Method for suppressing interference frequency components from a microphone signal recorded by means of an external microphone of a vehicle

Abstract

The invention relates to a method for suppressing interference frequency components from a microphone signal recorded by means of an external microphone of a vehicle, in which a driving speed (v) of the vehicle is determined during the recording of the microphone signal and in which a frequency range of the microphone signal is suppressed by means of a bandstop filter, wherein the bandstop filter has a variable center frequency (f0) which is set to a value proportional to the driving speed depending on the detected driving speed (v).

Description

[0001] The invention relates to a method for suppressing interference frequency components from a microphone signal recorded by an external microphone of a vehicle according to the preamble of patent claim 1.

[0002] Such external microphones are typically used in vehicles to capture and evaluate acoustic signals from the vehicle's surroundings, in particular to detect horns from emergency vehicles or to recognize voice instructions from people.

[0003] The microphone signal of such an external microphone has a frequency spectrum that, in addition to useful frequency components that are to be recorded and evaluated, also contains interference frequency components that are undesirable because they make the evaluation of the useful frequency components more difficult.

[0004] The interference frequency components can come from ambient noise, but they can also come from noise generated by the vehicle itself while driving.

[0005] DE 10 2019 205 694 A1 discloses a method and arrangement for noise suppression in audio signals in a vehicle. The method comprises a noise suppression filter that at least partially suppresses a noise component in an audio signal, wherein at least one property of the noise suppression filter is determined as a function of a received vehicle speed value.

[0006] DE 10 2022 201 395 A1 discloses a method for determining the direction of origin of sound signals using a plurality of microphones. The sound detected by the microphones is filtered using a velocity-dependent filter.

[0007] The invention is based on the object of suppressing interference frequency components from the microphone signal that are generated by the vehicle itself.

[0008] The object is achieved according to the invention by the features of patent claim 1. Advantageous embodiments and further developments emerge from the subclaims.

[0009] According to the method, interference frequency components are suppressed from a microphone signal recorded by an external microphone of a vehicle by determining a driving speed of the vehicle during the recording of the microphone signal and suppressing a frequency range of the microphone signal by means of a band-stop filter, wherein the band-stop filter has a variable center frequency which is set to a value proportional to the driving speed as a function of the recorded driving speed.

[0010] The center frequency of a band-stop filter is understood to be the center of a frequency range to be blocked by the band-stop filter.

[0011] According to the invention, the center frequency of the band-stop filter is set such that it corresponds to a product of the driving speed of the vehicle and a vehicle-specific proportionality factor.

[0012] The determination of the vehicle-specific proportionality factor is preferably based on a recording and evaluation of a frequency spectrum of sound intensities, which are recorded with the external microphone during a measurement in a wind tunnel at different wind speeds.

[0013] Furthermore, the evaluation of the frequency spectrum according to the invention comprises an analysis of the frequency spectrum with regard to the presence of a predetermined frequency pattern in the frequency spectrum.

[0014] The predetermined frequency pattern is defined according to the invention as a pattern which has local amplitude peaks of the sound intensity, the frequency of which changes proportionally to the wind speed and the width of which remains essentially the same when the wind speed changes, wherein the proportionality factor (K) corresponds to a slope of a regression line interpolated to the amplitude peaks from this region of the frequency pattern.

[0015] In an advantageous development of the method, it is provided that upon identification of a range of the frequency spectrum corresponding to the predetermined frequency pattern, the frequencies of the amplitude peaks from this range and the associated wind speeds are selected, that an equation of a linear regression line is created for the selected frequencies and wind speeds, which approximately indicates how the selected frequencies depend on the wind speeds, and that the vehicle-specific proportionality factor is set to a value that corresponds to a slope of the regression line.

[0016] In a further advantageous embodiment of the method, it is provided that an average width of the amplitude peaks whose frequencies have been selected is determined and used as the basis for determining the bandwidth of the band-stop filter.

[0017] The method according to the invention is used in a vehicle having an external microphone for recording a microphone signal and a band-stop filter for filtering the microphone signal, in order to dynamically vary a frequency range to be blocked by the band-stop filter as a function of the driving speed of the vehicle.

[0018] An exemplary embodiment of the invention is described in more detail below with reference to a figure. The single figure shows a frequency spectrum of a microphone signal recorded with an external microphone of a vehicle in a wind tunnel at different wind speeds.

[0019] The invention is based on the finding that raised structural parts on the body of a vehicle, for example struts of a roof rail, holders of exterior mirrors or an antenna, can act as disruptive bodies which can lead to the formation of so-called Kärmännchen air vortices while driving, which in turn can lead to the formation of a whistling noise whose frequency is proportional to the driving speed.

[0020] With the method according to the invention, such whistling noises can be easily eliminated from the frequency spectrum of the microphone signal by filtering using a band-stop filter.

[0021] The band-stop filter is designed to block a frequency range that is dynamically varied depending on the detected vehicle speed. A center frequency of the frequency range to be blocked is determined according to the equation f0=K⋅v where f0 represents the center frequency of the band-stop filter, ie the center of the frequency range to be blocked by the band-stop filter, v represents the recorded vehicle speed, and K represents a vehicle-specific proportionality factor.

[0022] The vehicle-specific proportionality factor K is determined by a measurement in a wind tunnel.

[0023] For this purpose, the vehicle or a vehicle of the same type is brought into the wind tunnel and the microphone signal of the vehicle's external microphone is measured in the wind tunnel at different wind speeds. w recorded.

[0024] From the recorded microphone signal, the frequency spectrum of the microphone signal at the different wind speeds v w The frequency spectrum represents the wind speeds recorded with the outdoor microphone at different wind speeds v wand different frequencies f recorded sound intensities p. The recorded frequency spectrum is provided as a graphical representation.

[0025] The figure shows a simplified schematic representation of such a frequency spectrum. The wind speed v is shown on the abscissa of this graphical representation of the frequency spectrum. w The frequency f is plotted on the ordinate, and the areas F1 to F5 with the different shades indicate the sound intensities p recorded in these areas. The legend L shows which shade corresponds to which sound intensity p. The boundary lines of the areas F1 to F5 can be interpreted as contour lines representing positions with the same sound intensity p.

[0026] It can therefore be seen from the graphic representation that the sound intensity p is higher in area F5 than in area F4, that the sound intensity p is higher in area F4 than in area F3, that the sound intensity p is higher in area F3 than in area F2 and that the sound intensity p is higher in area F2 than in area F1.

[0027] It can be seen from the figure that the area F4 has a narrow, strip-like, and linearly increasing profile. This means that the area F4 shows a frequency pattern characterized by local amplitude peaks of the sound intensities, whose frequency f is proportional to the wind speed v. w changes and whose width changes over the wind speed v w remains essentially constant. This is the typical signature of whistling sounds caused by Käräman air vortices.

[0028] To determine the proportionality factor K, the frequency spectrum is analyzed for the presence of the above-mentioned signature. This means that it is determined whether the frequency spectrum exhibits a frequency pattern characterized by local amplitude peaks of the sound intensities p, whose frequency f is proportional to the wind speed v. w changes and whose width remains essentially constant over speed.

[0029] If a region with such a frequency pattern is detected (in the figure, this is region F4), a regression line G (shown as a dashed line in the figure) is interpolated to the amplitude peaks from this region. The slope of this line G corresponds to the desired proportionality factor K.

[0030] In other words: In the graphical representation of the frequency spectrum, areas with high sound intensities are sought, which are strip-shaped and have a relative wind speed v w have a linearly increasing curve. If such a region is found (in the figure, this is region F4), a regression line G is interpolated into this region, and the slope of this line is used as the vehicle-specific proportionality factor K to determine the center frequency f0 of the band-stop filter according to equation (1).

[0031] The searched area and the regression line G can be found using image processing methods.

[0032] It is further proposed that the mean width B of this detected region F4 be determined, and based on this, the bandwidth of the bandstop filter be determined. The bandwidth of the bandstop filter can be set to the value of the mean width B or to a value proportional to it, depending on how narrow the suppression of the whistling noise of the Käräman air vortices is desired.

[0033] The parameters of the band-stop filter, i.e. the center frequency and the bandwidth, are thus determined by evaluating the frequency spectrum of the microphone signal recorded in the wind tunnel.

[0034] If the identified area shows the above-mentioned signature only in a certain speed range, for example only at a wind speed v wof more than 50 km / h, the band-stop filter is advantageously controlled in such a way that it is only active at driving speeds v that lie within this specific speed range.

[0035] If several ranges with the above-mentioned typical signature of whistling noises can be identified in the frequency spectrum, then it is advantageous to consider only the range with the more dominant amplitude peaks, i.e., with the higher sound intensities p. However, it is also conceivable that a separate band-stop filter is provided for each of these ranges to suppress the interfering frequency components.

[0036] The figure shows that the sound intensities p are maximum in the F5 range. However, this F5 range does not exhibit the typical signature mentioned above and is therefore not considered in this method. A different method would have to be used to treat such interference frequencies, but this method is not the subject of this invention.

Claims

[1] Method for suppressing interference frequency components from a microphone signal recorded by an external microphone of a vehicle, wherein - a driving speed (v) of the vehicle is determined during the recording of the microphone signal - a frequency range of the microphone signal is suppressed by means of a band-stop filter, wherein the band-stop filter has a variable center frequency (f0) which is set to a value proportional to the driving speed as a function of the detected driving speed (v), characterized bythat the center frequency (f0) is set such that it corresponds to a product of the driving speed (v) and a vehicle-specific proportionality factor (K), wherein the determination of the vehicle-specific proportionality factor (K) is based on a recording and evaluation of a frequency spectrum of sound intensities (p) which are recorded with the external microphone during a measurement in a wind tunnel at different wind speeds (v w ), wherein the evaluation of the frequency spectrum comprises an analysis of the frequency spectrum with regard to the presence of a predetermined frequency pattern in the frequency spectrum, and wherein the predetermined frequency pattern is defined as a pattern having local amplitude peaks of the sound intensity (p) whose frequency (f) is proportional to the wind speed (v w ) and whose width changes with a change in wind speed (v w) remains essentially the same, whereby upon identification of a range (F4) of the frequency spectrum corresponding to the given frequency pattern, the frequencies (f) of the amplitude peaks from this range (F4) and the associated wind speeds (v w ) and for the selected frequencies (f) and wind speeds (v w ) an equation of a linear regression line (G) is created, which approximately shows how the selected frequencies (f) depend on the wind speeds (v w ), whereby the vehicle-specific proportionality factor (K) is set to a value corresponding to a slope of the regression line (G). [2] Method according to claim 1, characterized by that a mean width (B) of the amplitude peaks whose frequencies (f) have been selected is determined and used as the basis for determining the bandwidth of the band-stop filter. [3] Vehicle with an external microphone for receiving a microphone signal and with a band-stop filter for filtering the microphone signal, wherein the frequency range to be blocked by the band-stop filter is varied dynamically as a function of the driving speed (v) of the vehicle using a method according to one of the preceding claims.

Images