|Title of Invention||
A PROCESS FOR TRANSMITTING DIGITISED AUDIO INFORMATION OF HIGH QUALITY AND WITH A SHORT DELAY
|Abstract||The invention relates to a process for the transmission of digitised audio information of high quality and with a short delay, in particular processes for the transmission of digitised audio information in an audio pickup (microphone) and/or playback path. According to the invention there is proposed a channel filter which is used to shape the high frequency spectrum of the transmission, wherein the spectrum has no attenuation in a first range of about 100 to 300 kHz useful bandwidth and in a range outside the first range it has a stop-band attenuation of particularly preferably more than 60 dB or more than 80 dB .|
The present invention preferably concerns a microphone, in particular a digital microphone, that is to say a microphone in which the signal which comes from the microphone transducer is already digitised in the microphone and the corresponding microphone signal is delivered digitally by the microphone, either by way of a cable or wirelessly, for example in a microphone receiver, pre-amplifier etc.
High-quality audio data produce large amounts of data after A/D conversion. If, as is usual for example in studio technology, the procedure involves sampling an audio signal at 48 kHz with a word width of for example 16 bits, the result is a data stream of 768000 bit/s per channel, which in the case of a stereo channel already signifies more than 1.5 Mbit/s.
In order to protect the data generated in that way from transmission errors, redundancy is added in the form of additional bits which are used at the receiving end for error recognition and error correction. As a result the amount of data to be transmitted is considerably increased once again, depending on the respectively desired correction capacity.
In order to reduce the amount of data to be transmitted audio data compression processes, for example in accordance with the MPEG standard, are frequently used. They reduce the redundancy contained in the audio signal insofar as they process longer blocks of audio data in accordance with various processes. With an increasing block size the compression rate which can be achieved increases and thus the amount of data falls. Processes which suffer with losses and loss-less processes are known, the former achieving a higher compression rate.
That block formation results in the generation of a delay in the transmission system as a data block can only be processed when it has
completely arrived. That applies both in regard to the transmitting and also the receiving end. That effect is unwanted in some uses, in particular in relation to studio applications in which microphones of the present kind are to be particularly used.
If the transmission path is at the beginning of an audio production chain, the reduction in redundancy in the audio signal is disadvantageous. Subsequent production steps could influence the audio signal in such a way that the quality thereof suffers markedly. In addition the free creative options of sound mixers would be restricted. Compression processes are therefore unwanted in high-quality transmission situations, that is to say for example in the studio, on the path from the microphone to the pre¬amplifier, mixer, recording unit etc.
Compressed audio signals are more sensitive to transmission errors. Disturbing a few bits within a compressed data block can render the entire block useless. Therefore enhanced precautions are to be taken for error correction, whereby the amount of data to be transmitted increases again. With certain error scenarios, the reduction in the amount of data can be completely nullified by the necessary error protection and even made into just the opposite.
The above-described aspects cause the use of audio data compression procedures to be fundamentally problematical so that as far as possible the use thereof should be entirely abandoned, at least in terms of recording audio in a studio.
Presentday systems for digital wireless radio transmission of audio data which must manage without compression processes or only with low compression factors operate in frequency ranges which admit high occupied bandwidths. The ISM bands at 900 MHz and 2400 MHz are popular. There the bandwidth occupied can be several MHz. It will be noted however that those frequency ranges are also used by many other radio systems. Prioritisation of given uses is not possible by virtue of authorisation regulations in those frequency ranges. The transmission of audio data from microphones to the receiver is accordingly in danger of being disturbed in
an unwanted fashion at any time because other transmitters are possibly transmitting in the proximity, on the same frequency band.
Frequency ranges outside the ISM bands without general clearance are available for professional uses. Transmission would be possible there, without involving interference. It will be noted that historically, due to analog radio transmission, authorisation regulations which require low occupied bandwidths apply to those frequency ranges. Thus for example an occupied bandwidth of a maximum of 200 kHz is prescribed for radio microphones in the UHF range (in Europe).
High-quality and thus high data-rate digital transmission systems with slight transmission delay in such narrow bandwidths are however not available.
Accordingly the aim of the invention is to develop a digital transmission process as such, but in particular for microphones or radio paths, with which high data streams of audio data can be transmitted, audio data compression processes do not necessarily have to be used, an adequate transmission rate is afforded even for error correction processes, which has a very low delay, and allows a battery mode of operation both for transmitters and also for receivers, and which has a low occupied bandwidth in the radio channel.
According to the invention those aims are achieved by a transmission process having the features set forth in the independent claims and by a microphone system/radio path for carrying out the transmission process. Advantageous developments are disclosed in the appendant claims and the description and also in the drawings.
A particular property of the transmission process according to the invention and the microphone according to the invention is that transmission is possible at a high data rate with at the same time a small bandwidth in real time. Thus for example with about 2 to 8 bit/symbol, preferably 4 bit/symbol, it is possible to provide a high-quality transmission of the audio signals of the microphone. In that respect it is not just wireless
audio microphone transmission that is in the foreground, but also wired audio microphone transmission.
The particular advantages and the essence of the invention are described in greater detail in the description hereinafter.
In that respect in the accompanying drawings: Figure 1 shows components of a transmitting device, Figure 2 shows components of a receiving device, Figure 3 shows a block circuit diagram of the audio source, Figure 4 shows a special non-linear characteristic, and Figure 5 shows the asynchronous clocking of an A/D converter device.
In a transmitting device as shown in Figure 1 an audio source 10 delivers a digital data stream which is due for data transmission. That audio source can include a non-linear characteristic which implements re-encoding of the signal for shorter word lengths.
At regular spacings the frame generation means 20 introduces additional data 22 into the data stream. They serve for synchronisation of the data stream at the receiver end and/or for the transmission of status data of the transmitter and/or for transmission of a check sum or other data for controlling or monitoring the transmitter or the receiver respectively. The frame generation means is capable of outputting a data stream which is sorted in accordance with relevance of the data. In that way bits of high significance and bits of low significance can be sorted at given positions in the frame.
In the channel coder 30 redundancy is added to the useful data for error correction at the receiver end. For that purpose the incoming data are grouped in short packets of the length N. Code words of the length M>N are formed therefrom. The packets of the length N contain data of various relevance stages which can be supplied to autonomous error correction systems. The relevance stages thus remain in the packets of the length M,
A subsequent signal generator 35 forms a complex modulation symbol from a code word. In that respect it makes use of a symbol supply
of the amount 2M. The individual modulation symbols differ in that respect in terms of amplitude and/or phase. Each symbol represents M bits of the data stream involving redundancy. The various relevance stages can in that case be attributed to corresponding bit positions within the symbol. Data with a high relevance are associated with more secure positions and vice-versa.
Alternatively the symbol generator can be so designed that in each case the difference in relation to the preceding code word is converted to afford a modulation symbol.
A subsequent channel filter 40 limits the spectrum of the symbol stream and thus forms the transmission spectrum.
In an HF transmitter 45 the filtered symbol data stream is converted into an analog signal and modulated on to an HF carrier, amplified and radiated by way of an antenna arrangement 46. In that respect it is possible to use individual antennae as well as antenna arrangements comprising a plurality of segments which automatically preferably actuate that segment which involves optimum emission.
In a receiving device corresponding to Figure 2 a single-channel or multi-channel antenna arrangement 86 receives the HF signal.
It is amplified in a single-channel or multi-channel HF receiver 85, subjected to coarse pre-filtering, separated from the HF carrier, and converted into a digital data signal. Analog/digital conversion is effected in each case in a converter device which is operated with an autonomous and fixed clock which is asynchronous with respect to the transmitter. That minimises clock jitter. Subsequent digital signal processing also operates in each case in parts asynchronously with respect to the transmitter and is finally only synchronised in the audio sink 50.
A single-channel or multi-channel channel filter 80 selects the transmission spectrum and respectively reconstructs the transmitted symbol stream.
A symbol decoder 75 reconstructs, from one or more symbol streams, the complex modulation symbols comprising the symbol supply
2M. For combining a plurality of symbol streams, it is possible to have recourse to items of side information from the respective reception levels, the respective spacings of the symbol from the ideal position and/or the distortions of the transmission channel.
The code words of the length M which are ascertained in that way are checked for errors in the channel decoder 70 and possibly corrected. The redundancy added in the transmitter is removed and the packets of the length N assembled to afford an overall frame.
In a frame disassembler 60 the additional data 62 added at the transmission end are separated off, possibly evaluated and the transmitted audio data stream outputted. The bits previously sorted on the basis of relevance are brought back into their original position.
An audio sink 50 contains a process for the detection and masking of residual errors, which contains no items of secondary information from the data stream. Only a check sum tests the data stream for the existence of residual errors. The correction values ascertained by that process are only used when a received frame has certainly been recognised as erroneous.
Some of the described components, their function and/or their combination are already known from various technical areas.
The object already described above is however attained in accordance with the invention in that in part new ways of attaining the object are pursued for the components, the required parameters being so selected that highly advantageous solutions are afforded or new functionalities are incorporated, which are particularly advantageous for the use envisaged here.
An audio source, specifically for the transmission of digital audio data, is shown in Figure 3. The function thereof is to provide the digital data stream to be transmitted.
The arrangement therefore includes means for receiving an analog input signal and a digitisation device. For feeding the analog input signals, the arrangement includes either a plug connection 102 for a microphone
and/or a directly connected microphone 101 and/or a general analog high-level input 103.
The input signal is adapted by way of an adjustable amplifier 105 to the actuating range of an A/D converter 107. A limiting device 106 is provided for preventing overdriving.
The A/D converter delivers sampling values of the width K. In studio technology that is generally 24 bits.
A downstream-connected non-linear characteristic 109 reduces the word width of the sampling value to the width L. In that situation a worsening of the signal/noise ratio is deliberately accepted.
It will be noted that the segment characteristics known from the field of telecommunications are unsuitable here. Segmentation thereof results in a stepped change in the signal/noise ratio. That is audible in the standards of studio technology and is therefore to be avoided.
A characteristic in accordance with the invention as shown in Figure 4 avoids the above-described disadvantages.
A linear relationship between input and output signal is provided for small input signals 120. For larger input signals 121, there is a non-linear relationship between input and output signal. That could be for example a logarithmic relationship. At the intersection A between the segments 120 and 121 a segment 121 is so calculated that it blends into the segment 120 continuously, that is to say without a kink. With respect to the intersection A the segment 120 forms the tangent to the segment 121.
The properties of the characteristic can be varied within wide limits by virtue of the choice of the intersection A, the gradient of the segment 120 and the curvature of the segment 121. The segments 122 and 123 with the intersection A' correspondingly apply in symmetrical relationship for negative input values, as shown in Figure 4.
It is further possible for the segment 121 or 123 respectively to be subdivided into subsegments of differing curvature, wherein the curvatures are to be so selected that a steady transition is guaranteed.
Alternatively the word width can be reduced by cutting off the less significant bits, in which case preferably a dithering algorithm is used for white noise formation.
In addition or alternatively the audio source can have a digital input 104. In that case the signal which is already digital is passed directly to the non-linear characteristic 109. A change-over switch 108 is possibly necessary in order to switch over between different inputs, whether of analog or digital nature.
In regard to parameter design of the frame generating means, the channel coder and the symbol generator, the arrangement fundamentally involves freedom in terms of the choice of the frame length, the number of redundant bits for error correction and the number of bits which are to be transmitted per symbol.
An essential limitation here is the low bandwidth in HF transmission. Therefore the number of bits which have to be transmitted must be kept as low as possible. It is therefore not possible to add a large amount of redundant bits. In addition it is necessary to transmit a plurality of bits per symbol. The number of bits which are necessary to organise or synchronise the data stream at the receiver end also has a marked influence.
A further limitation is the admissible delay time. Long frame lengths can require large input and output buffers at the transmitting and receiving ends, thereby increasing the delay time. The same applies to channel codes of great block length.
In order to be able to transmit the required amount of data with a less occupied bandwidth, M bits are transmitted simultaneously in a modulation symbol.
In accordance with the invention in dependence thereon a channel coder 30 is so selected that, starting from data blocks of the length N, it produces output words of the same length M as the modulation symbol.
In addition the frame length is selected as an integral multiple of the word length L of the sampling values and at the same time as an integral multiple of the length N.
If additional data, for example for synchronisation, are required in the frame, they are to be embedded in the frame, with the omission of sampling values.
The complication and expenditure for data synchronisation is minimised by parameter design in accordance with the above-indicated rules. The channel code words are automatically synchronised by the modulation symbol. A short bit pattern suffices for synchronisation of the entire frame. As the entire frame passes through a channel coder all bits are protected against transmission errors, including all additional bits. The number of additional bits necessary is minimised.
Alternatively some bits can be excluded from that protection if they are viewed as being less relevant or less worthy of protection.
A particularly advantageous characteristic is afforded if the number of bits per modulation symbol is selected as an integral multiple of the output word width of the channel coder.
The foregoing rules also apply if the basis considered is not just a frame but a plurality of frames combined to afford a group, and if that group is then viewed as a frame in the above-indicated sense.
Usually, in a digital receiver for audio signals, all components are operated with a clock which is derived from a common master clock by division. That also includes in particular the clock for the A/D converter device which digitises the analog signals arriving from the HF portion. That master clock is so adjusted by a clock recovery circuit that it is exactly tuned to the transmission signal.
That arrangement always involves a certain clock jitter which upon A/D conversion leads to a sub-optimal result.
In accordance with the invention that above-indicated disadvantage is avoided by the arrangement shown in Figure 5.
The incoming analog signals which represent a filtered symbol data
stream are applied to an A/D converter device 210.
The sampling clock 215 operates independently of all the remaining
clocking of the system. It is fixedly set to a frequency which is a multiple of
the necessary baseband sampling rate. The digitised symbols pass into an interpolation device 240 and a symbol clock recovery device 220.
The symbol clock recovery means, on the basis of the digitised symbol stream, calculates the true symbol and also passes it to the interpolation device 240.
In the interpolation device 240 the symbols are sampled again at the true symbol rate and thus reduced to the true symbol rate.
The symbols obtained in that way then pass through the further receiver circuit 250, namely the channel decoder and the frame disassembling means and then pass into a short FIFO memory 260. The read-out clock for the FIFO memory is based on the true symbol clock which is obtained by the clock multiplier and/or clock divider 225 and a jitter suppression means 230.
Supplemental to or alternatively to the above-described transmission process the present application also concerns a touch-sensitive antenna diversity for a radio transmission microphone. That antenna diversity can be constructed together with a digital microphone (as described hereinbefore) or also can be independent thereof. Such a radio microphone with connected antenna emits a modulated HF signal.
If then the aim is to increase the transmission reliability of the transmitted HF signal and thus to achieve a high-quality signal transmission, it must be pointed out in this context that an antenna diversity is usual in the case of microphone receivers in order to minimise field strength collapses caused by multipath propagation.
Usually radio microphones comprise a microphone, a housing with an internally disposed transmitter unit and a transmitting antenna. Those functional units are usually so assembled that the precise position of the transmitting antenna is not necessarily recognisable. By virtue thereof it is usual that a hand-held microphone is not only held at its housing but also at the antenna or at the microphone cage.
Gripping the microphone at the transmitting antenna means that the antenna is shadowed and the base impedance thereof is altered. At the
same time the hand will absorb a not inconsiderable part of the transmission energy. Those effects considerably reduce the transmission power emitted and lead to a marked reduction in range.
It is here that the invention now proposes equipping a microphone hand transmitter with two antennae which are connected selectively or weightedly to the transmitter unit. Those antenna are disposed at different positions on the microphone, which are as far away from each other as possible, for example on the microphone cage on the one hand and at the end of the microphone handle on the other hand. For that purpose the antennae must be electrically insulated from the microphone housing and galvanically or capacitively connected to a change-over switching unit. That change-over switching unit which is disposed between the transmitter and the antenna can comprise an HF switch or a circulator which is connected to the diversity antennae.
When using a change-over switch the switching criterion in the detector is obtained from the reflected power detected in the line coupler. Complete shadowing of both diversity antennae at the same time is highly improbable due to the large spacing between the two antennae. If nonetheless that situation occurs, it can be assumed that one of the two antennae, due to the larger geometrical dimensions thereof, can be less screened than the other antenna.
The switching criterion can also be taken from a contact sensor which evaluates a change in resistance of the sensor relative to the housing or a hum voltage increase in the sensor and triggers the switching operation.
The above-described solution, which is also to be used on its own, of the touch-sensitive antenna diversity for radio transmitting microphones is shown in further Figures 6, 7, 8 and 9. In those Figures the individual references denote as follows:
1 - microphone
2 - diversity antenna at the rear end of the microphone
3 - diversity antenna at the front end (cage)
4 - change-over switching unit
5 - transmitter
6 - housing sensor for changes in resistance or hum voltage
11 - circulator
12 - amplifier
21 - HF switch
22 - amplifier
25 - line coupler
26 - detector.
Additionally to but also separately from the above-described transmission processes or described microphone it is also possible to use a process for relevance-adapted error protection coding of digital audio data in order thereby to improve the overall quality of audio data transmission.
It can be used in relation to digital wireless transmission systems (but also elsewhere) which include a combination of a high-value digital modulation process with transmission data of different relevance stages. As an example 16 bit PCM-coded audio data also have 16 relevance stages. In the case of digital modulation with 4 bits per symbol, depending on the respective association of the bits with the modulation system, there are now also up to 4 quality stages of differing bit error rates. The PCM bits with the highest information content are now associated with the symbol bits of high quality, and vice-versa.
The relevance-adapted error protection coding process is based on the utilisation of differences inherent in the system in the uncorrected bit error rate between the individual output bit positions of higher-value digital modulation modes. By virtue of association of those different quality stages with autonomous error correction systems, those differences are retained and thus permit relevance-adapted distribution of the useful data. Higher quality stages with a lower bit error rate receive the data with the highest information content while data with a lower information content are distributed to lower quality stages.
The advantages over an individual error correction system are afforded firstly by a markedly improved aural impression with the same channel bit error rate. That effect is afforded with a clever choice of the parameters although the bit error rate after correction under some circumstances even turns out to be higher than in the case of an individual error correction system. A further effect is afforded in automatic and continuous adaptation of the final quality to the channel properties. In contrast to an abrupt interruption in the connection, there is firstly a continuous deterioration in quality ('graceful degradation'). The aural perception shifts from a sudden 'crash' to a continuous rise in background noise, which the hearer perceives as being substantially more pleasant. The relationship between higher noise with a poor radio connection rather corresponds to the previous accustomed hearing conditions in the case of analog transmission.
The various options can be set forth by reference to some further examples:
Figure 10 shows a structural diagram for a transmission system without relevance-adapted error protection coding; and
Figure 11 shows a structural diagram for a transmission system with relevance-adapted error protection coding.
Prerequisites for the process:
- after digital demodulation there are more than k=l quality stages, for which reason a high-grade digital modulation process with at least three bits per symbol is necessary as otherwise different quality stages in the uncoded data are not to be expected (k=1),
- the data to be transmitted contain more than n=l relevance stages, for which reason the process can be applied for example to PCM coded audio data,
- of the k quality stages present a plurality can be combined to afford the desired number m, for which reason the following applies: m≤k, and
- of the n relevance stages present a plurality can be combined to afford the desired number m, for which reason the following applies: m≤n.
The foregoing description of relevance-adapted error protection coding as shown in Figure 11, besides the described advantages, suffers from disadvantages in particular in regard to implementation complication and expenditure as a plurality of coders and decoders must be operated in parallel relationship. That requires on the one hand an increased level of hardware complication and expenditure in comparison with simple coding, while on the other hand under some circumstances the delay in the data also increases. More specifically many decoders operate with a fixed delay (calculated in bits). If now the overall data rate is divided up to a plurality of parallel data streams, the data rate per decoder falls and the overall delay (measured in time) rises. In that situation the overall delay is determined by the decoder with the lowest data rate.
A numerical example may illustrate this:
If the decoder delay is at 400 bits and the overall data rate is at 1 Mbit/s, the delay in the case of a decoder of tdecoder is at 400 bit/1 Mbit/s, that is to say 400 µs.
If the data rate in the first decoder is at 0.4 Mbit/s and the data rate in the decoder 2 is at 0.6 Mbit/s (0.4 Mbit/s + 0.6 Mbit/s = 1 Mbit/s, that gives a delay at two decoders of:
and in accordance with the foregoing description that signifies a delay of 1 ms because the decoder with the lowest data rate determines the overall delay.
A way of resolving that problem in accordance with the invention while at the same time involving relevance-adapted coding lies in the use of a folding code with subsequent 'variable pointing'. In that case the various quality stages are no longer divided into parallel data streams but processed by a respective common coder or decoder of high quality (and due to the higher data rate thus also rapidly). Pointing is subsequently effected in order to reduce the redundancy to the desired degree again.
In contrast to the process known in the state of the art that pointing however is not effected constantly but variably and in adapted relationship with the data to be processed. That pointing is correspondingly reversed again in the receiver and then decoding can take place in a common decoder. In that situation the delay remains as short as possible.
That is also clearly illustrated by the diagram shown in Figure 12.
In this connection attention is to be directed to the state of the art disclosed in WO 03/069918, US No 6 170 073, EP 0 798 888 and US No 5 751 739. To sum up it can be said that the above-indicated state of the art admittedly discloses overall an unequal error protection, but does not propose adaptation to high-value modulation processes. In contrast thereto, in the case of the process according to the invention, parallel unequal coding is also used to acquire the quality stage and the modulation symbol. Unequal error protection is generally used in the above-indicated documents from the state of the art in order to produce several on a channel with one quality stage, while in the case of the invention it serves to preserve the stages.
Alternatively to or additionally to the relevance-adapted bit mapping, an artificial noise signal can be added to the analog signal in dependence on the residual error rate or other quality criteria, in order to mask residual errors. The usual aural impression of analog transmission is further enhanced in that way.
Error masking and/or addition of noise can be used independently of each other at different outputs of the receiver. In that way for example the production path and a listening-in path can be treated to differing degrees by means of different processes and/or with different parameters.
If necessary an item of additional information can be added to given outputs, which signals an acoustic reference to given operating conditions to the user or monitoring personnel. Thus for example a buzzing noise can be mixed in at the limit of the range, to signal impending breakdown in the connection. Signalling can be effected by the most widely varying tones or tone sequences but also in plain text.
The present inventions and in particular the described transmission process can be particularly preferably used not only for microphones but also in the opposite transmission direction, that is to say for loudspeaker systems, headphone systems and in particular for in-ear monitor systems. In particular when the described wireless transmission process is used for headphones or earphones of any kind and stereo transmission is necessary, a slight level of audio data compression is used at the analog input of the transmission system, and that compression is then correspondingly decompressed at the receiver end.
If the individual aspects of the invention have already been described in a highly specific context in the present application, the individual aspects of the invention will also be set forth in more general form hereinafter. It is to be emphasised once again at this juncture that the individual aspects of the invention can already in themselves represent an independent invention and not just in conjunction with the whole of the further disclosure, even if combined consideration of all features is an extremely advantageous configuration in order to provide a digital microphone which is adequate in terms of quality requirements and which is still viable in terms of complication and expenditure.
In Figure 13, the basis adopted is the illustrated spectrum mask with a 200 KHz useful bandwidth.
Based on the available bandwidth (BW), that affords the symbol rate (fSym) in accordance with the following formula:
A meaningful range of parameters can be specified in that respect as follows: BW ≤ 200 kHz, a pulse shaping rolloff factor a: 0.1 ≤. α ≤. 0.5 and a symbol rate in the range between 130 kHz and 185 kHz.
A specific example of values in that respect would be a bandwidth BW of 200 kHz, an a of 0.3 and an fsym of 150 kHz.
By virtue of a high-value modulation process with M bit per symbol (with a symbol supply of 2M) there follows for the coded bit rate (fCOded) in accordance with the formula:
An appropriate parameter range is specified by M ≥. 1 bit and ≤. 8 bits, a symbol supply of 2M ≥. 2 and ≤ 256 and a coded bit rate ('gross') of fcoded ≥ 130 kbit/s and ≤ 1.48 Mbit/s.
A specific example of values can be specified by M = 4 bits (for example QAM-16), 2M = 16 and fCOded = 600 kbit/s.
With a code rate C which is defined by the quotients of Cin and Cout the uncoded bit rate funcoded follows therefrom in accordance with the following formulae:
An appropriate parameter range for the code rate is here >. 1/2 and ≤ 1 and for an uncoded bit rate ('nett') ≥ 65 kbit/s and ≤ 1.48 Mbit/s. An excellent parameter example for C in that case is at 3/4, that is to say 0.75 and for funcoded at 450 kbit/s.
Finally for adroit pointing synchronisation it is possible to describe a factor between the code output length Cout and the modulation value M, in accordance with the following formulae:
An appropriate parameter range for k (synchronisation coupling) is in that respect as follows:
A specific example of values for k = Cout/M can be specified at 4/4, that is to say 1.
By way of the data structuring used that affords an audio data rate fAudio and in that respect in accordance with the above-indicated formula BW is represented as follows:
In that respect an appropriate parameter range for the frame rate FR is ≥ 1 and ;≤ 1.5 and the audio data rate fAudio should usually be ≥ 43 kbit/s and ≤ 1.48 Mbit/s. A specific numerical example can be specified for FR at 1.125 and fAUdio at 400 kbit/s.
On the further assumption of PCM coding of the audio data, there hereby follows an audio sampling rate fs, which is already involved in the above-indicated formula for bandwidth, as follows:
An appropriate parameter range for PCM resolution N is in this case ≥ 12 and ≤ 24 and for the audio sampling rate ≥ 1.8 kHz and ≤ 124 kHz. A specific numerical example can be specified for N = 16 and fs = 25 kHz.
The necessary bandwidth efficiency in that respect is in accordance with the following formula:
so that with the above-specified parameter ranges for a and M, there is a bandwidth efficiency of ≥ 0.1375 Hz/bit/s and ≤ 1.5 Hz/bit/s.
On the assumption of a value for a = 0.3 and M = 4, that gives a bandwidth efficiency as follows:
The described digital receiver according to the invention, besides various digital modulation modes, can also demodulate conventional analog modulation such as for example FM (frequency modulation). In that respect any compression or pre-emphase which has been implemented is reversed at the digital level. The receiver according to the invention is thus fully rearwardly compatible with the various systems which have already become established.
For the purposes of comfortably finding various modulated transmitters and/or free transmission channels, the receiver according to the invention can have a frequency scan function which indicates band occupation to the user graphically and/or also in tabular form. In that respect modulation of the transmitter and further parameters of the specific modulation such as for example the compander used of an FM path can be automatically ascertained and displayed. The result of the scan function can be used to implement proposals for the choice of frequencies for a specific audio transmission system. The system is thereby enabled to implement self-configuring.
Sampling (AD conversion) of a higher reception bandwidth in the receiver than is required in a single transmission path provides that more than one transmitter can be simultaneously received and demodulated due to the following channel selection in the digital mode. For that purpose it is necessary to increase the IF bandwidth in the analog mode and the dynamics of AD conversion.
Separation of the channels is then effected in the digital plane by digital filters or other necessary computing operations. In that respect each channel can have its own parameters, even different modulation or varying preprocessing.
Audio signals of the various channels can be locally mixed in the
receiver, in particular in a bodypack receiver. That mixing is controlled by
suitable operating means locally in the bodypack or is implemented locally
automatically in accordance with predetermined parameters or
implemented under remote control by remote operating console (similarly
to a mixing desk) by a sound engineer. In the latter case only control signals are transmitted to the bodypack. The above-described processing variants are not limited to a bodypack receiver but also apply in a corresponding manner to stationary receivers.
By virtue of the software-defined functionality of transmitter and receiver the properties of digital transmission can be highly flexibly adapted to the demands of various situations of use (flexibility due to scalability). Thus it can be useful to scale down the audio quality to the benefit of a longer operating time or a greater transmission range in a manner appropriate to the use involved, for example for speech/reporting. On the other hand the audio quality, dynamics and/or sampling rate can be increased if the situation involves bridging over only a short distance and/or a limited operating time is acceptable.
Mono, stereo and double mono audio transmission can also be selected as options for various situations of use.
The digital data stream can be transmitted in encrypted form, for example by means of previously known encryption mechanisms and processes, whereby bugging of the radio path can be effectively prevented. It is only if the transmitter and the receiver have the same key that audio transmission and in particular reception is possible.
Encryption has the advantage that the audio signal of the microphone cannot be easily received by a third party, but only by the receiver which is associated with the microphone and thus even in the case of a presentation on stage, in a television studio or elsewhere it is not possible for somebody to receive the microphone signal in decrypted form and thus make illicit recordings.
In contrast to the previous diversity processes, in the case of the digital receiver according to the invention, it is not only possible to switch over between the individual reception channels, but simultaneous processing of all incoming data streams can take place. As the bits of the data streams are provided with an item of quality information (soft decision), it is possible, for each bit, to freshly decide on the reception
branch from which it is selected. The selection algorithm can both act in a switching-over manner and also implement a weighted junction. That means that the most advantageous weighting of the data streams from a plurality of channels is possible and the information content is put to optimum use. The selection algorithm can be supported by further side information such as for example field strength of the antenna signal or the prevailing bit error rate. The algorithm can be executed adaptively or in a switch-over mode in order to take account of the HF characteristics of the space involved.
A display is selected for the representation of continuous weighting, that display appropriately representing that channel balance for the users. To represent the weighting of two reception channels, preferably a bar which starts from the centre is adopted, which represents the tendency to one channel or the other to the two sides thereof. The receiver according to the invention thus has a diversity weighting display.
A process which is optimised for efficient processing is used to conceal residual errors which have remained. It substantially comprises a number N of fixed filters of a filter depth k. They are optimised for typical audio signals and can be algorithmically very easily implemented. An adaptive filter is adjusted to the filter which is optimum at the time by an adaptation algorithm. That ensures continuous adaptation to the audio signal. By means of the CRC check sum, it is decided whether the adaptive filter is trained at the time or is used for audio sample replacement. It is only if the prediction error at the time is markedly above the mean prediction error and the CRC sum indicates an error that replacement is effected. That threshold or the underlying algorithm can also be adapted. In addition, feedback of the replaced samples is necessary to avoid error propagation.
The post-error concealment just outlined above is shown in Figure
Digital implementation of microphone transmitters and/or receivers
makes it possible in a simple fashion to integrate additional functions which
make it possible to set up or check the radio path and further, possibly wired connecting paths for the audio signal. An audio test signal of known amplitude which is generated automatically in the transmitter and/or receiver (for example sine tone, noise etc) makes it possible to check the audio connecting paths and in particular also to correctly set the level for the transmission paths and processing equipment. A modulation signal of known symbol succession which is generated automatically in the transmitter (for example PRBS sequence, sequential sequence of all symbols etc) permits measuring acquisition and testing of the quality of the emitted signal and the radio path.
For a microphone receiver, it is also highly advantageous if it is provided with a filter which can be switched over (pluggingly interchangeable) at the receiver input. In addition it is also particularly advantageous if the filter is integrated at the receiver input in the receiver antenna, for example it can also be screwed to or attached to the antenna of the receiver. Accordingly the input filter of the receiver is then a unit with the antenna and such an electronically shieldable antenna is particularly advantageous because mounting of the antenna provides that the filter properties of the receiver are also adjustable at the same time.
It is also advantageous if the antenna and the receiver contain devices which enable the receiver to adjust to antenna-specific parameters or to show that the antenna does not suit the desired reception frequency range.
Finally it is also particularly advantageous for a microphone system according to the invention if there are provided means for altering frequency response characteristic or sound characteristic so that, by virtue of a suitable sound design, the microphone can have a desired characteristic, for example the characteristic of a microphone like the famous U87 from Georg Neumann, Berlin, or MD 421 from Sennheiser, and others. Therefore, by way of the frequency response characteristic or sound characteristic altering means, a customer can set a desired characteristic and in particular can also select such a characteristic which is perhaps not
optimum in the acoustic sense but is characteristic of a given generation of microphones or a given type of microphone.
The invention overall achieves better frequency economy and the feedback characteristic is also improved.
1. A process for transmitting digitised audio information of high quality and with a short delay, in particular a process for transmitting digitised audio
information in an audio pickup (microphone) and/or playback path, wherein a channel filter is used to shape the high frequency spectrum of the transmission, wherein the spectrum has no attenuation in a first range of about 100 to 300 kHz useful bandwidth and in a range outside the first range it has a stop-band attenuation of particularly preferably more than 60 dB or more than 80 dB (Figure 13).
2. A process as claimed in claim 1, wherein the audio information is
processed in blocks and the audio bandwidth (HF) is proportional to the symbol rate
3. A process as claimed in claim 2, wherein no audio data compression is used in the process.
4. A process as claimed in claim 2, wherein with a bandwidth of about 150 to 250 kHz (or a value from that range of values) the symbol rate is typically approximately in the range between 130 and 220 kHz, preferably at 150 kHz, and the pulse shaping rolloff factor (a) is typically in the range between 0.1 and 0.5, preferably at about 0.3 and the bandwidth is calculated in accordance with the formula:
5. A process as claimed in claim 1, wherein the audio data are processed
in blocks, wherein a high-grade modulation process is used for the transmission of
the audio information, in which a plurality of bits are assembled in a modulation
symbol and transmitted, wherein the bandwidth is inversely proportional to the
modulation value M which is preferably in the range of 1 to 8 bits, in particular in
the range of M = 4 bits, and the bandwidth is also proportional to the coded bit rate
which is preferably in the range of about 130 kbit/s and 1480 kbit/s, preferably
being about 600 kbit/s, and the symbol supply is 2M and is in the range of 1 to 256.
6. A process as claimed in claim 1, wherein the bandwidth is inversely proportional to the code rate, wherein the code rate is preferably ≥ 1/2 and ≤ 1, preferably the code rate corresponds to the value 0.75 and/or the bandwidth is also proportional to the uncoded bit rate (funcoded), wherein the uncoded bit rate is preferably ≥ 64 kbit/s and ≤ 1480 kbit/s and in particular assumes a value of 450 kbit/s.
7. A process as claimed in claim 1, wherein the bandwidth is proportional to the factor k which describes the quotient of the code output length (Cout) and the modulation value M, wherein k preferably assumes the value 1 but can also be of values from the range 1/n;...; 1/2 or 1, 2,.,.n, wherein n is a natural number.
8. A process as claimed in claim 1, wherein for the transmission of digital
audio information of high quality with a short delay, wherein no audio data
compression is used, the audio data are processed in blocks and wherein the
bandwidth is proportional to the audio data rate which is preferably in the range
between 43 kbit/s and 1480 bit/s, preferably at about 400 kbit/s and the bandwidth
is also proportional to the frame rate which is preferably greater than 1 and less
than 1.5, preferably at a value of 1.125.
9. A process as claimed in claim 1, wherein for the transmission of digital audio information of high quality with a short delay, wherein no audio data compression is used, the audio data are processed in blocks, the audio data are subjected to PCM coding with a PCM resolution N, wherein N is preferably ≥ 12 and ≤ 24, for example at 16, and the audio sampling rate fs is about ≥ 1.8 kHz and ≤ 124 kHz, preferably approximately at 25 kHz.
10. A process as claimed in claim 1, wherein the audio data after coding are subjected to pointing, preferably variable pointing, which is adapted to the data to be processed, and said pointing upon reception of the coded data is correspondingly reversed prior to the decoding operation.
11. A process as claimed in claim 1, wherein for the transmission of digital
audio data of high quality with a short delay, wherein the digital audio data are
disassembled at the transmission end into frames, wherein each frame is of a frame
length of Runcoded bits, wherein said frame contains an integral number of sampling
values (samples) of the bit length N, there are provided E bits for error recognition
or error correction, S bits are used for synchronisation and A bits are used for other purposes, and that the frame is extended by one or more channel coders and application of an overall code rate C in the range between 0.5 and 1 to a frame length of Rcoded = Rcoded/C bits, and/or RCOded is an integral multiple of the symbol length M, wherein the value for Funcoded is a value from the range between 20 and 500 bits and the value for Rcoded is a value from the range between 20 and 100 bits.
12. A process as claimed in any one of the preceding claims, wherein the modulation symbols differ in respect of amplitude and/or phase.
13. A process as claimed in any one of the preceding claims, wherein in the transmitter the functionalities of frame generation, channel coding, symbol generator and in the receiver the functionalities for symbol decoding, channel decoding and frame disassembling are so designed that they can sort the data to be transmitted in accordance with various relevance stages, starting from the lowest and/or highest relevance stage to the respective higher/lower relevance stage.
14. A process as claimed in claim 13, wherein the various relevance stages can be processed differently and independently of each other.
15. A process as claimed in any one of the preceding claims, wherein in the symbol generator at the transmission end an association of data of high relevance is effected with the more secure bit positions and low relevance with more insecure bit positions.
16. A process as claimed in any one of the preceding claims, wherein at the transmission end a plurality of antennae or antenna segments are used or provided for the emission of a high frequency (HF) signal and that those antennae or antenna segments which have the best adaptation at the time are preferably used for emission.
17. A process as claimed in any one of the preceding claims, wherein the received signal is digitised by an A/D converter device which is operated at a clock of constant frequency and the sampling rate of which corresponds to a multiple of the baseband Nyquist rate and the true symbol rate is recovered in a separate independent device and/or the actual symbol is calculated in an interpolation circuit and/or signal processing is effected in asynchronous relationship with the
transmission clock and synchronisation to the precise transmission clock is effected only at the end of the signal processing chain.
18. A process as claimed in any one of the preceding claims, wherein one or more reception paths are provided at the reception end and that in the symbol decoder items of side information are obtained from the symbol distortions and/or the reception field strengths of each reception path and that the incoming information in the channel decoder contains contributions from all reception paths.
19. A process as claimed in any one of the preceding claims, wherein an algorithm which preferably does not have any precise information about the error position is used for concealing residual errors.
20. A process as claimed in claim 19, wherein the data from the concealing algorithm are used only when a transmitted data frame was recognised as defective.
21. A process as claimed in any one of the preceding claims, wherein a continuous, non-linear characteristic is used as the source coding (Figure 4) which comprises at least two segments, wherein the first segment is linear and the second (and/or all further) segment are non-linear.
22. A process as claimed in claim 21, wherein an inverse characteristic to Figure 4 is used in the associated receiving device.
23. A process as claimed in any one of the preceding claims, wherein with an analog input circuit comprising an adjustable input amplifier, a limiting device for preventing overdriving, a digitisation device with the word width K and a computing device for computing a non-linear characteristic, wherein the non-linear characteristic for positive values comprises at least two segments, the segment for low input values (120) is linear and the segment or segments (121) are non-linear, the intersection between the segments is so calculated that the tangents to the curves coincide when approaching the intersection from both sides, the characteristic for negative input values (122, 123) is of a symmetrical configuration relative to corresponding positive input values and the output word width L of the non-linear characteristic is less than the word width of the digitisation device.
24. A process as claimed in claim 23, wherein a microphone is connected directly to the input circuit or a microphone can be plug-connected thereto or analog high-level signals can be connected.
25. A process as claimed in any one of the claims 23 and 24, wherein the input circuit has a change-over switching means (108) in order alternatively to select a digital input.
26. A process as claimed in any one of the claims 23, 24 and 25, wherein the gradient of the first segment (120 or 122) is unequal to 1.
27. A process as claimed in any one of the preceding claims, wherein the said limiting circuit is already provided in a digitisation device.
28. A process as claimed in any one of the preceding claim,s wherein frame length, channel coder and word width of the modulation symbols are selected in dependence on each other, the frame length is ah integral multiple of the word width of the sampling values, the frame length is an integral multiple of the input word width of the channel coder and/or the number of bits per modulation symbol is equal to the output word width of the channel coder and/or the necessary additional data are inserted into the frame with the omission of sampling values.
29. A process as claimed in claim 28, wherein the number of bits per modulation symbol or the inverse thereof is an integral multiple of the output word width of the channel coder.
30. A process as claimed in claim 29, wherein a plurality of subframes are combined in a frame.
31. A process as claimed in any one of the preceding claims, wherein embedded in the frame are additional data in the form of a synchronisation pattern and/or a check sum, wherein the sum of the bits of said additional data is an integral multiple of the word width of the sampling values.
32. A process as claimed in any one of the preceding claims, wherein
comfort noise is mixed with the audio information for error concealment.
33. A process as claimed in any one of the preceding claims, wherein an audio signal is outputted at more than one output, wherein each output can be processed with its own concealment algorithm.
34. A process as claimed in any one of the preceding claims, wherein an audio signal is outputted at more than one output, wherein if required signalling information can be individually mixed with each output, the signalling information signalling given operating conditions. The signalling information can be any kind of acoustic information, in particular sounds, sound sequences or plain text information.
35. A process as claimed in any one of the preceding claims, wherein the digital data stream contained in the microphone signal is encrypted and a corresponding decrypter is present in the receiver in order to decrypt the encrypted digital data stream.
36. A process as claimed in any one of the preceding claims, wherein the microphone is a digitally implemented radio microphone and the said microphone is switchable by a switch into a test mode in which a defined audio signal (for example sine tone, noise etc) is generated at the microphone end and emitted instead of the microphone signal.
37. A process as claimed in claim 36, wherein the digitally implemented radio microphone is a microphone receiver, said receiver being switchable by a switch into a test mode in which a defined audio signal (for example sine tone, noise etc) is generated and individually made available at the outputs of the receiver.
38. A process as claimed in claim 36 or 37, wherein the microphone is switchable by a switch into a test mode in which at the microphone end a defined modulation signal is emitted with known modulation symbols (for example PRBS sequence, sequential sequence etc).
The invention relates to a process for the transmission of digitised audio information of high quality and with a short delay, in particular processes for the transmission of digitised audio information in an audio pickup (microphone) and/or playback path. According to the invention there is proposed a channel filter which is used to shape the high frequency spectrum of the transmission, wherein the spectrum has no attenuation in a first range of about 100 to 300 kHz useful bandwidth and in a range outside the first range it has a stop-band attenuation of particularly preferably more than 60 dB or more than 80 dB .
|Indian Patent Application Number||1947/KOLNP/2006|
|PG Journal Number||20/2012|
|Date of Filing||11-Jul-2006|
|Name of Patentee||SENNHEISER ELECTRONIC GMBH & CO. KG|
|Applicant Address||AM LABOR 1, 30900, WEDEMARK, GERMANY|
|PCT International Classification Number||H04H1/00; H04R1/00|
|PCT International Application Number||PCT/EP2004/014833|
|PCT International Filing date||2004-12-30|