EP3606098A1 - Method and circuit arrangement for operating a condenser microphone - Google Patents

Abstract

A method and a circuit arrangement for operating a condenser microphone according to a high-frequency method with double-sideband amplitude modulation are described. The signal processing is carried out largely by means of a digital circuit part (DC), which is preferably implemented in the form of an FPGA or a digital signal processor, and an RF generator (CG) for generating a digitally oversampled RF carrier signal (x _{c, h} (k); x _c (k)), preferably a unit (NS) for noise shaping, and a digital demodulator (Dm) for demodulating a digital DSB-AM audio signal (x <sub > DSB </sub> (m)) and for generating a digital baseband audio signal (x _the (m)). A decimator (Dc) for decimating the sampling rate of the digital baseband audio signal (x _dem (m)) and for generating a digital baseband audio signal (x _AF ( n)) provided with a standard audio sampling rate. A major advantage here is that a high dynamic range can be achieved without the use of multi-channel A / D converters and a high signal / noise ratio in the digital DSB-AM audio signal. (Fig. 2)

Description

The invention relates to a method and a circuit arrangement for operating a condenser microphone according to a high-frequency method with amplitude modulation.

In the case of a condenser microphone, sound waves are known to be recorded in the form of changes in the capacitance of a condenser capsule which has a membrane moved by the sound waves and at least one rigid counterelectrode. A distinction is made here. a. between standard capacitor converters, which have only one counter electrode, and symmetrical capacitor converters or push-pull converters, in which the membrane is arranged between two counter electrodes, so that two capacitors are formed. A membrane deflection in one direction caused by incident sound waves leads to a reduction in the air gap to one counter electrode and a simultaneous enlargement of the air gap to the other counter electrode. This results in a corresponding opposite increase or decrease in the capacitance of the two capacitors (push-pull converter).

Since the mass of the membrane is very small, it can follow the incident sound waves very precisely and quickly, so that condenser microphones have superior physical properties, in particular compared to dynamic microphones, and thus low-frequency output signals with high sound quality, impulse fidelity, sensitivity and low noise can be generated.

However, condenser microphones require an external power supply for the condenser capsule in order to be able to convert the changes in capacitance into an output signal. A distinction is made between two fundamentally different processes:
In the low-frequency method, the capacitor capsule is charged with a bias voltage via a high-resistance resistor. Which is caused by the impinging sound waves Changing capacitance of the capacitor capsule leads to a corresponding proportional change in the output voltage at the capacitor capsule when the capacitor charge is essentially constant. However, this method has several disadvantages. On the one hand, leakage currents can occur, particularly in environments with high humidity, which lead to a poor signal / noise ratio, and on the other hand, due to the small capacitance of the capacitor capsule, an impedance converter with a very high input resistance is required for further processing of the output voltage.

To avoid these disadvantages, one of the known high-frequency methods is preferably used, in which the changing capacitance of the capacitor capsule is used to modulate the frequency, the phase or the amplitude of a high-frequency electrical carrier signal (RF signal).

In frequency modulation, the capacitor capsule is part of a resonance circuit, so that the changes in the capacitance of the capacitor capsule caused by the sound waves lead to corresponding changes in the resonance frequency. However, this method requires a very frequency-stable oscillator so that the audio signal obtained after demodulation is not disturbed.

In phase modulation, the capacitor capsule is part of an LC parallel resonance circuit, so that the changes in the capacitance of the capacitor capsule lead to corresponding changes in the resonance frequency of the LC circuit and thus to phase shifts in the externally generated and coupled RF signal. This method also requires a very frequency-stable RF signal and a high Q factor of the LC circuit to achieve a large phase shift or microphone sensitivity. Another disadvantage is that the phase response has only a relatively small linear range, so that in the case of a high sound pressure level, the change in capacitance and the resulting phase shifts are no longer proportional to one another.

In the case of amplitude modulation, the capacitor capsule finally forms, together with two identical windings, an RF transformer bridge circuit which is tuned to the frequency of an RF carrier signal which is generated by an RF oscillator coupled via a further winding. Due to the movement of the membrane and the resulting change in capacitance, a proportional double-sideband amplitude-modulated (DSB-AM) audio output voltage with suppressed RF carrier signal can be generated at the bridge branch of the circuit.

This method is particularly suitable for the symmetrical capacitor converters or push-pull converters mentioned at the outset, in which the two capacitors are connected as capacitive voltage dividers in one of the bridge arms.

From the DE 4300379A1 and the DE 102010000686A1 Examples of such RF transformer bridge circuits with symmetrical capacitor converters are known.

In the case of standard capacitor converters, however, a corresponding reference capacitor for the voltage divider must be connected in the bridge arm.

From the US 6,697,493 B1 A method and a circuit arrangement are known with which an acoustic signal impinging on a capacitive sound transducer is converted into a digital signal. For this purpose, a counter signal is applied to the sound transducer in such a way that it remains essentially in its idle state when an acoustic signal strikes it. Each deflection from the idle state in one direction or the other is converted into a digital signal "1" or "0", from which the counter signal is derived and which contains information about the acoustic signal.

In the US 2008/0219474 A1 describes an integrated circuit which has a preamplifier for an input signal generated with a microphone membrane or a microphone element, in relation to which the membrane is movable, a charge pump and a low-pass filter for generating a microphone output signal, the low-pass filter is provided for filtering the pump voltage and for generating a bias for the membrane or the microphone element.

In " IEEE Sensors Journal ", Vol. 11, No. 2, February 2011, pages 296 to 304 describes a digital CMOS silicon condenser microphone in which the acoustic sensor is integrated on a chip together with the electrical circuits. All components are operated with the same supply voltage, whereby a charge pump is not required.

The invention has for its object to provide a method and a circuit arrangement for operating a condenser microphone with which the above-mentioned superior physical properties of such a microphone can be exploited even better and converted into LF output signals with high fidelity. In particular, a digital microphone with the least possible analog effort is to be created, in which the dynamic range and the signal / noise ratio of the LF output signals is at least equivalent or even better than known analog microphones.

This object is achieved with a method according to claim 1 and a circuit arrangement according to claim 11.

An essential principle of the invention is therefore to obtain an LF output signal by digital signal processing in combination with a suitable analog circuit for detecting the capacitance changes generated by incident sound waves, which is particularly characterized by a high dynamic range and a high signal / noise ratio.

Advantages of the solutions according to the invention are that the reduced number of analog components and in particular the use of a digital synchronous demodulator means that no noticeable non-linearities are introduced. Furthermore, that too Electrical 1 / f noise largely reduced or avoided using analog bandpass processing.

Through the digital generation of a (digital) RF carrier signal, in particular in combination with the highest possible sampling rate and / or with the highest possible resolution or large word width, the frequency of the analog RF carrier signal can be matched very precisely to the resonance of the analog circuit part, and a spectrally very pure RF carrier signal can also be generated.

The dependent claims contain advantageous developments of the invention.

Since digital / analog converters usually have a limited resolution, the digital RF carrier signal is preferably subjected to noise shaping before the digital / analog conversion, with which not only the resolution or word width is reduced, but also the quantization noise is shifted into frequency ranges which are used for Detection of the relevant AF bandwidth of the microphone is not required. This maintains a high signal / noise ratio in the relevant frequency band.

The amplitude of the digital HF carrier signal is preferably adjustable in order to change or adapt the sensitivity of the microphone, in particular as a function of an occurring sound pressure level. The sensitivity can thus be set digitally and no multi-channel analog / digital converters are required in order to achieve a high dynamic range or a high gain.

This amplitude is preferably changed automatically by means of a sound pressure level detector / predictor as a function of a detected or predicted sound pressure level in such a way that the sensitivity of the microphone is adjusted accordingly and overdriving of the analog circuit part is avoided.

Finally, an additional signal-to-noise ratio and dynamic range can also be obtained by digitizing the acquired audio signal, preferably over-sampled and clocking it down to a lower sampling frequency after demodulation (for example to a standard sampling frequency of 48 kHz).

Fig. 1

eine bekannte HF-AM-Schaltung mit einem Kondensatormikrofon mit Gegentaktwandler;

Fig. 2

ein Blockschaltbild einer ersten Ausführungsform einer erfindungsgemäßen Schaltungsanordnung;

Fig. 3

eine erste Komponente des digitalen Schaltungsteils;

Fig. 4

eine zweite Komponente des digitalen Schaltungsteils;

Fig. 5

ein Blockschaltbild einer zweiten Ausführungsform einer erfindungsgemäßen Schaltungsanordnung;

Fig. 6

eine erste Ausführungsform eines analogen Schaltungsteils mit GegentaktKondensatormikrofon;

Fig. 7

eine zweite Ausführungsform eines analogen Schaltungsteils mit GegentaktKondensatormikrofon;

Fig. 8

eine dritte Ausführungsform eines analogen Schaltungsteils mit GegentaktKondensatormikrofon; und

Fig. 9

ein alternativ in den Schaltungsteilen gemäß den Figuren 6 bis 8 einsetzbares Kondensatormikrofon mit Standardwandler.

Further details, features and advantages of the invention result from the following description of exemplary and preferred embodiments with reference to the drawing. It shows:

Fig. 1

a known HF-AM circuit with a condenser microphone with push-pull converter;

Fig. 2

a block diagram of a first embodiment of a circuit arrangement according to the invention;

Fig. 3

a first component of the digital circuit part;

Fig. 4

a second component of the digital circuit part;

Fig. 5

a block diagram of a second embodiment of a circuit arrangement according to the invention;

Fig. 6

a first embodiment of an analog circuit part with push-pull condenser microphone;

Fig. 7

a second embodiment of an analog circuit part with push-pull condenser microphone;

Fig. 8

a third embodiment of an analog circuit part with push-pull condenser microphone; and

Fig. 9

an alternative in the circuit parts according to the Figures 6 to 8 Applicable condenser microphone with standard converter.

Since a condenser microphone is preferably used in the analog circuit part of the circuit arrangement according to the invention and this is preferably operated using the high-frequency method mentioned at the beginning with amplitude modulation, the basic function and implementation of this operating mode should first be based on the Figure 1 are explained.

Figure 1 shows an exemplary circuit arrangement with a condenser microphone with a symmetrical converter (push-pull converter), which has a first and a second capacitor C ₁ (t), C ₂ (t), the capacities of which, as explained above, are changed in opposite directions by incoming sound waves s (t).

The circuit arrangement operates according to the high-frequency method with amplitude modulation mentioned at the outset, which is preferably used in connection with the invention, and comprises a first transmitter Tr1, a second transmitter Tr2, an HF generator HF, a first and a second diode D1, D2 and a charging capacitor Ca.

On the primary side of the first transformer Tr1 there is a first winding W1, via which the HF voltage (HF carrier signal) generated by the HF generator HF is coupled.

On a first secondary side, the first transformer Tr1 has a second and an identical third winding W2, W3. The two windings W2, W3 together with the two capacitors C ₁ (t), C ₂ (t) form a bridge circuit. On a second secondary side of the first transformer Tr1 there is a fourth and an identical fifth winding W4, W5, the outer ends of which are connected to one another via two diodes D1, D2 connected in series and which together form a synchronous rectifier.

The second transformer Tr2 is finally connected with one side between the center tap between the second and third windings W2, W3 and ground, while the other side is connected between the center tap between the fourth and fifth windings W4, W5 and one of the output connections. The other output connection is tapped between the two diodes D1, D2. The charging capacitor Ca, to which the output voltage x _AF (t) in the form of the demodulated audio signal (baseband signal) is applied, is connected in parallel with the output connections.

The HF bridge circuit formed by the first and the second capacitor C ₁ (t), C ₂ (t) and the second and the third winding W2, W3 is tuned to the frequency of the HF carrier signal generated by the HF generator HF. As long as the membrane of the capacitor capsule is in the middle rest position (ie no sound waves are incident), the bridge is balanced and there is no RF voltage at its output. When the membrane is deflected by incoming sound waves s (t), an HF voltage is generated which is proportional to the membrane deflection and which is fed to the charging capacitor Ca for demodulation via the synchronous rectifier D1, D2 and then to the membrane in magnitude and phase proportional output voltage x _AF (t) is present.

Figure 2 shows a schematic block diagram of an exemplary first embodiment of a circuit arrangement according to the invention, which is composed essentially of three components, namely a digital circuit part DC, an analog circuit part AC and a mixed analog / digital circuit part MC, the latter being connected between the digital and the analog circuit part and connects the two.

The digital circuit part DC comprises the digital components of the circuit arrangement. These are essentially an HF generator CG for generating a digital oversampled HF carrier signal x _{c, h} (k) with a first high sampling rate f _{s, 1} and with a first high resolution, which corresponds to a first large word width of w ₀ bits , an optional unit NS for noise shaping of the RF carrier signal x _{c, h} (k) and for generating an RF carrier signal x _c (k) with the same first sampling rate f _{s, 1} and with a lower resolution than the first resolution, the corresponds to a correspondingly smaller second word width of w ₁ bit.

The digital circuit part DC further comprises a demodulator Dm for demodulating an applied digital DSB-AM audio output signal x _DSB (m), which has a second sampling rate f _{s, 2} and a third resolution or third word length of w ₂ bits, and for Generation of a demodulated digital (baseband) audio signal x _dem (m) with unchanged (Second) sampling rate f _{s, 2} and a fourth resolution or fourth word width of w ₃ bits. Finally, the digital circuit part comprises an optional decimator Dc for decimating or reducing the second sampling rate f _{s, 2 of} the demodulated audio signal x _dem (m) and for generating a digital audio output signal x _AF (n) with a standard audio sampling frequency of for example 48 kHz (third sampling rate fs) with a desired fifth resolution or a fifth word width of w ₄ bits.

This digital circuit part D-C is preferably realized in the form of an FPGA (field programmable gate array) and represents a digital signal processor (DSP). The implementation is preferably carried out by a corresponding program that is executed on a system for data processing.

The analog / digital circuit part MC comprises a digital / analog converter D / A for generating an analog RF carrier signal x _c (t) from the supplied digital RF carrier signal x _c (k), and an analog / digital converter A / D for generating a digital DSB-AM audio signal x _DSB (m) with said third resolution or third word length of w ₂ bits from a supplied analog DSB-AM audio signal x _DSB (t), which is generated with the microphone unit Cm.

The analog circuit part AC finally includes the microphone unit Cm to a capacitor microphone and serves the changes in the capacitance of the microphone by means of the supplied analog RF carrier signal x _c (t) generated by the incident sound wave s (t) (NF-audio signal) in the to implement analog DSB-AM audio signal x _DSB (t) with suppressed RF carrier signal x _c (t).

This circuit arrangement works in detail as follows:
The HF generator CG generates an oversampled digital HF carrier signal x _{c, h} (k) with the aforementioned first high sampling rate f _{s, 1} and the first high resolution, which corresponds to the first large word width of w ₀ bits ,

The level of the carrier frequency f _{r of} the HF carrier signal x _{c, h} (k) is chosen in a suitable manner, in particular depending on the electrical properties of the analog circuit part AC (in particular its resonance frequency, see below).

The term “oversampling” is to be understood to mean that the signal is generated at a sampling rate that is higher than is required according to the sampling theorem to detect the RF carrier frequency f _r (also called oversampling). The (first) sampling rate f _{s, 1 of} the RF carrier signal x _{c, h} (k) with the frequency f _r is therefore greater than the sampling rate of 2 * f _r required according to the sampling theorem.

The aforementioned first high resolution of the RF carrier signal x _{c, h} (k) (and the associated first large word width of w _o bit) is chosen to be large enough that the signal / noise ratio in the relevant frequency range is sufficiently high and then still large is enough if there is a gain adjustment of the RF carrier signal. This principle is explained below using the Figure 5 still to be explained.

Such an HF generator CG can be, for example, a synthesizer that works according to the known DDS (direct digital synthesis) method, or can be implemented in the form of a look-up table (LUT). A DDS synthesizer has the advantage that the carrier frequency can be easily adapted to the electrical properties, such as the resonance frequency of the analog circuit part A-C, while using a look-up table reduces the digital circuitry and thus the hardware costs. In addition, further HF generators are conceivable that meet the requirements mentioned.

Since digital / analog converters with a high sampling rate usually have a relatively limited resolution, the relatively large first word width of w ₀ bits of the RF carrier signal x _{c, h} (k) is preferably reduced accordingly to the second word width of w ₁ bit.

In order to achieve this without significantly reducing the signal-to-noise ratio in the relevant frequency band, the digital RF carrier signal x _{c, h} (k) is first preferably fed to the unit NS for noise shaping, with which the quantization noise in in is known to concentrate in areas outside the relevant frequency band and is thus shifted into frequency ranges that are not required to record the relevant LF bandwidth (ie usually about 40 kHz symmetrically about the carrier frequency). This maintains a high signal-to-noise ratio in the relevant frequency band and generates an RF carrier signal x _c (k) with the reduced second word width of w ₁ bit, which is re-quantified in accordance with the resolution of the digital / analog converter D / A ,

Figure 3 shows the HF generator CG and a possible equivalent circuit diagram of the unit NS for noise shaping. The z-transfer function realized with this is: $${\mathrm{X}}_{\mathrm{c}}\left(\mathrm{z}\right)={\mathrm{X}}_{\mathrm{c}.\mathrm{H}}\left(\mathrm{z}\right)+\mathrm{e}\left(\mathrm{z}\right)*\left(\mathrm{1}-\mathrm{H}\left(\mathrm{z}\right)\right)$$

This means that the quantization error e (k) is spectrally weighted by H (z) and thus the re-quantization error for the strongly over-sampled RF carrier signal can be minimized in the relevant frequency band.

If, on the other hand, one can (depending on the application) accept a somewhat lower signal / noise ratio in the relevant frequency band, then alternatively the HF generator CG can also be dimensioned so that it has a digital HF carrier signal x _c (k) with a reduced, adapted to the resolution of the digital / analog converter second word width of w ₁ bit, so that one can then do without the unit NS for noise shaping.

In both cases, the digital RF carrier signal x _c (k) is fed to the digital / analog converter D / A, with which an analog RF carrier signal x _c (t) is generated therefrom in a manner known per se.

This analog RF carrier signal x _c (t) is fed to the microphone unit Cm and serves to convert the changes in the capacitance of the microphone produced by the incident sound waves s (t) (LF audio signal) into a DSB-AM audio signal x _DSB (t) implement, preferably by a high frequency method with amplitude modulation and suppressed carrier frequency. For this purpose, reference is made to the explanations below in connection with the Figures 6 to 9 ,

The analog DSB-AM audio signal x _DSB (t) generated by the microphone unit Cm is then fed to the analog / digital converter A / D. In order to generate a high-quality digital audio signal, it has the aforementioned highest possible (third) resolution with the third word width of w ₂ bits. Furthermore, since it is a bandpass signal, the DSB-AM audio signal x _DSB (t) is (oversampled) sampled at the second sampling rate f _{s, 2} , which in turn is higher than it is after the Sampling theorem is required for the bandwidth of 40 kHz occurring in the bandpass signal.

This is preferably implemented using an analog / digital converter which works according to the known successive approximation (SAR) method.

SAR analog / digital converters of this type have the advantage that they have a relatively low converter latency, which for the following are also based on FIG Figure 5 Gain control described is of particular importance. In addition, a very high signal-to-noise ratio can be achieved by using such high-performance analog / digital converters.

Designs of such converters are known and suitable, for example with a resolution of up to 18 bits and a sampling rate of up to 15 MHz or more.

The digital DSB-AM audio signal x _DSB (m) present at the output of the analog / digital converter A / D with the said high (third) resolution with the third word width of w ₂ bits and the said second sampling rate f _{s, 2} then becomes supplied to the demodulator Dm, with which it is synchronously demodulated, so that a digital baseband signal x _dem (m) with the second sampling rate f _{s, 2} and the fourth resolution or fourth word width of w ₃ bits results.

The demodulator Dm is preferably a coherent demodulator whose RF demodulator signal x _{d, h} (m) with the second sampling rate f _{s, 2} and a desired sixth resolution or sixth word width of w ₅ bits is digitally generated and used for demodulation. Since the frequency of the RF carrier signal x _{c, h} (k) is known, the RF demodulator signal can be generated in a frequency-synchronous manner.

Figure 4 shows an example of a block diagram of such a demodulator, which has a demodulator / generator unit DmG and preferably a unit PhE for phase estimation, at the input of which the demodulated digital baseband signal x _dem (m) is present and whose output is connected to the demodulator / generator unit DmG is to shift or adjust the phase of the RF demodulator signal x _{d, h} (m) generated by the demodulator / generator unit DmG, with which the DSB-AM audio signal x _DSB (m) is demodulated, in order to achieve coherent demodulation , This is necessary because the phase of the HF carrier signal x _DSB (t) generated by the microphone unit Cm and modulated with the audio signal is usually not known.

However, if the exact phase of the RF carrier signal x _DSB (m) at the input of the demodulator / generator unit DmG is determined, for example, by prior phase measurement of the overall system, the unit PhE for phase estimation can alternatively be dispensed with.

The frequency of the RF demodulator signal x _{d, h} (m) generated by the demodulator / generator unit DmG is equal to the frequency of that generated by the RF generator CG RF carrier signal x _{c, h} (k), while the sampling rate f _{s, 2} corresponds to that of the analog / digital converter A / D.

This digital demodulation of the RF carrier signal (bandpass signal) x _DSB (m) enables an almost ideal linearity to be achieved. The accuracy of this linearity only depends on the internal bit resolution, whereby all disadvantageous analog influences are avoided.

Since the demodulated digital baseband signal x _dem (m) now has a bandwidth of approximately 20 kHz and is largely over-sampled at the (second) sampling rate f _{s, 2} , it is preferably filtered and clocked down by the following decimator Dc to generate a digital NF audio signal x _AF (n) with a usual or standard sampling frequency of, for example, 48 kHz (third sampling rate fs).

An additional signal-to-noise ratio and dynamic range are achieved at the same time. For example, if you assume oversampling with a factor of 100 (which corresponds to a sampling frequency of the analog / digital converter A / D of 4.8 MHz) and white quantization noise, there is an additional signal / noise ratio and dynamic range of 20 dB, which corresponds to an additional resolution of about 3 bits.

The decimation process can be carried out in several stages. The decimator Dc can be implemented, for example, by using FIR (finite impulse response) filter topologies such as, for example, CIC (cascaded integrator comb) filters or polyphase filters. To minimize the decimation latency, IIR (infinite impulse response) filters in particular can be used. In this case, a major criterion in the design of these filters is to maintain a linear phase in the audio frequency band.

Figure 5 shows a schematic block diagram of an exemplary second embodiment of a circuit arrangement according to the invention, which also results from the three above mentioned components, namely a digital circuit part DC, an analog circuit part AC and a mixed analog / digital circuit part MC.

The same or corresponding components and signals are given the same names in this block diagram as in the block diagram of Figure 2 ,

Compared to the digital circuit part DC the in Figure 2 In the circuit arrangement shown, this embodiment additionally comprises an optional first correction unit Corr1, connected between the output of the analog / digital converter A / D and the input of the demodulator Dm, a sound pressure level detector / predictor SPL-DP, which is also connected to the output of the analog / Digital converter A / D and further connected to the first correction unit Corrl, and a unit GC for gain control, which is controlled by the sound pressure level detector / predictor SPL-DP and by means of a first multiplier M1 that generated by the HF generator CG RF carrier signal x _{c, h} (k) with a first gain factor g ₁ (k) and by means of a second multiplier M2, the output signal of the first correction unit Corrl with a second gain factor g ₂ (m).

Furthermore, an optional second correction unit Corr2 is provided, which is connected to the output of the optional decimator Dc. The unit NS for noise shaping is again optional here.

This embodiment takes into account in particular the fact that high sound pressure levels can lead to an overload of the analog circuit part A-C and the analog / digital converter A / D. For this reason, the analog output level of microphones can usually be set, e.g. the recorded LF signal is attenuated.

A particular advantage of the embodiment according to the invention is that such adjustability or attenuation of the LF signal is not necessary.

Since the sensitivity of the microphone is directly proportional to the amplitude of the analog RF carrier signal x _c (t) and according to the invention the RF carrier signal is generated in the digital plane, the sensitivity can be increased and decreased digitally without additional analog circuits and thus the dynamic range of the microphone can be expanded or optimized accordingly.

The ability to digitally change the RF carrier signal is one of the major advantages of generating the RF carrier signal in the digital plane. Since the RF carrier signal is generated in the frequency range of interest with a high signal-to-noise ratio as explained above, its intensity can be reduced before it is subsequently subjected to noise shaping in the unit NS, thereby maintaining an optimal signal-to-noise ratio.

Thus, as in the Figures 3 and 5 is shown, the setting of the sensitivity preferably before the noise shaping, namely by the said application of the digital RF carrier signal x _{c, h} (k) with the first gain factor g ₁ (k) ≤ 1 (and of course also the RF generator CG can be controlled directly with the first gain factor g ₁ (k) accordingly).

Assuming a voltage level of -20 dBFS at a 94 dBSPL input level, the maximum processable sound pressure level would be 114 dBSPL. In order to extend this range by 30 dB to a level of 144 dBSPL (if the microphone in question can detect such a sound pressure level), the amplitude of the RF carrier signal can be set in such a way that an amplification factor g ₁ ( k) of about 1/32.

ist, wird die Empfindlichkeit des Mikrofons vermindert, was zu dem erforderlichen zusätzlichen Dynamikbereich für den Analog/Digital-Wandler A/D führt.By using such a factor g ₁ (k), the output resolution is reduced by about 5 bits. However, this loss can largely be compensated for again by the noise shaping explained above, so that the RF carrier signal continues to have a sufficiently high signal / noise ratio. There $$\mathrm{G}{\left(\mathrm{k}\right)}_{\mathrm{dB}}=-\mathrm{20}{\log }_{\mathrm{10}}\left(\mathrm{32}\right)\approx -\mathrm{30}\mathrm{dB}$$

the sensitivity of the microphone is reduced, which leads to the required additional dynamic range for the analog / digital converter A / D.

In this case, the above-mentioned first high resolution or the first large word width of w ₀ bits of the RF carrier signal x _{c, h} (k) can, for example, be selected such that it already provides the lost 5 bits before the noise is formed, ie the RF carrier signal x _{c, h} (k) before the noise shaping is resolved by 5 bits higher than the RF carrier signal x _c (k) after the noise shaping.

In addition, however, further measures are preferably taken in order to be able to process the dynamic range of modern or future microphones of between 140 and 150 dB or more, which is very high compared to the analog / digital converter A / D, largely without loss.

According to the invention, use is made of the so-called "gain ranging" in which the level range of the microphone is automatically adapted to a recorded sound pressure level and the dynamic range of the entire system is thus automatically changed or increased and decreased accordingly.

It is known to use analog / digital converters with several channels for this purpose, each of which has different amplification factors, so that several analog / digital converted signals are produced which then have to be combined accordingly on the digital level and postprocessed to produce an audio signal to generate an increased dynamic range, ie a higher bit resolution. However, the signal / noise ratio is not improved, which is considered to be disadvantageous.

According to the invention, this problem is solved in the circuit arrangement according to Figure 5 solved by an automatic sensitivity adjustment or adjustment of the microphone when occurrence or the prediction of high sound pressure levels.

For this purpose, the sound pressure level detector / predictor SPL-DP is provided, which unit GC for gain control, the first gain factor g ₁ (k) mentioned to act on the RF generator CG or the RF carrier signal x _{c, h} ( k) generated, controlled.

The SPL-DP sound pressure level detector / predictor is used to record or predict particularly high sound pressure levels. Since, as has already been mentioned above, a SAR analog / digital converter A / D is preferably used, and the digital DSB-AM bandpass signal x _DSB (m) is strongly oversampled, the system can go to high largely without delay Sound pressure levels react.

The unit GC not only generates the first amplification factor g ₁ (k) for the application of the RF carrier signal x _{c, h} (k), but also the second amplification factor g ₂ (m) with which the digital multiplier M2 is used DSB-AM audio signal x _DSB (m) is amplified or multiplied at the input of the demodulator Dm. The unit GC generates the second gain factor g ₂ (m) in such a way that the amplitude adjustment or multiplication of the RF carrier signal x _{c, h} (k) is compensated for by the first gain factor g ₁ (k) and thus a corrected digital one DSB-AM audio signal x ^∼ _DSB (m) is generated, which is supplied to the demodulator Dm.

In order to achieve the most exact possible compensation, the second gain factor g ₂ (m) is preferably determined not only by inverting the first gain factor g ₁ (k), but also the internal delays and filter properties of the circuit and the change in the sampling rate in the signal path between the first multiplier M1 or the unit NS for noise shaping and the second multiplier M2 is also included.

As a result of this automatic adjustment of the sensitivity in both branches, the bits lost in the digital / analog converter D / A are transferred to the analog / digital converter A / D, which leads to an increased dynamic range. The signal / noise ratio after the analog / digital conversion remains largely unchanged, as in the known approach described above. The analog / digital converter A / D can thus also be referred to as a floating point A / D converter.

The optional first correction unit Corrl is preferably used to, before the digital DSB-AM audio signal x _DSB (m) is loaded with the aforementioned second amplification factor g ₂ (m), possibly remaining erroneous samples due to different filter characteristics between the digital / analog conversion and to compensate for the analog / digital conversion, in particular if the factor g ₂ (m) is determined only by inverting the first gain factor g ₁ (k) or is not accurate enough the delayed, filtered and inverted value of g ₁ (k) equivalent. For this purpose, the circuit can be measured precisely, for example, so that the required correction can be calculated and applied. Another possibility is to work with predicted samples at the critical points.

The first amplification factor g ₁ (k), with which the digital RF carrier signal x _{c, h} (k) is applied, can be both a constant factor and a monotonically decreasing factor over time. It only has to be ensured that no clipping occurs in the analog circuit part AC due to the high sound pressure level.

If necessary, the optional second correction unit Corr2 can finally be provided, with which the temporary LF audio signal x ^∼ _AF (m) output by the decimator Dc is further corrected and the digital LF audio signal x _AF (n) is generated.

To explain the analog circuit part AC is finally on the Figures 6 to 8 Referred. These figures each show the preferred use of a condenser microphone with a push-pull converter. As an alternative to using such a push-pull microphone shows Figure 9 a condenser microphone with standard converter C ₁ (t) and reference condenser C ₂ . This standard microphone can be used instead of the push-pull microphone Figures 6 to 8 can be used by connection with the respective identical contact points K1, K2, K3. The following explanations therefore apply to both types of microphones.

As already mentioned, the condenser microphone is preferably operated using a high-frequency method with amplitude modulation, in particular double-sideband amplitude modulation (DSB-AM).

The specific circuit with which the microphone is operated is selected depending on the properties of the microphone, the intended application, the required recording quality and other criteria.

An (analog) DSB-AM audio signal x _DSB (t) is therefore preferably generated with the analog circuit part AC. For this purpose, the analog circuit part AC comprises a transformer Tr with a primary side L _p and a secondary side L _s , which has a center tap or is formed by two identical inductors L _s1 , L _s2 connected in series.

Parallel to the secondary side L _s is a capacitive voltage divider, namely the series circuit made up of the two capacitors C ₁ (t) and C ₂ (t) of the push-pull microphone or the series circuit made up of the capacitor C ₁ (t) of the standard microphone and the reference capacitor C ₂ , so that an RF bridge circuit is formed in this way.

The RF carrier signal x _c (t) generated as explained above is applied to the primary side L _{p of} the transmitter Tr.

The DSB-AM audio signal x _DSB (t) (microphone signal with suppressed HF carrier signal x _c (t)) present at the output connection of the bridge is processed in accordance with Figures 6 to 8 on the membrane between the two capacitors C ₁ (t), C ₂ (t) of the (push-pull) microphone or on the Contact point K3 decoupled. However, it is also possible to carry out the coupling at the center tap of the secondary side L _{s of} the transformer Tr, while the membrane or the contact point K3 itself is coupled to a reference voltage.

In the simplest version according to Figure 6 the decoupled DSB-AM audio signal x _DSB (t) after impedance conversion and amplification is fed directly to the analog / digital converter A / D of the mixed analog / digital circuit part MC.

However, the DSB-AM audio signal x _DSB (t) is preferably coupled out via a resonance circuit which is matched to the output connection of the bridge.

When executed according to Figure 7 the DSB-AM audio signal x _DSB (t) is coupled out as a current signal i _x (t) via a resonance circuit with a series connection of an inductance L and a resistor R connected in series with the output connection of the bridge.

The generated current i _x (t) is according to the in Figure 7 The embodiment shown is preferably detected or measured by means of a transimpedance amplifier which is implemented by an amplifier V (operational amplifier) which is connected in an inverting manner and which is fed back through the parallel connection of a capacitance C _f with a resistor R _f , and at whose output the DSB-AM audio signal x _DSB (t) (voltage signal) is present.

The inductance L determines the resonance frequency f _{r of} the circuit (which corresponds to the lowest source impedance), while the frequency bandwidth (which is 40 kHz for the DSB-AM audio signal) and thus the Q factor of the circuit is determined only by the resistance R. becomes.

An advantage of this embodiment is that, due to the low source impedance, a DSB-AM audio signal x _DSB (t) with improved noise properties can be achieved.

When executed according to Figure 8 the DSB-AM audio signal x _DSB (t) is coupled out as a voltage signal u _x (t) via a resonance circuit with a series connection of an inductance L and a resistor R connected in parallel to the output connection of the bridge.

The generated voltage u _x (t) is according to the in Figure 8 The embodiment shown is preferably detected or measured by means of an amplifier which is implemented by an amplifier V (operational amplifier) connected in a non-inverting manner and fed back by a voltage divider R ₁ / R ₂ , and at the output of which the DSB-AM audio signal X _DSB (T ) is present.

One advantage of this embodiment is that the voltage signal u _x (t) has a higher sensitivity (depending on the Q factor), but the source impedance is also higher.

Claims (15)

Method for operating a condenser microphone using a high-frequency method with amplitude modulation, characterized by the following steps: Generating a digital RF carrier signal (x _{c, h} (k); x _c (k)), Generating an analog RF carrier signal (x _c (t)) by digital / analog conversion of the digital RF carrier signal (x _{c, h} (k); xc (k)), Generating an analog DSB-AM audio signal (x _DSB (t)) by double-sideband amplitude modulation (DSB-AM) of the analog RF carrier signal (x _c (t)) using the microphone, Generating a digital DSB-AM audio signal (x _DSB (m)) by analog / digital conversion of the analog DSB-AM audio signal (x _DSB (t)), and - Demodulating the digital DSB-AM audio signal (x _DSB (m)) and generating a digital baseband audio signal (x _dem (m)). Method according to claim 1,
in which the digital RF carrier signal (x _{c, h} (k); x _c (k)) is over-sampled at a first sampling rate (f _{s, 1} ). Method according to claim 1,
in which the digital RF carrier signal (x _{c, h} (k)) is generated with a first high resolution which is greater than a maximum resolution in the digital / analog conversion, and in which a digital RF carrier signal before the digital / analog conversion (x _c (k)) is generated which has a reduced second resolution adapted to the maximum resolution in the digital / analog conversion and in which the signal / noise ratio in the DSB-AM- Frequency band of the audio signal is at least largely compensated. Method according to claim 1,
in which the amplitude of the digital RF carrier signal (x _{c, h} (k); x _c (k)) is digitally adjustable to change the sensitivity of the microphone. Method according to claim 1,
in which a sound pressure level is detected or predicted in the digital DSB-AM audio signal (x _DSB (m)) and a first amplification factor (g ₁ (k)) is generated as a function thereof, with which the amplitude of the digital RF carrier signal (x _{c , h} (k); x _c (k)) to adapt the sensitivity of the microphone to the recorded or predicted sound pressure level. Method according to claim 5,
in which a second amplification factor (g ₂ (m)) is generated by inverting the first amplification factor (g ₁ (k)) in such a way that by changing the digital DSB-AM audio signal (x _DSB (m)) Amplitude of the digital RF carrier signal (x _{c, h} (k); x _c (k)) is at least largely compensated for by the first amplification factor (g ₁ (k)). Method according to claim 1,
in which the digital DSB-AM audio signal (x _DSB (m)) is generated by analog / digital conversion of the analog DSB-AM audio signal (x _DSB (t)) oversampled with a second sampling rate (f _{s, 2} ). A method implemented in hardware or software for performing at least one of the steps according to one of claims 1 to 7. System for data processing with a digital signal processor or an FPGA for performing at least one of the steps according to one of Claims 1 to 7. Digital signal processing unit with instructions for executing the method according to claim 8 with a system according to claim 9. Circuit arrangement for operating a condenser microphone according to a method according to one of claims 1 to 7, with: a digital RF generator (CG) for generating a digital RF carrier signal (x _{c, h} (k); x _c (k)), a digital / analog converter (D / A) for generating an analog RF carrier signal x _c (t) from the digital RF carrier signal (x _{c, h} (k); xc (k)), a microphone unit (Cm) with a condenser microphone for generating an analog DSB-AM audio signal (x _DSB (t)) by double-sideband amplitude modulation (DSB-AM) of the analog RF carrier signal (x _c (t)) by means of the microphone, - an analog / digital converter (A / D) for generating a digital DSB-AM audio signal (x _DSB (m)) from the analog DSB-AM audio signal (x _DSB (t)), and - A digital demodulator (Dm) for demodulating the digital DSB-AM audio signal (x _DSB (m)) and for generating a digital baseband audio signal (x _dem (m)). Circuit arrangement according to claim 11,
with a unit (NS) for noise shaping, for generating a digital RF carrier signal (x _c (k)) with a reduced resolution adapted to the digital / analog converter (A / D), and for noise shaping the RF carrier signal for at least extensive compensation of a signal / noise ratio in the DSB-AM frequency band of the audio signal which is impaired by the reduction in the resolution. Circuit arrangement according to claim 11, with: a sound pressure level detector / predictor (SPL-DP) for detecting or predicting a sound pressure level in the digital DSB-AM audio signal ((x _DSB (m)), and - A unit (GC) for gain control, which is controlled by the sound pressure level detector / predictor (SPL-DP) and generates a first gain factor (g ₁ (k)) with which the amplitude of the digital RF carrier signal (x _{c, h} (k); x _c (k)) for adapting the sensitivity of the microphone to the detected or predicted sound pressure level can be changed by means of a first multiplier (M1). Circuit arrangement according to claim 13,
in which the unit (GC) for gain control generates a second gain factor (g ₂ (m)) by inverting the first gain factor (g ₁ (k)) in such a way that by applying the digital DSB-AM audio signal (x _DSB ( m)) by means of a second multiplier (M2) the change in the amplitude of the digital RF carrier signal (x _{c, h} (k); x _c (k)) is at least largely compensated for by the first amplification factor (g ₁ (k)). Circuit arrangement according to claim 11,
in which the microphone unit (Cm) has an RF transmitter bridge circuit (L _s1 , L _s2 ) with a capacitive voltage divider formed by the microphone for generating the analog DSB-AM audio signal (x _DSB (t)), the DSB-AM Audio signal (x _DSB (t)) is coupled out via a resonance circuit matched to an output connection of the bridge.

Images