Sound reproducing method

Abstract

A sound-reproducing method for converting a digital audio signal to an analog audio output signal is characterized in that an audio input signal is modulated by a DSP to give a frequency-modulated audio output signal and the frequency-modulated audio output signal is demodulated.

Description

[0001] TO NWIEDE RG AB EV ER FA HREN

[0002] The invention relates to a sound reproduction method for converting a digital audio signal into an analog audio output signal, characterized in that an audio input signal is modulated by a DSP and / or an ASIC into a frequency-modulated audio output signal and the frequency-modulated audio output signal is demodulated unchanged.

[0003] Current models of digital-to-analog converters (D / A converters) use common semiconductor technology and standard components. This limits the designer's design options and thus the ability to influence the sound.

[0004] In already known methods for FM broadcasting, an analog stereo signal is digitized, then mixed to an MPX signal and then provided as high-frequency FM.

[0005] It is therefore an object of the present invention to provide a sound reproduction method and a sound reproduction DA converter which provides an improved sound signal in the digital / analog conversion (D / A conversion) by means of the combination of frequency modulation and the use of a demodulator, has a high accuracy, a low circuit complexity and is basically suitable for all known data rates, preferably without the use of semiconductors.

[0006] Of particular note are the digital zero-crossing interpolation for frequency generation and the filter linearization of the demodulator. Furthermore, the use of electron tubes in direct connection with the conversion process or a DAC also represents an innovation. This task is solved by a sound reproduction method for converting a digital audio signal into an analog audio output signal, characterized in that an audio input signal is modulated by a digital signal processor (DSP) and / or an ASIC into a frequency-modulated audio output signal, the frequency-modulated audio output signal is demodulated unchanged, and the audio output signal is then output as an audio signal. This sound reproduction method differs from the existing sigma-delta or R2R methods, since the accuracy of the conversion depends solely on the hardware and software of the DSP.

[0007] The invention discloses a novel converter and a novel conversion method, namely True Frequency Modulation (TFM). In the TFM method, the digital signal processor (DSP) and / or the ASIC converts the digital signals into a frequency-modulated audio output signal that can be demodulated by a downstream functional unit. This sound reproduction method promises very high accuracy and reduced circuit complexity in the analog section, since the actual D / A conversion already takes place in the DSP and / or ASIC. Furthermore, the analog section or demodulator does not need to be modified for higher data rates. DSPs and / or ASICs have the advantage of lower costs while simultaneously offering excellent properties, such as high frequency and hardware specifically suited for audio signal processing.This advantage is especially evident over FPGAs (Field Programmable Gate Arrays), whose hardware can be configured by a program, but whose logic generation requires languages ​​such as VHDL (Very High Speed ​​Integrated Circuit Hardware Description Language). In particular, there are special processors for processing stereo signals that provide the necessary hardware for two audio channels, which is why the faster DSP and / or ASIC are the better alternative.

[0008] In one embodiment, the audio input signals are processed using fixed-point processors and / or floating-point processors, preferably a combination of fixed-point and floating-point processors, preferably in the audio signal processing area within a single unit. Even more preferably, decoded audio input signals are processed accordingly. The floating-point processor is a specialized coprocessor that manipulates numbers faster than the basic microprocessor circuitry used in the prior art. In addition, the signal values ​​are more accurate than with the fixed-point processors used in this area. This results in a more accurate signal that can be converted into a better audio signal. Additionally, there is less time delay, and fewer noise signals occur.

[0009] In one embodiment, an audio input signal, preferably an audio input signal, is input to a digital signal processor (DSP) and / or ASIC and can be transmitted to various modules of the DSP and / or ASIC.

[0010] A DSP is a specialized microprocessor chip whose architecture is optimized for the requirements of digital signal processing. DSPs are widely used in audio signal processing, telecommunications, digital image processing, radar, sonar, and speech recognition systems, as well as in common consumer electronics devices such as CD players, amplifiers, mobile phones, hard disk drives, and high-definition television products. The goal of using a DSP is typically to measure, filter, or compress continuous, real-world analog signals. Most general-purpose microprocessors can also successfully execute digital signal processing algorithms, but may not be able to keep up with such processing continuously in real time. Dedicated DSPs also generally have better power efficiency.

[0011] An ASIC (application-specific integrated circuit) as defined in this document is a microchip designed for a specific task, for example, a specific transmission protocol, for data center hardware such as switches, or for mobile computers such as smartphones or wearables. In contrast, there are general-purpose integrated circuits, such as the microprocessor or the main memory modules (RAM - Random Access Memory) in a computer. ASICs are used in a variety of applications, including vehicle emission control or environmental measurements. An ASIC can be prefabricated for a specific application or manufactured as a custom product (usually using prefabricated individual components as building block solutions) for a specific customer application. For the purposes of this document, an ASIC is also an FPGA (Field Programmable Gate Array), i.e.A digital integrated circuit (IC) into which a logic circuit can be loaded. The term can be translated as a field-programmable (logic) gate arrangement (i.e., on-site at the customer's site). The essential basic structure of an FPGA is an array of basic blocks, each with a simple programmable lookup table (LUT) and a 1-bit register (flip-flop). Depending on the number of available inputs, the LUTs can implement any n-digit binary function. The desired function is programmed by storing the defining truth table in the SRAM cells of the LUT, and the function is calculated by reading the memory address determined by the inputs. For a long time, LUT structures with four binary inputs were common.To reduce the complexity of LUT-to-LUT connections, newer FPGAs are switching to LUTs with up to six inputs to implement functions with multiple inputs. In addition to the LUTs, the interconnection of components on an FPGA is also largely freely configurable. Multiplexer structures in the basic blocks often enable very fast local signal paths for integrating or bypassing the flip-flop, for feedback from its output, for connecting neighboring blocks, and the like. For more distant connections, a grid of immense bus structures lies between the basic blocks, to which inputs and outputs can be connected. Additional programmable switching components at the grid intersections allow signal distribution across the entire chip.

[0012] The audio input signal transmitted to the frequency modulator (FM modulator) is preferably a digital audio signal and is modulated into a frequency-modulated audio output signal. Digital frequency modulation (FM) is a form of digital modulation or line coding and, unlike frequency modulation, does not produce an analog, continuously modulated signal, but rather a time-discrete digital bit sequence. Specifically, digital frequency modulation (FM) is the encoding of information in a carrier wave by changing the wave's instantaneous frequency. This technology is used in telecommunications, broadcasting, signal processing, and data processing. One advantage of digital frequency modulation is that it has a higher signal-to-noise ratio and therefore suppresses high-frequency interference better than an amplitude modulation (AM) signal of the same power.Under sufficiently strong conditions, the FM signal contains pronounced high-frequency noise artifacts that are audible even at low volumes and with less complex tones. This form of modulation contrasts with amplitude modulation, in which the amplitude of the carrier wave changes while the frequency and phase remain constant.

[0013] In another embodiment, the audio input signal is converted by an input decoder before FM modulation. Since the audio input signal can be introduced from multiple inputs, the audio input signal is converted accordingly and preferably channel-separated in the signal processing before being modulated into an FM signal by the FM modulator. The bit rate and sampling frequency are detected in the decoder and passed to the display via the input / output unit. To avoid noise when switching or when a zero signal is present, there is a mute function that switches off the output stage.

[0014] In a preferred embodiment, the frequency modulated audio output signal is routed out of the DSP and / or ASIC and is not input back to the DSP and / or ASIC.

[0015] The frequency-modulated audio output signal is then transmitted unchanged and demodulated into an analog audio output signal. According to the invention, the signal remains unchanged during transmission, forwarding, input, output, and similar signal movements.

[0016] In a more preferred embodiment, the DSP or ASIC and the FM demodulator are separated from each other by being divided into functional units in the form of a digital part and an analog part. In an even more preferred embodiment, the DSP or ASIC and the FM demodulator are separated from each other by their respective housings.

[0017] In another embodiment, the separate functional units have their own power supply. Separate power supplies can prevent signal interference caused by crosstalk from certain functional units and can better distribute the voltages as needed.

[0018] The demodulation circuit converts an FM signal into an amplitude response. FM demodulators are designed for narrowband or wideband operation, depending on the signal bandwidth. A narrowband demodulator (e.g., for voice radio) only captures a frequency deviation of approximately 10 kHz, while a wideband demodulator (as is common in FM radio) must process a frequency deviation of approximately 150 kHz.

[0019] For FM demodulation within the meaning of the invention, various circuits can be used, such as edge and push-pull edge discriminators, Foster-Seeley and Riegger phase discriminators, ratio detectors or ratio discriminators, coincidence demodulators, and counting discriminators. The edge discriminator is particularly preferred, and a modified single-ended edge discriminator is even more preferred. Edge discriminators can only convert narrow-bandwidth FM signals into AM signals with low distortion. The downstream AM demodulator generates the analog signal (AF signal) from this. However, it has surprisingly been found that, within the framework of "true frequency modulation," all known methods are suitable for demodulation without reducing sound quality. However, thanks to digital linearization within the meaning of the invention, higher bandwidths can also be used in this case.In addition, an edge discriminator, and even more so a modified single-ended edge discriminator, allows a demodulator to be constructed as discretely as possible.

[0020] In a preferred embodiment, a frequency measurement is performed at a specific time. An equivalent output voltage can be generated from this measurement.

[0021] In another embodiment, the edge discriminator and, for example, an electron tube are connected in parallel in a grid-base configuration. This circuit is suitable for high-frequency applications and offers high gain.

[0022] In another embodiment, a resonance curve and correction values ​​for frequency generation are generated from the output voltage values ​​as a function of frequency. A further advantage of the tuning circuit is that deviations due to component tolerances or aging processes can also be compensated.

[0023] In a preferred embodiment, digital linearization of the resonance curve is performed. During edge demodulation, the FM reception signal is fed to a slightly detuned resonant circuit and converted at its resonance edge into an amplitude-modulated RF signal. After subsequent AM demodulation, however, this signal is only approximately linear with respect to the frequency change. Digital linearization can compensate for the non-linearity of the resonance curve, for example, in the range from 9 MHz to 11 MHz, which would otherwise lead to erroneous amplitude values.

[0024] A sound reproduction method within the meaning of this document is a method for reproducing, preferably music or other audio formats, in entertainment electronics for the home and / or professional applications, such as recording studios or sound systems. The sound reproduction method according to the invention is intended and suitable for implementation on entertainment electronics devices or on devices in professional recording studios. The sound reproduction method according to the invention is explicitly not a method for transmitting audio signals in the broadcasting sector. No MPX signal is mixed and then provided as high-frequency FM for the VHF range.

[0025] In the method according to the invention, an analog frequency modulation is generated from a digital input signal. The frequency modulation process is used with a pure DAC. To obtain an audio signal, the generated analog FM signal only needs to be demodulated. The audio signal is directly suitable for playback on consumer electronics devices or professional applications.

[0026] In a further embodiment, the sound reproduction method for converting a digital audio signal into an analog audio output signal is characterized in that the audio output signal is transmitted from the digital signal processor (DSP) and / or ASIC by means of one or more high-frequency transmitters, preferably by means of one or more electron tubes. In addition to this new conversion sound reproduction method, various techniques can be used to process high input and operating voltages. These allow greater design freedom and optimal configuration for a natural sound. With electron tubes, significantly higher voltages can be used for conversion. The harmonic spectrum and intermodulation distortion behave much more favorably when signal processing is carried out with electron tubes, resulting in a more harmonious and musical sound.

[0027] Electron tubes can also be used in the output stage to make the perceived "digital" sound of transistor devices somewhat more natural. The novel use of tubes within the conversion device, rather than in the device's periphery, enhances this effect, thus enabling a significantly more natural sound.

[0028] In a further embodiment, the audio output signal from the FM modulator is transmitted unchanged and / or directly to the FM demodulator via electron tubes.

[0029] In another embodiment, the separate functional units, digital part and analog part, are connected to each other only via electron tubes.

[0030] In one embodiment, electron tubes have a high voltage gain (approximately 20 to 100) and sensitivity. The input sensitivity and gain can be adjusted using current or voltage feedback. This allows for a wider frequency response and lower distortion.

[0031] In another embodiment, a bias adjustment is performed. This also significantly influences the distortion spectrum, which can even have a positive effect in the case of a slight mistuning, as the K2 component is slightly increased. Even with higher levels of distortion, a more pleasant listening experience is achieved.

[0032] In a further embodiment, the sound reproduction method for converting a digital audio signal into an analog audio output signal according to one of the preceding claims is characterized in that an audio input signal is channel-separated, wherein the audio input signal is separated into at least two channels, preferably into at least four channels, more preferably into four channels. By separating the signals into different channels, the separate signals can have a higher signal intensity without signal loss occurring during transmission.In a further embodiment, the sound reproduction method for converting a digital audio signal into an analog audio output signal according to one of the preceding claims is characterized in that the channel-separated signals are processed separately from one another, preferably by a dual mono architecture, wherein the separation is preferably carried out in the DSP and / or ASIC and / or wherein the separation preferably takes place before transmission to the FM modulator.

[0033] In one embodiment, each audio input signal is an audio signal that is transmitted to the DSP, preferably an audio signal that is transmitted to the DSP and / or ASIC for the first time. In a further embodiment, the audio input signal is characterized in that it is transmitted to the sound reproduction DAC according to the invention for the first time. This input is preferably transmitted via an interface in the form of WLAN, Ethernet, USB, etc. 2 S, AUX, S / PDIF transmitted.

[0034] A dual mono architecture is characterized by an audio signal that is separated into at least two mono channels and / or is independent of each other, but which have a common source and are output independently of each other as an audio output signal. This avoids any interference and influences that are still present with other signal types such as a stereo signal. This improves the channel separation between the stereo channels, and the dual mono architecture is intended to prevent crosstalk in the circuit. A stereo signal is separated into two channels that have a common source and are always dependent on each other. According to the invention, a preferably digital audio signal is separated in the DSP and / or ASIC into at least two mono channels and the separated audio signals are processed separately from each other. The separated audio signals are preferably not modified and / or transmitted by any functional unit of the other separate audio signal.Preferably, the separate audio signals have their own functional units that process and / or transmit only the audio signal from a mono channel. In one embodiment, the channel separation takes place in the signal processing functional unit. The signal processing is preferably part of the DSP and / or ASIC.

[0035] In a further embodiment, the signal is transmitted to the FM modulator functional unit after separation by the signal processing functional unit, wherein preferably one FM modulator is available for each separate audio signal.

[0036] In a further embodiment, digital interfaces such as l 2 S, S / PDIF, and USB on-chip peripherals can be connected as external functional units. For the actual implementation, an algorithm is available in which the modulation index and carrier frequency can initially be freely selected to achieve the best acoustic properties.

[0037] In a further embodiment, the sound reproduction method for converting a digital audio signal into an analog audio output signal according to one or more of the preceding claims is characterized in that the resolution and / or frequency range are determined by the data rate and the frequency range is above 30 kHz, preferably above 50 kHz and particularly preferably above 100 kHz.

[0038] In a further embodiment, the sound reproduction method for converting a digital audio signal into an analog audio output signal is according to one or more of the preceding claims, wherein the frequency modulation is carried out with a maximized FM deviation and / or modulation index.

[0039] If the FM deviation is maximized, the output voltage at the demodulator can become very high, which allows a well-measured range to be achieved with a resolution of 24 bits. In a further embodiment, the sound reproduction method for converting a digital audio signal into an analog audio output signal is characterized in that the FM signal is transmitted from the analog domain to the digital domain via galvanic isolation, preferably from the FM demodulator to the analog domain via one or more RF transmitters coupled to the DSP. Galvanic isolation reduces interference from the DSP and / or ASIC transmitted via the power supplies. Furthermore, no significant signal loss is to be expected on this signal path.

[0040] Preferably, the FM signal is transmitted separately by channel via two RF transformers, which forward the FM signal to preferably two FM demodulation units. By separating the signals into different channels, the separate signals can have a higher signal size without signal loss during transmission.

[0041] In a further embodiment, the sound reproduction method for converting a digital audio signal into an analog audio output signal is characterized in that the audio input signal is frequency-modulated by the FM modulator, wherein preferably the frequency of the applied amplitude value is processed by a voltage-controlled oscillator, preferably by at least two voltage-controlled oscillators, more preferably by at least four voltage-controlled oscillators, even more preferably by one voltage-controlled oscillator each, depending on the amount of channel-separated signals.

[0042] In a further development of the invention, a DDS is used as the voltage-controlled oscillator, the frequency value of which is related to a digital voltage value.

[0043] In signal processing according to the present invention, the transients are preferably calculated from the digital audio input signal and processed by a voltage-controlled oscillator (VCO). VCOs change frequencies, preferably depending on the applied amplitude value, thereby resulting in FM modulation. Channel-separated FM modulators are preferably clocked synchronously, and the channel-separated, frequency-modulated audio output signal is transmitted via a transmitter, such as an RF transmitter.

[0044] In a further embodiment, the sound reproduction method for converting a digital audio signal into an analog audio output signal according to one or more of the preceding claims is characterized in that no semiconductor technology is used for standard components.

[0045] Current models use common semiconductor technology and standard components. This limits the possibilities for modifying the design and thus the sound. Instead, complex algorithms are used and the sampling rate is continually increased (oversampling). However, the digital processing of the original signal (oversampling) results in effects such as oscillations. This is perceived as unnatural and harsh. By dispensing with semiconductor technology, a better musical experience can be ensured.

[0046] In a further embodiment, the sound reproduction method for converting a digital audio signal into an analog audio output signal is characterized in that a maximum relative error of the frequency for the FM modulation is at most 2 -24 This low relative error is preferably ensured by a stable clock supply, preferably enabled by a digital signal processor (DSP) or an application-specific integrated circuit (ASIC).

[0047] In a further embodiment, the sound reproduction method for converting a digital audio signal into an analog audio output signal is characterized in that a signal-to-noise ratio of at least 100 dB, preferably at least 110 dB and particularly preferably at least 120 dB, preferably around a dynamic range of 24 bits. This is achieved by the combination of RF demodulator and DSP and / or ASIC. In particular, the modulation index for the FM is set accordingly high. In conjunction with the transmission bandwidth of the useful signal (NF), this directly influences the frequency deviation of the FM, which additionally depends on the DSP and / or the ASIC, its clock frequency, the clock and the corresponding FM carrier frequency, because the choice of carrier frequency is directly related to the accuracy of the transmission. If this is very high in relation to the clock frequency of the DSP, only a few clock pulses are generated in one period of the carrier frequency.This reduces the frequency resolution. What's important is a high and extremely precise clock frequency, as ensured by the inventive sound reproduction method and the inventive sound reproduction DAC.

[0048] In a further embodiment, the sound reproduction method for converting digital data into an audio output signal is characterized in that the digital data and / or the audio output signal are not processed according to a superheterodyne principle. In a superheterodyne, the reception frequency is first converted to a lower fixed frequency, the so-called intermediate frequency. For this purpose, the signal is mixed with the frequency of an oscillator. The frequency of the oscillator is higher by the intermediate frequency, which is the difference between the two frequencies.

[0049] In a further embodiment, the sound reproduction method for converting digital data into an audio output signal is characterized in that the least significant bit (LSB) can be easily measured with up to a data rate of 24 bits.

[0050] In a further embodiment, the FM signal is extracted in the one or more demodulator units using a phase discriminator and passed on to a low-pass filter, the coefficients of which are preferably based, among other things, on the sampling rate (sampling frequency). The sampling rate is the frequency with which an analog signal is measured and converted into a time-discrete signal. In a further embodiment, the volume control takes place downstream of the RF modulator functional unit, preferably downstream of the low-pass filter, which in this embodiment is downstream of the FM demodulator. Implementation at the digital level has the disadvantage of significant losses in the resolution of the summation of the noise components of the subsequent stages, since these are not controlled.After the low-pass filter, the signal is sufficiently large, and any noise components from the demodulator are attenuated by the control, thus maintaining a fundamentally constant signal-to-noise ratio. Furthermore, overloading of the subsequent tube amplifier units is avoided and a wide dynamic range is provided, allowing even quieter recordings to be fully controlled.

[0051] In a further embodiment, the sound reproduction method is carried out on a consumer electronics device.

[0052] In a further embodiment of the invention, the sound reproduction method is carried out on a device in a recording studio.

[0053] Preferably, the filtered analog signal is amplified to a standardized line level via the output stage to drive an audio amplifier. This provides the option of volume control. This also allows the D / A converter to be used to drive power amplifiers.

[0054] Because the transmission and storage of audio data is primarily digital, the present invention utilizes a subsystem of FM signal transmission to implement a D / A conversion process in a single device. For this purpose, the I2S bus / DSD is used as a standard digital input interface. Its data is used to generate digital frequency modulation. An analog FM signal is then generated using DDS and an RF D / A converter. All signal processing steps take place in a single device. Therefore, no radio transmission standards need to be adhered to.

[0055] In an optional embodiment, the FM parameters are selected to the limits of what is physically possible, for example, to increase the frequency response of the desired signal several times over. This places the frequency range at least partially outside of 30 Hz and 15 kHz, whereas in radio transmission the audio frequency range is limited to 30 Hz and 15 kHz. The signal-to-noise ratio exceeds the signal-to-noise ratio required by the FM standard. This is made possible by maximizing the frequency deviation and the resulting FM gain.

[0056] In a further embodiment of the invention, zero-crossing interpolation is performed. This avoids phase and amplitude jumps, resulting in the digitally generated frequency modulation being virtually identical to analog FM. Therefore, reconstructing the useful signal requires no oversampling, as in sigma / delta converters, nor heavy filtering, as in DA converters, which, for example, only operate at a sampling frequency of 44.1 kHz.

[0057] In an optional embodiment of the invention, the FM demodulation method uses the concept of our modified single-ended edge discriminator to gain more influence on sound-relevant components using a discrete circuit design. This option is not available with integrated semiconductor circuits. Furthermore, the circuit design and dimensions are determined only by the FM parameters, not by the data rates of the input signal to be converted. This eliminates the need to modify the hardware for expected higher data rates.

[0058] In principle, all technically feasible demodulation methods are suitable for the inventive sound reproduction method, although they are not always practical. The discrete design with electron tubes was chosen as an example because it achieves a significant improvement in sound potential. The disadvantage of the nonlinearity of the resonance curve of the modified single-ended slope discriminator was eliminated by the digital characteristic linearization process. This also allows component tolerances and aging processes to be compensated within certain limits. This process can also be applied to the characteristics of other electronic components.

[0059] The problem is further solved by means of a sound reproduction DA converter for converting a digital audio signal into an analog audio output signal, comprising a digital signal processor (DSP) and / or an ASIC comprising an FM modulator, characterized in that the audio output signal is transmitted unchanged to the FM demodulator.

[0060] In a further embodiment, the audio reproduction DAC for converting a digital audio signal into an analog audio output signal is characterized in that the audio reproduction DAC is discretely constructed from the FM demodulation stage onward. A discrete design of the functional units allows for significant influence on sound-relevant components. This is only possible to a limited extent with chip converters.

[0061] A sound reproduction DAC within the meaning of this document is a DAC for the reproduction of music or other audio formats in consumer electronics for the home and / or in a professional recording studio. The sound reproduction DAC according to the invention is intended and suitable for use in consumer electronics devices or on devices in professional devices for outputting audio signals. These devices do not include devices for transmitting audio signals. The sound reproduction DAC according to the invention is explicitly not a device for use in the broadcasting sector for transmitting audio signals.In a further embodiment, the sound reproduction DA converter for converting a digital audio signal into an analog audio output signal is characterized in that the sound reproduction DA converter is separated into a digital part and an analog part (functional upper units), wherein the FM modulator and FM demodulator transmit signals via a functional unit provided for this purpose, preferably directly via at least one RF transmitter, more preferably via at least two RF transmitters, wherein the transmitter preferably comprises at least one electron tube, more preferably at least two electron tubes.

[0062] In a further embodiment, the DSP and / or ASIC comprises an input selector, an input and output unit and an FM modulator.

[0063] In a preferred embodiment, the DSP and / or ASIC additionally comprises an input decoder and signal processing. The input selection provides several digital inputs (ports) for the various source signals with their respective data formats. The interfaces l are preferred. 2 S, S / PDIF, and USB are implemented. Ethernet and Wi-Fi connections are also optional.

[0064] The various ports can be selected via an external controller and routed to the input decoder via the DSP and / or ASIC's internal input selection. The decoder detects the bit rate and sampling frequency and forwards them to the display via the input / output unit. To avoid noise when switching or when a zero signal is present, there is a mute function that disables the output stage.

[0065] In one embodiment, fixed-point processors and / or floating-point processors, preferably a combination of fixed-point and floating-point processors, are used in the field of audio signal processing, preferably within a single unit. The floating-point processor is a specialized coprocessor that manipulates numbers faster than the basic microprocessor circuitry used in the prior art. Additionally, the signal values ​​are more accurate than with the fixed-point processors used in this field. This results in a more accurate signal that can be converted into a better audio signal. Additionally, there is less delay and fewer noise signals.

[0066] In a preferred embodiment, the signal processing section calculates the transients from the digital audio input signal and forwards them to the two channel-separated voltage-controlled oscillators (VCOs). The VCOs change their frequency depending on the incoming amplitude value, resulting in FM modulation.

[0067] For the purposes of this invention, voltage-controlled oscillators (VCOs) also refer to frequency generators using direct digital synthesis (DDS), which operate like a VCO in that a digital voltage value produces an analog output frequency. Direct digital synthesis (DDS) is used to output a continuous frequency for the duration of each sample. This is a method for digital oscillation generation with arbitrarily fine frequency resolution. The precision of the system depends on the accuracy of the clock supply and the size of the frequency register. For example, with a 32-bit frequency register for the maximum transmission frequency of 11 MHz, a frequency resolution of approximately 2.56 millihertz (mHz) results.

[0068] In a further embodiment, both modulators are clocked synchronously and output the useful signal via one port each.

[0069] In another embodiment, an input decoder is interposed between the audio input signal and the FM modulation. A decoder is a logic circuit with multiple inputs and multiple outputs that converts coded inputs into coded outputs, where the input and output codes are different. The input code has fewer bits than the output code, and there is a one-to-one mapping of input codewords to output codewords.

[0070] In a further embodiment, the audio reproduction DAC for converting a digital audio signal into an analog audio output signal comprises a digital signal processor (DSP) and / or ASIC, a control unit, and / or a power supply, wherein these components are preferably located in the digital part. The digital signal processor (DSP) and / or ASIC includes the modular units input selection, input unit and output unit, input decoder, signal processing, and / or FM modulator.

[0071] In a further embodiment, the sound reproduction DA converter for converting a digital audio signal into an analog audio output signal comprises at least one FM demulator, preferably two FM demulators, one low-pass filter, preferably two low-pass filters, more preferably two different low-pass filters, one audio amplifier, preferably two audio amplifiers, one analog output stage, preferably two analog output stages and / or a separate voltage supply, wherein these components are preferably arranged in the analog part.

[0072] In a further embodiment, the sound reproduction DA converter according to the invention comprises a high-precision clock generator (ultra high precision clock).

[0073] An ASIC (Application-Specific Integrated Circuit) is preferably used as the basis, as it is particularly suitable for the present applications. In another embodiment, the sound reproduction DAC according to the invention is long-term and temperature-stable, since electron tubes generate considerable heat. For this purpose, ventilation slots are integrated into the housing.

[0074] The main task of the digital unit is to convert the amplitude values ​​of the digital audio input signal into an equivalent output frequency. The change in the amplitude values ​​per sample automatically creates frequency modulation. The minimum frequency change is determined by the resolution of the data word (e.g., 16 bits or 24 bits), and the change interval is determined by the sampling rate of the digital audio input signal.

[0075] The analog component comprises an FM demodulator, a TP filter, and preferably an output stage per channel, with signal processing taking place separately for each channel. By processing the signal separately for each channel, the signal can be processed more precisely due to the size of the frequency-modulated signal according to the invention.

[0076] In one embodiment, the individual modules of the analog section can be implemented using either semiconductors or tube technology, with tube technology being preferred. Although the sound is improved with tube technology, almost all alternative devices use semiconductor components due to their lower cost. The harmonic spectrum and intermodulation distortion are significantly more favorable when processing signals with electron tubes compared to semiconductors, resulting in a more harmonious and musical sound.

[0077] In another embodiment, vacuum tubes are also used in the output stage to make the perceived "digital" sound of transistor devices somewhat more natural. The novel use of tubes within the conversion device, rather than in the device's periphery, according to the invention, enhances this effect and thus enables a significantly more natural sound.

[0078] In one embodiment, electron tubes have a high voltage gain (approximately 20 to 100). The input sensitivity and gain can be adjusted using current or voltage feedback. This enables a wider frequency response and lower distortion.

[0079] In a further embodiment, the audio reproduction DAC for converting a digital audio signal into an analog audio output signal is characterized in that the FM demodulator is coupled to the digital signal processor (DSP) and / or the ASIC by means of one or more RF transmitters, with the digital and analog components preferably being galvanically isolated. No signal loss occurs through RF transmission. Galvanic isolation reduces interference from the DSP and / or ASIC transmitted via the power supplies. Furthermore, no significant signal loss is to be expected on this signal path, despite the high modulation index at the FM threshold.

[0080] In one embodiment, the analog and digital domains are galvanically isolated. Two RF transformers are preferably used for this separation, which forward the FM signal to two FM demodulation units. By separating the signals into different channels, the separate signals can have a higher signal magnitude without signal loss occurring during transmission.

[0081] In a further embodiment, demodulators are included that extract the wanted signal using a phase discriminator and forward it to one or more low-pass filters, preferably channel-separated low-pass filters. These are one or more low-pass filters whose coefficients are based, among other things, on the sampling frequency. In a further embodiment, the sound reproduction DAC comprises an output unit of the DSP and / or ASIC with a connection to one or more low-pass filters to improve the quality of the wanted signal.

[0082] In a further embodiment, the audio reproduction DAC for converting a digital audio signal into an analog audio output signal is characterized in that the one or more low-pass filters are connected downstream of the demodulators. Preferably, several filters are included per channel, which have different filter properties depending on the data rate. Channel separation and low-pass filters with different filter properties in relation to the data rate enable sound reproduction at the highest level, which also allows for sound customization.

[0083] Description of the embodiments

[0084] Embodiments of the sound reproduction method according to the invention and the sound reproduction DAC according to the invention are shown schematically in simplified form in the following drawings and are explained in more detail in the following description. They show:

[0085] Fig. 1 Principle diagram of the sound reproduction process

[0086] Fig. 2 Principle diagram of the sound reproduction process

[0087] Fig. 3 Principle diagram of the sound reproduction process

[0088] Fig. 4 Block diagram of the digital / analog converter

[0089] Fig. 5 Basic circuit of the single-ended edge discriminator

[0090] Fig. 6 Schematic diagram of the entire demodulator circuit. With regard to the invention, a minimized function of the digital unit was used, without filter and compensation algorithms. Of course, any functional units not described in the exemplary embodiments, such as filters, etc., are still included.

[0091] Fig. 1 shows an embodiment of the inventive sound reproduction method for converting a digital audio signal into an analog audio output signal. According to the invention, the audio input signal 110 is input via the interface 11 into the digital part 1, specifically into the DSP and / or ASIC 10. The audio input signal 110 is then transmitted channel-separated to the respective FM modulators 14. By means of the FM modulators 14, preferably voltage-controlled oscillators (VCOs) 14, the channel-separated signals are modulated into an FM signal 140. As frequency-modulated audio output signals 150, the signals are transmitted unchanged to the FM demodulators 21 in the analog part 2, which is separated from the digital part 1, demodulated into an analog signal 210, and output via the output stages 23 as an audio output signal 230.

[0092] Fig. 2 shows a further embodiment of the inventive sound reproduction method for converting a digital audio signal into an analog audio output signal, which is based on the previous embodiment. In this embodiment, the frequency-modulated audio output signals 150 are transmitted unchanged by means of RF transmitters 15 to the FM demodulators 21 in the analog section 2, which is separated from the digital section 1.

[0093] Fig. 3 shows a further embodiment of the inventive sound reproduction method for converting a digital audio signal into an analog audio output signal, which is based on the previous embodiment. In this embodiment, the audio input signal 110 is converted in an input decoder 12, since the audio input signal 110 can be present in various formats, and processed in the signal processing unit 13 into channel-separated audio input signals 130 before these are forwarded to the FM modulators 140. In addition, the analog signal 210 demodulated by the demodulator 21 is conditioned by a TP filter 22, and the conditioned analog signal 220 is output via the output stages 23 as an audio output signal 230.

[0094] Fig. 4 shows an embodiment of the sound reproduction DA converter according to the invention. The sound reproduction DA converter is separated into digital part 1 and analog part 2. Digital part 1 is the first part of the TFM digital / analog conversion. Preferably, an ASIC (Application-Specific Integrated Circuit) is used as the basis, as it is especially suitable for the present applications, but other chips can also be used. The interface 11 has, among other things, the required l 2 S interface and a bus connector for, for example, an ADA-HSMC board, which can contain a digital-to-analog converter for frequency generation. The l 2In this embodiment, the S interface consists of three signal lines, although more or fewer are also conceivable. In one embodiment, interface 11 comprises today's standard interfaces, since the TFM development primarily involves a D / A conversion process. Input unit 16 allows, for example, the connection of Ethernet, USB storage devices, access to online music providers, UPnP, or Airplay. These components can also preferably be operated with an app or via a browser, which, in conjunction with the TMF D / A converter, allows for the creation of a complete music streamer.

[0095] In this embodiment, the Serial Data (SD) line alternately contains the sample information of the left and right audio channels, with Word Select (WS) defining the time frame for the data words of the left or right channel. Accordingly, there is no disadvantageous bidirectional communication, as with USB, and asynchronous transmission is prevented. Thus, the frequency corresponds exactly to the sampling rate. The Serial Clock signal SCK is required for bit synchronization of the serial data stream. The frequency results from the sampling rate and the resolution of the data word (e.g., 16 bits or 24 bits). Changing the amplitude values ​​per sample automatically creates frequency modulation. The minimum frequency change is specified by the resolution of the data word (e.g., 16 bits or 24 bits), and the change interval is specified by the sampling rate of the digital audio input signal.

[0096] To output a continuous frequency for the duration of each sample, direct digital synthesis (DDS) is used. This is a method for digital oscillation generation with arbitrarily fine frequency resolution. The precision of the system depends on the accuracy of the clock supply and the size of the frequency register. For example, a 32-bit frequency register for the maximum transmission frequency of 11 MHz results in a frequency resolution of approximately 2.56 millihertz (mHz).

[0097] In this embodiment, the output values ​​for a sinusoidal signal waveform are generated using a customized CORDIC algorithm and passed to a dual high-frequency D / A converter. The CORDIC (Coordinate Rotation Digital Computer) algorithm is an efficient iterative algorithm that can be used to implement many mathematical functions.

[0098] The digital part 1 has a separate voltage source 17 of, for example, 12 VDC / 3.5 A. To operate the entire system, individual parts of the audio reproduction DAC have separate voltage sources that are designed according to their applications to avoid interference and / or signal loss.

[0099] The power amplifier can be tuned with additional filter elements so that no interference from the power supply is perceived and the noise floor is minimal. The power supply's capacity is also designed to easily drive another power amplifier for four-channel operation. At the same time, the power supply is used to power the tube output stage, saving additional effort and space.

[0100] By using a data buffer for the digital audio input signal, the frequency for the subsequent sample can be preset, allowing for interpolation during the transition. The phase angle is also taken into account. The harmonics generated by the interpolation are significantly reduced compared to the jumps. The interpolation affects only a small time range of the consecutive frequency (approximately 30 to 50 nanoseconds), which has only a minor impact on the rise time (slew rate), even with large dynamic jumps. This means that the signal curve almost corresponds to an analog frequency modulation.

[0101] Both channel transformers are connected downstream of the high-speed DAC, providing electrical isolation and matching the 50-ohm transmission path to the demodulator 21. The FM output voltage in this embodiment is 3 Vpp. This corresponds to the maximum possible amplitude and ensures the maximum possible signal-to-noise ratio (SNR) and minimal harmonic distortion. Additionally, a low-pass filter 22 can be connected downstream of the D / A converter to better suppress interference from the DDS clock frequency.

[0102] In this embodiment, the sound reproduction DAC comprises electron tubes. These can be based, for example, on vacuum fluorescent displays. They can be operated with operating voltages from 5 VDC to 80 VDC. The gain is between 3 and 9, depending on the circuitry and operating voltage. This allows for a circuit without negative feedback. Furthermore, the low operating voltage and low power consumption of the heater are a great advantage. To compensate for the output impedance of the electron tube, an impedance converter in the form of a transistor can be included in an advantageous embodiment in the sound reproduction DAC. For example, by selecting an output driver with which an output impedance of approximately 650 ohms can be achieved. In addition, the electron tube is preferably mechanically decoupled and shielded to counteract microphony and the absorption of vibrations.

[0103] Optionally, additional features such as digital filter curve linearization for demodulator 21 and the option of four-channel operation with digital frequency separation can be added. Additionally, the option to compensate the speaker frequency response using a compensation filter for specific speakers is also implemented.

[0104] This embodiment includes a two-channel demodulator 21 with subsequent analog signal processing for forwarding to a connected integrated amplifier and preferably integrated volume control for driving power amplifiers. The demodulator 21 and the analog signal processing are preferably constructed using discrete components in order to influence the sound-relevant components.

[0105] Further features such as digital filter curve linearization for demodulator 21 and the option of four-channel operation with digital frequency separation may also be included. Additionally, the option to compensate the speaker frequency response using a compensation filter for specific speakers may also be included.

[0106] Fig. 5 shows an embodiment of a basic circuit of a single-ended edge discriminator 21 of the sound reproduction DAC according to the invention. For the sound reproduction DAC according to the invention, almost all known methods for FM demodulation are suitable. Basically, a frequency measurement is first performed at a specific time. From this, an equivalent output voltage can be generated. In order to realize the most discrete design possible for the demodulator 21, a modified single-ended edge discriminator 21 based on an electron tube double triode is preferably used. This allows a grid-base circuit to be realized, with the two systems of the double triode connected in parallel. This circuit is well suited for high-frequency applications and offers high gain.Preferably, a resonant circuit is implemented for the edge discriminator 21, with the resonant circuit being calibrated to 8.7 MHz. This may result in a non-linear resonance curve in the operating range (FM deviation) from 9 MHz to 11 MHz, which normally leads to erroneous amplitude values. For this purpose, a digital characteristic linearization has been developed with which the low-frequency output of the analog signal of the entire demodulator circuit, including the downstream stages, can be connected to the input of the TFM-DAC via an A / D converter, and the amplitude values ​​can be measured. This results in correction values ​​for frequency generation. A further advantage of the calibrating circuit is that deviations due to component tolerances or aging processes can also be compensated within certain limits.

[0107] Fig. 6 shows an embodiment of a demodulator circuit 21 according to the invention. Due to the frequency change caused by the FM, an amplitude-modulated signal (AM) is generated at the coil, the envelope of which contains the audio signal. Capacitor C18 is provided for DC decoupling for the downstream rectifier circuit. Diodes D3 and D4, which function as an envelope detector, can then extract a symmetrical AF audio output signal, with capacitors C21 and C22 serving to initially smooth the analog audio signals. This is followed by a further RF filter, comprising the RC elements R27, C24 and R26, C23. In the known prior art, this circuit only has one diode D3 for rectification. By adding the oppositely polarized diode D4 and the downstream components, an additional AF signal rotated by 180° is generated. The LF signals at C19 and C20 are still superimposed with remnants of the carrier frequency.However, these are present in both signals in the same phase. When the signals are combined, for example in the form of an operational amplifier or audio transformer, the sign of the audio signal is inverted and then added to the signal. This doubles the useful signal at the output, and interference is canceled out. The signal at the output of IC1 can be used for all further processing applications with the appropriate scaling. For this purpose, operation with an integrated amplifier can be used, or a volume control can be implemented that allows the connection of active loudspeakers or power amplifiers. The demodulators preferably have their own voltage source of 120 VDC / 100 mA (anode voltage) and 6.3 VDC / 1.5 A (heater voltage). An additional small toroidal transformer can be encapsulated in the housing for the demodulator unit.This ensures good shielding for the RF section and minimizes mechanical vibrations of the transformer. In addition to the anode and heater voltage windings, the symmetrical voltage supply for the filter unit (+ / -15 V) can also be provided.

[0108] Volume control

[0109] The audio reproduction DAC includes a relay-based attenuator, in this case consisting of seven relays, each with an attenuation of 1 dB, 2 dB, 4 dB, 8 dB, 16 dB, 32 dB, and 64 dB. This results in a maximum attenuation of 128 dB, with a minimum step size of 1 dB. This allows for fine control. Another major advantage is the ability to work with different resistor materials and designs, which also influence the sound. A pController was programmed for control, which, in addition to being operated via two up / down buttons, also allows the connection of an IR sensor or communication via the i2C bus. The pController also allows individual step sizes to be implemented across the attenuation curve, which increases ease of use. For the prototype, the attenuators were equipped with various resistors (THT, SMT, metal film). These resistors are built into the attenuator.In terms of sound, for example, the 0.6-watt metal resistors in THT design exhibited a good bass foundation and a somewhat stronger structure or roughness in the midrange. The SMT metal-film resistors, on the other hand, had a very clean sound with finer resolution. Resistors can, of course, also be used in other circuit components, which then also have an influence on the sound. However, this must also be taken into account when selecting resistors.

[0110] Dielectric strength, performance and other properties must be taken into account.

[0111] To operate the entire system, numerous voltage sources are necessary, which are preferably installed separately according to their application in order to avoid interference.

[0112] - Digital part: 12 VDC / 3.5 A

[0113] Demodulator 120 VDC / 100 mA (anode voltage) and 6.3 VDC / 1.5 A (heater voltage)

[0114] - Analog part / filter / OP-Amps (small signal) + 15 VDC / 0.5 A and -15 VDC / 0.5 A

[0115] Power amplifier + 35 VDC / 3.3 A and -35 VDC / 3.3 A

[0116] A small toroidal transformer for the demodulator unit is encapsulated in the housing. This ensures good shielding for the RF section and minimizes mechanical vibrations of the transformer. In addition to the anode and filament windings, it also provides the symmetrical voltage supply for the filter unit (+ / -15 V).

[0117] The power amplifier has been tuned with additional filter elements to ensure no interference from the power supply is perceived and the noise floor is minimal. The power supply's capacity is also designed to easily drive another power amplifier for four-channel operation. At the same time, the power supply is used to power the tube output stage, saving additional effort and space.

[0118] 1 digital part

[0119] 2 Analog part

[0120] 10 DSP / ASIC

[0121] 11 Interface

[0122] 12 input decoders

[0123] 13 Signal processing

[0124] 14 FM modulator

[0125] 15 RF transmitters

[0126] 16 Input / output unit

[0127] 17 Voltage source digital part

[0128] 18 Control panel and display

[0129] 21 FM demodulator

[0130] 22 TP filters

[0131] 23 Output stage

[0132] 27 Voltage source analog part

[0133] 110 Audio input signal (i2S signal)

[0134] 120 data words

[0135] 130 channel-separated audio input signal for the modulator

[0136] 150 Frequency modulated audio output signal

[0137] 210 analog signal

[0138] 220 processed analog signal

[0139] 230 audio output signal

[0140] C18 capacitor

[0141] D3 diode

[0142] D4 diode

[0143] C19 capacitor

[0144] C20 capacitor

[0145] C21 capacitor C22 capacitor

[0146] R27 resistor

[0147] C24 capacitor

[0148] R26 resistor

[0149] C23 capacitor

Claims

PATENT CLAIMS 1. A sound reproduction method for converting a digital audio signal into an analog audio output signal, characterized in that an audio input signal is modulated by a DSP and / or by an ASIC into a frequency-modulated audio output signal and the frequency-modulated audio output signal is demodulated, the audio output signal being output as a sound signal.

2. Sound reproduction method for converting a digital audio signal into an analog audio output signal according to claim 1, characterized in that the audio output signal is transmitted unchanged from the digital signal processor (DSP) and / or the ASIC to an FM demodulator, which is preferably separate from the digital signal processor (DSP) and / or ASIC.

3. Sound reproduction method for converting a digital audio signal into an analog audio output signal according to claim 1 or 2, characterized in that the audio output signal is transmitted from the digital signal processor (DSP) and / or ASIC by means of one or more high-frequency transmitters, preferably by means of one or more electron tubes.

4. Sound reproduction method for converting a digital audio signal into an analog audio output signal according to one of the preceding claims, characterized in that an audio input signal is channel-separated, where the audio input signal is separated into two channels, preferably four channels.

5. Sound reproduction method for converting a digital audio signal into an analog audio output signal according to one or more of the preceding claims, characterized in that an input signal is channel-separated and both channels are processed separately from one another, preferably by a dual mono architecture, wherein the separation is preferably carried out in the DSP and / or ASIC and wherein the separation preferably takes place before transmission to the FM modulator.

6. Sound reproduction method for converting a digital audio signal into an analog audio output signal according to one or more of the preceding claims, characterized in that the resolution and / or the frequency range is at least 80 kHz, preferably 100 kHz and particularly preferably 120 kHz.

7. Sound reproduction method for converting a digital audio signal into an analog audio output signal according to one or more of the preceding claims, characterized in that a maximized FM deviation and / or modulation index.

8. Sound reproduction method for converting a digital audio signal into an analog audio output signal according to one or more of the preceding claims, characterized in that the input voltages are processed by means of electron tubes 9. Sound reproduction method for converting a digital audio signal into an analog audio output signal according to one or more of the preceding claims, characterized in that the input signals are processed by means of fixed-point processors and / or floating-point processors, preferably a combination of fixed-point and floating-point processors.

10. Sound reproduction method for converting a digital audio signal into an analog audio output signal according to one or more of the preceding claims, characterized in that the FM demodulator is coupled to the digital signal processor (DSP) and / or the ASIC by means of one or more RF transmitters, wherein the digital part and analog part are preferably galvanically isolated.

11. A sound reproduction method for converting a digital audio signal into an analog audio output signal according to one or more of the preceding claims, characterized in that the harmonic spectrum and / or the intermodulation distortion preferably behave more favorably when signal processing is carried out using electron tubes.

12. Sound reproduction method for converting a digital audio signal into an analog audio output signal according to one or more of the preceding claims, characterized in that the audio input signal is frequency modulated by the FM modulator, wherein preferably the frequency of the applied amplitude value is determined by a voltage-controlled oscillator (VCO), preferably by two voltage-controlled oscillators (VCOs), particularly preferably by four voltage-controlled oscillators (VCOs), particularly preferably depending on A number of channel-separated signals are processed by a voltage-controlled oscillator (VCO).

13. Sound reproduction method for converting a digital audio signal into an analog audio output signal according to one of the preceding claims, characterized in that a maximum relative error of the frequency for the FM modulation is at most 2- 24 and / or a signal-to-noise ratio of at least 120 dB, preferably a dynamic range of 24 bits.

14. Sound reproduction DA converter for converting a digital into an analog audio output signal, characterized in that the sound reproduction DA converter is separated into a digital part and an analog part, wherein an FM modulator and an FM demodulator transmit signals via a functional unit provided for this purpose, preferably directly via an RF transmitter, more preferably via two RF transmitters, wherein the transmitter preferably comprises at least one electron tube, more preferably at least two electron tubes.

15. A sound reproduction DA converter for converting a digital audio signal into an analog audio output signal according to claim 14, comprising • a digital signal processor (DSP) and / or ASIC • a control panel and / or • a power supply, wherein these components are preferably located in the digital part, characterized in that in the digital signal processor (DSP) and / or ASIC the module units • Input selection and / or • Input unit and output unit and / or • Input decoder and / or Signal processing and / or FM modulator is included.

16. A sound reproduction DA converter for converting a digital audio signal into an analog audio output signal according to one of claims 14 or 15, comprising • at least one FM demodulator, preferably two FM demodulators and / or • one low-pass filter, preferably two low-pass filters, more preferably two different low-pass filters and / or • one NF amplifier, preferably two NF amplifiers and / or • one analog output stage, preferably two analog output stages and / or • a separate power supply, whereby these components are preferably located in the analogue part.

17. Sound reproduction DA converter for converting a digital audio signal into an analog audio output signal according to one of claims 14-16, characterized in that in the field of audio signal processing, fixed-point processors and / or floating-point processors, preferably a combination of fixed-point and floating-point processors, are used, preferably within one unit.

Images