Method and apparatus for compressing and transmitting high speed data

Description: TECHNICAL FIELD [Detailed description of the invention]

Methods and equipment for compressing and transmitting high-speed data Technical field of invention The present invention relates to communication systems and, more specifically, transmits high speed data communication signals. It relates to a signal processing technique for compressing to improve performance and increase communication capacity. Background of the invention Communication systems are well known to those skilled in the art, and current telephone systems are used by a large number of users. A species that transmits telephone signals over a single transmission line, such as an electric wire or fiber optic cable. It uses various multiplexing techniques. Most of these "wired" systems are time division multiplexing (TDM) format, that is, a large number of channels faster than the channel information speed It uses a format in which it is sent sequentially by degree. Normal telephone multiplexing is the sampling of telephone signals and the output of that sampling. It requires transmission to be at a frequency much higher than the frequency of the telephone signal. That To this end, the system samples and encodes telephone signals and multiplexes the signals. It is converted and transmitted, and after receiving and demultiplexing, it is decrypted. That kind of The sampling and coding system is pulse code modulation (PCM) and analog Sampling of audio band signals 8 km / s with 8-bit display for each sample Do at the speed of. Therefore, the audio band signal is 64 kilobits per second (kb / s). Converted to a digital signal. Another form of communication system is a radiotelephone system. Radiotelephone system Is a group of preselected radio frequencies (RF) for transporting telephone communication signals between two or more points. ) Is used, and the frequency division multiple access (FDMA) format is usually used. Wireless communication system These wireless systems, called mu, are, for example, regional telephone services in rural areas. Used for mobile phone services in mobile units or portable units. In one category of RF communication systems, the multidimensional information formed by the modulation of RF carriers Use TDM to allow the user access to the reporting time slot. this If a large number of users are competing for a small group of information time slots, then this Shi The stem is called Time Division Multiple Access (TDMA). Allows TDMA for FDMA RF communication channels US Pat. No. 4,675,863 (listed here) called FDMA / TDMA. The method described (incorporated in the specification of) is adopted to increase the communication capacity of the RF communication system. Has been used. However, RF communication systems are wired or fiber optic cable systems. Capacity is constrained by frequency compared to Therefore, in order to further increase the communication capacity, the telephone signal via the RF channel Signal compression technology has been used to reduce the bandwidth required for transmission. Audio signal The techniques commonly used for this are subband coding, adaptive differential pulse code modulation (ADPCM). And residual linear predictive coding (RELP). RELP or similar audio compression al Gorism RFs a sampled quantized audio signal at 64 kilobits per second (kb / s) Transmit as a slow bit rate (eg 14.6 kb / s or less) signal over the channel Make it possible. The receiver converts this 64 kb / s audio signal from the bit rate deceleration signal. Reconstructed, the receiver will notice little or no degradation in signal quality. Basic audio compression methods such as RELP are encoded using known features of audio signals. And the decoding algorithm. One form of RELP method is a kind of human voice It is premised on the harmonic characteristics of. However, today, the communication signal in the telephone network is large. The part is a non-voice data communication signal such as fax (FAX) or voice band modem data. Is. Unfortunately, these data communication signals do not have the characteristics of voice signals. Audio compression algorithms are not compatible with these data communication signals. Therefore, RF communication that monitors telephone signals to detect incoming data communication signals. There is also a credit system. Usually 2.4 kb / s or less (slow data) fax or voice band model The data signal representing the data signal is detected and applied to a special compression algorithm. Vedanā The credit machine reconstructs the data signal without slowing down the transmitted data rate. That kind of sis The system and method are described, for example, in U.S. Pat. No. 4,974,099 (listed herein in this specification). It is described in (see). However, the current telephone data signal is 9.6 kb / s (high-speed de). Data) or higher (14.4 kb / s, 28.8 kb / s or higher or lower) Is becoming more common, and the above techniques are sufficient for compressing these high-speed data. It shouldn't be added. Compression of these high speed data, especially of these high speed data Multiple coding results in quality degradation of the modem or fax signal, resulting in RF communication of the signal. It was necessary to reduce the transmission speed at the modem or fax stage when passing through the stem. Outline of the invention This communication system is a set of data signals, each with one form of coding. Receive telephone signals and transfer those telephone signals to at least one radio frequency (RF) carrier Send with. Each of these RF carriers has a group of information slots, the telephone signal. Each of the is assigned to at least one information slot to modulate its RF carrier Can be done. This system monitors and identifies data signals, as well as data signals. Includes the process of compressing each data signal to reduce the required transmission bandwidth. A brief description of the drawing The present invention will be best understood from the following detailed description associated with the accompanying drawings. FIG. 1 is a block diagram of a wireless communication system. Figure 2 is a high-level block diagram of the practical application of the compression system of the present invention, and the dynamic bandwidth. Includes width allocation feature terms, and high-speed ultra-high-speed data codecs. FIG. 3A shows high-speed data-coded data detection according to an exemplary embodiment of the present invention. And selection, as well as top-level illustrations of radio channel slot calculations and allocations It is a flowchart. FIG. 3B responds to a high speed data channel request according to one embodiment of the present invention. Channel formation A high-level processor that shows the process of channel allocation performed by the processor. It is a low chart. FIG. 4A is a graph showing the characteristics of the A-law quantization device. Figure 4B shows the contrast between the signal-to-quantization noise characteristics of PCM and that of uniform quantization. F. Figure 4C shows the mapping of signal samples from one quantization to another. Illustrate the method of compression by. FIG. 5A shows a high-level block of a high-speed data coding device according to an exemplary embodiment of the present invention. It is a figure. FIG. 5B shows the transmission coding of a high speed data encoding device according to an exemplary embodiment of the present invention. Illustrate the process. FIG. 6A shows a higher block of a high speed data decoding device according to an exemplary embodiment of the present invention. It is a figure. FIG. 6B shows a transmission decoding process of a high speed data decoding apparatus according to an exemplary embodiment of the present invention. Illustrate Seth. FIG. 7A shows an upper block of an ultra-high speed data coding device according to an exemplary embodiment of the present invention. It is a figure. FIG. 7B shows the line code of the ultra-high speed data coding device according to the exemplary embodiment of the present invention. Illustrate the conversion process. FIG. 8A shows the upper block of the ultra-high speed data decoding device according to the exemplary embodiment of the present invention. It is a figure. FIG. 8B shows the transmission decoding of the ultra-high speed data decoding device according to the exemplary embodiment of the present invention. Illustrate the roses. FIG. 9 shows a compression quantization sump of a PCM quantization sample according to an exemplary embodiment of the present invention. High-level float that illustrates the ultra-high-speed quantization algorithm used for mapping to It's a jar. Overview Communication devices and methods receive telephone signals and no corresponding transmission in each of those telephone signals Modulates a radio frequency (RF) carrier. Each of the transmitting RF carriers has a predetermined number of information slots And assign each of the telephone signals to at least one of those information slots Modulate the RF carrier with the telephone signal. This communication device and method is the above telephone signal Receive each of those phone signals to detect the data signals contained in one of the A detector that squeezes and a coding that encodes the data signal into a compressed coded signal. Includes equipment. This device and method allocate each information slot when detecting a data signal. While checking the guessing situation, the required band for transmission of the compressed coded signal Position a predetermined number of unallocated sequential information slots for the region width. This quota The status shows the assigned / unallocated status of each information slot to other telephone signals. Also, This device and method is a communication channel from an unassigned sequential information slot that has been located. The process of forming the flannel and the modulation of the communication channel signal with the coded signal Including with Roses. According to one aspect of the invention, high speed data compression transmission systems are high speed data. The signal is transmitted as a compressed coded signal via the communication channel. This fast The degree data signal is received as at least one data signal sample block, This system includes a high-speed coding device and a high-speed data decoding device. This high speed The data coding device is (1) at least one data signal sun representing the amplitude peak value. The receiver of the data signal block, each containing a pull, and (2) the amplitude peak value. A calculation method for calculating the data signal block gain value proportional to (3) and (3) the gain value Includes a quantizer selector that selects the corresponding quantizer. The quantization device is a plurality of quantities at predetermined (for example, uniform) intervals determined by the gain value. Each data in the above data signal block has a quantization level value and is selected by the quantization device. Quantize the sample into a compressed data sample. This gain value and multiple compressions The data sample forms a compressed coded signal. This high speed data compression transmission The system is a transmitter that transmits a compressed coded signal over a communication channel. Includes a receiver that receives the signal via its communication channel. The high-speed data decoding device of this high-speed data compression transmission system is (1) compressed. Data sample and corresponding gain value receiver and (2) its based on its gain value Select a uniform inverse quantizer with multiple uniformly spaced output values determined by the gain value. Includes a quantity predictor selection device. This inverse quantization device is compressed based on the above gain value. Each of the data samples is processed to produce a block of reconstructed data signal samples. According to another aspect of the invention, ultra-high speed data compression transmission systems are ultra-high. The speed data signal is transmitted via the communication channel. This ultra-high speed data signal is Received as at least one data signal sample block containing one quantization , This system is equipped with an ultra-high speed data encoder and an ultra-high speed data decoder To. This ultra-high speed data coding device has (1) at least one having an amplitude peak value. A receiver of a data signal block containing two data signal samples and (2) this amplitude A calculation method for calculating the data signal block gain value proportional to the peak value, and (3) its A new set of quantization levels corresponding to the gain value of the sample block, that is, each Select a new set of quantization levels, which is the selected quantization level of the first quantization Quantizer selection device to be used, and (4) the quantization level set of the first quantization and a new quantity. Signal sample for each signal sample value based on the relationship with the child level set Includes a quantization level mapping processor that maps values to compressed level values. The gain value and the compressed data sample form a coded signal. This system Is a transmitter that transmits this coded signal via a communication channel and its coded signal. Is provided with a receiving device for receiving from a communication channel. Communication system of wireless communication system An embodiment will be described below with reference to the channel. However, this invention is wireless or it It is not limited to RF carrier communication other than. This invention is a wired communication system communication channel It can also be applied to increase communication capacity. The ultra-high-speed data decoding device of this ultra-high-speed data compression transmission system is (1) compressed. A receiver of a data sample and its corresponding gain value, and (2) its corresponding advantage Based on the obtained value, the output value determined by the gain value and the quantity having a new quantization level Includes an inverse quantizer selector that selects the child device. This inverse quantization device has the above gain value Based on the process and reconstruction of each of the compressed data samples Produces a sample block. According to another aspect of the invention, the ultrafast data quantization method is the first plurality. Quantization signal sample, i.e. each corresponding one corresponding quantization amplitude value and quantization amplitude A first plurality of quantized signal sensors having at least one signal sample of peak value. Map the sample to a second quantized compressed sample and gain value. this The methods are as follows: (1) Check each amplitude value to calculate the amplitude peak value and correspond to the amplitude peak value. And (2) for each of the first plurality of quantization signal samples described above. A predetermined number of successive segments, each with a large number of quantization level values Includes setting a predetermined number of segments with. These segs that follow each other The quantization level value for each of the ment is related to the above gain value, and they are predetermined. The first segment of a number of successive segments is the amplitude peak of the above-mentioned plurality of signal samples. Corresponds to the value. This quantization method maps the quantized signal samples to each quantized compressed sample. Ping, (1) Quantization level of each segment for each of those quantized signal values Retaining the chosen value number until extraction at the level of value 0, and (2) negative value Negative values are signed to indicate the amplitude of. Detailed description of the invention Data compression system Figure 1 shows a wireless communication system that can put into practice the features of high-speed data compression of the present invention. It is a schematic diagram. As shown, this communication system includes base station 11 and a group of subscriber stations. Including 10 and. Base station 11 is a communication partitioned within a range of radio frequencies selected in advance. Simultaneously and in parallel with the above subscriber station 10 by simultaneous transmission and reception via the channel Believe. Base station 11 interacts with the local telephone system equipment of the telephone office 12 of the telephone company. Also forms. Ordinary wireless communication equipment (for example, King of Pusher, PA) The location of the SLS-104 type device manufactured by Interdigital Communications Co., Ltd.) is in order. 24 channels (from base station to subscriber station) and 24 channels in the opposite direction 300-500 MHz spectrum range (from subscriber stations to base stations) Prepare inside. Signal transmission and reception from the base station to the subscriber station is a frequency within its spectrum area. It is done through a pair of communication channels (forward and reverse) that modulate the number. Normal In this system, base station 11 signals in parallel through the above 24 channel pairs. Give and receive. These 24 channels occupy a frequency band of, for example, 2 MHz. This 2M Supports 80 channels in the Hz frequency band, for example by adopting a channel interval of 25 KHz. Can be ported. In one embodiment of this system, base station 11 is on the above channel. The lower frequency of the pair is transmitted to the subscriber station 10, and the subscriber station 10 is the upper frequency of the base station 1. Can be sent to 1. Such a system was published on June 23, 1987. National Patent No. 4,675,863, "Singlex or multiplex audio and / or data signal communication Described in "Subscriber RF Telephone System for Simultaneous Use by Several RF Channels" , The patent is listed here and its contents are incorporated herein by reference. Time division multiple access technology is used for each carrier frequency to increase communication capacity. One In a good example system, each frequency of the above channel pair is divided into four time slots. In addition, the base station 11 is simultaneously connected to the subscriber stations 11 up to 4 stations via one carrier wave. Allows signals to be sent and received. That is, this base station uses 24 channel pairs. Enables modulation by telephone signals up to 95 channels and controls 1 channel Used for other incidental functions. One aspect of increasing communication capacity with the above method is the RF communication channel (or wired). It is to compress the communication channel to be transmitted via (channel). For voice As mentioned above, voice coding technology such as RELP can be used. Also low Rin and other names for speed data compression technology and low-speed facsimile data compression technology Used as described in National Patent No. 4,974,099 "Communication signal compression system and method" It is possible, the patent is listed here and its contents are incorporated herein by reference. In the system described above, there are three voice band coding devices: RELP, low speed data. Data and low-speed fax data encoders transmit 64 kb / s PCM signals to 14.5 kb / s Compress to issue. These three encoders have a single 116 RF slot at 14.5 kb / s Or it can operate within the period of a double-width 4-phase RF slot. Sound RELP encoder Uses for voice and uses a low-speed data encoder to transmit multiple voice bands at speeds of 2400 BPS or less. The low-speed fax coding device is used for the transmission of Group III fax transmission of 2400 BPS. Each of the transmission coding devices has a corresponding decoding device in the receiving device, and the decoding device is installed. For example, it can be assigned via the system control channel. Allowing communication systems to accommodate high-speed voice band modems and high-speed fax machines To achieve this, two interrelated voice band compression techniques of the present invention are used. Up The above-mentioned coding device and decoding device (codec), that is, the high speed codec described herein. These devices, labeled as decks and ultra-high speed codecs, are data signal data. Low-speed data encoding by lightening compression and giving more bandwidth Achieve better compressed data transmission performance than stationary and low speed fax coders .. The above high-speed codec is a communication system that transmits voice band modems and faxes. Allow transmission to be transmitted even when the speed drops to 9.6 kb / s. Super high speed coordination Supports voice band modem transmission and fax transmission even if the speed increases to 14.4 kb / s To This high speed codec has three 16-phase RF slots or four 8-phase RF slots Works in lots. The above ultra-high speed codec uses four 16-phase RF slots Works. High-speed data compression algorithm and ultra-high-speed data compression algorithm Zum minimizes harmful distortion via digital channels with data speed constraints It is desirable to suppress the signal to pass the analog audio band waveform signal. These codecs use several RF slots, so RF communication channels Requires dynamic reassignment of slots within. Dynamic time slot / bandwidth of the present invention The width allocation feature term detects and monitors data transmission and data in the required number of slots. Form a channel, but low-speed data if the required number of slots are available Assign a coding device or a low-speed fax coding device to the call. This kind of allocation method Is, for example, issued November 15, 1988, US Pat. No. 4, in the name of DR Borziano, No. 785,450 "A device that achieves frequency switching agility in digital communication systems And Methods , and the patent is listed here and its contents are described in this specification. Incorporate. Figure 2 shows a dynamic time slot for high-speed data compression, which is a good example of a wireless communication system. / Top block diagram of the implementation of the compression system of the present invention including the bandwidth allocation feature term Is. This system includes compression selection including control unit 201 and monitoring unit 202. Selector processor (CSP) 200, channel forming processor 260, and compression Coder / Decoder (CODEC), ie RELP codec 210, low speed data Tacodec 220, low speed fax codec 230, high speed data codec 2 Includes 40 and 250 ultra-fast data codecs. The CSP200 receives a telephone signal from the private telephone exchange 260 and monitors the telephone signal. Therefore, various types of data signals can be identified from the modem response signals of those signals. It is designed to activate the communication channel settings. For communication between subscribers In another embodiment, the CSP200 delivers the telephone signal from a local source other than the above. receive. The monitoring unit 202 of the CSP200 notifies the control unit 201 of the incoming data signal. Let me. The control unit 201 is formed outside the RF communication channel, and the core Assign compression formats to decks 210, 220, 230, 240 and 250 Has an action. The channel forming processor 260 receives a transmission channel request from the CSP200 and receives an electric signal. Allocate available RF communication slots for spoken signals. This channel formation processor The service 260 keeps the current system channel allocation information in memory (not shown). Determine a time slot that you have and is not currently occupied by another phone signal. TDMA cis As is known in the system, each channel time slot has a guard time That is, a short time is provided to activate the receiver before data transmission. Channel formation for incoming data signals that require more than one RF time slot The Rossessa 260 forms channels in a predetermined number of time slots, and those time slots If the guards are adjacent to each other, set only one guard time. The channel forming processor 260 in one embodiment of the present invention is a network. It could be configured with a wireless processor unit (RPU) of a base station. RPU is the cis in Figure 1 Memory of channel time slot allocations for the entire system and channel time slots Responsible for the allocation of. The RELP codec 210 is a compression coding (and decoding) algorithm for audio signals. Put the rhythm into practice. Low speed data codec 220, low speed fax codec 23 0, High-speed data codec 240 and ultra-high-speed data codec 250 have been identified Practical implementation of each data compression algorithm for audio band data in different formats. CSP200 and these codecs 210, 220, 230, 240 and 250 Performs data signal monitoring operation, signal processing operation, and signal compression coding / decoding operation. As such, it is generally integrated into a single digital signal processor. This species The integrated signal processor of is, for example, TMS 320C5 manufactured by Texas Instruments. Choose from X-type digital signal processors. The operation of the compression system of the present invention will be described here. Further referring to Figure 2, When a voice call is first set up, the voice RELP codec 210 initially responds to the phone signal. Assigned. The CSP200 monitors the telephone signal through the monitoring unit 202 and controls it. Knit 201 determines the type of voice band signal based on the detection of the modem response signal. Each type of voice band data has a specific identifiable modem response signal. Table 1 is subordinate It is a compilation of some of the various typical modem outgoing response characteristics well known in the next technology. To. Table 1 is for illustration purposes and describes all possible modem characteristics. It's not a thing.<img file="JP2000504524A_D0001.tif" /> Returning to Fig. 2, the type of voice band data is determined and high-speed data compression or ultra-high When fast data compression is needed, the CSP200 activates voice channel reassignment, which The dynamic time slot allocation used in this case is described below. Control unit 201 An RF communication channel that sends a signal to the channel formation processor 260 and has a predetermined number of time slots. Form a tunnel. In one embodiment of the present invention, one time slot is called. Dynamically assigned, but this is not a requirement. Channel forming processor 260 Examines the memory mentioned above to determine the number of available RF time slots and the RF carrier position. Determine. When the channel forming processor finds the number of the predetermined slot, the RF communication channel The channel consists of a predetermined number of RF time slots, and the status is notified to the control unit. Be done. The control unit 201 then has a compatible high speed data codec or ultra high speed Assign the data codec to the data signal and assign the compressed data signal to the above form. Assign to the created multi-slot RF communication channel and modulate that channel To. If there are not enough time slots available, control unit 201 will be charged. Informed that, the RF communication channel is formed in a single RF time slot, and the control unit To 201, use a low-speed data codec or a low-speed fax codec to send the data. Assign to issue. As described above, in one embodiment of the present invention, a multi-slot communication device Automatically divides one time slot when a phone signal is received before channel formation Guess and make sure the phone signal is already assigned one slot at that point To. Dynamic time slot / bandwidth allocation Figure 2 summarizes the time slot requirements for signal compression types.<img file="JP2000504524A_D0002.tif" /> The high-speed coding device described above has a 3-slot 16-phase channel and a 4-slot 8-phase. Since both channels are modulated with data, the compressed data of this device is these It is desirable to fit in the narrower bandwidth of the two channels. In connection with Figure 1 Bits about the channel types of the examples in the wireless communication system described above. Table 3 shows the availability.<img file="JP2000504524A_D0003.tif" /> In Table 3, "zero value" indicates the unmodulated state, and "preamble" is the same as the bit. Indicates a period pattern, where "CW" is a codeword, namely call control, call processing and system. Represents a codeword that contains gunning information. "A block" and "B block" Is the block for the first and second 22.5 ms of the compressed audio band data sample. Represents As can be seen in Figure 3, the 4-slot 8-phase channel is a 3-slot 16-phase channel. The number of transport bits is smaller than that of flannel. Therefore, in one embodiment of the present invention The compressed output block of a high-speed data encoder is 1041 bits or less. .. Table 4A shows the bit allocation of the compressed output block of the high-speed data encoder. ..<img file="JP2000504524A_D0004.tif" /> In Table 4A, "protected" means forward error correction (FEC) in the bitstream. ) Is applied. 4 in the bitstream of an ultra-high speed data encoder Modulate slot 16-phase channels, thereby 14 within a period of 22.5 ms each 08 bits will be available for coding device data. Table 4B shows the bit allocation of the compressed output block of the ultra-high speed data encoder. Shown.<img file="JP2000504524A_D0005.tif" /> High-speed data compression technology and ultra-high-speed data compression technology, which will be described later, will be applied to communication channels. This is an embodiment of the present invention that requires a large number of time slots, but is the same as that described herein. Other types of compression methods based on the same concept are other types of data signals, that is, the above-mentioned sounds. It can be developed for types of data signals that do not necessarily have voice band modem characteristics. So The dynamic time slot / bandwidth allocation method used in the above examples in these other examples as well. Can be used. A general dynamic time slot / bandwidth allocation method is described below. Figure 3A is, for example, Illustrate the dynamic time slot / bandwidth allocation process in action on 2 CSP200 .. Referring to FIG. 3A, the voice monitoring step 301 is data as the voice call is set. Monitor the phone for signal detection. In step 301, first RELP -Assign the deck 210 to a telephone signal. However, if there is a data signal, the judgment Tep 302 determines the type of voice band signal based on the detection of the modem response signal .. If the data is low speed data or low speed fax data, step 303 Is a low speed allocation process, eg single RF carrier slot allocation Assign process. Step 304 is if the data signal is fax data or low Determine if it is speed data, low speed fax codec 230 or low speed data Assign algorithm steps 305 and 306 for codec 220. If the signal is of the high velocity data type in step 302, then In the next step 307 of the high speed data from the channel forming processor 260 Request a channel. In one embodiment of the invention, this channel forming processor Su 260 requires user / subscriber conditioning information for channel type requests To do. In another embodiment of the invention, for the requirement of the correct channel type. Data signal requires either high-speed data compression method or ultra-high-speed data compression method It is determined from the modem signal whether or not to do so. Figure 3B responds to the high speed data channel request from step 307 of Figure 3A. The process of channel allocation performed by the channel forming processor 260 is shown. The channel forming processor 260 is the base station radio of the prior art system described above. It can be configured with a processing unit (RPU), which is connected to the subscriber station via the communication channel. RF carrier time slots can be assigned to communications. Starting from step 320 in Figure 3B, this processor normally sounds on a phone call. Assign a voice channel, but any that is also described in U.S. Pat. No. 4,675,863. You can choose process assignment. Next, step 321 is step 307 of FIG. 3A. Search for high-speed data channel requests from. If there is no such request, Chan The flannel allocation is the default mode, ie the audio mode in this example. Stay in ortho mode. If so, step 322 is high for subscriber stations Subscriber conditioning to determine if the speed data channel is ready for acceptance To explore. Subscriber stations are not conditioned to accept high speed data channels If not, use the required number of slots in step 323 for low-speed data / fax. Allocate a tunnel. If the subscriber station is ready to accept high-speed data channels, step The 324 is a high-speed data channel (UHSD Chan) whose subscriber station is an ultra-high-speed type. Determine if you are ready (or have been requested) to accept flannel) To. If ready, step 325 is the availability of a given number of RF carrier slots Check and step 326 forms a UHSD channel if available. Step 325 is the required number of 16-phase RF time slots (4 in this example). Remembers the current status of system channel allocation to determine if is available It can be performed by a processor that checks the memory. Slot requirements available If not, a high-speed data type channel as described later in step 328. This process checks if the channel can be formed as (HSD channel) To. In step 324, the subscriber conditioning (request) of the high-speed data channel If it is instructed that it cannot be formed as an ultra-high speed type (UHSD) channel, it will be de-high speed. It is required that the data channel should be formed as a high-speed type HSD channel. Alternatively, step 327 checks whether the subscriber conditioning dictates. finger If not indicated, a low velocity data channel is formed in step 323 as described above. However, if the HSD channel is required or conditioned Step 328 uses a given number of RF carrier time slots for HSD channels Check if it is possible. Step 328 is to check the memory that remembers the current state of channel allocation. The first required number of time slots (16-phase RF time slots) (3 in this example) is Detection of availability and time slot (8 phases) if not available Whether the second requirement (4 in this example) of the RF time slot) is available It can be embodied by a processor that enables detection of. Use the required number of slots If possible, assign those slots in step 329 to form the HSD channel To be done. The above high-speed channel availability check step is the required number of cha If no channel is found, step 323 assigns a low speed channel I can only. Returning to Figure 3A, responding to high-speed data channel requests in step 308. This process checks the answer. The request is rejected in this step 308 Assign a low-speed algorithm if a high-speed data channel is not formed Step 303 and the sequence are performed so as to be performed. High speed data channel If the request is accepted, divide by high speed channel availability step 309 Determine the type of channel you have assigned. High speed data channel is super high speed Coding algorithm of ultra-high speed data codec 250 when dealing with data In step 310, the high-speed data channel corresponds to the high-speed data. If you have a high speed data codec 240 coding algorithm in step 311 Execute. High speed data codec and ultra high speed data codec High speed codec 240 and ultra high speed codec 250 were sampled Input and output telephone signals (Pulse Code Modulation (PCM) telephone signals in the examples) The bidirectional data channel of the present invention as a signal is compressed. Sample compression pro The incoming call signal to Seth is usually a 64 kb / s A-law or μ-law PCM signal, but 1 28kb / s 16-bit integer sample or other types of sample for conversion process Can be used. 64kb / s (or 128kb / s) sample in this compression process Compress the bitstream to a lower data rate. RF cha with this low speed data Send to the decompression process via the tunnel and decompress the above low velocity data in that process Reconstruct into a 64 kb / s (or 128 kb / s) sample bitstream. Coding device The purpose of is the original sampled signal from the synthesized or reconstructed sample. Is to be approximated. In the PCM system, the analog audio band signal is sampled at a sampling speed of 8 km / Convert to digital sample sequence in seconds. These samples are 8-bit capable , Quantization level 256. The index of goodness when sampling an analog signal is Signal-to-quantization noise ratio (SQNR). For uniformly spaced quantizers, SQNR Expressed as 6B-1.24dB, where B is the number of bits per quantized sample. .. Therefore, the SQNR of the 8-bit uniform interval quantizer is 46.76 dB, which is an audio signal. Is excellent against. In SQNR, the original analog signal is the dynamic lens of the quantization device. Obtained only if it has an amplitude over the entire di. The dynamic range of the original signal Clipping occurs when the dynamic range of the quantization device is exceeded. this Is highly undesirable signal distortion for both voice signals and voice band modem signals Only. If the dynamic range of the original signal is smaller than that of the quantizer , Less than the achievable maximum SQNR value of 46.76 dB. Dynamic range of the signal Is below the dynamic range of the quantizer. SQNR decreases by 1 dB for every 1 dB. Become. Since the dynamic range of voice band signals used in telephones is large, uniform intervals Quantizers may not be the best choice. Therefore, non-uniformly spaced quanta It adopts a chemical device. There are two standards for non-uniform interval quantizer for PCM. There are μ law and A law, and these standards are well known to those skilled in the art, but Simon It is described in Chapter 8 of "Communications Systems" by Hakin, and the book is listed here. Incorporate into this specification. Both of these laws are the dynamic lens of the quantization device. Quantization levels at logarithmic intervals are used for the extension of the di. Figure 4A shows A law quantization The characteristics of the device are shown. The spacing between quantization levels at high signal levels is the spacing at low signal levels Greater than As a result, the SQNR becomes more uniform between the samples. These quanta The highest SQNR value for the computer is smaller than that for the 8-bit uniform quantizer, The SQNR of these quantizers increases for a wider range of signal levels. Figure 4B compares the SQNR value vs. signal level characteristics for the A-law 8-bit uniform quantizer. I'm comparing. Uniform quantizers perform better at high signal levels, but A The law quantizer maintains high SQNR values over a wide dynamic range. A voice band modem using the above μ law or A law 64 kb / s PCM is a telephone signal die. It works well in the telephone network due to its wide dynamic range. Of these modems The transmission output level is set high to maximize channel usage, but the telephone is Channels are associated with various signal level drops. As a result, the modem output level is increased. Even if you have it, its level can drop significantly elsewhere in the telephone network. PC The dynamic range of M is the compensation in such a case. Data rate 64kb / s No PCM per sample when compressed to lower data rate The number of bits is reduced, and the SQNR is usually significantly reduced. Distortion due to compression is quantized By dynamically designing the device to fit the dynamic range of the input signal It is minimized in this invention. When the two dynamic ranges match, a new ward Quantize the sample values using a screened level interval quantizer. Figure 4C maps signal sample values from one quantization to another. Illustrate a simple example of the compression method. The block of signal sample 410 It consists of three samples 411, 413 and 415. First set of quantization levels 420 displays approximations for sample amplitudes 412, 414 and 416. However , These quantization levels are shown for a given number of information bits, i.e. the first quantization. For each of the 20 levels, 5 bits represent one of its first quantization levels. It is a requirement to transmit to the receiver. Three samples 411, 413 and 15 bits is desirable to send the three sample values corresponding to 415. A good example of the invention is the amplitude peak value for each of the signal sample value blocks. Draw a new level of tuples based on. As shown in Figure 4C, the sample Lock 410 contains sample 413 with an amplitude peak value of 414. This method peaks Define a new quantization level set by defining the value 414 as the highest level value , Determine a predetermined number of level values below this amplitude. As shown in Figure 4C, there are five Corresponds to the level value of. The bit required to define the level value for this new quantization The number of bits is only 3 bits, but the amplitude peak value is also the new quantization level value and the original amount. It must be sent as a ratio coefficient indicating the relationship with the child conversion level value. Therefore , 5 bits and 9 bits corresponding to the original amplitude peak value (3 bits per sample) Will be sent for sample block 410, which requires 14 bits. this The example shows transmission one bit less, but originally if there are 10 samples in the block Where 50 bits are required for the quantization method of, the transmission bits required for the new quantization device Is only 35 bits. Examples designed for the μ-law and A-law standards are described below. But here The method described in is an arbitrary system that receives a sample quantized by a non-uniform compression quantization device. Applicable to stems. High speed data codec FIG. 5A is a high-level block diagram of the high-speed data encoding device. Reference numeral of this embodiment The controller is 64 kb / s PCM and 46.58 kb / s forward error correction (FEC) coded compressed data storage. Perform data conversion with the system. The data rate of compressed data is 40.267 kb / s Therefore, the remaining transmission bitstream is used for error correction. As shown in FIG. 5A, the high-speed data coding apparatus of the present invention can be adopted or rejected. The uffer 510, the PCM decompressor 520, the gain calculator 522, and the delay means 521. , Data sample quantizer 523 and transmission coding means 530 with selectable acceptance / rejection including. Transmission coding means 530 provides FEC coding device 532 and insertion means 531. Including. Acceptable / rejectable buffer 510 sample for high-speed data compression process Hold a predetermined number of samples to form a block of. Block sample It can also be received in the format of. PCM stretcher 520 is A law or μ method Rule Converts a PCM sample to a linear sample. Gain calculator 522 is a sample block The quantized gain value for the data is calculated, and the data sample quantization device 523 is a quantum. To form a uniform interval quantization device in which the quantization level value is graduated with the quantization gain value. This quantized gain value is used for. The delay is the compression process where this quantized gain value is It is shown that it is determined before the formation of the quantized sample, which is coded by the transmission coding means. 530 is an error in preparing for the transmission of coded quantization gain and coded quantization sample. It is used to perform correction coding. The operation of the high-speed data compression coding device will be described below. 64kb as shown in Figure 5A / s Receives a PCM sample (A law or μ law) in buffer 510. Buffer 5 10 produces a PCM sample as a 22.5 ms sample block. 8 km of PCM At a speed of sample / sec, each block contains 180 samples. Received PCM frame Is sent to the PCM stretcher 520, and 16 μ-law or A-law samples are sampled with this stretcher. Convert to a bit straight line sample (16-bit integer sample). The block of the linear sample thus obtained, that is, 16 bi in the examples. Sends a block of integer samples to gain calculation processing means 522, which means amplitude Detect a sample of the block with the maximum value (absolute value). The amplitude of the sample is its blow Determine the quantized gain value of the clock. This quantized gain value is the amplitude value, that is, It can consist of the difference between the maximum sample value and the maximum block value, or a multiplication value. Quantization The accumulated gain value is quantized using a 64-level logarithmic interval quantizer. Gain calculation process The means 522 produces both a quantized gain value and a coded quantized gain value. .. The coded-coded quantization gain value is 6 bits, and is used in a logarithmic interval gain quantization device. Represents one of the 64 levels in. Quantized gain value from gain calculator 522 and support from PCM decompressor 520 A block of samples is supplied to the data sample quantizer 523. Delay means 52 1 is, the gain calculation processing means 522 is a sump by the data sample quantizer 523. That the task for that block must be completed before it can be compressed Shown for display. Data sample quantizer 523 is 32 levels uniform A remote quantization device is used to quantize 180 samples in a block. Quantization level It is dynamically adjusted block by block using the quantization gain value. Therefore, between uniform The dequantization level is from + quantization gain value for the above 180 sample set-quantum It is in the range of the gain value. The sample quantizer is a 5-bit code of the above 180 samples. Output only the conversion display. That is, compression does not require an actual quantization value. .. Inserting Coded Quantization Gain and Coded Quantization Samples 531 and It can also be sent to a transmission code process means 530 consisting of a FEC coder. FEC The coding device 532 is a (64,57) extended Hamming coding device, which is a Hamming code. 1-bit error correction and 2-bit error detection in each 64-bit block To enable. FEC coding device 532 has coded quantization gain and coded Takes the quantized samples and feeds them to the inserter 531 which the inserter 531 encodes. Output compressed data. The inserter in one embodiment of the present invention is 16^*64 It is an inserter. Figure 5B shows transmission coding including inserter 531 and FEC Hamming coding device 532. An embodiment of means 530 is shown. A 64-bit x 16-bit block is shown. 16 Each of the lines in represents a single 64-bit extended Hamming code word. In the coding device Data from left to right Codeword 0 bits Starting with 0 Codeword 15 bits 63 Read into the insert block so that it ends with. Bit position (column) 0,1,2,4,8 , 16 and 32 are skipped and zero is inserted. Humming after filling the inserter 531 Coding is performed by the FEC coding device 532 with 57 data bits in each line. Humming pa Lite Insert the bits at bit positions 1,2,4,8,16 and 32 as shown in the figure. Insert the parity check bit at bit position 0. For all 16 codecs Parity bit and parity bit P_iCan be calculated as follows. P_i= XOR codeword bit [k] i = 0.6 (k-1) &amp; 2ⁱ 0; However, &amp; is a bitwise binary code AND function. .. After inserting the parity bit at each bit position, the parity check bit Calculate the PC (1 bit for each code) as follows.<img file="JP2000504524A_D0006.tif" />After calculating the parity bit and completing the insertion, the data is codeword 0, bit 0. Read the column from top to bottom from the inserter, starting with codeword 15 and ending with bit 63. Squeeze out. FIG. 6A is an upper block diagram of the high-speed data decoding apparatus according to the embodiment of the present invention. To. This high-speed data decoding device is the data compression process of the high-speed data coding device. The reverse of the above is put into practice, and the transmission / decryption processing means 601 that can be selected for acceptance / rejection and the frame Memory gain decoding device 610, data sample dequantization device 620, and PCM stretcher Includes 630 and buffer 640. Transmission decoding processing means 801 is a reverse inserter 603 And the FEC decoder 602. The operation of this high-speed data decoder will be described below with reference to FIG. Received compression Data is 16^*It can also be supplied to the reverse inserter 603, which is a 64-bit reverse insertion processing means. Wear. The output of the reverse inserter 603, FEC, which is a (64,57) extended humming decoder Add to decoder 602. This humming decoder corrects 1-bit error per block It is possible to detect positive and 2-bit errors. FIG. 6B shows one embodiment of the present invention. The reverse insertion process and the humming decoding process are shown. Reading data into the reverse inserter 603 Starting from codeword 1, bit 1 and ending with codeword 15, bit 63 Do it in the order below. The syndrome is calculated as follows. Calculate Parity Bits: P_i= XOR codeword bit [k] i = 0.5 (k-1) &amp; 2ⁱ 0; However, &amp; is a bit-wise binary AND function Syndrome = chain product P5 | P4 | P3 | P2 | P1 | P0 The parity check bit (1 bit for each code) is calculated as follows: Ru:<img file="JP2000504524A_D0007.tif" /> The numerical display of this syndrome is a bit error (if it occurs). Indicates the position. When a bit error occurs, the parity check bit in the code If a bit is inserted, the bit is inverted (corrected). Otherwise, There are two or more bit errors in the do, and the syndrome is considered to be incorrect. Syndro If the number is zero, no bit error has occurred. As with the encoder, 16 16 parity bits and 16 parity check bits for all codewords It can be calculated at the same time by the exclusive OR operation of the bit width. Returning to FIG. 6A, the decoded data from the FEC decoder 602 is the encoded quantum. It consists of a coded sample and a coded quantization gain. Encoded quantization gain is profitable It is supplied to the obtained decoder 610, and the encoded quantization gain is used as an index in this decoder. And read the quantization gain from the table. As mentioned above, this coded amount The child gain represents the level value of the 64-level logarithmic interval quantizer. This quantization gain value is supplied to the data sample inverse quantization device 620, where it is 32 quantized. Bell Uniform spacing Used to scale the level values in the quantization level table. Scale In the quantization level table, the coded quantization sample of the linear quantization sample Decrypt into a block. A block of coded quantized sample is PCM sunned by PCM compression processing means 630. Convert to a block of pulls (A law or μ law). Then of this PCM sample The block is supplied as a PCM sample as a 64 kb / s output signal. Supply to Fa 640. Super fast codec FIG. 7A is a high-level block diagram of the ultra-high-speed data coding device. This super high speed day The data coding device performs data compression and data decompression of the ultra-high speed voice band modem signal. U. This encoder has 64 kb / s PCM and 62.58 kb / s FEC coded compressed data strips. Perform data conversion with the system. The actual compressed data rate is 56.311 kb / s Therefore, the remaining bitstream is used for error correction data. Ultra high speed codec Is high It is similar to the speed codec. As shown in FIG. 7A, the ultra-high speed data coding of the present invention can be adopted or rejected. With Fa 710 and sample format preprocessing means 720, which can also be selected for acceptance , Gain calculation processing means 722, delay means 721, and data sample quantizer 7 23 and a transmission coding processing means 730 that can be selected for acceptance or rejection are included. Line code processor Stage 730 further comprises an FEC coding device 732 and an inserter 731. Acceptable / rejectable buffer 710 is a sample buffer for ultra-high-speed data compression processing. Hold a predetermined number of samples for lock formation. Sample format pre-processing Reason 710 excludes the formatting of PCM sample A law and other transmission standards. Leave the sample value and set it to a given number, such as a decimal equivalent that is convenient for subsequent processing. Convert to omat. Gain calculation processing means 722 is attached to this sample block The quantization gain value is calculated, and this quantization gain value is used by the data sample quantization device. A set of quantization levels that have a quantization level value proportional to the quantization gain at predetermined intervals. Raise the bell. The above delay causes the compression process to generate a coded quantization sample. It is shown that the quantization gain is determined before the process, and the transmission coding processing means 730 has been coded. Quantization Gain and Coded Error Correction Coding for Transmission of Quantized Samples Is used to do. The operation of the ultra-high speed data compression process is described below. 64kb / s PCM sample above (A method) (Law or μ law) is supplied to buffer 710. This buffer 710 is a PCM sun The pull occurs as a 22.5 ms sample block. Each at 8 kb / s speed of PCM The block contains 180 samples. Unlike high-speed codecs, ultra-high-speed codecs straight-line PCM samples. Do not convert to sample. Instead, 8-bit PCM data is formatted as a given type. Converted to a sample display. In one example, this is the case for the μ law. No conversion operation to format is required, but for Law A, a sample former Pre-processing means 720 to a predetermined level value format before subsequent quantization processing Convert the sample. As will be apparent to those skilled in the art, the μ law sample is the A law sample. Both of these formats can be converted to a simple display, and in another embodiment both of these formats are third. Can be converted to a given format. Ultra-high-speed codecs have the same type of PCM compression on both sides of the link. so It is desirable to have. If they are not of the same type, additional processing must be added , Μ-law characteristic and A-law characteristic from one end to the other due to the difference Non-linearity occurs in. The sample block received in the predetermined sample format is a gain calculation processing means. Supplied to 722, this means 722 sun of maximum amplitude value (absolute value) in the block Detect pulls. The amplitude of this sample is the quantized advantage of the block Determine the benefits. The gain of this quantization is 7 bits by not using the sign bit of the amplitude. It takes. Table 5 shows how to represent numbers according to the A law and the μ law. Each of these displays The absolute value of the corresponding sample is determined, and the maximum amplitude value is calculated.<img file="JP2000504524A_D0008.tif" /> The quantized gain from the gain calculation processing means 722 and the block of 2's complement are Data after calculation of quantized gain value, as shown in Adoption of Delay Means 721 It is supplied to the sample quantizer 723. Data sample quantizer 723 is from A-law or μ-law sample block Form a new quantization device with a quantization level set. One sample block The following describes how to set up this new quantization device. A law quantization The device divides the range of input amplitude into seven segments, and the μ law quantizer divides the input amplitude into seven segments. Divide the width into eight segments. The following explanation is the A law with seven segments The process is taken up, but this description of the A law is the compression of the μ law sample. It is clear to those skilled in the art that it can be extended to. Each segment (except the first segment) has half the amplitude range of the adjacent segment And each segment (excluding the first segment) has 16 quantization level values Have. Therefore, the size of the quantization step in each segment is adjacent. Twice that of Tep. Table 6 shows the A law segment in one example. Shown along with the range of width magnitudes and the magnitude of the quantization step.<img file="JP2000504524A_D0009.tif" /> The sample representing the input data signal is the entire dynamic range of the A-law quantizer. The A-law quantization device is the selected one of the A-law quantization level. Is converted into a new quantization device by removing. This new quantization device has a uniform level Pros when all segments are used to display sample blocks Seth will be described next. The step width of the last segment, 1/32, is this quantum Since it is the largest step width in the chemical equipment, it is in the last segment. The quantization level value is maintained. The sixth segment has a quantization level value step width It is 1/64. 1/32 step width in 7th segment is the amount in 6th segment The step width is reduced to 1/32 by erasing every other childization level. This process Repeat for the 5th to 6th segments in the same way. 2nd and 1st segment Even if you combine them, it only reaches the range of 1/32, so both of the quantization levels are maintained. Not done. As a result, 31 positive quantization levels and 31 negative quantization levels are formed. And the zero level that separates the first positive segment from the first negative segment is maintained This gives a 63-level uniform interval quantizer. This process then calculates the amplitude peak value of the sample block and that amplitude value. Is included in which of the A law segments. To that sample block All segments higher than this "peak value segment" are ignored. .. The step width of the peak value segment is uniform. The step width of the quantizer Define. Therefore, the average prepared in this way for this sample block. In a one-step quantization device, all the quantization levels of the peak value segment Maintain and maintain half the quantization level in the next lowest amplitude-enabled segment And the amount until the last segment is reached or the availability of the quantization level value is interrupted Assign child level values. The operation of the ultra-high speed quantizer of the embodiment of the present invention, that is, the 128-level quantizer. The method of production is shown in Fig. 9. In step 904, this method is performed on a stretched sample (A law or μ law stretch). Receive the block. Sample amplitude peak values in the block and their correspondence in step 906 The segment of is determined, and the one with the amplitude peak value is defined as the peak value segment. In step 910, maintain each quantization level value of the peak value segment. If the zero level is not reached in step 912, all 16 levels in the next segment will be restored. To have. If the zero level is not reached in step 914, all 16 levels in the next segment To maintain. If the zero level is not reached in step 916, the quantization level in the next segment Maintain every other value (eight level values). If the zero level is not reached in step 918, the four in the next lowest amplitude segment Maintain one level. If the zero level is not reached in step 920, two of the next lowest amplitude segments Maintain one level. If no zero level is found in step 922, then in the next lowest amplitude segment Maintain one level of. Maintain zero level at step 924. Finally, in step 926, the positive quantization level of the negative sign of equal magnitude to each other. And set the sign value to generate a negative level. Above amplitude peak value (7 bits) and 180 7-bit coded sumps Make up the compressed output from the ultra-high speed coded compression process. Returning to Figure 7A, the coded-coded quantization gain and the coded-coded quantization sample are transmitted. It is supplied to the transmission coding processing means 730. A good example of this transmission coding processing means 730 is For example, (87,80) FEC coding device 732 consisting of Hamming coding device Including. Hamming code can correct single-bit errors in 87-bit blocks. This FEC coding device performs forward error correction coding and quantities at uniform quantization level intervals. Insert the childized compressed data sample into the inserter 731, i.e. 16^*87 bits It is supplied to the inserter 731, which is a block insertion device. This inserter 731 is an RF communication device Provides encoded compressed data for channel modulation. Figure 7B illustrates a good example of a line code processing means for this ultra-high speed data encoder. To. The figure shows an 87 x 16-bit block. 87-bit Hamin, each of 16 lines Represents a codeword. The encoder reads the data into the inserter block. Codeword 15 starting from bit 1 of codeword 0 across the line from left to right Do so so that it ends with bit 86 of. Bit position (column) 1,2,4,8,16,32 and And 64 are skipped and filled with zero. The last column / word of the insert block receives special treatment Kake. That is, the column / word stores data only in the first three rows / bit positions. It is housed. Hamming coding is performed for 80 data bits in each row after the inserter is full. .. Humming parity bits are shown at bit positions 1,2,4,8,16,32. And insert at 64. 16-bit width of DSP with parity bits for 6 codes Can be calculated simultaneously using the exclusive OR function of. Parity bit P_iIs shown in Figure 7 Sutoo Calculate with the following formula. P_i= XOR codeword bit [k] i = 0.6 (k-1) &amp; 2ⁱ 0; However, "&amp;" is a bit-wise OR function<img file="JP2000504524A_D0010.tif" /> After calculating and inserting the parity bit, read the data from the inserter in the upper row. Starting with bit 1 of codeword 0 and bit 87 of codeword 15 in the order from to the bottom row Do so so that it ends with. Figure 8 shows the inserter block. There are 88 words numbered 0-87. No. The word 1 is unused but maintained for identity with the HSD. The first word is Do not transmit. The numbers 0 to 1266 represent 1267 bits from 181 words. In Table 8 "P" means parity.<img file="JP2000504524A_D0011.tif" /><img file="JP2000504524A_D0012.tif" /> FIG. 8A is a block diagram of the ultra-high speed data decoding device of the present invention. Data expansion The process is the reverse of the data compression process, and the decoding device is a transmission decoding with selectable acceptance / rejection. Processing means 801, gain decoding device 810, and data sample dequantization device 820 And the sample format reprocessor 830 that can be selected or rejected, and the sample format that can be selected or rejected Includes a buffer of 840. The transmission / decoding processing means 801 that can be selected for acceptance / rejection is a reverse inserter. Includes 803 and FEC detector 802. As shown in Figure 8A, the received coded compressed data is decrypted for transmission coding and And to correct the transmission error, it is sent to the transmission decoding processing means 801. In the embodiment of the present invention Transmission decoding processing means 801 in^*Reverse consisting of 87 bit block reverse inserters Includes inserter 803. The output of this reverse inserter 803 is (87,80) humming decoding It is supplied to the FEC decoding device 802 composed of the devices. This humming decryption device is It is possible to correct 1 bit per tick. FIG. 8 shows the transmission decoding process of the ultra-high speed data decoding apparatus according to the embodiment of the present invention. That is, an example of a decoding process including reverse insertion and humming demodulation is shown. Coded pressure Reading the reduced data into the reverse insertion device is performed by bits of codeword 0 in order from top to bottom. Start with 1 and end with bit 86 of codeword 15. Last column / word Requires special treatment. Bit position where a bit error occurred (if any) in the numerical display of the syndrome Is shown. When a bit error occurs, the bit is inverted (corrected). Sindh If the loam is zero, no bit error has occurred. Ultra-high speed data coding device As in the case, up to 16 codewords using 16-bit exclusive OR operation Parity bit can be calculated. The syndrome is calculated as follows. Calculate Parity Bits: P_i= XOR codeword bit [k] i = 0.6 (k-1) &amp; 2ⁱ 0; However, &amp; is a bit binary AND function. Syndrome = chain product P6 | P5 | P4 | P3 | P2 | P1 | P0 The decrypted data from the FEC decoder 801 is coded with the coded quantized sample. It consists of a quantized quantization gain. The coded gain is obtained in addition to the gain decoding device. The device sends the quantized gain value to the data sample inverse quantization device 820. The data sample quantization device uses this quantized gain value (amplitude pin of the block). A law (or μ) for a 7-bit coded sample using a sample of values Rule) Generate a reference table containing the quantization level. This quantization device is super fast Exactly the same procedure as described above for the degree data coding device, that is, 128 Refer to the 256 accumulated items corresponding to one of the possible coded quantized sample values. It is realized by the procedure included in the lighting table. However, this reference table is used in reverse. reference Form a table on the accumulation of 128 possible coded quantized sample values And the corresponding coded quantized sample (7-bit code) to the table storage item You can find the corresponding PCM sample as an index of. As shown in Fig. 8A, if the A law compression is required, the sample format can be selected or rejected. -Decoded sample block by Matte Reprocessor 830 A Law Four Convert to a desired sample format such as matte. In the case of A law, the field of μ law In that case, the decrypted sample block corresponding to the reconstructed ultra-high-speed data sample is It is supplied to the output buffer 840, and a 64 kb / s PCM compression signal is output from this buffer. Occurs as a number. Preferable examples of the present invention have been illustrated and described, but these examples are only examples. It will be understood that it is the purpose. Without deviating from the true meaning of this invention Those skilled in the art will come up with numerous modifications, modifications and replacements. Therefore, the attachment The claims are intended to include all of these variations within the true meaning and scope of the invention. It is a diagram.