Method and apparatus for detection and bypass of tandem vocoding

Background of the Invention I. Field of Invention The present invention relates to communication. In particular, the present invention relates to new and improved methods for detecting and avoiding serial vocoder connections. II. Description of related technologies In digital communication systems, the transmitter uses a vocoder processing system that efficiently encodes voice information. One method of efficient coding uses variable data rate coding to reduce the amount of data sent in the absence of pauses and other voice activities, thereby other than the destination receiver. The level of interference caused by this transmitter is reduced for the receiver of. In the receiver, a vocoder processing system is used to reproduce the voice information. In addition to the voice information, only the non-voice information or a mixture of the voice information and the non-voice information is transmitted to the receiver. A suitable vocoder for this environment is described in US Pat. No. 5,414,796, entitled "Variable Rate Vocoder," issued May 9, 1995 and assigned to the assignee of the present invention. Based on voice activity between 20 milliseconds (ms) frames, this vocabulary provides digital samples of voice information at four different rates, eg, about 8,000 bits per second (bps), 4,000 bps, 2,000 bps and 1,000 Convert to bps encoded data. Each frame of vocoder data is formatted as 9,600bps, 4,800bps, 2,400bps, 1,200pbs data frames with overhead bits. The highest rate data frame corresponding to 9,600 bps frames is called "full rate" frame, 4,800 bps data frame is called "half rate" frame, 2,400 bps data frame is called "quota rate", 1, 200bps data frames are called "1/8 rate" frames. Additional details of formatting the Bocoder data into data frames were published on April 23, 1996 and transferred to the assignee of the invention, "Methods and Devices for Formatting Transmitted Data". It is explained in the title US Pat. No. 5,511,073. The United States, entitled "Systems and Methods for Producing Waveforms in CDMA Cellular Telephone Systems," published April 7, 1992, assigned to the assignee of the invention, the disclosure of which is incorporated herein by reference. Patent No. 5,103, As described in issue 459, the data frame is further processed, spread spectrum modulated and transmitted. Systems that use the types of efficient vocoders described in the patents mentioned above provide high quality when combined with systems that do not use efficient vocoder processing, such as standard ground line systems. .. However, when a vocoder system is combined with another vocoder system so that it has a tandem (back-to-back) vocoder configuration, the voice quality is lower than with a single vocoder system. The present invention attempts to avoid the drawbacks of tandem bocode processing by detecting and avoiding this configuration. Abstract of the Invention The present invention is a method and an apparatus for preventing deterioration of voice quality due to a tandem bocode processing operation. A general vocoder has an encoder part and a decoder part. The encoder portion receives the voice data encoded in the pulse code modulation (PCM) format and supplies the packetized data. The decoder part receives the packetized data and generates a PCM output. The PCM signal is a multi-bit signal representing an analog voltage value corresponding to an audio signal. The packetized bocoder data format is an efficient way to encode voice information, which is required to identify the same voice signal in PCM format in order to represent voice information using the packetized format. This means that less data bits can be used than the number of bits. In a typical radiotelephone connection, the voice of the first user is encoded in the mobile unit by a remote vocoder to generate packetized vocoder data. Packetized vocoder data is sent over a wireless connection. At the base station, the local vocoder converts the packetized data into the PCM format used by the public switched telephone network (PSTN). The PSTN conveys this PCM signal to a second end user. Tandem vocoder if the two connected end users each have a common type of vocoder If you avoid the input processing configuration, avoid conversion to PCM format, and send the packetized data directly through the PSTN, the resulting audio output will sound better. The tandem bocode processing reduces the quality of the audio output due to the post-filtering processing performed in the audio decoding process. Post-filtering improves the perceived quality of the output speech, but causes distortion in the spectral content of the reproduced speech when compared to the original speech. The present invention is a method and apparatus for detecting a tandem vocoder configuration and avoiding conversion to PCM and reconversion to packetized data. The present invention is also a method and apparatus for detecting a tandem vocoder configuration and modifying the filtering in the vocoder to prevent quality deterioration due to post-filtering. In a preferred embodiment of the invention, a first remote vocoder receives analog voice and produces packetized vocoder data transmitted over a wireless link. The first local vocoder receives packetized vocoder data from the wireless link. The first local vocoder converts the packetized data into a multi-bit PCM output. The first local vocoder also appends a detection code to one of the least significant bits (LSB) of the PCM output. The first local vocoder sends the PCM signal to the PSTN. The first local vocoder also receives PCM input through the PSTN. The first local vocoder constantly monitors the least significant bit of the PCM input to find the detection code that indicates that the second local vocoder is connected to the receiving end. If a detection code is detected, the modified filter mode is established. The modified filter mode modifies the filtering in the first local vocoder to prevent quality degradation of the output audio signal. If the first local vocoder detects a detection code from the second local vocoder, it begins to use packetized data and redundancy checking instead of the second of the outgoing PCM LSBs. The first local vocoder is the incoming PCM Start monitoring the second of the LSBs. If the redundancy check indicates that valid packetized data has been received, the first local vocoder will stop converting the PCM output to packetized data and simply be on the second of the LSBs. Send the packetized data of to the first remote vocoder. Whenever the redundancy check fails and no detection code is detected, the first local vocoder returns to convert the incoming PCM into packetized data. In this way, the tandem bocode processing configuration is avoided.

BRIEF DESCRIPTION OF THE DRAWINGS To allow a more complete understanding of the present invention, the following detailed description of embodiments illustrated in the accompanying drawings will be referred to. Figure 1 illustrates signal transmission between two remote system users. Figure 2 illustrates the connection between two remote system users. Figure 3 shows a common PCM signal transmission format. FIG. 4 shows an exemplary PCM signal transmission format that includes a vocoder packet and a detection code. FIG. 5 is a flowchart showing a method of detecting a remote vocoder and bypassing PCM conversion. FIG. 6 illustrates signal transmission between two remote system users when a third user enters the connection. FIG. 7 illustrates the connection between two remote system users when a third user enters the connection. Figure 8 shows another method of bypassing the PCM conversion, including remote vocoder detection and confirmation modes. Figure 9 shows four formal values for both confirmation and bypass modes. Figure 10 shows how to make improvements while detecting remote vocoders and outputting PCM packets. Description of Preferred Embodiment In modern communication systems that transmit audio signals, analog audio signals are digitized and sent to the network as digital signals. Examples of such modern communication systems are cellular telephone systems, personal communication systems (PCS), wireless local loop (WLL) systems, and satellite communication systems such as GLOBAL STAR. In each of these systems, a connection is established between the two remote users. The present invention is applicable when two remote users are connected through one or more modern communication systems in which both remote users use similar vocoders. Figure 1 illustrates the signal transmission between two remote users using a similar vocoder, and Figure 2 illustrates the connection between two remote users using a similar vocoder. .. Remote unit 10 is analog A mobile phone that receives voice 170 and produces analog voice 80. The remote unit 10 includes a remote vocoder 15 including an encoder 180 and a decoder 90. The encoder 180 receives the analog voice 170, converts it into a digital form, encodes the digital voice, and packets it for transmission over the wireless link 20 as a packet 190. The decoder 90 receives the packet 100 from the wireless link 20, decodes the packet 100 to generate digital voice, and converts the digital voice into an analog voice signal 80. This process involves US Pat. No. 5,103,459, US Pat. No. 5,511,073, and US Pat. No. 5,414, as described above. It will be realized according to the teaching of No. 796. At base station 30, the decoder 200 of the local vocoder 35 receives packet 190 from the wireless link 20. The decoder 200 converts the packetized data into a standard pulse code modulated signal (PCM) 210 for transmission to the public switched telephone network (PSTN) 40. The encoder 110 of the local vocoder 35 at the base station 30 receives the PCM 120 from the PSTN 40 and generates a packet 100 for transmission over the wireless link 20. PCM data is a multi-bit digital value that represents the voltage level of the corresponding analog signal. For example, the most basic PCM conversion process has three steps. First, continuous analog audio signals are sampled at discrete time intervals. Then, the sample value is quantized to generate a digital value. The quantized value is then encoded into a digital representation of the quantized value. In any sample data system, the conversion process to PCM introduces quantization noise. For conventional linear coding schemes, digital codewords are truncated binary representations of analog samples. The effect of this truncation is most pronounced for small signals. The audio signal has a wide dynamic range. If a linear scheme with sufficient dynamic range to accept the maximum signal is used, the signal-to-noise ratio of the smaller amplitude signal will be reduced unless long word lengths are used. Linear coding of the quantized sample is not sufficient to provide an acceptable signal-to-noise ratio for smaller amplitude signals without using an excessively large number of digital bits. The standard PCM format used for PSTN systems uses an alternative to linear coding, where the magnitude of the coding interval is proportional to the input signal level. The coding interval is narrower for small amplitude signals and wider for larger signals. As a result, smaller amplitude signals are represented by more coding levels and have finer resolution. The resulting coding method is almost It is logarithmic and has the property of producing the maximum dynamic range for a given signal-to-noise ratio and word length. The non-linear relationship between the codeword and the effectively represented sample compresses the number of bits required to encode an analog signal with a given dynamic range. A decoding process that uses a similar reverse process is called decompression. Therefore, the term "compression and decompression" (obtained from the terms compression and decompression) means the conversion between an analog sample and a logarithmic PCM value. Figure 3 shows the 8-bit format of PCM signals commonly used in PSTN systems. The most significant bit is the sign bit that represents the sign (+/- voltage level) of the sample value. The next 3 bits represent the exponent that reflects the amplitude of the sample value. The four low-order bits represent the actual value of the sample value, that is, the formal value. It is well known that within the four bits of the formal value, the least significant bit (LSB) can be partially compromised without significantly affecting the resulting system sound quality. In fact, telcos generally do not guarantee that the LSB is transmitted correctly through the PSTN and reserves the right to use the LSB for in-band signal transmission purposes. In a typical PSTN PCM format, a complete loss of LSB results in a loss of 6 dB in the resulting signal-to-noise ratio to the voice. If the LSB is compromised for some time, the signal-to-noise ratio drop is diminished. For example, if the LSB every 10th is compromised, the drop drops to about 1 dB. 1 and 2 illustrate the situation in which the present invention is most effective. Figures 1 and 2 show that the PSTN40 connects two similar base stations, each with a similar remote vocoder. At base station 50, the decoder 130 of the local vocoder 55 receives the packet 140 from the radio link 60. The decoder 130 converts the packet 140 to a standard PCM 120 for transmission to the PSTN 40. Local vocoder The 55 encoder 220 receives the PCM 210 from the PSTN 40 and generates a packet 230 for transmission to the wireless link 60. The remote unit 70 is a mobile telephone that receives analog voice 160 and produces analog voice 250. The remote unit 70 includes a remote vocoder 75 consisting of an encoder 150 and a decoder 240. The encoder 150 receives the analog voice 160, converts it into a digital form, encodes the digital voice, and packets the encoded data for transmission to the wireless link 60 as a packet 140. The decoder 240 receives the packet 230 from the wireless link 60, decodes the packet to generate digital voice, and converts the digital voice into an analog voice signal 250. In an ideal situation, packet 140 on unrelated link 60 is identical to packet 100 on wireless link 20. However, in reality, the tandem vocoder configuration of FIG. 1 and the process of converting packet 140 to PCM120 and PCM120 back to packet 100 do not produce ideal results. In fact, each remote unit 10, The voice quality at 70 is significantly lower than for a single vocoder. The present invention attempts to avoid this decrease. Figure 5 shows the process performed by the local vocoder 35 or 55. When the call is initiated, the vocoder 35 receives the PCM 120 and supplies the packet 100, receives the packet 190 and begins supplying the PCM 210. In addition to converting the packet 190 to the PCM 210, the decoder 200 replaces it with a detection code on the LSB of the formal value portion of the PCM data format and interleaves in a preferred embodiment. Since it is quite possible that the PSTN will use the LSB for signal transmission, the detection scheme must work independently of or with the in-band PSTN signal transmission, or even higher. Must be placed on one of the bits. A simple way to reduce the signal-to-noise ratio drop due to the addition of a detection code is to use a detection code that only compromises the nth LSB. However, a periodic code that impairs the nth LSB may cause in-band PSTN signal transmission and humming, which may not be detected by the receiving vocoder. The preferred method of putting the detection code in the LSB is to use a pseudo-random pattern detection code that, on average, spoils each nth LSB. In the pseudo-random pattern, the pulse position of the detection code changes with time. The detection process for pseudo-random code allows some errors to be tolerated, so interference from in-band PSTN signal transmission does not interfere with code detection. Block 300 in FIG. 5 represents the initial state call packet inactive mode. The decoder 200 generates PCM and a detection code. Encoder 110 is looking for an incoming detection code, as indicated by block 302. If the encoder 110 detects a detection code, the encoder 110 will packetize instead of the other one in the LSB of the formal value of the PCM120. A signal is transmitted to the decoder 200 so as to use data. FIG. 4 shows the four formal value bits of the PCM output when the packet is about to be sent in the first embodiment. The sign bit and exponential digit (not shown in Figure 4) continue to carry PCM coding information. Similarly, the formal value MSB290 and the second high-order bit 292 continue to carry PCM coding information. The third high-order bit 294 is replaced with vocoder packet data and some kind of redundancy check. A wide variety of redundant checks that can be used to provide code check capabilities are technically well known. LSB298 always conveys the detection code. Alternatively, the LSB298 may be used in addition to or in place of the detection code to convey part of the vocoder data or redundancy check bits. As mentioned earlier, packetized vocoder data is a fairly efficient way to encode audio information, which is the same audio in PCM format to represent the audio signal using the packetized format. This means that fewer data bits can be used than are needed to identify the signal. Therefore, in some cases, a single bit of 8-bit PCM code serially conveys vocoder packets sent to the wireless link with redundancy checking at the same rate. If a variable rate vocoder is used, it is only necessary to overlay a subset of the PCM code if the data rate is less than the full rate. Block 304 represents a state called call packet active mode in which the decoder 200 is sending packet 190 in the PCM 210. As shown by block 306 in FIG. 5, the encoder 110 monitors a redundancy check indicating that packet 140 is being received within the PCM 120. If the redundancy check is successful, the encoder 110 will take packet 140 contained in the PCM 120 as packet 100, as shown in block 308. And send directly, essentially ignoring the rest of the PCM input. If the redundancy check fails, the encoder 110 monitors the detection code, as indicated by block 310. If a detection code indicating the presence of the local vocoder 55 still exists, the encoder 110 begins again encoding the incoming PCM data into packets, as indicated by block 312, but is indicated by block 304. Continue to output packets and redundancy checks as done. If no detection code is detected, the encoder 110 begins again encoding the incoming PCM data into packets, as indicated by block 314, and outputs the packet and redundancy check as indicated by block 300. stop. The decoder 200 continues to generate the detection code, but no longer appends packet 190 to PCM210. Similarly, the encoder 110 stops trying to send the packet 140 directly and begins encoding the PCM 120 into the input packet 100. In an alternative embodiment shown in FIG. 5, the detection code is added to the PCM signal only if the data rate of the system is less than the full rate. For example, U.S. Pat. No. 5,414, mentioned earlier, When the variable rate decoder described in 796 is used, one-eighth rate data indicates that the user is not speaking and the vocoder packet is only transmitting background noise. Therefore, the significant degradation in the resulting signal caused by adding the detection code to the PCM while communicating at a 1/8 rate increases background noise, but the actual voice transmission packet. Does not affect. The asynchronous characteristics of the present invention often help reduce the transmission of detection codes. No other modification is required for the present invention. Decreasing the transmission of detection codes delays the system entering packet active mode, which is typically on the order of 50 milliseconds. The advantage of the packet active mode state of block 304 is that the direct response of packet 100 to packet 140 provides a much higher quality sound than a standard tandem bocode processing system. One advantage of this method is that each local vocoder enters packet active mode autonomously and asynchronously. There is no need to adjust the exact time for each vocoder to switch to packet active mode and return to packet inactive mode if necessary. Replacing two of the formal data bits of a PCM signal reduces the quality of the audible signal produced by decoding it as a PCM signal, but it does not erase the signal and the resulting decoding The analog audio signal is understandable. The first embodiment of the invention shown in FIGS. 4 and 5 may be modified to include an intermediate confirmation mode as shown in FIGS. 8 and 9. The confirmation mode allows the PCM to packet transition to be further synchronized between the two vocoders, reducing the signal-to-noise ratio reduction that is added to the PCM operation in the process of switching to bypass mode. Figure 8 shows another process performed by the local vocoder 35 or 55. When the call is initiated, the vocoder 35 receives the PCM120 and packet 10 Supply 0, receive packet 190 and start supplying PCM210. In addition to converting packet 190 to PCM 210, the decoder 200 puts a pseudo-random detection code into the LSB of the formal value portion of the PCM data format, which is shown in the packet-inactive column of FIG. Replaces one in each nth LSB on average. In the preferred embodiments of FIGS. 8 and 9, n is large enough so that the resulting reduction in audio is less than 1 dB, and n generally has a value of 8 or greater. The sign bit and exponential digit (not shown in Figure 9) continue to carry PCM coding information. Block 400 in FIG. 8 represents the initial packet-in-active mode state. The decoder 200 generates PCM and a detection code. Encoder 110 is looking for an incoming detection code, as indicated by block 402. The process of detecting the 1 / n detection code allows the presence of the detection code to be indicated while allowing some error. When the encoder 110 detects the detection code, the local vocoder 35 goes into confirmation mode, as indicated by block 404. In the confirmation mode state, the decoder 200 continues to generate the PCM 210, but the PCM data is shown in the confirmation column of FIG. 9 where the confirmation sequence is in the two higher bits in the preferred embodiment. , The quality deteriorates when mixed. One advantage of the confirmation mode is that the higher PCM bits are not used for PSTN in-band signal transmission and achieve a more reliable indication of the compatibility of the two vocoders before entering bypass mode. Is what you can do. The confirmation sequence does not have to use all the bits of the higher digit, and in a preferred embodiment only one of each n2 bit is used. The confirmation sequence may be a simple pattern or a pseudo-random sequence. The verification code provides information about the vocoder's operating type or mode of operation. Includes. The confirmation sequence may include a sample of the actual vocoder-ized audio information used to verify that the two vocoders are really compatible before the switch to bypass mode is made. The advantage of a confirmation sequence using one of every n2 bits is that in embodiments where lower signal-to-noise ratio in confirmation mode is preferred, it is on the order of 13 dB, rather than accidentally entering bypass mode. It is possible to make n2 large enough so that accidental entry causes less quality degradation to the system. Confirmation mode can be used to synchronize the transition of both vocoders to bypass mode. In another embodiment of FIGS. 8 and 9, two higher digits of the formal value are used to convey the packet and redundancy check in bypass mode. Using two of the higher digits results in a signal-to-noise ratio reduction on the order of 18 dB. A decrease in signal-to-noise ratio of 18 dB has a significant detrimental effect on the resulting voice quality. Therefore, it is advantageous to minimize the amount of time that only one of the vocoders is in bypass mode. Confirmation mode can be used to synchronize the transition to bypass mode so that the transitions to both vocoders, independent and autonomous, occur at about the same time. Confirmation mode also minimizes accidental entry into bypass mode. Block 404 represents the state where the decoder 200 is sending a 1 / n2 verification code in the PCM210. As shown by block 406 in FIG. 8, encoder 110 monitors a similar verification code in PCM120. If the verification code is conveniently detected, as indicated by block 408, the system goes into bypass mode. If the verification code fails to be detected, the decoder 200 stops sending the verification code, as indicated by block 412, and reenters the packet-in-active mode state, as indicated by block 400. .. Block 408 represents a bypass mode in which the decoder 200 sends the packet 190 in the PCM 210 and the encoder 110 directly sends the packet 140 contained in the PCM 120 as the packet 100. As shown by block 410 in FIG. 8, encoder 110 monitors the quality of received data. If quality is favorable, the system remains in bypass mode, as indicated by blocks 408 and 410. Quality checks can fail due to redundant check failures, lost links, high error rates, and poor quality. If such a failure occurs, block 410,418, As indicated by 400, the vocoder 35 exits bypass mode and reenters packet-in-active mode. Another major advantage of the invention's autonomous and asynchronous characteristics, shown in FIGS. 5 and 8, is that it brilliantly exits packet active mode in the event of a situation that disrupts its ability to send packets. .. Therefore, there is no need to inform the local vocoder that the connection has changed. The local vocoder can detect the change and autonomously return to packet-in-active mode. 6 and 7 illustrate a situation in which such a previously established packet active mode is compromised, causing the system to return to packet inactive mode. If a conference call is initiated by one of the parties to the original call, a third party joins the current connection shown in Figures 1 and 2. In this case, a third party ground line unit 280 is added to the connection. Third party unit 280 does not have a vocoder, produces standard PCM260 and receives PCM270 from PSTN40. PCM270 is the sum of PCM120 and PCM210, as produced through the summer 46. The encoder 110 receives the sum of PCM260 and PCM120 through the totalizer 42. Similarly, the encoder 220 receives the sum of PCM210 and PCM260 through the totalizer 44. When adding two logarithmically encoded PCM signals, each PCM signal must first be converted back to a linear quantized digital value. Linear digital values can be added directly to each other and recoded logarithmically as a sum. This process erases the packet and detection code. Packets and detection codes are erased even if a third party has a vocoder similar to that at base stations 30 and 50. In this case, the system in packet active mode (block 304 in Figure 5) or bypass mode (block 408 in Figure 8) begins to fail the redundancy test. Next, the encoder looks for a non-existent detection code and Boko Da enters packet-inactive mode (block 300 in Figure 5 or block 400 in Figure 8). This process takes some time, but the third party 280 can participate in the conversation while the process is taking place. It generally takes about 50 milliseconds for each vocoder to exit packet active mode asynchronously. The sign bit, amplitude bit, and high-order bit of the formal value are unaffected, so the third-party unit 280 receives a degraded signal until the respective local vocoders 35 and 55 enter packet-inactive mode. Similarly, for each local vocoder 35 and 55, as each local vocoder 35 and 55 independently and autonomously detects the impact on the system of the third party unit 280 and begins to re-encode the PCM packet. , A valid PCM signal is available. Local vocoders 35 and 55 continue to send detection codes even when the third party unit 280 is connected. If the third party unit 280 leaves the connection, the local vocoders 35 and 55 begin sending packets again through the process of Figure 5 or Figure 8. One of the main reasons tandem vocoders produce degraded quality audible signals is that modern vocoders use post-filters in the output of the voice decoding process. Post-filters block some of the quantization noise introduced by analog-to-digital conversion and digital coding processes. However, when the post-filter improves the perceived quality of the coded speech, the post-filter causes distortion in the spectral content of the reproduced speech compared to the original speech. When the vocoder processing process is repeated in the reproduced voice, the quality deteriorates due to this spectral distortion. The use of postfix filters in vocoders is described in US Pat. No. 5,414, mentioned above. It is explained in No. 796. One way to prevent quality degradation due to the use of post-filters is to modify the filtering inside the vocoder when the two vocoders are operating in tandem. For example, refer to FIG. 1 again and assume that the decoder 200 has a post-filter or other filter that produces a similar effect. If a tandem vocoder situation is detected using any one of the methods of the invention, several courses of operation are taken. The filter in the decoder 200 is completely disabled or bypassed so that it is unaffected by the filter. Alternatively, the filter in the decoder 200 may be modified, such as by making the frequency response of the filter more gradual. In yet another embodiment, an inverse filter may be added to the encoder 220 to compensate for the effects of the post-filter. In this case, the response of the inverse filter may be varied. FIG. 10 shows one embodiment of vocoder detection that incorporates disabling of the filter in the vocoder. The embodiment of FIG. 10 is based on the operation described above with FIG. Similar reference numbers are used to associate with similar behavior in Figure 8. The difference between FIG. 10 and FIG. 8 is that it replaces the bypass mode state as shown by block 408 with the modified filter mode state as shown by block 420. In the modified filter mode state, the system continues to send PCM-encoded data instead of sending packets. Seeing Figure 9 again, in modified filter mode, the four formal values have the same configuration as in packet-inactive mode. The difference lies in the fact that the filtering within at least one vocoder has changed. Moving on to Figure 10, starting with block 404, block 404 represents a situation where the decoder 200 is sending a 1 / n2 verification code in the PCM 210 just as in Figure 8. Encoder 110 has a similar confirmation in PCM 120, as shown by block 406 in FIG. Monitor the code. If the verification code is conveniently detected, the system enters the corrective filter mode state, as indicated by block 420. If the verification code fails to be detected, the decoder 200 stops sending the verification code, as indicated by block 412, and reenters the packet-in-active mode state, as indicated by block 400. .. Block 420 represents a modified filter mode in which, for example, the decoder 200 has a modified postfix filter and sends the resulting PCM data and detection code within the PCM210. Encoder 110 monitors the quality of received data, as indicated by block 410 in FIG. If quality is favorable, the system remains in modified filter mode, as indicated by blocks 420 and 410. Quality inspections can fail due to detection code failures, lost links, high error rates, and poor quality. If such a failure occurs, block 410,418, As indicated by 400, the vocoder 35 exits modified filter mode and reenters packet-in-active mode. The embodiment of FIG. 10 has several advantages. One advantage is that the difference between the PCM-encoded data in packet-inactive mode and the PCM-encoded data in modified filter mode is between the PCM-encoded data in packet-inactive mode and the PCM-encoded data in bypass mode. It is much less than the difference between them. Therefore, the modified filter mode state is unlikely to cause a significant audible effect if it accidentally enters this state or is in the process of transitioning between states. Modify filter mode can be used even when two different types of vocoders are connected. Each tandem vocoder can have a post-filter or other filter that causes spatial distortion, even if the method used to encode the audio signal is different between the two tandem vocoders. There is sex. In either case, the audible quality can be improved by eliminating the influence of the filter. In this case, one of the codes sent by the vocoder indicates the presence of a filter that has an impact that must be modified when working in tandem with other vocoders. The modified filter mode state can be very easily combined with other embodiments of the present invention. For example, the modified filter mode state can replace block 312 in FIG. The modified filter mode state can also be used as a transition state between the packet-inactive state and the confirmation state, and between other states such as between the confirmation state and the bypass state. The myriad variations of the present invention are readily apparent to those skilled in the art. For example, it is easy to reconstruct which bits of the formal value are used to accept the detection code and data packets. The previous description of a preferred embodiment is provided to allow one of ordinary skill in the art to manufacture or use the present invention. For these embodiments The various modifications are readily apparent to those skilled in the art, and the general principles set forth herein can be applied to other embodiments without the use of inventive power. Therefore, the present invention is not intended to be limited to the embodiments presented herein and should follow the broadest scope consistent with the principles and novel features disclosed herein.