Digital transmission system with transmitter and receiver, as well as data carrier

The present invention relates to a digital communication system F_s transmitting a wideband digital signal, such as an audio signal, at a sampling frequency through a transmission medium; the system having a transmitter having an input unit for receiving broadband digital signals, which input unit is connected to one of the inputs of a signal source forming part of the transmitter, which generates a second digital signal and outputs it at one of its outputs ; said second digital signal consisting of consecutive frames, each frame comprising a plurality of information packets, each information packet comprising N bits, wherein N> 1; the system further comprising a receiver comprising a decoder having an input for receiving a second digital signal, said decoder having an output connected to an output unit and providing a broadband digital signal, and wherein BR is the rate of the second digital signal, and N_s the number of samples of the broadband digital signal for which the second digital signal information is contained within a frame of the second digital signal.

The invention also relates to a transmitter and receiver for use in a data transmission system, a transmitter in the form of a means for recording a second digital signal on a track of a recording medium, a recording medium by a transmitter, and a second digital signal to a customer in the form of a device.

The type of data transmission system defined in the opening sentence is known as "The Digital Coding of Speech Signal Requirements for the Auditory System" of The Critical Bánd Coder. ME. Krasner, Proc. IEEE ICASSP 8, Vol 1, p. 327-331, April 9-11, 1980. This article deals with a data transmission system in which the transmitter uses a subband encoding system while the receiver uses a corresponding subband decoding system, but the present invention is not limited to such an encoding system as follows.

In the system known from that publication, the speech signal band is divided into a plurality of subbands having a bandwidth approximately equal to the critical band of the human ear in the appropriate frequency ranges (see Figure 2 of the Krasner article). This division was chosen because, based on psychoacoustic measurements, it was predicted that such a subband quantization error would be optimally suppressed (masked) by the signal of this subband if the quantization tolerance follows the noise suppression characteristics of the human ear; this characteristic gives the threshold value for noise suppression by the simple sound corresponding to the center of the critical band (see Figure 3 of Krasner).

It has been found that for a high quality digital music signal encoded by 16 bits at a sampling rate of 1 / T = 44.1 kHz per signal sample according to the CD standard, this known subband encoding system with properly selected bandwidth and properly quantized subbands provides an output signal at the encoder output which averages approx. It is described in 2.5 bits and the music signal being played is not perceptible different from the original in virtually any sound range of the various music signals.

The subbands do not necessarily have to match the bandwidth of the critical bands of the human ear (hearing). The sub-bands may have other bandwidths, for example, all sub-bands may have the same bandwidth provided that the masking threshold tolerances are adjusted accordingly.

It is an object of the present invention to provide methods for a data transmission system, in particular by selecting a very specific format for transferring a broadband digital signal through a data transmission medium after conversion to a second digital signal, thereby providing a highly flexible, highly versatile data transmission system. created. This means that the transmitter must be able to convert broadband digital signals of different formats into a second digital signal (which, among other things,_s differ in sampling frequency; This allows F to take different values ​​such as 32 kHz, 44.1 kHz, and 48 kHz as used by the AES and EBU digital audio interface - interface 2

EN 210 644 B.) Similarly, the receiver must be capable of generating a broadband signal of the correct format from the second signal mentioned above. In this respect, the data transmission system according to the invention is characterized in that, if P = (BR / N) (n_s/ F_s) in formula P is an integer, then (in formula BR is the bit rate of the second digital signal, and n_s the number of samples of the broadband digital signal whose information is contained within a frame of the second digital signal) the number of information packets B within a frame P, or that if P is not an integer, the number of information packets in each frame P '(where P 'is an integer smaller than P closest to P), and P' + 1 in other frames, so that the distribution exactly matches the requirement, that the average frame rate of the second digital signal is, in essence, F_s/ n_smust be equal to, and that a frame must contain at least one first frame portion containing the synchronization information. The division of the frames B into information packets serves the purpose of arbitrary F_s in the case of a broadband digital signal with a sampling frequency, the average frame rate of the second digital signal transmitted by the transmitter is such that the duration covered by a second digital signal frame corresponds to the_s pattern. This allows synchronization based on the information packet, which is simpler and more reliable than bit based synchronization. Thus, in cases where P is not an integer, the transmitter is capable of supplying a frame with an information block P '+ 1, instead of an information block P', so that the second digital signal has an average frame- rate F_s/ n_s value. Since, in this case, the distance between the synchronous information (synchronous signals, synchronous words) in the first frame of the following frames is also an integer multiple of the information packet length, the synchronization can still be based on information packets. Preferably, the first frame portion further includes information on the number of frame information packets. In a frame containing a B information packet, this information may be exactly the B value. That is, this value is P 'for a frame containing information packet P *, and P' + 1 for frame containing information packet P '+ 1. Alternatively, this information may be P for each frame⁷regardless of whether the frame contains a P 'or P' + l information packet. The information packet inserted as (P '+ lj-th) may, for example, contain only zero values. In this case, the packet does not carry any useful information. Of course, the information packet inserted as (P + l) -th can be filled with useful information. the frame portion may further include system information such as a broadband digital signal F transmitted to the transmitter_s sample rate, copy enable codes, the type of broadband digital signal being transmitted to the transmitter (for example, a stereo or mono signal or a digital signal containing two substantially independent audio signals). As shown below, other system information is possible. The insertion of system information provides the receiver with flexibility and also allows the second digital signal received to be correctly converted to a broadband digital signal. The second and third frame portions of the frame include the signal information. The transmitter may include an encoder comprising a signal splitting means responsive to the broadband signal that generates a second digital signal in the form of an M number (more than one) subband signals, the encoder further comprising means for quantizing each sub-signal. For this purpose, any transformation coding such as Fast Fourier Transform (FFT) can be used. In this case, the data transmission system is characterized in that the second frame portion contains allocation information which - for at least some sub-subsets, indicates the number of descriptive bits of samples of quantized subband signals from said subscripts, and ) On the receiver side, inverse transformation coding must be used to restore the broadband digital signal, for example, inverse fast Fourier transform (IFFT). A data transmission system in which the signal splitting device takes the form of an analysis filtering means responsive to a broadband digital signal and generates an M number of subband signals; divides into a tracking band, and wherein the quantizing device is configured to quantize the corresponding subband signals per block. Such a data transmission system is a system using the subband encoding described above. Such a data transmission system is further characterized in that, for at least some subband signals, the allocation information in the second frame portion of the frame provides the number of bits describing the sampling values ​​of the quantized subband signals from the aforementioned subband signals; present) contains the sampling values. This actually means that the allocation information is in the frames before the sampling values. This allocation information is needed to divide a continuous, serial bit stream of sample values ​​in the third frame portion into an appropriate number of separately sampled sample values ​​on the receive page. The allocation information may also be such that each sampling value is encoded with a certain number of bits per subband and frame. This is a stationary or static bit allocation based transmitter. The allocation information may also be indicative of the time-varying number of bits in the subbands. In this case, we are talking about adaptive or dynamic bit allocation based transmitters. Constant or adaptive bit allocation, among other things, is "Low bit rate coding of high quality audio signals; An introduction to the MASCAM system ”c. G. Theile and

The publication of IIU 210 644 B is published in the August 1988 issue of EBU Technical Review, No. 230. Placing the allocation information in a frame prior to the frame of the samples has the advantage that it is simpler, real-time on the receiver side, and can only be decoded with little signal delay. As a result of this sequence, it is not necessary for all information contained in the third frame to be stored in the receiver first. When the second digital signal is received, the allocation information is stored in the receiver. Since the information content of the allocation information is much smaller than that of the samples included in the third frame part, this requires a much smaller storage capacity than if all samples had to be stored in the receiver. Upon arrival of the serial data stream of the samples in the third frame portion (immediately), this data stream may be divided into each bit number pattern in the allocation information without the need to pre-store the signal information. All allocation information can be contained in a single frame.

However, this is not absolutely necessary, as will be seen from the following.

The data transmission system is further characterized in that the third frame portion further comprises information on scale factors, a scale factor is associated with at least one of the quantized subband signals in the third frame portion, and characterized in that the scale factor information in the third frame portion sub-band before the signs. In the transmitter, the samples can be encoded without normalization, i.e., without dividing the amplitude of an array of subband samples by the amplitude of the sample having the largest amplitude in this block. In this case, the transmission of scale factors is not required. If the samples are normalized during coding, information on the scale factors must be sent so that the maximum amplitude mentioned is available. In this case, if the information on the scale factors is in the third frame before the samples, it is possible to store in the memory the scale factors produced from said information and to reciprocate the samples and said scale factors immediately upon arrival of the samples. immediately, ie without delay. The information on the scale factors may consist of the scale factors themselves. However, it is to be understood that the value added to the third frame portion as the scale factor may also be the reciprocal of the amplitude of the largest sample in the block, in which case the receiver does not have to be reciprocal and consequently the decoding may be faster. Alternatively, the scale factor values ​​may be coded before being inserted into the third frame portion as scale factor information, after transmission. It is also clear that if a subband signal is zero after tax quantization, as is clearly apparent from the sub-band allocation information, then no scaling factor information is required for that sub-band. A data transmission system in which the receiver comprises a decoder comprising a synthesis filtering means responsive to a given quantized subband signal to produce a replica of the broadband signal, the synthesis filtering means combining the subbands to increase the bandwidth by increasing the sampling rate. the original signal band of the digital signal, characterized in that the samples of the subband signals (if any) are inserted into the third frame section in the order which corresponds to the order in which said samples are received in the synthesis filter device after being received by the customer. Placing the samples in the third frame section in the same order as they are placed in the receiver synthesis filtering device also results in fast decoding without the need to store the samples in the receiver before further processing. Consequently, the storage capacity required by the receiver is essentially limited to the storage capacity for system information, allocation information, and, if used, scale factor information. Signal delay is only limited to a limited extent, and this is mainly due to signal processing on samples. The allocation information for the various quantized subband signals is suitably inserted into the second frame portion in the same order as the samples of the subband signals are included in the third frame portion. The same applies to the order of the scale factors. If desired, the frame can be divided into four parts, wherein the first, second, and third frames are as described above. The last (fourth) frame portion of the frame may include error detection and / or error correction information. After receiving this information, errors in the second digital signal during transmission may be corrected in the receiver. As mentioned above, a broadband digital signal may be a mono signal. Or it may be a stereo signal consisting of a first (left) and a second (right) channel component. If the transmission system is based on a subband coding system, the transmitter provides subband signals that include a first and a second subband signal component, which, after quantization in the quantizing device, is converted to the first and second quantized subband signals. In this case, the frame should include allocation information and scale factor information - the latter if the samples were scaled in the transmitter. The order is also important here. Obviously, the system can be expanded to handle a broadband digital signal consisting of more than two components.

The invention is applicable to various digital transmission systems, such as systems for transmitting digital audio signals over the air (digital audio broadcasting). Of course, other applications are possible. For example, transmission via optical or magnetic media. An example of transmission by optical medium is through glass fiber transmission or by optical disc or .4

EN 210 644 B transmission by optical tape. An example of transmission using a magnetic medium is transmission via a magnetic disk or magnetic tape. The second digital signal is then stored in one or more tracks of the recording medium, such as optical or magnetic disk or magnetic disk, in the format of the present invention.

The versatility and flexibility of the transmission system lies in the special format by which the information is transmitted in the form of a second digital signal, for example via a data carrier. In addition, a special design of the transmitter is suitable to produce a special format for different types of input signals. The transmitter generates the system information required for the various signal types and inserts this information into the data stream to be transmitted. On the receiver side, a special receiver is used, the receiver disconnects the information from this system from the data stream and uses it for proper decoding.

Information packets are a kind of fictitious unit used to set the length of a frame. This means that these packets do not need to be directly visible during the second digital signal information. The relationship between the information packets and the existing digital audio interface standard is recorded in IEC 958. Since this standard is primarily applicable to consumer products, it defines frames that contain a pattern of both the left and right channels of the stereo signal. These patterns are described by binary complement numbers of 16 bits. For N = 32, a frame according to this digital audio interface interface interface can transmit exactly one packet of information of a second digital signal. By digital audio interface standard, the number of frames per unit time is equal to the sampling frequency. In this case, the number of frames per unit of time (the frame rate) should be chosen to be BR / N. This choice allows the use of integrated circuits currently used in devices with digital audio interface standards.

The invention by way of example! Embodiments thereof will now be described in more detail with reference to the drawings. The

First FIG. 2B is a view showing a second digital signal produced by the transmitter, said second digital signal being divided into frames, each frame consisting of information packets. THE

Second Fig. 4A shows the structure of the frame. THE

Third Figure 1B shows the structure of the first frame portion of the frame. THE

4th An example of a transmission system is shown in FIG. The

5th Figure by frame! Table B is a table specifying the number of information packets B for specific values ​​of the bit rate BR and the sampling rate Fs. THE

6th FIG. 3B shows the number of frames in a non-well-conditioned series and the number of frames containing the series "blank" information packet for different values ​​of the BR bitrate. THE

7th FIG. 2B shows system information located in the first frame portion of the frame. THE

8th Fig. 4A shows the distribution of digital information for different (two) channels in different operating modes. THE

9th Fig. 6A shows the importance of the allocation information inserted in the second frame part. THE

10th and the

11th FIG. 3B illustrates the order in which allocation information in the second frame part is stored in two formats, A and B, respectively. THE

12th Figure 4A shows an example of a customer. THE

13th FIG. 2B is a view showing a transmitter in the form of a means for recording a second digital signal on a magnetic medium. THE

14th FIG. 4A is a view showing a receiver in the form of a device for recovering a second digital signal from a magnetic medium. THE

15a-15d.

Figures 3 to 5 show further possibilities for fitting scale factors and samples into the third frame portion of the frame. THE

16th Fig. 4a shows a further modified version of the transmitter. THE

17th FIG. 4A shows another possible structure of the first frame part of the frame. THE

18th FIG. 6A shows information about a system mounted in the first frame part of FIG. 17. THE

19th and Fig. 20 shows in more detail the information contained in the first frame part of Fig. 17. THE

21st and FIGS. 22 and 22 illustrate the sequence in which the allocation information is located in the second frame portion following the first frame portion of FIG. 17. THE

23rd Fig. 2A shows the structure of a frame containing an auxiliary signal. THE

24th Figure 1A shows how scale factors are created. THE

25th Figure 5A shows quantization of scaled samples to q-bit numeric values. THE

26th Figure 3B shows a recovery from a q-bit digital value.

Figure 1 schematically shows the second digital signal as it is generated by the transmitter and transmitted as it is transmitted through the transmission medium. The second digital signal consists of frames, two such frames (the jth and the j + 1th) are shown in Figure 1a. The frames, like the j-th frame, consist of information packets IP1, IP2, IP3, ... (see Figure 16).₀, b_b ... bN__rmarked with - consists of bits. (see figure below).

The number of information packets in the frame depends on (a) the bit rate of the BR, at which rate the second digital signal is transmitted through the transmission medium, (b) N, i.e. the number of bits in an information packet, where N is greater than , (c) F_s, the sampling frequency of the broadband digital signal, and

HU 210 644 Β (d) n_scorresponding to the number of samples of the broadband digital signal corresponding to the second digital signal after conversion to the transmitters, and enclosed in a frame as follows.

The P parameter is calculated using the following formula:

P = (BR / N) {n_s/ F_s)

If this formula gives an integer P, then the number B of one frame information packet is equal to P. If the result of the calculation is not an integer, some frames will contain a P 'information packet, others will contain a P' + l information packet. The number of frames containing the information packet P 'or P' + l is obviously chosen such that the average number of frames per time unit (the frame rate) is F_n/ n_sis equal to. It is further assumed that N = 32 and n_s= 384th Figure 5 is a table of these N and n_svalues ​​plus four bitrates and three Fs_sfor sampling rate, enter the number of information frames (sections) per frame. Obviously, the 44.1 kHz F_s for the sampling rate, the P parameter is not always an integer; accordingly, some samples contain information packets 34 and others 35 (at BR = 128 kbit / s). This is shown in Figure 2 as a frame. The frame consists of an information packet P '; these are IP1, IP2, ..., IPP '. In some cases, the frame contains a P '+ l information packet. This is achieved by adding an additional information packet (an "empty" section) to the frame consisting of the information packet P '. 6. The second column of the table in Fig. 1A shows the number of frames in the non-conditioned, matched series for a sampling frequency of 44.1 kHz and the four bitrates mentioned above. The third column indicates how many of said number of frames in the sequence contain a P '+ 1 information packet. Subtracting the values ​​in the third column from the values ​​in the second column gives the number of frames in the series that P⁷ information package. In this case, the information packet (P '+ lj) does not have to carry information. In this case, the information packet (P' + lj), for example, contains all zeros. Other values ​​(such as intermediate values) are also possible Figure 2 shows that the frame consists of three frames, in the following order: FD1, FD2, FD3.

The first frame part of FD1 contains synchronous and system information. The second frame part of FD2 contains the allocation information. The third frame portion of the FD3 samples and, when present, the scale factors associated with the second digital signal. Before further explanation, the operation of the transmitter in the transmission system according to the invention must be described.

FIG. 4 schematically depicts a transmission system including an S_BB a transmitter 1 having an input unit 2 for receiving a broadband digital signal, such as a digital audio signal. In the case of an audio signal, it may be a mono signal or a stereo signal, in the latter case the digital signal consists of a first (left channel) and a second (right channel) signal component. It is assumed that the transmitter includes a subband encoder for encoding the broadband digital signal and, consequently, the receiver includes a subband decoder for restoring the broadband digital signal. The transmitter contains an S_BB 3 analysis filtering devices sensitive to broadband digital signal, which are M_SBi · · · S_SBM generates a subband signal, which is an analysis filtering tool in S_BB subdivides the full width signal band of a digital broadband signal into a successive m sub-band (l <m <M) with decreasing sampling rate. These subbands may have the same bandwidth, or else the subbands may have different bandwidths. In the latter case, the subbands may correspond, for example, to the bandwidth of critical bands of the human ear (hearing). The transmitter further includes means 9 for quantizing the corresponding subband signal per block. This quantizer 9 is depicted in Figure 4. Such a subband encoder is, of course, known and is described in, among others, the publications of Krasner and Theila et al. Reference is also made to European Patent Application EP 289 080.

Instead of further describing the operation of the subband encoder, reference is made to those publications. These publications are considered as part of this application. With such a subband encoder, significant data compression can be achieved, e.g._BB broadband digital signal per sample! The signal transmitted from 16 bits to the receiver 5 via the transmission medium 4 (see FIG. 4) is, for example, per sample. Can be compressed to 4 bits. In the former, n_s = 384 was assumed. This means that there are blocks of 384 patterns of the broadband digital signal, each of which is 16-bit. We now assume M = 32, that is, the broadband digital signal is subdivided into 32 subband signals in the analysis filter device 3. At this point, 32 subbands (blocks) appear on the 32 analysis outputs of the 3 filtering devices, each block containing 12 samples (equal to the width of the subbands) and each sample 16 bits. This means that the information duration at the outputs of the analysis filter device 3 is still the same as S_BB a block of 384 sample blocks of broadband digital signal at the input 2. The device 9 performs information reduction by using the prior art for masking to coarser quantize samples of 32 sub-bands containing 12 samples each, so that they can be described in fewer bits. In static bit allocation, each pattern is represented by a fixed number of bits per subband and frame. This fixed number may be different or the same for each subband, for example, the patterns may be represented uniformly by 4 bits. With dynamic bit allocation, the number of bits associated with each subband may vary over time, sometimes resulting in even greater data loss and sometimes better quality at the same bit rate. In device 9, quantized subband signals are transmitted to generator unit 6. Starting from the quantized subband signals, this generator unit 6 generates a second digital signal which is shown in Figures 1 and 2. This is the second

EN 210 644 Β digital signal, as previously described, can be transmitted directly through the medium. However, preferably, this second digital signal is first modified in a signal converter (not shown) for transmission through the transmission medium 4. Such a signal converter includes, for example, an 8 to 10 converter such as that described in European Patent Application EP 150 082. This converter converts 8-bit data words to 10-bit. Such a converter further enables the insertion of error detection bits. The purpose of everything is to enable the information to be received on the buyer side to be corrected.

Of course, the signal received by the receiver 5 from the transmission medium 4 must be freed from the interleaved bits and converted from 10 to 8. The structure and content of the frames will now be discussed in more detail. The first frame portion of FD1 of Figure 2 is shown in greater detail in Figure 3. Figure 3 clearly shows that in this case the first frame part is exactly 32 bits and it is exactly an information packet, this EP1 is the first information packet of the frame. The first 16 bits of the information packet form the sync signal (sync word). For example, a synchronous signal may consist of all "Γ" s. 16-31. bits carry system information. A16-31. bits represent the number of frame information packets. Consequently, this number is equal to P ', even if the frame contains an information packet P', even if the frames contain an additional information packet IP P '+ 1. P 'up to 254 (1111 1110 in binary form) to avoid similarity to the sync signal. 24-31. bits carry information about the frame format. Figure 7 gives an example of the arrangement and significance of this information. Bit 24 indicates the frame type. The length of the second frame part in format A (number of packets of information) differs from the length in format B. As will be apparent from the following, the second frame part of FD2 consists of 8 information packets, using format A, these IP2-IP9 information packets (including the IP9 information packet), and the second frame part of FD2, using format B, consists of 4 information packets, these are IP2 to IP5 information packets (including the IP5 information packet). Bits 25 and 26 indicate whether information is allowed to be copied. 27-31. bits indicate how to use it. This means:

a) the channel mode, which indicates the type of broadband signal (as mentioned earlier, the signal may be a stereo signal, a mono signal, or two different signal components, for example, the same text in two languages). Figure 8 shows the channel mode. It illustrates how the signal components are divided between the two channels (channels I and II) in the above cases.

b) Broadband signal F_s sampling frequency.

(c) the highlighting applicable to the broadband digital signal in the transmitter. Refers to a special standard for highlighting defined by CCITT (Comité Consultative Internationale de Télégraphie et Téléphonie).

The contents of the FD2 frame portion of Figure 2 will be described in more detail with reference to Figures 9, 10 and 11. The second frame - Using the format, contains eight information packages. This is because we assume that the SBB broadband digital signal (more precisely, each of its signal fragments) is converted to 32 subband signals. Each 4-bit allocation word is associated with each subband. This gives a total of 64 4-bit allocation words each, these 64 allocation words can be placed in exactly 8 information packets. Using format B, the second frame part contains allocation information for only half of the subbands, and thus the second frame part consists of only 4 information packets. Figure 9 illustrates the importance of AW four-bit allocation words. The allocation word for a subband specifies the number of bits by which the subband patterns, after quantization in the device 9, are described for that subband. For example: the value 0100 of the AW allocation word indicates that the patterns are written (represented) by 5-bit words. It follows from Figure 9 that the value 0100 of the AW allocation word indicates that no samples were generated in that band. This situation may occur, for example, if the subband of the adjacent subband is of such amplitude as to completely suppress (mask) the subband of that subband. The 1111 value of the AW allocation word is unused because it shows a strong similarity to the synchronous word of the first information packet IP1. In Figure 10. Figure A shows the order, In addition to the mode of operation (frame mode), in which the allocation words AWj, m, (where ja denotes one of two channels, that is, possible values ​​of j are: I, II, and today the 32th subband (i.e., m may be between 1 and 32) are located in the second frame portion. The allocation word AW I, 1, which is first inserted into the first subband signal component (channel I, subband 1) of the first and lowest (frequency) subband. Next, the allocation word AWII, 1 is inserted into the second frame part of FD2, which is the allocation word AWH, 1 to the second subband signal component (channel,, subband 1) of the first and the lowest (frequency) subband. Next, the allocation word AW I, 2 is inserted into the second frame part of FD2, which is the allocation word AW I, 2 to the first subband signal component (channel I, subband 2) of the second and lowest subband. This is followed by the allocation word AW Π, 2, which belongs to the second subband signal component of the second subband (channel Π, subband 2). Thus, it continues up to the allocation word AW Π, 4, which is inserted in the second frame part of FD2 and belongs to the second subband signal component (channel,, subband 4) of the fourth subband. This completes the frame IP2 information packet (section 2), which is the first information packet of the second frame part FD2. Next, the IP3 information packet (Figure 3). ) are filled with the allocation words AW I, 5, AW II, 5, ... AWΠ, 8, and thus continue in the order illustrated in Figure 10. Figure 10 shows only the indexes of the allocation words AWj, m, inserted. Figure 11 shows the order of allocation words for a B-format frame. In this case, only the allocation words of the subbands 1-16 are inserted.

HU 210 644 B

The order as illustrated in Figure 10 corresponds to the order in which j. and at the same time m. some of the samples in the subband are received in the synthesis filter device after they have been received by the receiver. This will be explained in more detail below. For example, the serial data stream contains only frames of A format. In the receiver, the allocation information in each frame is used to correctly generate samples from the information in the third frame portion of that frame. However, the serial data stream may consist of alternating A-format and B-format frames. However, frames in both formats may include, in the third frame section, samples for each channel and each subband. In this way, a B format frame does not have the features of FIGS. subband I. or II. with the allocation information needed to generate samples from the channel B from the third frame portion of the B format frame. The receiver includes a repository in which the allocation information contained in the second frame portion of the format A frame can be stored. If the next frame is in B format, then only frames 1-16 will be used. and sub-bands I and Π. The channel allocation spatial allocation information is replaced by the allocation information included in the second frame portion of the B format frame, and the allocation information for these subbands from the previous and still stored A format frames is used to produce samples for the B format subband 1732. The reason for the alternate use of format A frames and format B frames is that the allocation information for some subbands, in this case the higher subbands (17-32), does not change rapidly. Because allocation information for the various bands is available in the transmitter during quantization, this transmitter is able to decide to generate a B format frame instead of a A format, provided that it is in the range 17-32. the allocation information for the sub-band (including 32) will not change (significantly). In fact, this shows that there is now additional space for inserting samples into the third part of FD3. At a given value of P ', the third frame part of a format B frame is four times longer than the third frame part of a format A frame. Consequently, this allows the lower 1-16. The number of bits describing the samples of the subbands would increase to achieve higher transmission accuracy for these subbands. Furthermore, if more accurate quantization of the lower subbands is required, the transmitter will automatically switch to generate B format frames. This may be to the detriment of the accuracy of the quantization of the upper subbands.

The third frame portion of FD3 of FIG. 2 includes samples of quantized subband signal components for two channels. If the allocation word 0000 is not included in the second frame part FD2 for any of the channels, this means that in the present example, the FD3 third frame part - each of the 32 subbands and the two channels - has been assigned -12 samples each. This means that there are 768 samples in total. In tax, samples - before they are quantified - can be multiplied by some scale factor. The amplitudes of the twelve samples per subband and channel are divided by the amplitude of the sample having the highest amplitude. In this case, the scale factors per subband and channel must be transmitted to the receiver in order to allow the inverse operation to be performed on the samples on the receiver side. For this purpose, in this case, the third frame part FD3 contains the scale factors SF1, m, one for the quantized subband signal components of each subband. In the present example, the scale factors are represented by 6 bytes, the most valuable bit being first, values ​​ranging from OOOOOO to 111110. The scale factors assigned to the allocated subbands (i.e., those with non-zero allocation information) are transmitted before the samples are transmitted. This means that the scale factors are located at the beginning of the third part of FD3, before the samples. This (mode or placement) allows fast decoding in the receiver 5 without having to store all the samples in the receiver, as will be seen below. The SF j, m scale factor may denote the value by which the samples of the j-channel of the m-subband are multiplied. Alternatively, the reciprocal of said value can be stored as a scaling factor, so that the reciprocal of the scaling factors on the receiver side is not required to calculate the correct value of the sample. For frame format A, the maximum number of scale factors is 64. If the allocation word AW j, m for channel j and subband m is set to 0000 (which means that there are no samples for this channel and this subband in the third frame part of FD3), then the scale factor for this channel and this subband need not be ( in the frame). The number of scale factors is then less than 64. The order of the SF j, m scale factors in the third frame portion of FD3 is the same as the order of the allocation words in the second frame portion. The order is thus as follows: SF I, 1; SF Π, 1; SF I, 2; SF Π, 2; SF I, 3; SF H, 3, ... SF I, 32; SF II, 32.

If it is not necessary to place a scale factor in the (frame), the order will be incomplete. For example, the order may be as follows:

... SF 1.4; SF 1.5; SF II, 5; SF Π, 6; ... In this case, a. There is no scale factor for Channel 4 sub-band 4 and Channel I sub-band 6. For a B format frame, scale factors for all channels and subbands can be placed in the third frame section. However, this is not necessarily the case. In this case, it would be possible for the third frame part of the frame to be limited to the one shown in FIGS. sub-band scale factors. This fill requires the receiver to have a magazine in which all scale factors can be stored at the moment when the pre-arriving A-format frame is received. Then, when you receive a B-format frame, only the 1-16. the sub-band scale factors are replaced by the scale factors in the B format frame8

HU 210 644. Β some. 17 to 32 of the previously received A format frame. The sub-band scaling factors are used to restore these sub-bands to the appropriate scale in the third frame portion of the B format frame.

The order of the samples in the third frame of FD3 is the same as the order of the allocation words and the scale factors: one sample for each subband of each channel, one after the other. This means that there are first samples of quantized subband signals belonging to the subchannels of the two channels, followed by second samples, and so on. The (binary) numbering of the patterns is arbitrary, and it is advisable to avoid using the binary word "single" any more.

The second digital signal generated by the transmitter 1 passes through the output 7 to the transmission medium 4; the signal is transmitted to the receiver 5 via the transmission medium 4. Transmission via the transmission medium 4 may be a wireless transmission, such as a radio transmission channel. However, other transmission media are possible. In this context, optical transmission may also be considered, for example via fiber optics or optical media (CD-like device); or transmitting on magnetic media using recording and reproducing techniques similar to RDAT or SDAT; with reference to RDAT and SDAT recording and reproduction technology, reference is made to Watkinson's 5th book, The Art of Digital Audio (Focal Press, London, 1988).

The receiver 5 includes a decoder which decodes the signal encoded in the encoder of the transmitter 1 and produces a copy of the broadband digital signal which is output to the output 8.

Figure 12 shows in greater detail the receiver 5 of Figure 4. The encoded signal (the second digital signal) is transmitted to the unit 11 via the entry point 10. The essential information about the incoming signal lies in the scale factors and the samples. The information of the second digital signal beyond the essential information is only needed for correct decoding. The decoding process is repeated for each incoming frame. The transmitter first receives the synchronization and system information from the frame. Each time the unit 19 detects each synchronization word located on the first 16 bits of the first frame portion of the frames. Since the synchronization words of successive frames are each time packed with so many information packets that are an integer multiple of P 'or P' + 1, the synchronization words can be detected very accurately. Once synchronization has been achieved, the synchronization word can be detected in the unit 19 by opening the time window of the unit 19, such as a packet of information, for each packet of information P 'so that only the corresponding portion of incoming information is transmitted to the synchronous word detector . If the synchronization word is not detected, the time window remains open for an additional packet of information, since the previous frame may have been a frame containing the P '+ 1 information packet.

From these synchronization words, a PLL (Phase Closed Loop) in the unit 19 can generate a clock signal for controlling the CPU 18. From the foregoing, it is obvious that the customer must know how many information packages are in a frame. To this end, the system information is fed to the switching device 15 via an input of the processing unit 18, which switching device is then in the position shown in the figure. The system information can then be stored on the storage compartment 18a of the processing unit 18. Information on the number of frame information packets may be transmitted to the unit 19 via the control signal line 20 so that the "time window" begins to detect the sync word at the appropriate time. When receiving system information, the switching device 15 is moved to its lower position. The allocation information in the second frame portion of the frame may then be stored in the memory 18b. The fact that the allocation information in the inbound frame does not contain allocation words for each band and channel is already apparent from the detected system information. For example, it can be determined whether the frame is an A format or a B format frame. Thus, in accordance with the information contained in the system information, the processing unit 18 stores the received allocation words at the appropriate locations in the allocation store 18b. Obviously, in the present example, the allocation store 18b contains 64 storage locations. If scale factors are not transferred, elements 11, 12 and 17 may be omitted and the content of the third frame portion of the frame is passed to the synthesis filter device via input 10, which is connected to said filtering means via connection 16. The order in which the samples are received by the filtering device 21 is the same as the order in which the filtering device 21 processes the samples to reconstruct the broadband signal. The allocation information stored in the allocation store 18b is needed to break the serial data stream of samples into discrete samples in the filtering device 21 so that each sample bit number is correct. For this purpose, the allocation information is passed to the filtering device 21 via line 22. The receiver further comprises a suppression unit 23 which suppresses the reconstructed digital signal provided by the filtering means 21. For correct repression, the first frame portion is shown in FIGS. The relevant information contained in its bits must be transmitted from the storage 18a to the suppression unit 23 via line 24. If the third frame part also includes scale factors SF j, m, the receiver includes a switch 11, a magazine 12 and a multiplier 17. At the time of arrival of the third frame portion of the frame FD3, the all switch is in a down state due to a control signal transmitted by the processing unit 18 through the line 13 . The scale factors can be derived from the magazine 12. As a result of the data signals provided by the processing unit 18 through the line 14 and leading to the memory 12, the scale factors are stored in a corresponding portion of the memory 12. The A12 repository has 64 slots to store the 64 scale factor. When a B format frame is received, the processing unit 18 outputs address signals to the storage 12 such that only the

HU 210 644 Β

1-16. The scale factors of the subbands are overwritten by the scale factors to be placed in the B format frame. Subsequently, the control signal applied to the switch 11 via the line 13 (line) causes the switch to switch to the indicated (top) position, thereby transferring the samples to the multiplier 17. As a result of the allocation information that is received by the multiplier 17 on the line 22, the multiplier first generates each sample from the serial data stream on the line 16 with the appropriate number of bits. Subsequently, the samples are multiplied to restore the samples to their correct pre-tax value. If the scale factors stored in the library 12 are the scale factors with which the samples have become normal in the transmitter, then the reciprocals of these scale factors must be formed and then multiplied by a factor of 17. Obviously, the reciprocal of the scale factors - after buying the scale factors - can be formed before being placed in 12 libraries. If the scale factors in the frame are equal to the values ​​at which the samples are to be multiplied at the time of reception, these scale factors can be written directly to the memory 12 and directly to the multiplier 17. It can be seen that there is no need to store all of these samples before signal processing on the frame samples begins. By the time the sample arrives through the bottom line, all the information needed to process this sample is already available, so that processing can take place immediately. The whole process is effected by control signals and clock signals which control all parts of the transmitter by the processing unit 18. Only a portion of the control signals are shown. This is not necessarily because the customer's operation is obvious to the skilled person. The processing unit 18 controls the multiplier 17, which multiplies the samples by the appropriate multiplication factors. The samples, which have been restored to their true amplitude, are fed to a filtering device 21 in which the subband signals are converted to a broadband digital signal. No further description of the receiver is required, as such receivers are generally known, see, for example, "Introduction to the MASCAM system for low bit rate coding". Theile et al., EBU Technical Review, no. 230, August 1988.) It is also clear that if system information is also transmitted, the receiver can be very flexible and can decode signals correctly, even for second digital signals having other system information.

A13. FIG. 2B is a schematic representation of another embodiment of a transmitter. The transmitter takes the form of a recording device 27 for recording broadband digital signals on a medium. The media (information carrier) in this case is 25 magnetic media. The generator unit 6 (encoder) provides a second digital signal to the recording device 27, which recording device 27 includes a tag head 26, by means of which the signal is recorded on a track (track) of the magnetic medium 25. The second digital signal may be recorded on a single track on the magnetic medium 25, such as a spiral scan recording device, in which case the single track is actually divided into consecutive tracks; their orientation is different from the longitudinal (circular) direction of the 25 magnetic media. An example of this method is the RDAT-like recording procedure. Alternatively, the information is wrapped and the hacked information is simultaneously recorded on adjacent bands that are longitudinally (circumferentially) on the magnetic medium 25. For this purpose, an SDAT-like recording procedure may be appropriate. A comprehensive description of these two procedures can be found in the book already given (J. Watkinson, "The Art of Digital Audio").

Again, the signal provided by the generator unit 6 can first be encoded in a signal converter. This coding may again be a conversion from 8 to 10 followed by an intervening operation as described with reference to FIG. If the encoded information is recorded on adjacent parallel tracks on the magnetic medium 25, these signal converters must be capable of distributing the encoded information between the different tracks.

Figure 14 schematically shows an embodiment of the receiver 5, which in this case is a reading device for reading the magnetic medium 25, which has recorded a broadband digital signal on the magnetic medium 25 in the form of a second digital signal using the device shown in Figure 13. The second digital signal is read by a track head 29 of the magnetic media 25 and fed to the receiver 5, which may have, for example, the construction of FIG. 12. The reading device 28 may be constructed to perform an RDAT-like or SDAT-like reproduction process. Both of these procedures are described in detail in Watkinson's book. Subject to Article 13. If the signal provided by the generator unit 6 of the recording device of FIG. 6B is converted by means of, for example, a conversion from 8 to 10 and by an interleaving method, the inserted portions must first be removed from the encoded signal read from the magnetic medium 25 and then 10 to 8. needs to be converted. Further, if the encoded signal is recorded on multiple parallel bands, FIG. The reproduction unit according to FIG. 1B must arrange the information read from these tracks in the correct order before further processing.

A15. Figure 3B illustrates several other variations of scale factors and samples in the third frame portion of the frame FD3. Fig. 15a illustrates the above described procedure in which the subbands m and I and II are shown. SF scale factors for the channels were placed in the third frame part of FD3, before the samples. Fig. 15b shows the same situation as Fig. 15a except for the storage capacity reserved for the SFI, m and SFII, m scale factors and the samples associated with the scale factors x (in the two channels of said subband m). Fig. 15b shows samples of the two channels of subband m arranged in array, while FD3 is normally located in the third frame part (distributed). The samples are y-bit. In the example above, x = 12

EN 210 644 B and y = 8. Figure 15c illustrates another format. The scale factors for the first and second channels of the subband are still included in the third part of FD3; with the exception that instead of x samples per subband m (left and right channels of the stereo signal) (i.e. 2x total), only x samples for subband m are included in the third frame part of FD3. For example, this sample x is composed by adding the appropriate samples of the two channels. In fact, so in this m. a mono signal is created in the subband. In Fig. 15c, pattern x is individually z bit. If it is equal to y, this will save space in the third frame part of FD3, which can be used to place specimens that require more accurate information (denser quantization). Alternatively, the x pattern of the mono signal can be expressed in z = 2y (= 16) bits. Such signal processing (signaling) is used when the phase difference between the left and right signal components in a subband is unimportant while the mono (monophonic) waveform is significant. This is especially true for signals in the higher subbands, since the phase sensitivity of the human ear is lower at frequencies in these subbands. By depicting the x pattern of the mono signal in 16 bits, the waveform (waveform) is better described (more precisely quantized) such that the space occupied by these patterns in the third frame portion of FD3 is the same as the example illustrated in Figure 15b. Alternatively, the patterns in Figure 15 are represented in 12 bits. Signaling is then more accurate than in the example illustrated in FIG. 15b, but space can be saved by restoring the signals in the third frame portion of FD3 (FIG. 15c) to provide a stereo effect known as intensity stereo. In this case, only the intensity of the left and right signals (in sub-band m) can differ (due to the different values ​​of the scale factors SFI, m and SF Π, m).

Figure 15d shows another possibility. In this case the m. There is only one SFm scale factor for the two signal components of the subband. This is particularly the case for low frequency subbands. Another possibility, not shown, is that m. Subbands I and II channel sample x, as in Fig. 15b, does not have the associated SF I, m and SFII, m scale factors. Consequently, these scale factors are not included in the same third frame part of FD3. In this case, the scale factors SF I, m and SF Π, m in the third frame part of the previous frame must be used in the receiver to normalize the samples. All of the options shown with reference to FIG. 15 can be utilized in the transmitter to provide efficient data transmission over the transmission medium. That is, the frames of FIG. 15 may alternate in the data stream. It will be appreciated that if the receiver still needs to decode these different frames correctly, the structure information of these frames should be included in the system information.

Figure 16 shows the transmitter in more detail. The figure shows how the various information elements can be combined into the serial data stream of Figures 1, 2 and 3. Fig. 16 shows in detail the generator unit 6 of the transmitter 1 as an encoder. The generator unit 6 comprises a central unit 30 which controls a plurality of elements of the encoder. The encoder includes a generator 31 located in the central unit 30, which generates synchronization information and generates the synchronization information 3. and includes a generator 32 for providing allocation information, a generator 33 for specifying scale factors, and a generator 34 for specifying patterns (sample values) in the frame. for. Generator 35 is a unit capable of generating an additional IP P '+ 1 information packet. The output of these generators goes to the corresponding input of the switching device 40 (five-position switch), the output of the switching device forming the output 7 of the generating unit 6. The switching unit 40 is also controlled by the central unit 30. The generators 32, 33, 34, 35 are controlled via wires 41.1-41.4. The operation of the transmitter is described in the case of a mono signal divided into a subband signal M. In this S_SB1-S_sbm (M pieces) subband is supplied to terminals 45.1, 45.2, ... 45.M. For example, an array of 12 samples per subband is associated. In units 46.1 to 46M, if present, the probe sample 12 is scaled to the amplitude of the probe's largest sample. The scale factor M is fed to the generator 33, if any, via the 47.147.M lines. The subband signals are included partly in the 48.1-48.M (M) quantizer and partly in the 49 units. The unit 49 determines, for each subband, the number of bits for which that subband is to be quantized. This information is passed through the 50.1-50.M wires to the corresponding quantizers 48.1-48.M so that these 48.1-48.M quantizers correctly quantify the 12 patterns of each subband signal (to the correct bit number). This (allocation) information also passes to the generator 32. Samples of quantized subband signals are fed to generator 34 via lines 51.1-51.M. The generators 32, 33 and 34 arrange the allocation information, scale factors and samples in an appropriate order (i.e., as described above). The central unit 30 generates synchronization information and system information for the frame to be generated; the information already stored in the generators 32, 33 and 34 is stored in the frame to be generated. In the indicated position of the switching device 40, the synchronization and system information associated with the frame is provided by the generator 31 and is output to the output 7. Subsequently, the switching means 40, by the control signal provided by the central unit 30 via the line 53, moves from the top to the second position so that the output of the generator 32 is output to the output 7. The allocation information provided by the generator 32 is then output to the output 7. The order of the allocation information is shown in Figure 10 or 11. . The switching device 40 is then moved from the top to the third (middle) position by the central unit 30. This means that the output of the generator 33 is output to the output 7. The generator 33 gives the scale in the correct order11

HU 210 644 B factors for the 7 outputs. The switching device 40 is then moved to its next position by the central unit 30 so that the output of the generator 34 reaches the output 7. The generator 34 then outputs the patterns of the various subbands to the output 7 in the correct order. During this cycle, exactly one frame is applied to the output 7. The central unit 30 then resets the switch unit 40 to its uppermost position. This starts a new cycle in which the next block, block of 12 samples, is coded for the next block, and the next block is generated at the output 7. In some cases, such as when F_s a sampling frequency of 44.1 kHz (see Figure 5), an additional information packet (blank section, see Figure 2) shall be issued. The switching device 40 is then moved from its position at which the generator 34 reaches the output 7 to the lower position of the central unit 30. The output of the generator 35 is then output to the output 7. The generator 35 generates an additional information packet IP P '+ l and is output to the output 7. The switching device 40 is then moved to the uppermost position by the central unit 30 and a new cycle begins. Obviously, if the errors of the signals received by the transmitter 1 during the transmission are corrected, appropriate channel coding should be applied to the second digital signal. It is also necessary to modulate the second digital signal before transmitting the second digital signal. Thus, the digital signal transmitted on the transmission medium 4 may not be directly identifiable as, but derived from, the second digital signal. It should also be noted that, for example, when the subbands are of different widths, the number of samples of the different subbands in the third frame portion of FD3 may vary and are likely to differ. For example, it is assumed that the subdivision is made into three subbands (lower subbands SB1, middle subbands SB2, upper subbands SB3). The SB₃ for example, the bandwidth of the upper subband should be twice the bandwidth of each of the other two subbands. This means that SB in the third frame of FD3₃ the number of samples in subband will also be twice the number of samples in any one or the other two subbands. The order in which the samples are placed on the receiver's reconstruction filter can then be as follows. The SB₍ the first sample of subband, SB₃the first sample of subband, SB₂ the first sample of subband, SB₃subband second sample, SB, subband second sample, SB₃ the third sample of sub-band SB₂ second sample of subband B, SB₃ subband fourth sample, etc. The order in which the allocation information in the second frame part of FD2 is: SB, the allocation word for the subband, followed by SB₃ sub-band allocation word followed by SB₂ sub-band allocation word. The same order applies to scale factors. The receiver can determine from the system information that in this case the cycle contains groups of 4 samples each, each group is a sample of SB, a subband, SB₃a sub-band, SB₂ and SB₃ contains another pattern of subband.

Figure 17 shows a different structure of the first frame part FD1. Again, the first frame part of FD1 contains exactly 32 bits and thus corresponds to an information packet. Again, the first 16 bits are the sync signal (or sync word). The sync word may be the same as the sync word of the first frame part FD1 of FIG. 16-31. however, the information contained in the bits is different from FIGS. bits of information. A b,₆-b, g bits denote the BR bitrate index. The BR bitrate index is a 4-bit number whose meaning is shown in the table in Figure 18. If the BR bitrate index is a 4-bit number "0000", it indicates the free format, then the bitrate is unknown, and then the decodem must rely solely on the sync word to detect the start of the new frame. The decimal number corresponding to the binary "llll" is not included in the table to avoid disturbing the sync word. The 18th. The second column of the table in Figure 1A shows the decimal rewrite of the 4-bit binary number. The bitrates for the BR bitrate index are listed in the first column.

Bits 20 and 21 are F_s the sampling frequency (see Figure 18). FIG₂₀ and b₂, 2 bit bits (4) and the corresponding Fg sampling frequencies. Bit 22 indicates whether the frame contains an "empty section" or not. (If included, b₂₂= 'l', if not, then b₂₂= 'O'.) A b₁₆-b₂₂ the information displayed on the bits determines how many pieces of information are actually contained in the frame. This means that the first frame part FD1 still carries information about the number of frame information packets.

Since n_sthat is, the number of samples of the broadband signal for which the information related to the second digital signal is contained within a frame, in this case known as N_s = 384, so the data in the table in Figure 8, b₂₂ bits, and use the following formula to determine how many B-packets are in the frame:

P = (BR / N) {n_s/ F_s)

The b₂₃ bit is reserved for some future expansion of the system. This future extension is described below. At this moment, this bit is set to "0". The contents of the first frame part of FD1 are -b₂4-b₃₁ 19 and 20 with respect to bits. The b₂₄ and b bits refer to the distribution of the audio signal (mode designation). Figure 20 shows that ab₂₄, bits of two bits (4) indicate whether the digital broadband signal is stereo ("00"), mono ("11"), bilingual ("10") or intensity stereo audio signal ("01"). In the latter case, bits 26 and 27 indicate which subband was processed by the intensity stereo method. Figure 20 shows that for "00", "01", "10" and "11" two-bit numbers, respectively, see Figures 5-32. subbands, 9-32. subbands, 13-32. subbands and 1732; subbands were processed using intensity stereo. As mentioned earlier, the intensity stereo method can be used for higher (upper) subbands because the human ear is less phase sensitive to

EN 210 644 B for frequencies in these sub-bands. The b₂₈ bit can be used as a copyright bit. If the bit is set to "1", this means that the information may not be copied, ie it may not be copied or copied. ab₂9 bits may indicate that the information is original information (b₂₉ = "1"), such as for pre-recorded tapes, or the information is copied (b₂₉ = "0"). ab₃₀ and b₃₁bits indicate the highlight in the transmitter to the broadband signal (see also description in Figure 7).

Below, a different configuration of the second frame part of FD2 is updated for the various modes indicated by the bits 24- ^ 27 of the first frame part of FD1. Again, the second frame part of FD2 contains 4-bit allocation words, the meaning of which is described with reference to FIG. With stereo assignment (mode) (b₂₄, b₂₅= 00) and for the bilingual layout (b₂₄, b₂₅ = 10) The second frame part FD2 again consists of 8 information packets, which are assembled as described in FIG. In the stereo assignment of Figure 10, "I" denotes, for example, the left channel component, and "II" denotes, for example, the right channel component. For a bilingual layout, "I" for one language, "Π" for another language. In the case of mono allocation (b₂₄, b₂₅ = 11) the length of the second frame part of FD2 is of course only 4 information packets.

FIG. Specify the order of the allocation words in the (4) information packets for the different subbands (1-32). Thus, each M th quantity represents a 4-bit allocation word that represents the bit number of each sample of the subband i having a value between 1 and 32. Intensity in stereo mode (for allocation) (b₂₄, = 01) there are four options, a and b₂₇ bits indicate (see Figure 20). Each option results in different contents of the second frame part of FD2.

Figures 22a-22d illustrate 4 types of contents of the second frame part FD2. If ab ^, b₂₇ bit position "00" is in positions 1-4. the subband signals are (normal) stereo signals and the 532 subband signals are intensity stereo signals. This means that for subbands 1-4, the allocation words for the left and right link components for these subbands must be stored in the second frame portion of FD2. Figures 22a show successive AW (L, 1); AW (R, 1); AW (L, 2); AW (R, 2); ...; Allocation words AW (R, 4); they are included in the second stage (information packet) of the frame, i.e. the first stage (information packet) of the second frame part FD2.

Fig. 22a shows only the index of the allocation words (ij), L or R may have a "value" (L represents the left channel, R represents the right channel component), j represents a value between 1 and 4 and a subband number. For the subbands 5-32, the left and right channel components contain the same pattern sequence. The difference - in each sub-band - lies only in the scale factors for the left and right channel components. Consequently, each subband requires only one allocation word. The allocation words AW (i, j) for subbands 5-32 are indexed by Mj, i.e., i for each subband is "M"; and j varies between 5 and 32.

Figure 22a shows that an information packet 41/2 is required to place the allocation word 36 in the second frame portion of FD2. If ab ^, b₂₇ bits are set to "01", then bits 1-8. the sub-band signals are normal stereo signals, while those shown in FIGS. sub-bands are intensity stereo signals. This means that Figures 1-8. each subband requires two allocation words [AW (L, j)] and AW (R, j)); 9-32. each subband requires only one allocation word [AW (M, j)]. It follows that a total of 40 allocation words are required, which occupy five information packets (sections), i.e., packets of information IP2-IP6. This is illustrated in Figure 22b. In this case, the second frame part FD2 is an information packet length 5.

If ab₂₆, b₂₇ bits have a value of "10", then bits 1-16. subband signals are normal stereo signals, while 17-32. subband signals intensity stereo signals. In this case, 48 allocation words are needed, which are placed in the second frame part of FD2; the length of the second frame part FD2 is 6 information packets, see FIG. 22d.

Our earlier comment on scale factors applies here as well. Assuming no 0000 allocation words were present for the subbands or channels, then 64 scale factors would be required using both stereo and intensity stereo mapping (modes). This is because mono subbands used in intensity stereo mode require two scale factors to create the intensity stereo for the left and right channels of that band, cf. 15c.

Obviously, using mono mapping (mode), the number of scale factors is halved, that is, the number of scale factors is 32 (assuming, in this case, that no sub-band allocation word has a value of 0000).

The procedure for determining the 6-bit scale factors is described below.

As mentioned above, the highest absolute value of each of the 12 samples of the subband channel is determined.

Figure 24a illustrates these maxima S, ^ | shows an absolute sample. The first bit (denoted by SGN) is the sign bit; (it has a value of "0" because it refers to its absolute value). The patterns are represented in binary complement mode. Samples contain k "zeros" followed by "l". The other bits of the 24-bit number are unimportant (these bits can be either "0" or "1").

IS_up I -ot 2^k(see Figure 24b). Thereafter, Smax | 2^k is compared with the binary number'DVp 010100001100000000000000 and the binary number DV2 = 011001100000000000000000. If S, ^ | <2^k DV [. then the constant of p is 2. If DVi <| S ,, ^ [<2^k For DV2, p is set to 1. If | Smax | 2^kB> DV2 such that p = 0. For k, 0 <k <20. The scale factor is given by the following expressions k and p.

SF = 3k + p.

Consequently, the SF scale factor has a maximum value of 62. This means that the scale factor is a 6-bit number.

The mocha of HU 210 644 B can be described. (The 6-bit binary number 111111 - this is a decimal 63 - is not used.) In fact, the 6-bit binary numbers are not scale factors, but they have a clear relationship to the actual scale factors, as shown below. Each of the 12 S samples is multiplied by a number dependent on k and p. Each of the 12 S samples is multiplied by the following expression.

S '= S * 2^k* g (p), where g (p) is related to p: g (p) = 1, ah p = 0:

g (p) = 1 + 2^-2+2^-8+2^-1θ+2^-16+2^-18+ -23, hap = 1; g (p) = l + 2¹+2^Jt+2^+2-⁸+2-⁹+2-¹⁰+2-¹³+2-¹⁵+2-¹⁶+2-¹⁶+2^-17+2^-19+2^-2θ, if p = 2.

The parameter k gives the number of steps of 6 dB, the factors g (p) and g (p) being the most accurate approximations of the number of steps of 2 dB. The scaled S 'samples are quantized so that they can be represented as a q-bit, double complement number. Figure 25 shows this for q = 3. S 'scaled (normalized) samples have values ​​between +1 and -1, see FIG. 25a. In the quantizer, these patterns must be described with q bits, where q corresponds to the allocation value for that subband. Since, as mentioned earlier, the q-bit number containing all "l" is not used in the description of the samples, so the interval between -1 and +1 has to be subdivided into 2 ^ '. For this purpose, the scaled (normalized) S 'samples are transformed into S samples according to the following formula. S = S '(1 -2 ^) - 2^-%

The S samples are then truncated to q-bits. (See Figure 25c.) Since the shape "111" is not allowed, the sign bit is inverted, cf. 25d. The q (= 3) bit numbers of FIG. 25d are included in the third frame part of FD3, see FIG. 2. Samples of S 'satisfying the inequality of -0.71 <S' <- 0.14 are binary '001'. Similarly, it continues for larger S 'samples up to an inequality of 0.71 <S' <1, in which case S 'is represented by binary "110". The binary value "111" is thus unused.

On the receiver side, the digital / analog conversion is produced as the inverse of the transmit-side quantization, cf. 26. This means that the sign bit of the q-bit binary numbers is inverted first, thereby returning to the normal (normal) binary complement representation, see Figs. 26b.

The S 'samples are then restored from the transformed S samples according to the following formula.

S '= (S + 2 ^⁺¹) (1+2^+2^-¼ ...) as shown in Figures 26c, 26d.

The S 'values ​​thus obtained are (exactly) within the original intervals of Fig. 25a. On the receiving side, the samples S 'are then restored to their original amplitude by the transmitted values ​​of k and p (k and p are two values ​​related to scale factors). Thus, the number g '(p) on the receiver side satisfies the following: g' (p) = 1 if p = 0 g '(p) = 2- ^ 2-2 + 2-5 + 2^-6 ha p = 1 g '(p) = 2- ^ 2-5 + 2 ^ + 2^-9 if p = 2.

To return to the original amplitude, use the following formula:

S = S '* 2^s* G '(p).

In the two possible variants of the frame shown in Figures 2 and 3 and Figures 2, 17 and 19, the third frame part of FD3 cannot be (completely) filled with information. The more often and the sooner this situation occurs, the more the subband coding algorithm is improved, i.e. the process of decomposing the signal into subband signals and quantizing the samples sequentially in each band. This patch mainly consists in transmitting information with a smaller number of bits per smaller sample! average bit rate. The unused portion of the third frame of FD3 can be used to transmit additional information. Reference is made to this in the first frame part FD1 of FIG₂₃ For future use 'bit. Generally (normally), this bit is set to 0, as illustrated in FIG. If an additional signal is included in the third frame part FD3 of the frame, then the first frame part FD1 is B₂₃ Reserved for future use ', see p. aa

17th will have a value of "l".

This allows the receiver to detect that the frame contains additional information while reading the first frame portion of FD1. For allocation information and scale factors, see p. 23, informing the receiver that only a portion of the third frame portion of FD3, the portion designated FD4 in FIG. 23, has quantized patterns of subband signals. The other part of FD3, indicated by FD5 in Figure 23, contains additional information. The first bits of this FD5 frame part are called "EXT INFO" or expansion information. These bits specify the type of additional information. The additional information may be, for example, for transmitting an additional audio channel, such as a second stereo channel. Another possibility to use these two additional audio channels is to realize the ambient audio together with the audio subband signals in the FD4 frame. In this case, the front-to-back information required for ambient audio may also be included in the FD5 frame. In the section labeled FD6, the FD5 frame portion may again contain allocation information, scale factors, and samples (in that order), and the order of allocation words and scale factors may be similar

Second 2, 17, and 19 respectively.

In the case of "ambient audio", the simple receivers decode only the stereo information in the second frame part FD2 and the third frame part FD3, except for the FD frame part. More sophisticated receivers are able to recover ambient audio information using the information contained in the FD5 framework.

The expansion information bits may indicate that the information in the FD6 frame part refers to text (for example, ASC1 characters). Placing video or image information in the FD6 frame part may also be considered, as well as the expansion information bits.

It should be noted that the invention is not limited to the embodiments herein. The invention also relates to the embodiments known herein14

They differ from the embodiments of the invention only with respect to the invention as defined in the claims, which are inert.