WO2019243068A1

WO2019243068A1 - Thermoacoustic sound transducer device, audio signal processing apparatus for a thermoacoustic sound transducer, and method for controlling a sound transducer

Info

Publication number: WO2019243068A1
Application number: PCT/EP2019/064771
Authority: WO
Inventors: Lukas WEBER; Alexander Buschle; Hans-Martin Dudenhausen
Original assignee: Sonic Technology Ug (Hb)
Priority date: 2018-06-19
Filing date: 2019-06-06
Publication date: 2019-12-26
Also published as: DE102018114620A1

Abstract

The invention relates to an audio signal processing apparatus (10) for a thermoacoustic sound transducer (20), the audio signal processing apparatus being designed to receive an audio signal (11) and to generate a control signal (12) for the thermoacoustic sound transducer (20) on the basis of the audio signal, wherein: the control signal also has an offset (62); the offset is a time-variable offset (62); and the audio signal processing apparatus (10) is designed to determine the time-variable offset, at least in parts, on the basis of the amplitude curve of the audio signal. The invention further relates to a thermoacoustic sound transducer device (30), comprising an audio signal processing apparatus (10) of this type and a thermoacoustic sound transducer (20), and to a method (100) for controlling a thermoacoustic sound transducer (20).

Description

THERMOACOUSTIC SOUND TRANSDUCER DEVICE,

AUDIO SIGNAL PROCESSING DEVICE FOR A THERMOACOUSTIC SOUND TRANSDUCER AND METHOD FOR STARTING A SOUND TRANSDUCER

The present invention relates to a thermoacoustic transducer device, a

Audio signal processing device for a thermoacoustic sound converter, a method for controlling a thermoacoustic sound converter, and a corresponding computer program.

[0002] Thermoacoustic sound transducers are known. The physical basics of

Thermoacoustic sound conversion is described, for example, in Daschewski et al. “Physics of thermo-acoustic sound generation”, Journal of Applied Physics, 1 14, 1 14903 pages 1-12, 2013.

In a conventional loudspeaker, a membrane is set in vibration. The most widespread design is the so-called moving coil loudspeaker, in which the membrane is oscillated through a centrally positioned voice coil movement is moved up and down or back and forth. The coil is controlled electrically and vibrates in the field of a surrounding magnet. The vibrations are mechanically transmitted to the surrounding medium, typically air.

In contrast, a thermoacoustic transducer device uses the

thermoacoustic effect. Thermal energy is converted into the vibrational energy of a medium, typically air. However, it can also be another gaseous, liquid, gel or solid medium.

Advantages of thermoacoustic sound transducers can be that none

moving parts are required and that a very flat design can be achieved.

An exemplary thermoacoustic sound transducer is in US 8,019,099B2

described. The thermoacoustic sound transducer disclosed therein has a plurality of layers made of carbon nanotubes.

Another thermoacoustic sound transducer based on porous silicon is made

Koshida, "Porous Silicon Acoustic Devices", Handbook of Porous Silicon, Springer International Publishing AG, 2017. The low thermal conductivity and low volumetric heat capacity of porous silicon make it possible to provide thermoacoustic sound transducers with advantageous energy efficiency.

Against this background, an object of the present disclosure may be to further improve sound generation with thermoacoustic sound transducers. In particular, it is desirable to further improve the energy efficiency of a system for thermoacoustic sound conversion, in particular in the case of complex audio signals.

According to a first aspect of the present disclosure, an audio signal processing device for a thermoacoustic sound transducer is proposed, the audio signal processing device being designed for:

Receiving an audio signal; Generating a control signal for the thermoacoustic sound transducer based on the audio signal, the control signal also having an offset,

wherein the offset is a time-variable offset, and wherein the audio signal processing device is designed to determine or adapt the time-variable offset at least in sections based on an amplitude profile (in particular based on amplitude maxima) of the audio signal.

[0010] According to a second aspect of the present disclosure, a thermoacoustic

Sound transducer device with the above-mentioned audio signal processing device and a thermoacoustic sound transducer proposed.

[0011] According to a third aspect of the present disclosure, a (computer-implemented) method for controlling a thermoacoustic sound transducer is proposed, the method comprising the following steps:

Receiving an audio signal;

Generating a control signal for the thermoacoustic sound transducer based on the audio signal, the control signal also having an offset,

the offset being a time-variable offset which is determined at least in sections based on an amplitude profile of the audio signal.

[0012] According to further aspects of the present disclosure, a corresponding

A computer program comprising instructions that, when a computer executes the program, cause the computer to perform the above method; and a corresponding storage medium or a data carrier with such a computer program is proposed. It goes without saying that the proposed sound converter device, the method, computer program and storage medium can be further developed in accordance with the audio signal processing device according to the first aspect.

[0013] The principle of operation of thermoacoustic sound transducers differs

fundamentally from conventional voice coil speakers. In moving coil speakers, as described at the beginning, a membrane is set in motion by a voice coil in an oscillating up and down or back and forth movement. Without power, the voice coil is in a neutral position. A positive amplitude of the control signal causes a positive deflection of the loudspeaker diaphragm in a first direction, for example a forward movement. A negative amplitude of the control signal causes a negative deflection of the loudspeaker diaphragm in a second direction, for example a backward movement. A mean value-free control signal, which oscillates around the zero point, thus produces an oscillating forward and backward movement of the loudspeaker diaphragm around the neutral position and thus a sound signal which is directly proportional to the control signal.

The principle of action of a thermoacoustic sound transducer, however, is based on the fact that the surrounding or adjacent medium is alternately heated or cooled. When heated, the adjacent medium expands and thus generates a positive sound pressure or a positive proportion of sound waves. When it cools down, the medium contracts and generates a negative proportion of sound waves. In principle, the medium can be actively heated (using a heating element) and actively cooled again (using a cooling element). However, since active cooling would involve a great deal of effort, it is easier to simply actively heat the medium and allow it to cool passively. In this case, a maximum negative amplitude is reached when there is no active heating, for example when no control signal or current flows through the thermoacoustic sound transducer. For example, one from Koshida et al. known sound transducer with porous silicon can be used.

[0016] In a first approximation, a thermoacoustic sound transducer can be regarded as a resistance element. When current flows through the thermoacoustic sound transducer, the thermoacoustic sound transducer heats up (active) and leads to an expansion of the surrounding medium and thus to a positive sound wave component. If no current flows through the thermoacoustic transducer, the thermoacoustic transducer cools down (passively). A maximum positive amplitude and a maximum negative amplitude thus result from the difference between a maximum heating and a minimum heating. In this case, the minimum heating can be achieved without current supply or a control signal of zero. By the thermoacoustic see sound transducer is driven by a control signal with an offset, both positive and negative amplitudes of the sound signal can be achieved. In the case of a sinusoidal audio signal, the sound signal to be generated can oscillate between a maximum positive amplitude and a maximum negative amplitude.

The inventors have recognized that a (in terms of amount) maximum positive or negative

However, the amplitude of complex audio signals only occurs in limited time windows. In particular in the case of music reproduction with a high dynamic range, the sound pressure required to be generated can vary. In order to be able to cover the entire dynamic range, the offset must, however, be selected such that the maximum positive or negative amplitudes to be generated can be achieved.

The inventors have also recognized that the offset required for this at

thermoacoustic sound transducers - in contrast to a moving coil loudspeaker - would lead to considerable power loss. It goes without saying that this problem does not exist with conventional moving coil loudspeakers, since the loudspeaker membrane is in the middle in a neutral position and experiences a maximum positive or negative deflection from there. However, a high power loss is disadvantageous, in particular in battery-operated mobile terminal devices, in which a thermoacoustic loudspeaker could advantageously be used due to the lack of moving parts and its low overall height.

However, the inventors have also recognized that a maximum offset does not have to be permanently provided, however. Instead, it is proposed that the offset is a time-variable offset, which is determined at least in sections based on an amplitude profile of the audio signal.

In other words, the audio signal processing device can be set up to control the thermoacoustic sound transducer in such a way that the offset for the control signal is determined and adapted in sections based on the amplitude profile, in particular based on the amplitude maxima or minima, of the audio signal to be reproduced , For example, for sections of the Audio signal with (in terms of amount) smaller positive or negative amplitudes

lower offset can be selected. Accordingly, a larger offset can be selected for sections of the audio signal with larger positive or negative amplitudes.

A mean offset in the reproduction of a complex audio signal can thus preferably be reduced. Since the power loss is proportional to the offset, the power loss can be reduced. Particularly in the case of battery-operated devices with a thermoacoustic loudspeaker, the runtime can be extended and the energy efficiency improved.

[0021] This is particularly true when playing complex audio signals with high

Dynamic range. The offset can be a time-variable offset, which is adjusted based on the maximum amplitude of the audio signal in order to ensure increased efficiency.

Under an audio signal processing device within the scope of the present

Disclosure a sound signal processing device can be understood. An audio signal can be understood to mean a sound signal or information about a sound signal to be generated or a (complex) sound wave to be generated. In particular, an audio signal within the scope of the present disclosure can also include infrared and / or ultrasound, in particular a frequency range from 10 Hz to 800 MHz, in particular audible sound in the range between 20 Hz and 20 kHz. An audio signal can thus relate to both an infrasound band, an audible sound band and / or an ultrasound band.

Under a control signal can within the scope of the present disclosure

Control signal can be understood, on the basis of which the thermoacoustic sound converter generates a sound signal. It goes without saying that the control signal can be fed directly to a thermoacoustic sound transducer or that further signal processing elements, for example an amplifier stage, can be interposed. It can be an analog or digital audio signal and / or control signal. [0024] The audio signal processing device can be set up to adapt the time-variable offset based on local maxima and / or minima of the audio signal. In other words, the determination of the offset can be based on local maxima and / or minima of the sections. An advantage of this embodiment can be that the determination and adjustment of the offset can be carried out in a simple manner and requires only a small amount of calculation.

[0025] The audio signal processing device can be set up to generate the control signal based on a root from (a sum of) the audio signal and the time-variable offset. Distortion of the sound signal during the reproduction of the audio signal can thus be reduced or avoided. In particular, the occurrence of higher harmonics or harmonics can be reduced. Particularly in the case of complex audio signals, the approach known from the prior art of driving a thermoacoustic sound transducer with half the frequency of a desired sound signal to be generated no longer leads to a distortion-free sound signal.

[0026] The audio signal processing device can be set up in such a way that the

time-variable offset is selected such that the control signal only accepts positive values or only negative values. Since a thermoacoustic sound transducer can be regarded as a resistance element in a first approximation, both positive and negative values of the drive signal can lead to a positive power output, i.e. Warming and thus lead to a positive sound amplitude. This can result in a frequency doubling, for example. Such a frequency doubling can be avoided by the control signal assuming - in sections in each case - only positive values or only negative values. A value of zero can be contained in the positive or negative values.

[0027] The audio signal processing device can be set up in such a way that the

time-variable offset is determined at least in sections based on an envelope of the audio signal. An envelope can be determined based on the amplitude maxima or amplitude minima of the audio signal. [0028] The audio signal processing device can be set up in such a way that the

time-variable offset is determined such that an adaptation speed of the time-variable offset is below a first cut-off frequency (or lower useful frequency), in particular below a frequency range of the audio signal. The adaptation speed can thus preferably be selected such that a desired application area or a desired (useful) frequency spectrum is not adversely affected. If, for example, an adaptation with a frequency of less than 20 Hz or with a corresponding time constant is selected for hearing sound, the hearing experience is not adversely affected. Nevertheless, the offset can be reduced on average and thus the average power consumption and power loss. In other words, the offset can be adapted, in particular, with or based on a time constant, the reciprocal of which lies below a (useful) frequency range of the audio signal.

[0029] The audio signal processing device can be set up in such a way that the

time-variable offset is adjusted at least in sections such that the amount of the offset is smaller than a difference between an average value of the audio signal and a global maximum or minimum of the audio signal. A global maximum or minimum can be known for stored or known audio signals. The offset can be selected for each section such that it is not greater, preferably for a plurality of sections, smaller than the difference between the mean value of the audio signal and the global maximum or minimum of the audio signal. As a result, an average power loss can be reduced.

[0030] The audio signal processing device can be set up so that the

Control signal is mapped by 1-bit digital signals. In particular, the audio signal processing device can have a first-order or higher-order delta-sigma modulator.

[0031] The audio signal processing device can be a converter device for generating a PWM (pulse width modulation) control signal or PDM (pulse density modulation) control signal, optionally based on the original control signal. nal have. In particular, the converter device can be a pulse width modulator or pulse density modulator.

[0032] The audio signal processing device can have a filter, in particular a filter

Low-pass filter, to provide a filtered drive signal. If a 1-bit converter is used, a subsequent low-pass filter can be used.

[0033] The audio signal can be an audio signal free of mean values. The

Audio signal processing can thus be set up to process an average-free audio signal. In other words, an average value of the audio signal can preferably be zero. In particular, it can thus be an audio signal which is used to control a conventional loudspeaker or a non-thermoacoustic sound transducer. Optionally, the audio signal processing device is set up to first generate an average-free audio signal based on a received audio signal. The audio signal processing device can then be set up to generate the control signal based on the averaged audio signal.

[0034] The audio signal processing device can be set up so that the

Determining the offset comprises the following steps:

Determining a peak value, for example determining a local maximum or minimum, of the audio signal in a first time interval i;

Comparing the peak value with a peak value of the audio signal in a previous time interval i-1, for example comparing local maxima or minima; and

if the peak value in the first time interval is smaller than the peak value in the previous time interval, decrease the offset; or

if the peak value in the first time interval is greater than the peak value in the previous time interval, increase the offset. An adaptive offset can thus be provided. In particular, energy efficiency can be improved. The method can be implemented with little computational complexity. The audio signal processing device can be set up to process an audio signal in the audible frequency range, in particular between 20 Hz and 20 kHz, and / or in the ultrasound range.

[0036] The audio signal processing device can be set up so that the adaptation or determination of the offset can be set as a function of an operating mode. For example, the adaptation can take place based on an energy-saving mode, a dynamic maintenance mode and a reaction speed mode. A lower offset can be provided in an energy-saving mode than in at least one other operating mode. In an energy-saving mode, the offset can be increased more slowly than in at least one other operating mode. As a result, the energy consumption can be reduced or minimized. It can be accepted that fast and individual amplitudes are not reproduced or not reproduced correctly. In a dynamic mode or dynamic maintenance mode, for example, the offset can permanently be half a maximum amplitude. This enables fast and individual amplitudes or peaks to be reproduced. However, this increases the average power consumption. A response speed mode can lie between the energy-saving mode and the dynamic maintenance mode.

[0037] The audio signal processing device can also be set up so that a

Time constant for adapting the time-variable offset is adjustable. An advantage of this embodiment is that the speed of adjustment of the offset can be set depending on the desired operating mode. In particular, a first (adjustable) time constant can be provided for increasing the offset between two time intervals and a second (adjustable) time constant can be provided for reducing the offset between two time intervals. In particular, the audio signal processing device can be set up to increase the offset based on a first time constant and to reduce it based on a second time constant. The first time constant can be shorter than the second time constant. This ensures a quick reaction to increasing amplitudes of the audio signal. Alternatively, the first time constant can be longer than the second time constant. The energy saving can thus be further improved. [0038] The audio signal processing device can furthermore have a buffer memory for the

Have an audio signal and be set up to determine or adapt the time-variable offset based on an amplitude profile of a buffered section of the audio signal to be reproduced in the future. In other words, the offset can thus already be adapted in advance to a section of the audio signal to be reproduced. An advantage of this embodiment is that the offset can be better adapted to a section of the audio signal to be reproduced.

[0039] The audio signal processing device can also be set up to adapt the

Control signal to compensate for a non-linearity of the thermoacoustic sound transducer. In this way, for example, thermal and / or electrical properties of a layer structure of a thermoacoustic sound transducer can be compensated. For example, a thermoacoustic sound transducer can have a thermal capacity, which can already be taken into account as a non-linearity during the control.

It is understood that the method further includes the step of controlling the thermoacoustic

Sound transducer can have based on the control signal.

It goes without saying that the developments and advantages described in detail above for the first aspect of the invention can apply correspondingly to the further aspects of the invention.

Further advantages result from the description and the attached drawings.

It goes without saying that the features mentioned above and those yet to be explained below can be used not only in the combination specified in each case, but also in other combinations or on their own without departing from the scope of the present invention.

Exemplary embodiments of the disclosure are illustrated in the drawings and are explained in more detail in the description below. Show it: 1 shows an exemplary embodiment of an audio system with an audio signal source, an audio signal processing device and a thermoacoustic loudspeaker;

2 shows an exemplary flowchart of a method for controlling a thermoacoustic sound transducer;

FIG. 3 shows a voice coil loudspeaker with associated signals; FIG.

4 shows a representation of a thermoacoustic sound transducer with associated signals;

5 shows an exemplary representation of control signals and generated sound signal;

6 shows a diagram of a control signal;

7 shows further exemplary representations of control signals and generated sound signal;

8 shows an exemplary diagram with different time-variable offsets;

9A-C show various diagrams of an audio signal to be generated with constant offset, a power of a drive signal and a voltage of the drive signal;

10A-C show various diagrams of an audio signal to be generated with time-variable offset, a power of a control signal and a voltage of the control signal; 11 AC shows various diagrams of an audio signal to be generated with time-variable offset, a power of a control signal and a voltage of the control signal;

12 shows a simplified block diagram of an audio signal processing device in connection with a pulse width modulator;

13 shows an exemplary diagram with various associated signal profiles;

14 shows a simplified block diagram of an audio signal processing device in connection with a pulse density modulator;

15 shows an exemplary diagram with various associated signal profiles; and

16 shows a further exemplary diagram with various associated signal profiles for a complex audio signal.

1 shows an exemplary embodiment of an audio system 1 with a

Audio signal source 2, an audio signal processing device 10 and a thermoacoustic loudspeaker 20. Optionally, the audio signal processing device 10 and the thermoacoustic loudspeaker 20 can together form a thermoacoustic sound transducer device 30. For this purpose, the audio signal processing device 10 and the thermoacoustic loudspeaker 20 can be integrated in a common housing, for example.

Optionally, the thermoacoustic loudspeaker 20 and the signal processing device 10 can be implemented (in whole or in part) as an ASIC (Application Specific Integrated Circuit). The sound converter device 30 can thus be designed as an ASIC. Optionally, the surge converter device 30 and / or the audio signal processing device 10 can have an amplifier. The thermoacoustic loudspeaker 20 can optionally be embodied on a silicon basis. In particular, both the thermoacoustic loudspeakers 20, as well as the signal processing device 10 are based on silicon and are combined in particular on a (standard) silicon wafer. An advantage can be an efficient, inexpensive production.

[0046] The audio signal source 2 provides an audio signal. The audio signal 2 can be used as an analog

Signal or data signal, wireless or wired are provided.

[0047] The audio signal processing device 10 is designed to: receive one

Audio signal 1 1; and generating a drive signal 12 for the thermoacoustic sound converter 20 based on the audio signal, the drive signal furthermore having an offset, the offset being a time-variable offset, and the audio signal processing device 1 being designed to base the time-variable offset based at least in sections on one To determine or adapt the amplitude profile of the audio signal. The audio signal processing device 10 is further configured to control the thermoacoustic sound converter based on the control signal 12. Based on the control signal 12, the thermoacoustic sound converter 20 generates a sound wave or a sound signal 21.

The audio signal processing device 10 can have an audio signal input 13 for

Receiving the audio signal 1 1 have. The audio signal processing device 10 can furthermore have a drive signal output 14 for outputting the drive signal 12. Optionally, a driver circuit or amplifier circuit can be provided, which can be designed as part of the audio signal processing device, for example as part of the control signal output 14, or can be designed as a separate element which can be arranged between the audio signal processing device 10 and the thermoacoustic sound transducer.

[0049] The audio signal processing device 10 can have a processor 15, in particular one

Have a digital signal processor (DSP) and / or a microcontroller (pC), which is set up to carry out the aforementioned signal processing. Optionally, the audio signal processing device 10 can furthermore have a storage device 16. The storage device 16 can contain program code or commands, which, when the program code is executed, cause the processor to carry out the above-mentioned signal processing.

As an alternative or in addition, the memory device 16 can serve as a buffer memory (buffer) for the audio signal and / or the control signal. An advantage of this embodiment can be, in particular, that parts of the audio signal to be reproduced in the future can already be buffered and the offset of the control signal can be adjusted at a current time based on parts of the audio signal to be reproduced in the future. An improved adaptation of the offset to the audio signal curve can thus be provided.

[0051] The audio signal processing device 10 can furthermore have an input 17.

For example, control commands can be provided via the input 17 in order to select a desired operating mode of the audio signal processing device 10 or to modify the type of offset adjustment.

FIG. 2 shows an exemplary flow diagram of a method 100 for controlling a thermoacoustic sound transducer. In a first step S101, an audio signal is received, for example from an audio signal source, as shown in FIG. 1. In a second step S102, a control signal for the thermoacoustic sound converter is generated based on the audio signal, the control signal furthermore having an offset, and the offset being a time-variable offset which is determined at least in sections based on an amplitude profile of the audio signal , In a third step S103, a thermoacoustic sound transducer is activated based on the generated activation signal. The generation of the control signal and the actual control of the thermoacoustic sound transducer based on the control signal can differ in time or can take place at different locations. Therefore, step S103 can be an optional method step.

The different modes of operation of a conventional moving coil loudspeaker (FIG. 3) and a thermoacoustic sound transducer (FIG. 4) are explained with reference to FIGS. 3 and 4. 3 shows a moving coil loudspeaker 40 (also known as a voice coil speaker)

shown. The voice coil loudspeaker 40 has a voice coil 41 which oscillates or oscillates in the magnetic field of a magnet 42. The voice coil 41 is coupled to a membrane 43 such that movement of the voice coil 41 in the magnetic field causes the membrane 43 to move. This mechanical movement is passed on to the surrounding medium, in particular air, and thus generates the desired sound pressure or a sound signal.

In Fig. 3, the signal diagrams A) show the audio signal for playback by the

Sound transducer, B) the corresponding control signal for the sound transducer and C) the sound signal generated. In the simplified illustration in FIG. 3 A, the audio signal 11 is a mean-free sine wave with frequency f. 3B shows a voltage profile 51 and a current profile 52 of a corresponding control signal. The positive or negative control causes the voice coil 41 to oscillate back and forth in the field of the magnet 42 in accordance with the control signal. As shown in FIG. 3C, the generated sound signal 21 (the generated sound pressure) is also directly proportional to the control signal 51, 52. Positive and negative amplitudes of the sound pressure are generated, as shown in FIG. 3C.

4 shows a simplified illustration of a thermoacoustic sound transducer 20.

Thermoacoustic sound converters convert electrical control signals or current signals into sound pressure. Precise energization allows the transducer and its surroundings to be specifically heated or cooled. Since the environment, usually a gas or air, changes its volume when the temperature changes, pressure fluctuations and thus sound waves can be generated.

[0057] The electrical power supplied changes the temperature and, as a result, that

Volume of the medium surrounding the transducer. For a positive sound amplitude, energy is supplied to the surrounding medium in the form of heat, which leads to its expansion. With a negative sound amplitude, on the other hand, little or no energy is supplied, which leads to contraction of the surrounding medium. In the case of a thermoelectric sound transducer, the electrical power P _ei supplied is therefore proportional to the sound power or the sound pressure P _a generated. Both digital and analog control or

Current signals are used. The analog control is described, for example, in the document by Daschewski et al. described. Furthermore, in the document Koshida et al. found that PDM signals (pulse density modulation) can also be used for energization. It goes without saying that the following explanations can apply correspondingly to analog and digital control signals.

Analog sound signals of the hearing and ultrasound band [up to and including the GHz range] are generally transmitted as electrical alternating voltage and can be described by the function U _Siänal (x). The amplitude can fluctuate between a positive and a negative value. The origin of this signal form lies with classic loudspeakers or microphones, as shown in FIG. 4, which function according to the voice coil principle. Since thermoacoustic sound converters simply behave like purely ohmic loads, it cannot be differentiated between positive and negative current directions. Both a negative and a positive direction of current result in the loudspeaker and its surroundings heating up and causing the surrounding medium to expand (positive sound wave).

In contrast to a moving coil loudspeaker, the sound signal 21 generated in a thermoacoustic sound transducer 20 is thus proportional to the heating and thus proportional to the thermal or electrical power consumed:

P _ei ⁼ U _ei ^x I _ei where U _{el is} the voltage and I _{el is} the current according to the control signal.

4, the signal diagrams again show A) the audio signal for reproduction by the sound transducer, B) the corresponding control signal for the sound transducer and C) the sound signal generated. Since the thermoacoustic sound transducer can be regarded in a first approximation as a resistance element with resistance R, both cause positive as well as negative current flow a warming. Thus, in FIG. 4B (in contrast to FIG. 3B) the amount of current and voltage is relevant, which is shown in broken lines in FIG. 4B.

If a sinusoidal, analog control signal with a single frequency f is applied to a thermoacoustic sound transducer, this thus generates a sound frequency with double frequency f _SC haii ⁼ 2 ^* f electrical with a simultaneous phase shift of! 4 l. The relationship is described in FIGS. 4 A, B and 4 C

The conventional, analog sound signal 11, as shown in FIG. 4 A, results in an equivalent voltage curve 51 and current curve 52 with a purely ohmic load, as shown in FIG. 4 B,

1 = (where R = const.)

Positive voltages result in a positive current flow, or negative voltages result in a negative current flow. However, as described above, the power generated in it is relevant for heating the thermoacoustic loudspeaker. A negative voltage and thus a negative current flow in turn result in a positive performance.

4B, the negative components are therefore mirrored on the X-axis in the positive region for better clarity as dashed lines. The product of voltage and current flow results in a purely positive, sinusoidal power with twice the frequency of the electrical input signal:

where w = 2P c / indicates the angular frequency. There is also a phase shift of 14 ^* l of the output frequency between the electrical input voltage and the output power.

In order to generate a sinusoidal sound signal with a frequency f is therefore in the publication Daschewski et al. "Physics of thermo-acoustic sound generation", Journal of Applied Physics, 1 14, 1 14903 pages 1-12, 2013 proposed to control a thermoacoustic sound transducer with a sinusoidal control signal with half the desired frequency f / 2 without offset. In Daschweski's publication, control with offset is also indicated in principle, but is described as disadvantageous because of the associated distortions. As a solution for a distortion-free sound signal, the control without offset is therefore proposed with half the desired frequency.

With the Daschewski et al. known solution, it is therefore possible to reproduce individual frequencies without distortion by driving the thermoacoustic loudspeaker at half the frequency of the actual target frequency. Another advantage of this type of control is that existing components such as conventional (audio) amplifiers can be used.

However, the inventors have recognized that the Daschewski et al. described

Solutions especially for the reproduction of complex audio signals with multiple frequencies can lead to undesirable distortions. This is illustrated by way of example in FIG. 5. The lower graph in FIG. 5 shows a representation of a sound signal 21, which experiences a distortion with respect to the voltage curve 51 and current curve 52 of the control signal. In this case, the control signal corresponds to a reproduction of the audio signal shifted into the purely positive range. The sound signal 21 generated is not proportional to the control signal but rather proportional to its square.

As shown in Fig. 5 by reference numeral 53, the drive signal can also be as a

PDM signal, pulse density modulation signal (English Pulse Density Modulation), or a PWM signal, pulse width modulation signal (Engl. Pulse Width Modulation; German) are generated.

To the frequency doubling described above with reference to Fig. 4 B and C.

To avoid zero crossings, the audio signal or the control signal generated based on the audio signal can be shifted by an offset value “C” into the purely positive (or alternatively purely negative) area including zero. In contrast to thermoacoustic sound transducers, conventional voice coil loudspeakers without input power (U _in = 0) are in a middle position or neutral position. From this position, the loudspeaker can generate a maximum negative and a maximum positive sound amplitude without delay (high dynamic range). In order to ensure such maximum dynamics even with a thermoacoustic loudspeaker, the middle position would also have to be assumed in the absence of positive or negative signal amplitudes. This can be achieved in that the input power of the thermoacoustic loudspeaker corresponds to half the maximum power when it is silent, ie

P & (ijßiiwrLmLiiffrmax

If

then ^ A-nsteueri g

From this point, both a maximum negative and a maximum positive sound wave (or half wave) can be generated:

Maximum negative sound wave.

^ Awst _B uring 0

Maximum positive sound wave.

In contrast to conventional moving coil loudspeakers, it is also in silence

Consumes energy. However, a high power dissipation is disadvantageous, in particular in battery-powered mobile end devices, in which a thermoacoustic loudspeaker could advantageously be used due to the lack of moving parts and its low overall height.

An exemplary control signal is shown in FIG. 6, the power P of the

Control signal 55 is shown over time. To the in a second section T2 To be able to provide full dynamic range uses the generated signal in the first

Partial area T1 in the present example only 2/3 of the power spectrum of the thermoacoustic loudspeaker. As a result, the loudspeaker constantly develops a heat loss in the area T1, which is represented by area 56. Against this background, it would be desirable to provide an efficient way of controlling thermoacoustic loudspeakers, in particular for generating complex or superimposed sound waves. Furthermore, it would be desirable to provide an efficient possibility for both simple and superimposed sound waves to generate them without distortion by digital 1-bit signal control (such as PDM or PWM).

An exemplary audio signal processing device 10 was explained above with reference to FIG. 1. Exemplary steps for signal processing which can be carried out by the audio signal processing device in one exemplary embodiment are explained in detail below. On the one hand, this can improve the quality of the sound signal to be generated and / or energy efficiency.

An audio signal U _signal (x) U _SilSnai {x) to be reproduced can be described by means of sine functions, where x describes a sample of the audio signal at a time t. In this case, U _signal (x) can have one or more overlapping frequencies and thus reproduce simple or complex sound signals, particularly in the hearing or ultrasound range.

To determine the electrical power to be _supplied P _ctrl Pjina _t auanm _g can

Usi _gHBi (x) can be increased by a correction value C C. In particular, it can

An audio signal can be applied with an offset C such that the raised signal has no zero crossings. In this way, undesirable frequency doubling caused by zero crossings can be avoided. In particular, the raised signal has no negative values.

Based on the function

which describes the sound signal to be generated, a control signal U _ctrl (x) U _{An3i: euerun3} (x) e ec net can be:

U _ctri (x) =] P _ctrl (x) x R and thus proportional to ^ {U _signal (x) + C) x R, where R approximates the resistance of the thermoelectric transducer and is constant. So in a simplified way

and if U _signal (x) is written as a superposition of sine frequencies

7 shows in the upper graph an example of an audio signal 1 1 to be generated above the

Time t. The audio signal processing _device receives the corresponding audio signal U _signai ( ^x ) 1 1 as an input signal. Optionally, the audio signal 11 can be an audio signal free of mean values, as shown in the upper graph in FIG. 7. Due to the approximately purely ohmic properties of the thermoacoustic sound transducer, the audio signal 1 1 is initially subjected to an offset and thus raised to a purely positive range or alternatively lowered to a purely negative range, so that the power of the control _signal P _ctrl ~ U _signal ( x) + C is. For this purpose, the audio signal processing device can be set up to generate a control signal

y] P _ctrI 51 to be calculated such that the power output and thus the sound signal 21 generated by the thermoacoustic sound transducer corresponds to the desired audio signal. The corresponding current flow

of the control signal is designated in the lower graph in FIG. 7 with reference symbol 52. The vertical axis in the lower graph in FIG. 7 indicates the amplitudes of the voltage 51, the current 52, the power 21 and the sound pressure 21, respectively. In other words, the audio signal processing device can be set up to generate the drive signal based on a root of the audio signal and the offset. This can reduce disturbing harmonics and improve the audio quality.

With reference to FIGS. 8 to 11 C, the determination of the offset C, in particular a time-variable determination and adaptation of the offset based on an amplitude profile of the audio signal, will be explained below using examples.

8 shows an exemplary signal profile of an audio signal 11 to be reproduced, which can be received at an input 13 of the audio signal processing device 10 (as shown in FIG. 1). In order to avoid negative amplitudes (or a change of sign), the audio signal could be subjected to a constant offset 61, the absolute value of which is greater than a global maximum of the audio signal to be generated. However, this would lead to a not insignificant power loss, as already described with reference to FIG. 6 above.

Instead, the proposed audio signal processing device 10 is advantageously set up to receive an audio signal 1 1 and generate a control signal 12 for the thermoacoustic sound transducer based on the audio signal, the control signal also having an offset, the offset being a time-variable offset, and wherein the Audio signal processing device is designed to determine the time-variable offset at least in sections based on an amplitude profile of the audio signal. In particular, the audio signal processing device can be set up to adapt the time-variable offset based on local maxima and / or minima of the audio signal. In FIG. 8, local minima are exemplified by diamond symbols 63. Instead of providing a constant offset 61, the offset can be determined and adapted dynamically in a time-variable manner, C (t).

In the exemplary embodiment shown in FIG. 8, the audio signal processing device can have, for example, a peak value interrogation unit in order to determine the local maxima and / or minima 63. Optionally, an envelope of the audio signal nals can be determined and the time-variable offset can be adjusted based on the envelope.

The audio signal processing device can optionally adjust the offset taking into account a parameter, in particular taking into account a settling time T _r (English rise time or attack time) and / or a decay time T _f (English fall time). These parameters can be advantageously selected in the field of tension between reaction speed (dynamics retention) and energy efficiency. Optionally, the audio signal processing device can be set up so that the (parameter for) adapting the time-variable offset can be set as a function of an operating mode. For example, a selection can be entered via the interface 17 shown in FIG. 1.

The rate of adaptation of the settling time and decay time can advantageously be chosen such that there are no relevant impairments of the sound pressure level for the respective application. For hearing aid applications, the adaptation speed could be, for example, a maximum of 20 Hz, and thus below the frequency spectrum relevant for humans.

8 shows an example of the determination of three different time variables

Offset curves 62a, 62b, 62c on the basis of the negative amplitude excursions 63 of the audio signal 11 in various forms, the curve 62a being slow T _r and slow T _f ; curve 62b mean T _r and slow T _f ; and curve 63c has fast T _r and fast T _f .

In the example described in FIG. 8, the offset to the negative amplitude of the

Audio signal 11 related, since this is to be generated by the sound transducer as a negative sound amplitude. A phase shift of 180 ° is also possible if the sound signal were lowered by the offset or correction value, based on a positive amplitude deflection, into the purely negative range. 9A, 10A and 1 1A show exemplary diagrams of curves for a received audio signal 11, here U _signal (x). The same audio signal is shown in each case. However, the diagrams differ in the choice of offset 62.

9B, 10B and 11B show exemplary diagrams of control signals for a thermoacoustic sound transducer, in each case the power of the control signal

12 P _ctrl (t) = -jU _signal (t) + C (t) is shown. FIGS. 9C, 10C and 11 C each show the associated voltage profile of the control _signal 12 U _signal (t) + C (t) over time. It goes without saying that the power of the control signal for a given control voltage depends on the resistance of the thermoacoustic sound transducer.

In FIG. 9A, a constant offset 61 is selected, which corresponds to a global minimum 63 of the

Audio signal 1 1 corresponds. If the global minimum is not known, the constant offset can be selected as half of the maximum possible power for the thermoacoustic sound transducer, as already shown in FIG. 6. One advantage of this configuration is that the maximum dynamic range is always maintained. By adapting the offset to the global minimum 63, however, energy savings can be achieved for signals which do not exhaust the full dynamic range of the thermoacoustic sound transducer.

In contrast to FIG. 9A, in FIG. 10A there is no constant, but a time-variable offset

62 chosen. The audio signal processing device is designed to determine the time-variable offset 62 based at least in sections on an amplitude profile of the audio signal 11. In the example shown in FIG. 10A, the time-variable offset is adjusted based on the local minima 63 of the audio signal 11. In order to ensure that the high amplitudes of the audio signal 11 present at the beginning of the reproduction are correctly reproduced, a start value 64 of the offset can be used as a predefined offset value, in particular a large offset, in particular corresponding to half the maximum possible power for the thermoacoustic transducers can be selected. In contrast to the examples shown in FIG. 6 or FIG. 9A, remains however, the offset is not constant at this value, but is continuously adjusted or tracked based on the audio signal 11, as shown in FIG. 10A.

From the comparison of FIG. 10B with FIG. 9B (or correspondingly from FIG. 10C with FIG. 9C) it is clearly evident that the average power consumption of the thermoacoustic sound transducer can be significantly reduced. It goes without saying that the saving depends on the actual signal curve of the audio signal 1 1. In other words, the saving depends on the deflection of the sound signal in relation to the working range of the sound converter. In the event of silence, the adjustment can save up to 100% of the energy (the output can be completely reduced to 0). If the sound signal is 20% of the working range (amplitude +/- 0.2 with a max. Working range of +/- 1, 0), up to 80% can be saved. If the sound signal is 50% of the working area, up to 50% can be saved. If the sound signal is 70% of the work area, the maximum is 30%.

In the present example, the savings that can be achieved thus depend on the distance from the lower vertex to the lower limit of the work area (or the upper vertex to the upper limit). The calculated offset value can correspond to this distance.

[0090] The audio signal processing device can be set up to accommodate the time variable

Determine offset iteratively, for example with the steps

Determining a peak value (e.g. local minimum 63) in a first time interval i;

Comparing the peak value with a peak value in a previous time interval i-1; and

if the peak value in the first time interval is greater than the peak value in the previous time interval, increase the offset. Here, the amounts of peaks and / or offset can be used. [0091] The audio signal processing device can be set up to adjust the offset

increase based on a first time constant and decrease based on a second time constant, in particular wherein the first time constant is shorter than the second time constant. The first and second time constants can describe a rise time T _r and a fall time T _f . A short rise or decay time means that the time-variable offset 62 can quickly follow the audio signal 11. In the present example, linear increases and decay times were used to reduce the necessary computing power. It goes without saying that other types of adaptation or tracking of the offset can also be carried out. The type of adaptation can be linear, in the form of an e-function or another mathematically describable function. The time constants serve as parameters and influence the speed / dynamics of the function for offset adjustment. The advantage of a high adaptation speed is that you can react quickly to higher dynamics of the input signal. In the present example, short time constants of 0.001 rise time and 0.001 decay time were chosen. The values regarding the rise time can be a (maximum permissible) change in amplitude per sample. The advantage of short time constants is that you can react quickly to signal changes.

However, an adaptation speed of the time-variable offset is preferably selected such that it lies below a first cut-off frequency, in particular below a frequency range of the audio signal, in particular below 20 Hz. In this way, disruptive artifacts in the audible range, which could be caused by the adjustment of the time-variable offset, can be avoided. However, it may happen that signal peaks or peaks in the audio signal require a higher amplitude than can be reproduced with the reduced offset, as shown in FIG. 10A. In this case the offset is increased again. Such signal peaks are shown in FIG. 10B for illustration as powers less than zero. However, these signal peaks can be cut off, so-called signal clipping, for example to avoid frequency doubling, as shown in FIG. 10C.

11 AC relate to an optional further development, the audio signal processing device 10 (see FIG. 1) also having a buffer memory 16 for the audio signal 11 and the audio signal processing device is set up to determine the time-variable offset based on an amplitude profile of a buffered section of the audio signal to be reproduced in the future. In this case, the time-variable offset can be adjusted even before such signal peaks are reproduced, so that signal distortions by clipping can be avoided or reduced.

Optionally, alternatively or additionally, the audio signal processing device can also be set up to always provide a minimum offset. For example, the offset value can be set lower (or higher) by a predetermined value, for example 0.15 lower, as shown in FIG. 11A. This minimum offset can be understood as a safety margin in order to achieve improved dynamics.

Basically, like the offset, the buffer value can be dynamically determined and adjusted based on the past values and / or the pre-analyzed future values. In this way, the dynamics can be further improved.

A control signal provided by the audio signal processing device, as stated above, can now be fed to a thermoacoustic sound converter for control. Optionally, the audio signal processing device can have a converter device for generating a PWM (pulse width modulation) control signal or PDM (pulse density modulation) control signal based on the control signal. The control signal can thus be output as a 1-bit digital signal such as PDM or PWM.

12 shows an illustration of simplified block diagrams of a first and a second exemplary embodiment of an audio signal processing device 10 with a pulse width modulator 1200 (see FIGS. 12A and B).

[0097] The audio signal processing device 10 can be an audio signal processing device, as described above and shown by way of example in FIG. 1. The audio signal processing device is designed for receiving an audio signal 11 and for generating and outputting a control signal 12. The control signal 12 is fed to and from a converter device 1200 in the form of a pulse width modulator converted to a PWM signal 1212. It goes without saying that the converter device 1200 can be part of the audio signal processing device or can be designed as a separate component. The corresponding signals are shown in Fig. 13.

[0098] In the present exemplary embodiment, the converter device 1200 is set up to generate a PWM signal based on the control signal 12. The pulse width is determined based on the amplitude of the control signal 12. The converter device 1200 can have a comparator 1201 and a sawtooth generator 1202. The pulse mode results from a comparison of the amplitude of the control signal 12 with an amplitude of the sawtooth signal 1203 output by the sawtooth generator 1202. In a simple embodiment, a working range of the sawtooth signal 1203 of the sawtooth generator 1202 (as well as the control signal 12) can be in a purely positive or negative Range. The control signal 12 can thus be converted into a PWM signal 1212 in a simple manner on the basis of the sawtooth signal 1203. In order to avoid overdriving, an amplitude range of the sawtooth signal 1203 or a dynamic range of the PWM signal can advantageously be greater than or equal to that as a maximum value of the control signal, as shown by way of example in FIG. 13. An advantage of this embodiment can be a dynamic maintenance while ensuring energy efficiency.

12, diagram B, shows a further possibility of a PWM modulation unit 1200.

Here, a working range or amplitude maximum 1204 in FIG. 13 (for example provided via 1205 in FIG. 12B) of the sawtooth signal 1203 can be set based on a dynamic range, in particular based on an amplitude maximum or minimum of the calculated control signal. For this purpose, the sawtooth generator 1202 can have a signal input for receiving the control signal 12 and / or information from which the required dynamic range can be derived. The sawtooth signal can thus fluctuate between 0 and an amplitude maximum or minimum of the calculated control signal 12. An advantage of this solution can be a more extensive, in particular a complete compression of the sound signal to be generated. It goes without saying that this approach can be used not only in the simplified block diagrams shown in FIG. 12 but also in more complex PWM modulation units, for example PWM Modulation units whose working range is not limited to the purely positive or negative range can advantageously be used.

14 shows an illustration of simplified block diagrams of a first and a second exemplary embodiment of an audio signal processing device 10 with a delta-sigma modulator 1400 (see FIGS. 14 A and B).

[00101] As an alternative to control with a PWM signal, a thermoacoustic

Sound transducers can also be controlled with a PDM signal (pulse density modulation signal). The audio signal processing device 10 can be an audio signal processing device, as described above and shown by way of example in FIG. 1. The audio signal processing device is designed to receive an audio signal 1 1 and to generate and output a control signal 12. The control signal 12 is fed to a converter device 1400, here in the form of a pulse density modulator, and is converted by the latter into a PDM signal 1412. It goes without saying that the converter device 1400 can be part of the audio signal processing device or can be designed as a separate component. The corresponding signals are shown in FIGS. 15 and 16.

According to the exemplary embodiments shown in FIG. 14, the control signal 12 can be represented by a 1-bit digital signal 1412, the audio signal processing device having a converter device in the form of a delta-sigma modulator 1400 of first or higher order. The converter device can be set up so that the generation of the PDM signal 1412, analogously to the determination of the time-variable offset C, can also take place taking into account a further correction value K. Analogously to the determination of the time-variable offset C, the correction value K can be determined as a constant or, preferably, as a dynamic value, taking into account the speed of reaction and energy efficiency.

In the present exemplary embodiment, the audio signal processing device 10 can thus be used to determine the control signal on the basis of a time-variable offset in the form of a correction value K. Here K can be the difference between an amplitude be the maximum or minimum of the control signal and an amplitude maximum or minimum of a working range of the PDM unit, ie:

thus a drive signal for the pulse density modulator can be determined as

An advantage of this embodiment according to FIG. 14A can consist in maintaining the dynamics while at the same time ensuring energy efficiency. All energization signals can preferably contribute to the desired sound generation. In the example shown in FIG. 14A, an input signal U _c tri can be adjusted based on a working range of the delta-sigma modulator 1400.

Another embodiment of a first order delta sigma modulator 1400 is shown in FIG

14B. Here, half of the amplitude maximum or minimum of the control signal 12 can be defined as the working range of the delta-sigma modulator 1400,

An advantage of this embodiment according to FIG. 14B can be seen in a complete

Compression of the PDM signal 1412 to be output with simultaneous energy efficiency. The adjustment speed of the work area

^ Arb eitsbereichPDM A _b management B may in this case from the point of receipt and dynamics

Energy efficiency can be carried out advantageously. In the example shown in FIG. 14B, a working range of the delta-sigma modulator can be specified, as indicated by the arrow originating from element 10. An advantage of this embodiment can be an improved, in particular complete compression. It goes without saying that the proposed solution can be used in addition to the block diagrams of the embodiments A and B shown in FIG. 14, also in higher-order delta-sigma modulation units or in a modified form.

Figures 15 and 16 show exemplary waveforms for that shown in Figure 14A

Embodiment. Here, 11 denotes the audio signal to be generated, 12 the control signal provided by the audio signal processing device 10, 1412 the PDM signal provided by the converter device in the form of a delta-sigma modulator 1400, and 19 the (smoothed) power supplied to the thermoacoustic sound converter. For better clarity and to be able to distinguish the signal profiles, the audio signal 1 1 was shown with an offset of 0.25. It goes without saying that the audio signal can preferably be an average-free audio signal. The curve 11 in Fig. 16 is shown accordingly. 15 and 16, the power 19 supplied to the sound transducer, apart from an offset, corresponds to the audio signal 11 to be generated. In the present example, FIG. 15 shows an audio signal 11 of a frequency of the hearing or ultrasound band 16 represents a superimposed sound signal of the hearing or ultrasound band.

As can be seen from FIGS. 14, 15 and 16, the proposed solution can be used both for

single, as well as for superimposed and complex frequencies of the hearing and ultrasound range. Another advantage of the proposed solution can be that the amplification unit of the energization signals can be kept simple by digital 1-bit signals, such as adapted PDM signals 1212 or PWM signals 1412. Compared to conventional Class D amplifiers, only a simple amplification circuit is required in this case. In addition, the amplified digital signal can be smoothed by a low-pass filter, so that overtones and distortions can be avoided or reduced.

Basically, the concepts described here are not specific

Frequency range limited but can be used for the entire infrared, hearing and / or ultrasound range, for example from 10 Hz to the gigahertz range. The audio signals and / or control signals can be analog or digital (e.g. PDM, PWM). Digital processing such as in pulse code format (eg wave format) can also be carried out. PDM or PWM signals can be fed to the thermoacoustic transducer with or without smoothing. It is also possible to consider a frequency-dependent power adjustment in the signal generation, for example to compensate for fluctuations in the frequency response (for example, through fluctuations in the manufacturing process or in the material structure). It is also possible to carry out the sound generation advantageously by specifically adapting the duty factor and sampling rate of the 1-bit control.

It goes without saying that it is in the generation of the drive signal based on a

Root from the audio signal and the offset can be an optional feature if only an improvement in energy efficiency is desired. Accordingly, the determination of a time-variable offset can be an optional feature if, instead, only an improvement in the sound quality is desired, the control signal for the thermoacoustic sound transducer being generated based on a root from the audio signal and a (constant) offset.

In summary, the sound generation with thermoacoustic sound transducers can be further improved with the solutions proposed here. In particular, the energy efficiency of a system for thermoacoustic sound conversion, in particular with complex audio signals, can be further improved. This is of great benefit in particular in the case of battery-powered mobile end devices, in which a thermoacoustic loudspeaker could advantageously be used due to the lack of moving parts and its low overall height.

Claims

claims

1. Audio signal processing device (10) for a thermoacoustic sound transducer (20), the audio signal processing device being designed for:

Receiving an audio signal (1 1);

Generating a control signal (12) for the thermoacoustic sound transducer (20) based on the audio signal, the control signal also having an offset (62),

wherein the offset is a time-variable offset (62), and wherein the audio signal processing device (10) is designed to determine the time-variable offset at least in sections based on an amplitude profile of the audio signal.

2. Audio signal processing device according to one of the preceding claims, wherein the audio signal processing device (10) is set up to adapt the time-variable offset based on local maxima and / or minima (63) of the audio signal (1 1).

3. Audio signal processing device according to claim 1, wherein the audio signal processing device (10) is set up to generate the control signal (12) based on a root of the audio signal (11) and the time-variable offset (62).

4. Audio signal processing device according to one of the preceding claims, wherein the time-variable offset (62) is selected such that the control signal (12) only accepts positive values or only negative values.

5. Audio signal processing device according to one of the preceding claims, wherein the time-variable offset (62) is determined at least in sections based on an envelope of the audio signal (11).

6. Audio signal processing device according to one of the preceding claims, wherein the time-variable offset (62) is determined such that an adaptation speed of the time-variable offset is below a first cut-off frequency, in particular below a frequency range of the audio signal (11).

7. Audio signal processing device according to one of the preceding claims, wherein the time-variable offset (62) is adjusted at least in sections such that the amount of the offset is smaller than a difference between a mean value of the audio signal (11) and a global maximum or minimum of the audio signal ( 1 1).

8. Audio signal processing device according to one of the preceding claims, wherein the drive signal (12) is represented by 1-bit digital signals, in particular wherein the audio signal processing device has a delta-sigma modulator (1400) of the first or higher order.

9. Audio signal processing device according to one of the preceding claims, with a converter device (1200, 1400) for generating a PWM (pulse width modulation) control signal or PDM (pulse density modulation) control signal based on the control signal (1 1).

10. Audio signal processing device according to one of the preceding claims, further comprising a filter, in particular a low-pass filter, in order to provide a filtered control signal.

1 1. Audio signal processing device according to one of the preceding claims, wherein the audio signal (11) is an average-free audio signal.

12. Audio signal processing device according to one of the preceding claims, wherein the determination of the time-variable offset (62) comprises the following steps:

Determining a peak value (63) in a first time interval i;

Comparing the peak value with a peak value in a previous current time interval i-1; and

if the peak value in the first time interval is greater than the peak value in the previous time interval, increase the offset.

13. Audio signal processing device according to one of the preceding claims, further configured to process an audio signal (1 1) in the audible frequency range, in particular between 20 Hz and 20 kHz, or in the ultrasound range, in particular between 20 kHz and 1 GHz.

14. Audio signal processing device according to one of the preceding claims, wherein the adaptation of the time-variable offset (62) can be set as a function of an operating mode, in particular based on an energy-saving mode, a dynamic maintenance mode and a reaction speed mode.

15. Audio signal processing device according to one of the preceding claims, wherein a time constant for adapting the time-variable offset (62) is adjustable.

16. Audio signal processing device according to claim 15, wherein the audio signal processing device (10) is set up to increase the offset (62) based on a first time constant (T _r ) and to reduce it based on a second time constant (T _f ), in particular wherein the first time constant is shorter than the second time constant.

17. Audio signal processing device according to one of the preceding claims, wherein the audio signal processing device (10) further comprises a buffer memory

(16) for the audio signal (11) and the audio signal processing device is set up to determine the time-variable offset based on an amplitude profile of a buffered section of the audio signal to be reproduced in the future.

18. Audio signal processing device according to one of the preceding claims, further configured to adapt the control signal (12) to compensate for a non-linearity of the thermoacoustic sound transducer (20).

19. Thermoacoustic transducer device (30) with

an audio signal processing device (10) according to one of the preceding claims; and

a thermoacoustic sound transducer (20).

20. A method (100) for controlling a thermoacoustic sound transducer (20), the method comprising the following steps:

Receiving an audio signal (S101);

Generating a control signal (S102) for the thermoacoustic sound transducer based on the audio signal, the control signal also having an offset,

21. A computer program comprising commands which, when the program is executed by a computer, cause the computer to carry out the method according to claim 20.