HK1138131B - Device and method for generating a signal for transmission or a decoded signal - Google Patents

Description

The present invention relates to information encoders for use in broadcasting systems and in particular to information encoders for use in digital broadcasting systems.

Digital audio broadcasting is developing rapidly, with DAB (Digital Audio Broadcast) for FM frequencies having been in existence for some time, and DRM (Digital Radio Mondial) for long, medium or shortwave.

Such broadcasting systems are characterized by the fact that a certain amount of data must be buffered before the broadcaster can start transmitting data. When the broadcaster switches on his broadcaster, this is not a problem, because the broadcaster simply starts the transmission a little later, whereby this distance from the broadcaster is not perceived as particularly disturbing. However, when the broadcaster switches the program, the listener from earlier systems is used to immediately transmitting data. However, this is not possible with a digital broadcaster, since a certain amount of data must first be stored by the other program before a transmission of this other program can be started.

U.S. Patent No. 6,842,724 B1 discloses a method and device for reducing an initial delay in packet-based data streaming applications or in a telecommunications network, such as a local exchange carrier network or an inter-exchange carrier network or a local or global computer network. A program source, such as an audio and/or video data stream, is encoded and transmitted as two or more separate bit streams, such as sequences of data packets. The transmission of one of these bit streams is to produce a higher quality data stream delay in relation to the transmission of the other bit stream source, whereby the transmission of the data stream can be delayed more than once. When receiving the program, the signal is decoded by two or more different receivers using a correspondingly different bit rate.

Unlike the transmission modes described in the above document, where only the delay of a data packet is taken into account, digital broadcasting systems face additional challenges arising from the fact that the transmission channel is not a wired channel but a wireless transmission channel. A transmitter thus has a transmitting antenna that emits radio waves that can be received and processed by a receiver that has a receiving antenna.Err1:Expecting ',' delimiter: line 1 column 279 (char 278)

Err1:Expecting ',' delimiter: line 1 column 373 (char 372)

The present invention is intended to provide a concept for digital data transmission in which the number of connection failures is reduced or eliminated.

This task is solved by a device to generate a transmitted signal as claimed 1, a device to generate a decoded signal as claimed 17, a process to generate a transmitted signal as claimed 24, a process to generate a decoded signal as claimed 25, a computer program as claimed 26.

The present invention is based on the finding that a reduction in unpleasant data loss can be achieved by moving away from the one-channel idea of encoding and transmitting program material. According to the invention, in addition to the typically transmitted program material, which is the main channel, a secondary channel is generated and transmitted, with the main channel and the secondary channel being decodable separately from each other and both of course representing the information signal. However, the main channel and the secondary channel are designed so that the transmission channel is more robust than the main channel against small streams. Thus, if a situation arises in which the SNR of the system is more continuous than the SNR of the main channel, a secondary channel is decoded and transmitted so that the transmission channel is no longer connected to the main channel, the transmission channel is no longer able to transmit information, but the transmission channel is no longer able to transmit information, and therefore the transmission channel is no longer able to transmit data, while the transmission channel is decoded.

However, in preferred embodiments of the present invention, the secondary channel will be a much more compressed representation of the information than the primary channel, so that, especially in the context of loss-induced data compression, the reproduction quality of the secondary channel will be worse than the reproduction quality if the primary channel can be reproduced, but this is not so problematic from the broadcaster's point of view, since each broadcaster prefers a reproduction with reduced quality to no reproduction at all.

This embodiment is particularly advantageous in that the total data rate for the transmission of the main and secondary channels is not significantly higher than if the main channel alone is transmitted, since the encoding of a high-loss compressed data signal inherently requires fewer bits than the encoding of a higher quality data signal. In particular, it is preferred that the bits required for the secondary channel are at most half as many as the bits required for the main channel and, in particular, even less than one tenth of the bits required for the main channel.

Another preferred embodiment is where the main channel and the secondary channel are not transmitted synchronously to each other, but with a time shift so that the main channel is delayed relative to the secondary channel. On the receiver side, the secondary channel that arrives earlier at the receiver, in relation to a certain time of the original information data, buffers the receiver while no or only a minor buffer of the main channel is needed. This means that when the transmission channel situation is favorable, the main channel can be transmitted without delay and a switch from the main channel to another main channel is also possible without problems. A switch delay would occur exactly once, then at the time of the switch a switch would have to be made to the secondary channel to make the switch back on.

The delay of the main channel versus the secondary channel is also advantageous in that, if a transmission failure occurs, even if it is so large that not even the secondary channel has been received with sufficient quality, the secondary channel can be played back at least for the time the secondary channel has been stored. In preferred embodiments, a storage time of more than 10 seconds and in particular of more than 20 seconds, preferably more than 30 seconds, is chosen, which results in a certain storage requirement, but which is not critical in terms of the convenient availability of a large user-friendly channel. The only advantage, however, is that the receiver receives a 30-second signal of a high quality - the lowest definition - without being guaranteed to be useful for the receiver, which is a particularly interesting situation, especially if the secondary channel is not used, but is not guaranteed to be useful in the case of a specific user-friendly one.

The following are examples of preferred embodiments of the present invention, which are described in detail in the accompanying drawings: Fig. 1a block diagram of an embodiment of a device for generating a signal to be transmitted;Fig. 2different embodiments of the encoder of Fig. 1;Fig. 3an alternative embodiment of the encoder of Fig. 1;Fig. 4a special embodiment of the source encoder of Fig. 2;Fig. 5a device for generating a decoded signal; andFig. 6a more specific embodiment of the device of Fig. 5;Fig. 7a flowchart of the steps taken on the part of the broadcaster, e.g. in a digital receiver.

Err1:Expecting ',' delimiter: line 1 column 81 (char 80)

Furthermore, an embodiment of the invention performs a time shift between the main and auxiliary channels so that the main channel is delayed relative to the auxiliary channel, and a more robust modulation is performed for the auxiliary channel than for the main channel with the main audio data.

A mode in DRM technology for the main channel, also called the media service channel (MSC), has a bitrate of, for example, 14.5 kb/s. In the side channel, in this embodiment, additional data is transmitted at 1 kb/s in the SDC (Service Description Channel) at a more robust rate, and is encoded with an advance of 30 seconds, for example. The main channel is therefore delayed by 30 seconds compared to the additional channel.

Fig. 1 shows a device for generating a signal to be transmitted, which is ultimately transmitted by a transmitting antenna 10 which receives its transmitting signal from a transmitter 12. The transmitter 12 is preceded by a coding device 14 which receives as input information 16 which includes audio and/or video information and which provides a signal to transmitter 12 which includes both main channel data 18a and secondary channel data 18b. Depending on the implementation of the coding device 14 and transmitter 12, the data 18a, 18b are transmitted between the coding device 14 and transmitter 12 in two separate data streams, or the data are either duplicated or already generated separately after the data has been isolated in a transmitter mode, the transmission is then carried out on a single channel, which is typically a decoder and a decoder.

In order to obtain a decoded version of the information, either the subchannel or the main channel is sufficient, in other words, the subchannel data alone, without the main channel data, is sufficient to provide at least a low-quality representation of the information on the coding page.

According to the invention, the encoder 14 is further designed to generate the main channel 18a and the secondary channel 18b in such a way that the secondary channel is more robust against transmission channel influences than the main channel.

Fig. 2 shows a detailed representation of the coder 14 according to various embodiments of the present invention, all of which are optionally shown in Fig. 2. The coder 14 may include a source or source coder 20, a redundancy coder 22, a mapper 24, and/or a modulator 26. The source coder 20 is fed with information 16. Depending on the implementation, the source coder 20 already provides the secondary channel 18b and the main channel 18a, as shown by the crossed lines between source coder 20 and redundancy coder 22. Alternatively, however, the source coder may be so detailed that it represents only a single version, which is then delivered by the dashed line between block 20 and block 22 as shown in the figure.Alternatively, however, the redundancy coder could also be trained to output only a single redundancy coded data stream, which, as represented by the dashed line between block 22 and block 24, is fed into the mapper 24, which then generates main channel 18a and secondary channel 18a using different mapping rules.

Alternatively, the mapper 24 may also generate a single output data stream and feed it to the modulator 26, which performs, for example, an FDMA, TDMA or CDMA modulation procedure, i.e. one of the known frequency multiplex, time multiplex or codemultiplex procedures or a combination of these methods as is known in the art. Depending on the implementation, the modulator may output a single signal, which would then include both main and secondary channels already in a single data stream, or the modulator may deliver the secondary channel and the main channel as separate data streams, which are then combined together in the transmitter 12 before the first one emits a single signal, which includes both channels.

In principle, it is sufficient that in one of the blocks 20, 22, 24, 26 different channels with different robustness are generated. However, several different robustnesses can also be accumulated. For example, the secondary channel generated by the redundancy coder 22 which is per se more fault-resistant due to the greater redundancy can be subjected in the mapper to a mapping rule that is more robust than another mapping rule with which the main channel can be charged, while in addition a more fault-resistant modulation method can be used for the secondary channel in the modulator than for the main channel.

However, the invention prefers that the source encoder 20 already generates two different output data streams, with the secondary channel having a low bitrate and the main channel having a high bitrate, which can then be processed by a combination of blocks 22, 24, 26 or which, for example, can only be processed by the redundancy encoder 22 and which can be processed by the redundancy encoder 22 and which can then be processed by the redundancy encoder 22 alone, whereby the redundancy encoder combines both data streams according to a specific rule, so that only one data stream enters the mapper 24 and one data stream leaves.

Different robustness in the redundancy encoder can be achieved according to the invention by using, for example, a Reed-Solomon code or a FEC code, e.g. with back-coupled slider registers, which has a certain generator polynomial and works with or without puncture.

A mapping rule has a certain number of symbols in the complex plane, whereas for a QPSK mapping there are only four symbols in the complex plane, while for a 16-QAM mapping there are 16 symbols in the complex plane, for example. This means that a decoder at QPSK only needs to distinguish between four different symbols, while a decoder at 16-QAM needs to distinguish between 16 different symbols. The minimum SNR for a QPSK mapping is thus significantly lower than the minimum SNR for a 16-QAM mapping.

Alternative modulation methods can also be used, such as DPSK or 8-QAM. Hierarchical modulation methods, where a QPSK is overlaid with a 16-QAM, for example, can be implemented for the different channels.

Different robustness can also be achieved in modulator 26 if, for example, CDMA modulation uses code sequences of different lengths for the secondary channel or the main channel, or if FDMA modulation uses different frequency bandwidths for the different channels, or CDMA modulation uses different lengths of time slots.

Depending on the implementation, the source encoder of Fig. 2 may include an AudioCodicer 30 and a parallel-connected voice encoder 32. A delay is between the audio encoders 30 and a combiner such as a multiplexer 34 connected, with the delay device designated as 36. For example, it may be configured as a FIFO storage, which is sized to store more than 10 seconds, MSC, respectively, more than 20 seconds and in particular, more than 30 seconds of data. The delay device is therefore larger than the input of the audio encoder 34 kb, while the delay is larger than the input of the combiner 34 kb, while the delay rate for the SDC/SDC service is 18 kb, while the data rate for the main channel is also higher than the input rate for the secondary channel, which is in addition to the 18 kb/s.

Other data rates are also possible, with preference given in particular to MSC to SDC or main channel to secondary channel ratios of < 2, in particular < 5 and special < 10.

In the embodiment shown in Fig. 3, the coder 30 which generates the main channel 18a is thus an audio coder and is trained separately from the coder 32 which generates the secondary channel 18b and is trained as a speech coder only. The speech coder 32 can provide complete speech coder frames, but it can also alternatively only output coefficients as a secondary channel to describe the spectral envelope. In particular, 32 of the language coders will be trained in such a way that the coefficients of the description of the secondary channel 18b are so finely differentiated and so often transmitted that the receiver derives information from the LPCLPCLPCLPCLPCLPCLPCL using a predictive coefficient. For example, when the coding coefficients are calculated on the LPCLPCL or LPCLCOD, it is preferred to use the LPCLPCL 40 as the quantity coefficients, or the coding coefficients, which are derived from the LPCLPCL.

Alternatively, the source encoder 20 of Fig. 4 may be configured as shown in Fig. 4. Such an encoder is used, for example, according to MPEG-4 or MPEG-1, Layer III (MP3). A filter bank 41 converts the information signal 16 into a spectral representation fed to a quantizer 43. The filter bank 41 may be a subband filter bank with, for example, 16 or 32 filter bank channels, or it may be an MDCT filter bank with, for example, 512 coefficients or 1024 coefficients, with an overlap-and-add-on functionality for Time-Domain Aliasing-Cancellation (TDAC) in a corresponding decoder.

The spectrum or spectral representation emitted by the filter bank 41 is quantized in the quantizer 43. The quantizer 43 is controlled by a psychoacoustic model 44 which is trained to calculate the psychoacoustic masking threshold per band and to quantize the masking so roughly that the quantization noise is below the masking threshold. The quantized spectral values emitted by the quantizer 43 are fed to a Huffman coder 45. It should be noted that the quantizer 43 calculates not only quantized spectral values but also scale factors that reflect the rough spectral structure of the spectral representation. The spectral structure is contained in the quantized quantum values.

For Huffman coding, the Huffman encoder 45 uses a variety of predefined codebooks, using 12 different code tables according to the MPEG-AAC standard, all of which differ in the range of values of the elements or spectral values or groups of spectral values encoded by the code table. Each code table is identified by its code table number, which, like the scale factors, is fed to a bitstream formater 46 and is needed on the decoding side to perform decoding with the correct code table.

The output data stream generated by the bitstream formater 46 is then the main channel, while the secondary channel is generated using a secondary channel selector 47. The secondary channel selector is trained to select a certain proportion of the data coming into the main channel to fill the secondary channel with this data. The less data is selected, the lower the data rate in the secondary channel, which is desirable for reasons of a responsible relationship with the transmission bandwidth. However, a certain minimum amount of data is needed to not only produce a color noise on the receiver side, but to be better at producing, for example, a speech comprehensibility.

Depending on the implementation, the subchannel data may thus come from a lower-case voice encoder or a lower-case audio encoder. Thus, even some or all of the subchannel coefficients may come from coefficients of the main channel encoder. Especially when the main encoder is a subband-based audio encoder, it is preferred to select the coefficients of the main channel encoder, which represent the scale factors, into the subchannel.

In particular, it should be noted that, as indicated in Figure 1, additional auxiliary channels 18c may be used, which may be equipped with different time delays and/or different transmission robustness.

Figure 5 shows a special implementation of a device for generating a decoded signal according to an example of an embodiment. The device shown in Figure 5 includes a receiver level 50, which may be coupled, for example, with a receiver antenna not shown in Figure 5. Then the receiver level 50 includes a typical receiver front end, with, for example, a down mixer with coupled local oscillator, to transfer the transmitted spectrum to the base band or to an intermediate frequency band or downstream.

The receiving signal received by the receiver level 50 comprises a main channel and a secondary channel, which are decodable separately from each other. In particular, a minimum signal-to-noise ratio required to decode the secondary channel is smaller than a minimum signal-to-noise ratio required to decode the main channel. The secondary channel is therefore more robust to the transmission characteristics of the transmission channel than the main channel.

Depending on the implementation, the receiver is downlinked to channel level 51 to separate the secondary channel from the main channel already on the RF side, but depending on the implementation, this functionality can also be integrated into a decoder 52 directly coupled to the receiver, which generates the main channel separately from the secondary channel. The broadcasting apparatus of the invention, as shown in Figure 5, can also be used by a quality controller 53 trained to assess a quality of reception. In this case, a quality controller can either re-set the signal to different signals already transmitted by the signal transmitter or use a decoder 53 decode the main decoder in a number of decoders, so that the transmitter can use the decoder 53A in a specific direction or in a specific direction, or use a decoder 53A in addition to the one already transmitted, which may be used in the case of a transition or a transition, but with a different quality controller.

The quality control device 53 is designed to provide a switching signal, when a reception quality is detected below a reception quality source, which is generally fed to the decoder 52 which is then switched from main channel to secondary channel.

The switch, which is e.g. contained in the decoder 52 or may be separately trained, is thus controllable by the quality controller to deliver the secondary channel as a decoded signal when the reception quality is less than a threshold quality and to deliver the main channel as a decoded signal when the reception quality is greater than or equal to the threshold quality.

An alternative embodiment of the receiver according to the invention is shown below in relation to Figure 6. The functionality of elements 51, 52, 53 is implemented in Figure 6 by alternative or additional elements, while the delay 54 of Figure 5 used to delay the secondary channel in relation to the main channel, thus compensating for the end-encoder-side delay, is no longer shown in Figure 6. It should be noted that the delay can be built between any blocks to ensure that, if, for example, a total failure occurs, at least according to the data stored in the decoder, a data output can be performed with the data already stored in the decoder alone.

In Fig. 6, the receiver level 50 of Fig. 5 is a demodulator 60 downlinked, which can perform a demodulation of the underlying modulation process, such as a TDMA, FDMA or CDMA process. This can then be followed by a demapper 61 downlinked, depending on the implementation, which will typically work with soft information to recap the modulation symbols into bits. The bits represented by soft information are fed to a channel decoder 62 which can be trained, for example, as a V-bit decoder or as a Reed-Solomon decoder. The channel decoder is based on the fact that the source decoder is connected to the source signal by the Red-Decoder which is typically introduced by a Red-Decoder, which is then fed to a second channel decoder, in particular a 4-bit decoder, which is then reduced to a decoder, and finally a second channel decoder, which is considered to be a decoder, which is a decoder, which is a decoder, which is a decoder, which is a decoder, which is a decoder, which is a decoder, in turn, a decoder, which is a decoder, in turn, a decoder, which is a decoder, in turn, a decoder, which is a decoder, in turn, a decoder, which is a decoder, in turn, a decoder, which is a decoder, in turn, a decoder, which is a decoder, in turn, a decoder, in turn, a decoder, which is a decoder, in turn, in turn, a decoder, which is a decoder, in turn, in turn, a decoder, which is a decoder, in turn, in turn, in turn, in turn, a decoder, in turn, in turn, in turn, in turn, in turn, in turn, in turn, in turn, in turn, in turn, in turn, in turn, in turn, in turn, in turn, in turn, in turn, in turn, in turn, in turn, in turn, in turn, in turn, in turn, in turn, in turn, in turn, in turn, in turn, in turn, in turn, in turn, in turn, in turn, in turn, in turn, in turn,

When the source decoder receives 63 main channel data, it has all the data needed for decoding, including the spectral fine structure, and a high quality output is achieved. However, if only secondary channel data is transmitted, such as the spectral shell curve of the original information signal, the source decoder will, for example, perform a signal synthesis, where the spectral fine structure is synthesized and weighted with the transmitted spectral gross structure data, such that a synthesized spectrum is generated, which is then fed into the synthesis filter bank to produce a decoded audio channel that has at least a linguistic compatibility.

Alternatively, the subchannel data stream may also be a band-bounded representation (e.g. up to 4 kHz) of the source data, so that the subchannel and the main channel differ only in their bandwidth.

On the other hand, if the secondary data stream is the output signal of a voice encoder, such as a CELP encoder as used in GSM, the broadcast receiver of the invention shall also include a GSM voice decoder to generate the secondary data if the reception quality threshold is exceeded.

The quality controller 53 in Figure 6 can be supplied with information on the reception quality from different points in the processing chain, but the quality controller 53 is preferably fed directly by a channel estimator 64 typically present in a broadcast receiver, trained to estimate the wireless transmission channel with or without pilot sounds.

The representation shown in Figure 6 assumes that the source decoder always emits both main and secondary channels in parallel, and this with compensated delay. This has the advantage that when the reception quality is lower, it is simply necessary to switch to switch 66 so that the decoded secondary signal is already present. Alternatively, however, especially in mobile devices, only the main channel can be separated to be decoded, decoding of the secondary channel only begins when the reception quality goes below the threshold.In this respect, the switch 66 can also be viewed schematically, as it can already be integrated into the functionality of the decoder and can in principle also exist anywhere where the main channel and the secondary channel are already separately present. For example, if the main channel and the secondary channel are already located at the output of the demapper 61, the switch 66 can already be located there to feed either the main channel or the secondary channel into the secondary channel decoder.

With regard to quality monitoring 53 it should be noted that, when the channel estimator 64 is not accessed or when additional data on the reception quality is desired, the main channel or the combined main channel/side channel signal can be accessed at any point to obtain an impression of the current reception quality and in particular the reception quality of the main channel.

The following is a basic outline of the process, as shown in Fig. 7, according to a preferred embodiment of the present invention on the receiver side. It is assumed that a main channel playback is taking place, as indicated in step 70. In parallel with the playback of the main channel in step 70, an assessment of the quality of the main channel takes place in step 72. Alternatively or additionally, as shown in Fig. 6, the quality of reception can of course also be assessed in general, i.e. without direct reference to the main channel, for example if a channel estimator is used, working with or without a pilot.

Err1:Expecting ',' delimiter: line 1 column 107 (char 106)

If the signal is a video signal, the side channel may be a downsampled version of the main channel. Data decimation is done by spatial decimation, e.g. every second pixel per image vertically and horizontally, and/or by temporal decimation, e.g. every second image of a sequence, or by other decimation measures.

Depending on the circumstances, the method of the invention may be implemented in hardware or software. The implementation may be on a digital storage medium, in particular a floppy disk or CD with electronically readable control signals, which can interact with a programmable computer system in such a way that the method is executed. In general, the invention thus also consists of a computer program product with a program code stored on a machine-readable medium to perform the method of the invention, if the computer program product runs on a computer. In other words, the invention may thus be realized as a computer program with a program code to perform the method, if the computer program runs on a computer.

Claims (26)

1. A device for generating a signal to be transmitted, comprising:

   an encoder (14) for generating an encoded signal from an information signal (16), wherein the encoder (14) is implemented

   to generate data for a main channel (18a) and data for a side channel (18b) which may be decoded separate from each other and represent the information signal, and

   to generate the data for the main channel (18a) and the data for the side channel (18b) so that the side channel is more robust against transmission-channel influences than the main channel (18a),

   wherein the encoder comprises a source encoder (20) which is implemented to generated both the data for the main channel (18a) and also the data for the side channel (18b), wherein the source encoder (20) further comprises a selector (47) to select a part of the data generated by the source encoder to form the data for the side channel (18b) using this data,

   wherein the side channel includes scale factors representing energies per different subbands of an information signal, or wherein the side channel comprises code book indices of code books which were used for encoding the main channel, wherein the code books represent values of different value ranges.
2. The device according to claim 1, wherein the encoder (14) is implemented to generate the data for the side channel (18b) so that it is a qualitatively worse representation of the information signal than the main channel (18a).
3. The device according to claims 1 or 2, wherein the encoder (14) is implemented to execute a lossy data compression with the information signal (16), wherein a compression factor in a generation of the data of the side channel (18b) is higher than in a generation of the data of the main channel (18a), or wherein for the main channel a lossless data compression may be executed, while for the side channel a lossy data compression may be executed.
4. The device according to one of the preceding claims, wherein the encoder (14) is implemented to generate the data for the side channel (18b) such that it requires a lower bit rate for transmission than the main channel (18a).
5. The device according to one of the preceding claims, wherein the encoder (14) comprises a source encoder (20) and a downstream redundancy encoder (22), wherein the redundancy encoder (22) is implemented to generate the data for the main channel (18a) with a first code rate and to generate the data for the side channel with a second code rate which is smaller than the first code rate, wherein the code rates are smaller than 1 and are defined by the ratio of a number of input bits with regard to a number of output bits derived from the input bits.
6. The device according to one of the preceding claims, wherein the encoder (14) comprises a mapper (24) which is implemented to generate the data for the main channel (18a) using a first mapping rule and to generate the data for the side channel (18b) using a second mapping rule, wherein a signal/noise ratio for decoding the main channel (18a) is higher than a signal/noise ratio for decoding the side channel (18b).
7. The device according to claim 6, wherein the second mapping rule defines more different mapping states than the first mapping rule.
8. The device according to claims 6 or 7, wherein a mapping state of the second mapping rule represents more bits than a mapping state of the first mapping rule.
9. The device according to one of the preceding claims, comprising:

   a modulator (26) for modulating information of the main channel and the side channel to a plurality of signal carriers, wherein the signal carriers are frequency carriers, time carriers and/or code carriers in an FDMA, TDMA and/or CDMA system.
10. The device according to one of the preceding claims, wherein the encoder (14) comprises a source encoder (20) comprising a speech encoder (32) for generating the data for the side channel (18b) and an audio encoder (30) for generating the data for the main channel, wherein the audio encoder requires more bits for a given time length of the information signal (16) than the speech encoder.
11. The device according to one of claims 1 to 9, wherein the encoder (14) is a source encoder which is implemented to execute a time-range/spectral-range conversion (41) and which provides a description of a spectral envelope for generating the sidechannel coefficients which is at least quantized so finely and is at least transmitted so often that, on the decoder side with a given signal/noise ratio, a speech intelligibility of more than 50% of the total transmitted words may be achieved.
12. The device according to claim 11, wherein the coefficients are LPC coefficients or derived from LPC coefficients.
13. The device according to claim 1, wherein the encoder (14) comprises a subband-based audio encoder configured to generate one scale factor per band, wherein the scale factors represent a spectral coarse structure of a spectral representation of the encoded signal, and wherein the selector (47) is implemented to select the existing scale factors per band.
14. The device according to claims 1 or 13, wherein the encoder (14) comprises a subband-based audio encoder which is implemented to encode quantized spectral values using different code books, wherein the code books represent value ranges of quantized spectral values of different sizes, and wherein code book indices are generated which each represent used code books for the quantized spectral values, and wherein the code book indices are generated representing the code books used for the quantized spectral values, and wherein the selector (47) is implemented to select the code book indices.
15. The device according to one of the preceding claims which is implemented to feed a DAB or DRM transmitter or to feed a package-oriented data network.
16. The device according to one of the preceding claims, wherein the encoder (14) is implemented to generate a further side channel (18c) with a different time offset and/or different robustness against channel influences which may also be decoded separately from the main channel (18a) and from the side channel (18b).
17. A device for generating a decoded signal, comprising:

    a receiver (50) for receiving a receive signal comprising a main channel and a side channel which may be decoded separate from each other, wherein a minimal signal/noise ratio for decoding the side channel is smaller than a minimal signal/noise ratio for decoding the main channel;

    a decoder (52) for generating a main channel which is separate from the side channel;

    a quality observer (53) for assessing a receive quality; and

    a changeover switch (66) which is controllable by the quality observer (53) to provide the side channel as a decoded signal when the receive quality is lower than a threshold quality and to provide the main channel as a decoded signal when the receive quality is higher than or equal to the threshold quality,

    wherein the side channel includes scale factors representing energies per different subbands of an information signal, and wherein the decoder (52) comprises a source decoder which is implemented, for generating the side channel, to synthesize spectral values in subbands for which scale factors were transmitted in the side channel and to weight the synthesized spectral values using the transmitted scale factors, or

    wherein the side channel comprises code book indices of code books which were used for encoding the main channel, wherein the code books represent values of different value ranges, and wherein the decoder (52) comprises a source decoder (63) which is implemented to synthesize code words using the code book indices, wherein for each code book index only one code word is synthesized belonging to the value range which the code book indicated by the code book index represents.
18. The device according to claim 17, comprising a channel estimator (64) for estimating a transmission channel with or without pilot symbols, and wherein the channel estimator (64) is implemented to feed the quality observer (53) with channel data.
19. The device according to claim 17 or claim 18, comprising a demodulator (60), wherein the quality observer (53) is implemented to assess data before and after the demodulator for quality observation.
20. The device according to one of claims 17to 19, wherein a demapper (61) is further implemented and wherein the quality observer (53) is implemented to evaluate an output signal of the demapper (61) for quality observation.
21. The device according to one of claims 17 to 20, wherein the decoder further comprises a channel decoder (62) and wherein the quality observer (53) is implemented to use the channel decoder output data for quality observation.
22. The device according to one of claims 19 to 21, wherein the quality observer (53) is implemented to execute quality observation using the main channel.
23. The device according to one of claims 17 to 22, wherein the decoder comprises a source decoder (63) which is implemented to detect invalid code words representing quantized values, and wherein the quality observer (53) is implemented to execute a quality observation on the basis of a number of detected invalid code words.
24. A method for generating a signal to be transmitted, comprising:

    generating (14) an encoded signal from an information signal (16), such that

    data for a main channel (18a) and data for a side channel (18b) are generated which may be decoded separate from each other and represent the information signal, and

    wherein the data for the main channel (18a) and the data for the side channel (18b) are generated such that the side channel is more robust against transmission-channel influences than the main channel (18a),

    wherein generating comprises using a source encoder (20) which is implemented to generated both the data for the main channel (18a) and also the data for the side channel (18b), wherein the source encoder (20) further comprises a selector (47) to select a part of the data generated by the source encoder to form the data for the side channel (18b) using this data
25. A method for generating a decoded signal, comprising:

    receiving (50) a receive signal comprising a main channel and a side channel which may be decoded separate from each other, wherein a minimal signal/noise ratio for decoding the side channel is smaller than a minimal signal/noise ratio for decoding the main channel;

    generating (52), by a decoder (52), a main channel separate from the side channel;

    assessing (53) a receive quality; and

    providing, as a decoded signal, the side channel, when the receive quality is lower than a threshold quality, or providing, as a decoded signal, the main channel, when the receive quality is higher than or equal to the threshold quality,

    wherein the side channel includes scale factors representing energies per different subbands of an information signal, and wherein the decoder (52) comprises a source decoder which is implemented, for generating the side channel, to synthesize spectral values in subbands for which scale factors were transmitted in the side channel and to weight the synthesized spectral values using the transmitted scale factors, or

    wherein the side channel comprises code book indices of code books which were used for encoding the main channel, wherein the code books represent values of different value ranges, and wherein the decoder (52) comprises a source decoder (63) which is implemented to synthesize code words using the code book indices, wherein for each code book index only one code word is synthesized belonging to the value range which the code book indicated by the code book index represents.
26. A computer program having a program code for executing the method of claims 24 or 25, when the computer program runs on a computer.