional gain values from at least the gain values, the speed gain values and the noise margin values of the determined functional filter data set and deriving functional filter data sets therefrom, A device is provided.
3. 3. The device according to claim 2, wherein the calculation unit (220) is arranged to calculate the velocity gain value by retaining a minimum value among the gain values and the noise margin value of the determined velocity filter data set, and to calculate the functional gain value by selecting a minimum value among the gain values of the determined functional filter data set and a difference between the noise margin value and the velocity gain value.
4. 4. The device according to claim 2 or 3, wherein the calculation unit (220) is adapted to calculate, in the presence of the engine speed function parameter and the ventilation data set function parameter, determining a primary functional filter data set and a second secondary functional filter data set from said engine speed functional parameters and said ventilation data set functional parameters such that a gain value of said primary functional filter data set is greater than a gain value of said secondary functional filter data set, - determining first feature gain values from at least the gain values, the speed gain values and the noise margin values of the main feature filter data set and deriving a first feature filter data set therefrom; - determining second function gain values from at least the gain values of the secondary function filter data set, the first function gain values, the speed gain values and the noise margin value, and deriving a second function filter data set therefrom; - calculating a feature gain value and a feature filter data set from said first feature gain value, said second feature gain value, said first feature filter data set and said second feature filter data set, A device is provided.
5. 2. The apparatus according to claim 1, wherein the calculation unit (220) performs the following when the speed parameter is less than a predetermined threshold: - determining a functional filter data set and a velocity filter data set in said storage unit (200) from said at least two parameters determined by said collection unit (210), - determining functional gain values from at least the gain values and the noise margin values of the determined functional filter data set and deriving functional filter data sets therefrom from the functional gain values, determining a velocity gain value from the determined gain values, the functional gain value and the noise margin value of the velocity filter data set and deriving a velocity filter data set therefrom from said velocity gains, A device is provided.
6. 6. The apparatus according to claim 5, wherein the calculation unit (220) is arranged to calculate the functional gain value by selecting a minimum value among the gain values and the noise margin value of the determined functional filter data set, and to calculate the velocity gain value by selecting a minimum value among the gain values and the difference between the noise margin value and the functional gain value of the determined velocity filter data set.
7. 7. The device according to claim 5 or 6, wherein the calculation unit (220) is adapted to calculate, in the presence of the engine speed function parameter and the ventilation data set function parameter, determining a primary functional filter data set and a secondary functional filter data set from said engine speed functional parameters and said ventilation data set functional parameters such that a gain value of said primary functional filter data set is greater than a gain value of said secondary functional filter data set, - determining first feature gain values from at least the gain values and the noise margin values of the primary feature filter data set and deriving a first feature filter data set therefrom, - determining second function gain values from at least the gain values of the secondary function filter data set, the first function gain values and the noise margin value, and deriving a second function filter data set therefrom; - calculating a feature gain value and a feature filter data set from said first feature gain value, said second feature gain value, said first feature filter data set and said second feature filter data set, A device is provided.
8. 8. The device according to claim 1, wherein the calculation unit (220) is arranged to apply a high pass filter to the masking equalisation filter depending on the difference between the noise margin, the velocity gain value and the function gain value.
9. 1. A method for equalizing an audio signal in a vehicle using a data communication bus, comprising: a) storing a speed filter data set, which consists of three items relating a speed value to a gain value associated with a coefficient defining a speed masking filter, and a functional filter data set, which consists of three items relating an engine speed value or ventilation value to a gain value associated with a coefficient defining an engine masking filter or ventilation masking filter; b) receiving at least two parameters available on the communication bus defining the speed of the vehicle and at least one functional parameter selected from the group consisting of an engine speed of the vehicle and a ventilation data set, as well as a noise margin value; c) deriving at least one velocity filter data set and at least one functional filter data set from the parameters of step b) and said noise margin value and combining them as a masking equalization filter to be applied to the speech signal; A method for equalizing an audio signal comprising:
10. 10. The method for equalizing an audio signal according to claim 9, wherein step c) comprises: c1) a sub-step of determining a velocity filter data set and a function filter data set from said at least two parameters determined by said collection unit (210); c2) at least the sub-step of calculating a velocity gain value from the determined gain value and the noise margin value of the velocity filter data set and deriving a velocity filter data set therefrom; and c3) the sub-step of determining functional gain values from at least the gain values, the velocity gain values and the noise margin values of the determined functional filter data set and deriving a functional filter data set therefrom; 23. A method for equalizing an audio signal, comprising:
11. 11. A method for equalizing an audio signal as claimed in claim 10, wherein the sub-step c2) comprises calculating the velocity gain value by selecting the minimum value of the gain values of the determined velocity filter data set and the noise margin value, and wherein the sub-step c3) comprises calculating the functional gain value by selecting the minimum value of the gain values of the determined functional filter data set and the difference between the noise margin value and the velocity gain value.
12. 12. The method for equalizing an audio signal according to claim 10 or 11, wherein the sub-step c3) comprises, if the step b) returns an engine speed function parameter and a ventilation data set function parameter, c3a) determining a primary functional filter data set and a second secondary functional filter data set from said engine speed functional parameters and said ventilation data set functional parameters such that a gain value of said primary functional filter data set is greater than a gain value of said secondary functional filter data set; c3b) determining first feature gain values from at least the gain values, the speed gain values and the noise margin values of the primary feature filter data set and deriving a first feature filter data set therefrom; c3c) determining second function gain values from at least the gain values of the secondary function filter data set, the first function gain value, the speed gain value and the noise margin value, and deriving a second function filter data set therefrom; and c3d) calculating a feature gain value and a feature filter data set from the first feature gain value, the second feature gain value, the first feature filter data set and the second feature filter data set; 13. A method for equalizing an audio signal comprising:
13. 10. The method for equalizing an audio signal according to claim 9, wherein, when the speed parameter of the step b) is less than a predetermined threshold, the step c) further comprises: c1) a sub-step of determining a functional filter data set and a velocity filter data set in said storage unit (200) from said at least two parameters determined by said collection unit (210); c2) at least the sub-step of determining functional gain values from the gain values and the noise margin values of the determined functional filter data set, and deriving functional filter data sets therefrom from the functional gain values; and c3) at least the sub-step of determining a velocity gain value from the determined gain values of the velocity filter data set, the function gain value and the noise margin value, and deriving a velocity filter data set therefrom from the velocity gain value; 23. A method for equalizing an audio signal, comprising:
14. 14. A method for equalizing an audio signal as claimed in claim 13, wherein the sub-step c1) comprises calculating the functional gain value by selecting the minimum value among the gain values of the determined functional filter data set and the noise margin value, and wherein the sub-step c2) comprises calculating the velocity gain value by selecting the minimum value among the gain values of the determined velocity filter data set and the difference between the noise margin value and the functional gain value.
15. 15. The method for equalizing an audio signal according to claim 13 or 14, wherein the sub-step c2) comprises, if the step b) returns an engine speed function parameter and a ventilation data set function parameter, c2a) determining a primary functional filter data set and a secondary functional filter data set from said engine speed functional parameters and said ventilation data set functional parameters such that a gain value of said primary functional filter data set is greater than a gain value of said secondary functional filter data set; c2b) determining first feature gain values from at least the gain values and the noise margin values of the primary feature filter data set and deriving a first feature filter data set therefrom; c2c) determining second function gain values from at least the gain values of the secondary function filter data set, the first function gain values and the noise margin value, and deriving a second function filter data set therefrom; and c2d) calculating a feature gain value and a feature filter data set from the first feature gain value, the second feature gain value, the first feature filter data set and the second feature filter data set; 13. A method for equalizing an audio signal comprising:
16. The method for equalizing an audio signal according to any one of claims 9 to 15, further comprising: d) applying a high pass filter to the masking equalization function in response to the noise margin and the difference between the speed gain value and the function gain value; A method for equalizing an audio signal comprising:
17. A computer program for use by a computer comprising instructions for carrying out the method according to any one of claims 9 to 16.
18. 18. A data storage medium having a computer program according to claim 17 recorded thereon.
