JP4888865B2 - Streaming data compensation method and digital signal receiving apparatus - Google Patents

Description

  The present invention relates to a streaming data compensation method for asynchronous communication that does not synchronize sampling frequencies of a transmission side and a reception side in a communication protocol, and in particular, a streaming data compensation method that compensates for excess and shortage of packets at the time of packet communication on the receiver side. About. The present invention also relates to a digital signal receiving apparatus using such a streaming data compensation method.

  In digital communication in which streaming data is transmitted and received asynchronously, it is common to use a crystal oscillator that outputs the same sampling frequency on the transmission side and the reception side. However, there may be a difference between the data transmission speed on the transmission side and the data processing speed on the reception side due to variations in the frequency accuracy of the crystal oscillator. In this case, if the data processing speed on the receiving side is faster than the data transmission speed on the transmitting side, data to be processed will eventually disappear, and data shortage will occur on the receiving side. On the other hand, if the data processing speed on the reception side is slower than the data transmission speed on the transmission side, the data to be processed is accumulated, so that excessive data occurs on the reception side.

Particularly when streaming data is transmitted and received in units of packets, the receiving side stores data in a buffer in units of packets, so that excessive data or insufficient data on the receiving side occurs in units of packets. For example, in a voice signal delivered with streaming data, if the data is insufficient for one packet, a silent part is generated for one packet. On the other hand, if the data is excessive for one packet, it is necessary to discard the information for one packet, which results in unnatural sound before and after the discarded packet. In either case, quality degradation occurs when streaming data is played back. In order to solve such a problem, a method of compensating for a sampling frequency deviation between the transmission side and the reception side by providing a sampling rate converter (hereinafter referred to as “SRC”) on the reception side is conceivable. In general, the price of SRC is high, and this increases the product cost of the receiver. Also, as a conventional technology, at the time of transmission, according to the order of appearance of data, a packet is created by separating it into N types of time-series data, and the packet is reverse processed at the receiving side to re-unify it into time-series data The technique which performs is proposed (for example, patent document 1). When there is a packet shortage due to this method, a data shortage sample occurs in the restored time series data, so that the packet shortage can be easily compensated and the compensation accuracy can be increased. It is described that it can be expected.
Open 2007-274568

  However, since the method according to Patent Document 1 needs to perform special processing on both the transmission side and the reception side, it can be employed only when a dedicated transmission / reception system is constructed. Further, when a shortage packet occurs, it is necessary to perform many interpolations in one packet after restoration, and there is a doubt about the quality of streaming data after the interpolation. Moreover, it cannot cope with an excessive number of packets.

  The present invention has been made in view of such circumstances, and is a streaming data compensation method that does not require special processing on the transmission side and can be performed only on the reception side, and can cope with either packet shortage or packet excess. An object of the present invention is to provide a streaming data compensation method that does not require many interpolations in one packet.

The streaming data compensation method according to the present invention is a streaming data compensation method that compensates for a packet shortage and packet excess caused by asynchronous transmission and reception in streaming data asynchronous packet communication on the receiving side. A difference in which the same number of data as the data included in the excess packets is distributed to a plurality of packets, and in the continuous data string in the same packet, the sum of the absolute difference values of two adjacent data in the data string is minimized. An absolute value minimum data string is calculated, and data interpolation or data erasure is performed in the difference absolute value minimum data string to compensate for the insufficient packet or the excessive packet.

  According to the above configuration, in the asynchronous packet communication of streaming data, since it is a streaming data compensation method that compensates for a packet shortage and packet excess that occur due to asynchronous transmission and reception on the reception side, special processing is performed on the transmission side. Thus, packet compensation can be performed only on the receiving side. Therefore, there is no need to construct a dedicated transmission / reception system. In addition, since the streaming data compensation method compensates for packet shortage and packet excess, it is possible to cope with both packet shortage and packet excess. In addition, it is a streaming data compensation method that compensates for packets by distributing the data as many as the data contained in the missing packets or excess packets to multiple packets and interpolating or erasing the data. No need.

  In the streaming data compensation method according to the present invention, the data interpolation is an absolute value of a difference between two adjacent data in the data string in a data string consisting of n consecutive data (n is an even number) in the same packet. It is preferable to perform the data interpolation by calculating a differential absolute value minimum data sequence that minimizes the sum of the two and inserting a weighted addition value of two adjacent data in the center of the differential absolute value minimum data sequence.

  According to the above configuration, since the data interpolation is performed by inserting the weighted addition value of two adjacent data at the center of the difference absolute value minimum data string, the data interpolation is performed so that the influence of the data interpolation is minimized. Interpolation is possible. Here, for example, an average value can be used as the weighted addition value. This method is effective when n data, that is, an even number of data is used as a data string.

  In the streaming data compensation method according to the present invention, in a data string composed of consecutive n + 1 data (n is an even number) in the same packet, the sum of absolute differences between two adjacent data in the data string is minimized. The difference absolute value minimum data string is calculated, and the median of the difference absolute value minimum data string is calculated by adding a weighted addition value between the data immediately before the median and the median, and data immediately after the median and the data It is preferable to perform the data interpolation by performing the data interpolation by replacing with two data consisting of weighted addition values with the median.

  According to the above configuration, the data interpolation is performed by calculating the median of the difference absolute value minimum data string, the weighted addition value between the data immediately before the median and the median, and the weighted addition value between the data immediately after the median and the median Since the data interpolation is performed by replacing the two data, the data interpolation can be performed so that the influence of the data interpolation is minimized. Here, for example, an average value can be used as the weighted addition value. This method is effective when n + 1 data, that is, an odd number of data is used as a data string.

  In the streaming data compensation method according to the present invention, in the data erasure, the sum of absolute differences of two data adjacent to each other in the data sequence is minimized in the data sequence including n consecutive data in the same packet. The difference absolute value minimum data string is calculated, and the data erasure is performed by replacing the two adjacent data at the center of the difference absolute value minimum data string with the weighted addition value of the two adjacent data at the center. Is preferred.

  According to the above configuration, data erasure is performed by replacing the two adjacent data at the center of the difference absolute value minimum data string with the weighted addition value of the two adjacent data at the center, so that the influence is minimized. Data can be erased. Here, for example, an average value can be used as the weighted addition value. This method is effective when n data, that is, an even number of data is used as a data string.

  In the streaming data compensation method according to the present invention, in the data erasure, in the data string composed of consecutive n + 1 data (n is an even number) in the same packet, the difference absolute value of two adjacent data in the data string The difference absolute value minimum data sequence in which the sum of the values is the smallest is calculated, and three data consisting of the median of the difference absolute value minimum data sequence, the data immediately before the median and the data immediately after the median are It is preferable to perform the data erasure by replacing the data immediately before the value with the weighted addition value of the median value and the two data consisting of the data immediately after the median value and the weighted addition value of the median value.

  According to the above configuration, the data erasure is performed by weighting the data immediately before the median and the median by weighting the 3 data including the median of the minimum difference absolute value data string, the data immediately before the median, and the data immediately after the median. Data erasure is performed by erasing the data by replacing the two values consisting of the addition value and the weighted addition value of the data immediately after the median and the median. It becomes possible. Here, for example, an average value can be used as the weighted addition value. This method is effective when n + 1 data, that is, an odd number of data is used as a data string.

  The streaming data compensation method according to the present invention is preferably used for a digital signal receiving apparatus.

  According to the present invention, a streaming data compensation method that does not require special processing on the transmission side and can be performed only on the reception side, can handle both packet shortage and packet excess, and can perform many interpolations in one packet. It is possible to provide a streaming data compensation method that does not need to be performed.

(First embodiment)
An embodiment of a streaming data compensation method embodying the present invention and a digital signal receiving apparatus using the streaming data compensation method will be described below with reference to FIGS.

  As shown in FIG. 1, which is a schematic diagram of a digital signal receiver, this receiver is a packet composed of a receiver 1 that receives packetized data transmitted from a transmitter and a DSP 2 (Digital Signal Processor). A digital-analog converter (hereinafter simply referred to as DAC3) that converts the digital signal output from the compensation unit and the packet compensation unit into sound is provided. More specifically, as shown in FIG. 2 which is a conceptual diagram of the digital signal receiving apparatus, the packet compensator sequentially writes packetized data received from the receiver 1 into a write pointer (hereinafter simply referred to as WP). And 3 types of buffers and a read pointer (hereinafter simply referred to as RP) for reading the buffer data and transmitting it to the DAC 3. These three types of buffers r_data_buf21, x_data_buf22, and p_data_buf23 each have eight packet storage areas from packet storage number 0 to packet storage number 7.

  When the receiver 1 of this digital signal receiving apparatus starts signal reception of packetized data, the received packetized data is written to r_data_buf 21 in units of packets via WP. At the start of reception, the WP writes the packetized data in the packet storage number 4 of r_data_buf21. At the same timing, the data of packet storage number 0 of r_data_buf 21 is written to the packet storage number 4 of x_data_buf 22 through the filter 24. Further, at the same timing, the value of the packet storage number 0 of r_data_buf 21 is read by the RP and sent to the DAC 3. Since the DAC 3 processes each data, the data reading of the r_data_buf 21 by the RP is also performed for each data.

  Here, since the data writing speed to r_data_buf 21 by the WP is faster than the receiving speed of the receiver 1, the moving speed of the packet storage area of the WP is the signal receiving speed of packetized data by the receiver 1, that is, transmission (not shown). Depends on the transmission speed of the machine. Similarly, since the data read speed of x_data_buf 22 by RP is faster than the conversion processing speed of DAC 3, the moving speed of the packet storage area of RP depends on the conversion processing speed of DAC 3. Normally, the transmission speed of the transmitter and the conversion processing speed of the DAC 3 are adjusted to be the same speed by a crystal transmitter or the like. Therefore, the transmission speed of the transmitter and the conversion processing speed of the DAC 3 are normally the same speed even if they are not particularly synchronized.

  When the transmission speed of the transmitter and the conversion processing speed of the DAC 3 are equal, when the receiver 1 receives the next packetized data, the WP writes the packetized data in the packet storage number 5 of r_data_buf21. At the same timing, the data of the packet storage number 1 of r_data_buf 21 is written to the packet storage number 5 of x_data_buf 22 via the filter 24. Further, at the same timing, the value of the packet storage number 1 of x_data_buf 22 is read by the RP and sent to the DAC 3. In the same repetition, as shown in FIG. 3, when WP writes data to the packet storage number 7 of r_data_buf21, when the data of packet storage number 3 of r_data_buf21 is copied to packet storage number 7 of x_data_buf22, x_data_buf22 At the same time, the data of the packet storage number 3 is read by the RP and sent to the DAC 3. Since 0 is input to r_data_buf21 and x_data_buf22 in the initial state, 0 is input to packet storage number 0 to packet storage number 3 of x_data_buf22 when WP writes data to packet storage number 7 of r_data_buf21. Has been. Further, the data of packet storage number 0 to packet storage number 3 of r_data_buf21 is copied to packet storage number 4 to packet storage number 7 of x_data_buf22, but the packet storage number 0 to packet storage in the initial state of r_data_buf21 is copied. Since the data of the number 3 is 0, 0 is input to all of the packet storage numbers 0 to 7 of the x_data_buf 22 after all.

  As shown in FIG. 4, when the receiver 1 receives the fifth packetized data, the WP writes the packetized data to the packet storage number 0 of r_data_buf21. At the same timing, the data of the packet storage number 4 of r_data_buf 21 is written to the packet storage number 0 of x_data_buf 22 via the filter 24. Further, at the same timing, the value of the packet storage number 4 of x_data_buf 22 is read by the RP and sent to the DAC 3.

  Thus, when data is written up to the packet storage number 7 of r_data_buf21 and x_data_buf22, it returns to the packet storage number 0 and is overwritten. Thereafter, this process is repeated. Note that the packetized data written to the packet storage number 4 of r_data_buf 21 for the first time in the state of FIG. 4 is copied to the packet storage number 0 of x_data_buf 22. However, the RP is still reading 0, which is the initial value of r_data_buf21 written in the packet storage number 4.

  After that, as shown in FIG. 5, when the receiver 1 receives the eighth packetized data, the WP writes the packetized data to the packet storage number 3 of r_data_buf21. At the same timing, the data of the packet storage number 7 of r_data_buf 21 is written to the packet storage number 3 of x_data_buf 22 via the filter 24. Further, at the same timing, the value of the packet storage number 7 of x_data_buf 22 is read by the RP and sent to the DAC 3.

  By writing the packetized data to the packet storage number 3 of r_data_buf 21 by the WP, all the initial values 0 of the packet storage number 0 to the packet storage number 7 of r_data_buf 21 have been updated. Further, in the subsequent processing, the first packetized data is updated to new data in order.

  At the same timing when the WP finishes writing the packetized data in the packet storage number 3 of r_data_buf21, copying of the data of packet storage number 7 of r_data_buf21 to packet storage number 3 of x_data_buf22 is completed, and the packet storage number of x_data_buf22 7 data reading by RP and transmission to the DAC 3 are completed. Eventually, the packetized data received by the receiver 1 is copied to the x_data_buf 22 with a delay of 4 packets, read further with a delay of 4 packets by the RP, and transmitted to the DAC 3. That is, the digital / analog conversion processing is performed by the DAC 3 with a delay of 8 packets. Further, the packet storage number copied to x_data_buf 22 and the packet storage number read by RP are advanced by 4 packet storage numbers, thereby providing a margin for 4 packet storage numbers forward and backward in the moving direction of RP. Note that a low-pass filter, a high-pass filter, or the like can be considered as the filter 24 provided between the r_data_buf 21 and the x_data_buf 22, but may be omitted if not particularly necessary.

  By the way, although the transmission speed of the transmitter and the conversion processing speed of the DAC 3 are usually the same speed, they are not completely synchronized because they are not synchronized. Accordingly, the moving speeds of WP and RP are also strictly different. Here, since there is room for four packet storage numbers in the front and rear in the moving direction of the RP, even if the moving speeds of the WP and the RP are strictly different, they are absorbed in a short time. be able to. However, for example, if the state in which the moving speed of the RP is faster than the moving speed of the WP continues, the margin for the 4 packet storage number provided in front of the moving direction of the RP is reduced, and finally the 0 packet storage number is obtained. There is no more data to do. On the other hand, if the RP moving speed continues to be slower than the WP moving speed, the margin for the 4-packet storage number provided in the rear of the RP moving direction will be reduced, and finally the 0-packet storage number will be received. Data processing is not in time. In order to prevent such a situation, it is confirmed whether the moving speed of the RP is faster or slower than the moving speed of the WP. If the moving speed of the RP is faster than the moving speed of the WP, the data is interpolated and the moving speed of the RP is WP. This digital signal receiving apparatus is provided with a streaming data compensation method for erasing data when the moving speed is slower than this. This will be specifically described below.

  As shown in FIG. 2, when the receiver 1 of the digital signal receiving apparatus starts receiving the signal of the packetized data, the WP writes the packetized data in the packet storage number 4 of r_data_buf21 at the same timing. The value of the packet storage number 0 of x_data_buf 22 is read by the RP and sent to the DAC 3. When the packet storage number 0 of x_data_buf 22 read by the RP is subtracted from the packet storage number 4 of r_data_buf 21 written by the WP at this time, it is 4. Here, when the moving speed of the RP and the moving speed of the WP are equal, as shown in FIG. 3, when the receiver 1 of the digital signal receiving apparatus starts receiving the signal of the fourth packetized data. The WP reads the value of the packet storage number 3 of x_data_buf 22 by RP at the same timing as writing the packetized data to the packet storage number 7 of r_data_buf 21, and sends it to the DAC 3. If the packet storage number 3 of x_data_buf 22 read by the RP is subtracted from the packet storage number 7 of r_data_buf 21 written by the WP at this time, it is 4 again. That is, when the packet storage number of x_data_buf 22 read by RP is subtracted from the packet storage number of r_data_buf 21 written by WP, it is 4 and does not change.

  However, when the moving speed of the RP is faster than the moving speed of the WP, for example, as shown in FIG. 6, the WP writes the packetized data to the packet storage number 7 of the r_data_buf 21 at the same timing as the x_data_buf 22 The value of the packet storage number 4 may be read by the RP and sent to the DAC 3. If the packet storage number 4 of x_data_buf 22 read by the RP is subtracted from the packet storage number 7 of r_data_buf 21 written by the WP at this time, it becomes 3. That is, when the moving speed of RP is faster than the moving speed of WP, the packet storage number difference is smaller than 4.

  On the other hand, when the moving speed of the RP is slower than the moving speed of the WP, for example, as shown in FIG. 7, the WP writes the packetized data to the packet storage number 7 of the r_data_buf 21 at the same timing as the x_data_buf 22 The value of the packet storage number 2 may be read by the RP and sent to the DAC 3. When the packet storage number 2 of x_data_buf 22 read by the RP is subtracted from the packet storage number 7 of r_data_buf 21 written by the WP at this time, the result is 5. In other words, when the moving speed of the RP is slower than the moving speed of the WP, the packet storage number difference is larger than 4.

  In other words, whether the moving speed of the RP is faster or slower than the moving speed of the WP can be confirmed depending on whether the packet storage number difference is larger or smaller than 4. More specifically, the packet storage number difference is calculated every time the WP moves the packet storage number. If the packet storage number difference is smaller than 4, it can be confirmed that the moving speed of the RP is faster than the moving speed of the WP. If the packet storage number difference is larger than 4, it can be confirmed that the moving speed of RP is slower than the moving speed of WP. As shown in FIG. 3, after the WP writes to the packet storage number 7 of the r_data_buf 21, the WP writes to the packet storage number 0 of the r_data_buf 21 as shown in FIG. 4, so the packet storage number of the r_data_buf 21 written by the WP. If the packet storage number 4 of x_data_buf22 read by RP is subtracted from 0, the packet storage number difference becomes -4. Therefore, when the writing of r_data_buf21 to packet storage number 7 by WP is completed, the calculation method is to subtract the packet storage number of x_data_buf22 read by RP after adding 8 to the packet storage number of r_data_buf21 written by WP. To change. For example, after adding 8 to the packet storage number 0 of r_data_buf 21 written by WP and subtracting the packet storage number 4 of x_data_buf 22 read by RP, the packet storage number difference becomes 4. On the other hand, when the reading of the packet storage number 7 of x_data_buf22 by RP is completed, the calculation method is changed so that the packet storage number of x_data_buf22 is subtracted without adding 8 to the packet storage number of r_data_buf21 written by WP. That is, a measure is taken to restore the packet storage number difference calculation method again. By doing so, the packet storage number difference is always a positive value.

  Packet compensation is performed when the packet storage number difference thus determined changes more than a certain value, for example, 4 + 2 or more (ie, 6 or more) or 4-2 or less (ie, 2 or less). When the packet storage number difference is 2 or less, packet compensation is performed to reduce the moving speed of the RP by increasing the number of data read by the RP. When the packet storage number difference is recovered to 3 or more by this packet compensation, the compensation is terminated. On the other hand, even if compensation for increasing the data read by the RP by 1 is not recovered to a difference of 3 or more in the packet storage number difference, conversely, if it becomes 1 or less, the number of data read by the RP is increased by 2 per packet. Compensate. Further, if the packet storage number difference does not recover to 3 or more by this processing, the number of data read by the RP is increased by 3 or more per packet.

  Conversely, when the packet storage number difference is 6 or more, packet compensation is performed to increase the moving speed of the RP by reducing the number of data read by the RP. If the packet storage number difference is restored to 5 or less by this packet compensation, the compensation is terminated. On the other hand, even if compensation for reducing the data read by RP by 1 per packet does not recover the difference in packet storage number to 5 or less, conversely, if it becomes 7 or more, the number of data read by RP is reduced by 2 per packet. Compensate. Furthermore, if the packet storage number difference does not recover to 5 or less even by this processing, the number of data read by the RP is similarly reduced by 3 or more per packet.

  This streaming data compensation method is a streaming data compensation method that compensates for a packet shortage and packet excess in a signal transmitted by an asynchronous packet on the receiving side, so that no special processing is required on the transmitting side, and streaming is performed only on the receiving side. Data compensation methods can be performed. In addition, since the streaming data compensation method compensates for packet shortage and packet excess, it is possible to cope with both packet shortage and packet excess. Furthermore, since this is a streaming data compensation method in which packets are compensated by distributing data to a plurality of packets to increase or decrease data, it is a streaming data compensation method that eliminates many data interpolations and data erasures in one packet.

First, a specific method for increasing data will be described. As shown in FIG. 8, when the moving speed of the RP is higher than the moving speed of the WP, the RP reads data of one packet or more while the WP writes data for one packet. Therefore, if this state continues, WP will be delayed by one packet from the standard state. This is a state where the above-described packet storage number difference is 3 (4-1). If this state further advances and there is no packet storage number difference, the result is that the data read by the RP is exhausted. Therefore, as described above, data is increased by one sample per packet, and this increase is performed by interpolating data between data adjacent to each other in one packet. Such data interpolation is preferably performed so that the influence of data interpolation is minimized. Therefore, a method for selecting a place to interpolate data becomes a problem. One way, as shown in FIG. 9, between the two differences is the smallest x ₁₄ and x ₁₅ data adjacent to insert the data x ₂₀ to be interpolated, as shown in FIG. 10. More specifically, first, all the absolute differences between two adjacent data in one packet are obtained. That is, all of a ₀ = | x ₁ −x ₀ | to a ₁₈ = | x ₁₉ −x ₁₈ | are obtained. Next, the minimum value a _x of a _{0 to} a ₁₈ is searched. Then, an average value (x _x + x _{x + 1} ) / 2 is inserted between x _x and x _{x + 1} as a weighted addition value. By such an operation, data can be inserted into a place where the difference absolute value is the smallest, in other words, a place where the change in the data value is small. Since the amount of data read by the RP is increased by such processing, the moving speed of the RP is slowed down. When the advance of the moving speed of the RP is eliminated as shown in FIG. If the progress of the RP movement speed is not resolved even if the number of data is increased by one per packet, the number of data is increased by two per packet. _{Specifically,} a 0 _~a next difference absolute value of the minimum value _{a x} plus minimum value _{a x 18 'looking} for, x _x' is less a _x weighting addition between the x _{x '+ 1} and The average value (x _{x ′} + x _{x ′ + 1} ) / 2 is further inserted as a value. Similar processing is performed when the number of data is increased by 3 or more per packet.

Next, the data reduction method will be described. As shown in FIG. 12, when the moving speed of the RP is smaller than the moving speed of the WP, the RP reads data of one packet or less while the WP writes data for one packet. Therefore, if this state continues, the WP will advance by one packet from the standard state. This is a state in which the packet storage number difference described above is 5 (4 + 1). If this state further advances and the packet storage number difference is 8, the result is that the data is updated before the RP is read. Therefore, as described above, the data is decreased by one sample per packet. This increase in data is achieved by replacing two adjacent data with one other data in one packet. Do. Such data replacement is preferably performed so that the influence of the data replacement is minimized. Therefore, a method for selecting a place to replace data becomes a problem. As one method, as shown in FIG. 13, x ₁₄ and x ₁₅ having the smallest difference between two adjacent data are replaced with data x _{20 as} shown in FIG. More specifically, all the absolute differences between two adjacent data in the packet 1 are obtained. That is, all of a ₀ = | x ₁ −x ₀ | to a ₁₈ = | x ₁₉ −x ₁₈ | are obtained. Next, the minimum value a _x of a _{0 to} a ₁₈ is searched. Then, the average value (x _x + x _{x + 1} ) / 2 is inserted between x _x and x _{x + 1} as a weighted addition value, and x _x and x _{x + 1} are deleted. Since the amount of data read from the RP is reduced by such processing, the moving speed of the RP is increased. When the delay of the moving speed of the RP is eliminated as shown in FIG. 15, the packet compensation is finished. If the delay in the RP movement speed is not resolved even if the number of data is decreased by one per packet, the number of data is decreased by two per packet. _{Specifically,} a 0 _~a next difference absolute value of the minimum value _{a x} plus minimum value _{a x 18 'looking} for, x _x' is less a _x weighting addition between the x _{x '+ 1} and The average value (x _{x ′} + x _{x ′ + 1} ) / 2 is further inserted as a value, and x _{x ′} and x _{x ′ + 1} are deleted. Similar processing is performed when the number of data is reduced by 3 or more per packet.

  Next, a method for causing the RP to read data with the number of data increased or decreased will be described. The above-described packet compensation is performed as follows using p_data_buf23 among the three types of buffers included in the packet compensation unit. As shown in FIG. 16, p_data_buf 23 has eight packet storage areas from 0 to 7, and records data for one packet in one packet storage area. The data of the packet whose data has been increased or decreased as described above is recorded in this p_data_buf 23. When packet compensation is performed, the RP reads the data after packet compensation recorded in p_data_buf 23 instead of the copy data of the packet recorded in x_data_buf 22, as shown in FIG. When the packet compensation is completed, the RP reads the data of x_data_buf 22 again and sends it to the DAC 3 instead of the p_data_buf 23.

    According to the digital signal receiving apparatus of the above embodiment, the following effects can be obtained.

  (1) The streaming data compensation method of the above embodiment is a streaming data compensation method that compensates for a packet shortage and packet excess in a signal transmitted asynchronously on the reception side, and thus does not require special processing on the transmission side. Packet compensation can be performed only on the receiving side. Therefore, there is no need to construct a dedicated transmission / reception system. In addition, since the streaming data compensation method compensates for packet shortage and packet excess, it is possible to cope with both packet shortage and packet excess. Further, since the streaming data compensation method compensates the packet by dispersing or erasing data in a plurality of packets, the streaming data compensation method does not require much compensation for one packet.

(2) According to the streaming data compensation method of the above-described embodiment, data interpolation is performed by using two adjacent data in a data sequence of minimum absolute value a _x consisting of two consecutive data, that is, a data sequence consisting of x _x and x _{x + 1.} Since data interpolation is performed by inserting an average value (x _x + x _{x + 1} ) of x _x and x _{x + 1} , it is possible to perform data interpolation so that the influence of data interpolation is minimized.

(3) According to the streaming data compensation method of the above embodiment, the data interpolation is performed on the data sequence of the differential absolute value minimum a _x composed of two consecutive data, that is, two data data adjacent to the data sequence composed of x _x and x _{x + 1} Since x _x and x _{x + 1} are replaced with the average value (x _x + x _{x + 1} ) of x _x and x _{x + 1} , data can be erased so that the influence is minimized.

  (4) Moreover, since the digital signal receiving apparatus of the said embodiment uses the streaming data compensation method concerning this invention, it is a digital signal receiving apparatus which has all the effects of said (1)-(3).

(Second Embodiment)
Next, a second embodiment of the digital signal receiving apparatus embodying the present invention will be described. Since the second embodiment has a configuration in which the data compensation method of the first embodiment is changed, detailed description of the same parts is omitted.

In the second embodiment, data compensation is performed as follows. As shown in FIG. 8, when the moving speed of the RP is higher than the moving speed of the WP, the RP reads data of one packet or more while the WP writes data for one packet. Therefore, if this state continues, WP will be delayed by one packet from the standard state. This is a state where the above-described packet storage number difference is 3 (4-1). If this state further advances and there is no packet storage number difference, the result is that the data read by the RP is exhausted. Therefore, as described above, data is increased by one sample per packet, and this increase is performed by interpolating data between data adjacent to each other in one packet. Such data interpolation is preferably performed so that the influence of data interpolation is minimized. Therefore, a method for selecting a place to interpolate data becomes a problem. In the present embodiment, as shown in FIG. 18, between x ₁₄ and x ₁₅ , which is the middle of four data of x _{13 to} x ₁₆ having the smallest value change between four consecutive data, FIG. inserting the data x ₂₀ to be interpolated as shown in. More specifically, first, all the sums of absolute differences of two adjacent data of four consecutive data in the packet 1 are obtained. _{That, a 0 = | x 1 -x} 0 | + | x 2 -x 1 | + | x 3 -x 2 | from _{a 16 = | x 17 -x 16} | + | x 18 -x 17 | + | x All of up to ₁₉ −x ₁₈ | _Next, search for minimum value _{a x} of a 0 _{~a 16.} Then, an average value (x _{x + 1} + x _{x + 2} ) / 2 is inserted as a weighted addition value between x _{x + 1} and x _{x + 2} . By such processing, data can be inserted at a location where the sum of absolute difference values is smallest, in other words, at a location where the change in data value is small. Incidentally, in the case of increasing two per packet data number, look for a ₀ ~a ₁₆ minimum value a _x plus the minimum value a next total sum of the difference absolute value of _x is less a _x of _', x _x the average value as a weighted sum value between _'+1 and _{x x' +2 (x x '} + 1 + x x' +2) / 2 is further inserted. Similar processing is performed when the number of data is increased by 3 or more per packet.

A method for packet compensation by reducing data will be described next. As shown in FIG. 12, when the moving speed of the RP is smaller than the moving speed of the WP, the RP reads data of one packet or less while the WP writes data for one packet. Therefore, if this state continues, the WP will advance by one packet from the standard state. This is a state in which the packet storage number difference described above is 5 (4 + 1). If this state further advances and the packet storage number difference is 8, the result is that the data is updated before the RP is read. Therefore, as described above, the data is decreased by one sample per packet. This increase in data is achieved by replacing two adjacent data with one other data in one packet. Do. Such data replacement is preferably performed so that the influence of the data replacement is minimized. Therefore, a method for selecting a place to replace data becomes a problem. In the present embodiment, as shown in FIG. 20, x ₁₄ and x ₁₅ which are the middle of four data of x _{13 to} x _{16 with} the smallest change in value between the four data are represented as data x as shown in FIG. Replace with ₂₀ . More specifically, all the sums of adjacent difference absolute values of four consecutive data in packet 1 are obtained. _{That, a 0 = | x 1 -x} 0 | + | x 2 -x 1 | + | x 3 -x 2 | from _{a 16 = | x 17 -x 16} | + | x 18 -x 17 | + | x All of up to ₁₉ −x ₁₈ | _Next, search for minimum value _{a x} of a 0 _{~a 16.} Then, the average value (x _{x + 1} + x _{x + 2} ) / 2 is inserted between x _{x + 1} and x _{x + 2} as a weighted addition value, and x _{x + 1} and x _{x + 2} are deleted. By such processing, it is possible to replace the two data of the portion where the sum of the absolute difference values is smallest, in other words, the portion where the change of the data value is small, with 1 data which is the average value as the weighted addition value. Incidentally, in the case of reducing two per packet data number, look for a ₀ ~a ₁₆ minimum value a _x plus the minimum value a next total sum of the difference absolute value of _x is less a _x of _', x The average value (x _{x ′ + 1} + x _{x ′ +2} ) / 2 is further inserted between _{x ′ +1} and x _{x ′ +2} as a weighted addition value, and x _{x ′ +1} and x _{x ′ +2} are deleted. . Similar processing is performed when the number of data is reduced by 3 or more per packet.

  Note that a method for causing the RP to read data with the number of data increased or decreased is the same as that in the first embodiment, and thus description thereof is omitted.

  Therefore, according to the second embodiment, in addition to the effects described in the first embodiment, the following effects can be obtained.

(1) In the second embodiment, determining the minimum value a _x of the sum of absolute difference values adjacent four consecutive data on packet 1, the data sequence x _x having a minimum value a _x of the _sum, x x + _1, An average value (x _{x + 1} + x _{x + 2} ) / 2 is inserted as a weighted addition value between x _{x + 1} and x _{x + 2} which is the center of x _{x + 2} and x _{x + 3} . Therefore, the minimum value a _x of the difference absolute value of two adjacent data in the packet 1 is obtained, and the average value (x _x + x _{x + 1} ) / 2 is inserted as a weighted addition value between x _x and x _{x + 1.} Compared to the first embodiment, it is possible to supplement the data to a place where the macro change of the data value is small. Therefore, it can be used for applications in which the influence of macro changes is more important than the influence of micro changes.

(2) In the second embodiment, determining the minimum value a _x of the sum of absolute difference values adjacent four consecutive data on packet 1, the data sequence x _x having a minimum value a _x of the _sum, x x + _1, The two adjacent data x _{x + 1} and x _{x + 2} in the center of x _{x + 2} and x _{x + 3} are replaced with the average value (x _{x + 1} + x _{x + 2} ) / 2 as a weighted addition value. Therefore, a first value a _x of the difference absolute value of two adjacent data in packet 1 is obtained, and x _x and x _{x + 1} are replaced with the average value (x _x + x _{x + 1} ) / 2 as a weighted addition value. Compared to the first embodiment, the data can be erased at a place where the macro change of the data value is small. Therefore, it can be used for applications in which the influence of macro changes is more important than the influence of micro changes.

(Third embodiment)
Next, a third embodiment of the digital signal receiving apparatus embodying the present invention will be described. Note that the third embodiment has a configuration in which only the data compensation method of the first embodiment and the second embodiment is changed. Therefore, detailed description of the same parts is omitted.

In the third embodiment, data compensation is performed as follows. As shown in FIG. 8, when the moving speed of the RP is higher than the moving speed of the WP, the RP reads data of one packet or more while the WP writes data for one packet. Therefore, if this state continues, WP will be delayed by one packet from the standard state. This is a state where the above-described packet storage number difference is 3 (4-1). If this state further advances and there is no packet storage number difference, the result is that the data read by the RP is exhausted. Therefore, as described above, data is increased by one sample per packet, and this increase is performed by interpolating data between data adjacent to each other in one packet. Such data interpolation is preferably performed so that the influence of data interpolation is minimized. Therefore, a method for selecting a place to interpolate data becomes a problem. In the present embodiment, as shown in FIG. 22, to interpolate x ₁₄ change in value between the three successive data is the smallest x ₁₃ 3 Data intermediate ~x _15, as shown in FIG. 23 replacing the data _{x 20} and _{x 21.} More specifically, first, all sums of absolute differences of two adjacent data of three consecutive data in the packet 1 are obtained. _{That, a 0 = | x 1 -x} 0 | + | x 2 -x 1 | from _{a 17 = | x 18 -x 17} | + | x 19 -x 18 | all to seek. Next, the minimum value a _x of a _{0 to} a ₁₇ is searched. Then _{x x,} a _{x x + 1} is the median of _{x x + 1, x x +} 2, the average value of the data _{x x} and median _{x x + 1} immediately before the median _{(x x + x x + 1} ) / 2, and after the median replacing data _{x x + 2} and a median average of the _{x x + 1} and _{(x x + 1 + x x} + 2) / 2 of the two data. By such processing, data can be interpolated at a location where the sum of absolute difference values is the smallest, in other words, at a location where the change in data value is small. Incidentally, in the case of increasing two per packet data number, look for a ₀ ~a ₁₇ minimum a _x plus minimum a _x following the difference absolute value sum is small a _x of the _', x _{x ',} x _{x' +} 1, x _{x '+ 2} of the median x _x' average value immediately before the _'median x _x' data x _x a _{+ 1 +1} median (x _{x '+} x _{x' +1)} / 2 and the average of the data x _{x '+ 2} and the median x _{x' + 1} immediately after the median _{(x x '+ 1 + x} x' + 2) substituted and / 2 of the two data. Similar processing is performed when the number of data is increased by 3 or more per packet.

A method for packet compensation by reducing data will be described next. As shown in FIG. 12, when the moving speed of the RP is smaller than the moving speed of the WP, the RP reads data of one packet or less while the WP writes data for one packet. Therefore, if this state continues, the WP will advance by one packet from the standard state. This is a state in which the packet storage number difference described above is 5 (4 + 1). If this state further advances and the packet storage number difference is 8, the result is that the data is updated before the RP is read. Therefore, as described above, the data is decreased by one sample per packet. This increase in data is achieved by replacing the three adjacent data with the other two data in one packet. Do. Such data replacement is preferably performed so that the influence of the data replacement is minimized. Therefore, a method for selecting a place to replace data becomes a problem. In the present embodiment, as shown in FIG. 24, replaces the data _{x 20} and data _{x 21,} as shown in FIG. 25 to 3 data of the changing smallest _x 13 _{~x 15} values between three data. More specifically, all the sums of adjacent difference absolute values of three consecutive data in the packet 1 are obtained. _{That, a 0 = | x 1 -x} 0 | + | x 2 -x 1 | from _{a 17 = | x 18 -x 17} | + | x 19 -x 18 | all to seek. _Next, search for minimum value _{a x} of a 0 _{~a 16.} Then, the three data x _x , x _{x + 1} , and x _{x + 2} are converted into the average value (x _x + x _{x + 1} ) / 2 of the data x _x immediately before the median and the median x _{x + 1} and the data x _{x + 2} immediately after the median average value of the median _{x x + 1 (x x +} 1 + x x + 2) substituted and / 2 of the two data. By such processing, it is possible to replace the three data of the portion where the sum of the absolute difference values is the smallest, in other words, the portion where the change in the data value is small, with the two data which are the respective average values. Incidentally, in the case of reducing two per packet data number, look for a ₀ ~a ₁₆ minimum value a _x plus the minimum value a next total sum of the difference absolute value of _x is less a _x of _', x The average value (x _{x '} + x _{x' + 1} ) / 2 of the data x _{x '} immediately before the median value and the median value x _{x' +1} of the three data _{x '} , x _{x' + 1} , x _{x '+2} and Replace with the average value (x _{x ′ + 1} + x _{x ′ +2} ) / 2 of the data x _{x ′ +2} immediately after the median and the median x _{x ′ +1} . Similar processing is performed when the number of data is reduced by 3 or more per packet.

  Note that a method for causing the RP to read data with the number of data increased or decreased is the same as that in the first embodiment, and thus description thereof is omitted.

  Therefore, according to the third embodiment, the following effects can be obtained in addition to the effects described in the first embodiment and the effects described in the second embodiment.

(1) In the third embodiment, data interpolation, the mean value of the median _{x x + 1} of the difference absolute value the minimum data string, the data _{x x} and median _{x x + 1} immediately before the median _{(x x} + _{x x + 1} ) / 2, and for performing the average value _{(x x + 1 + x x} + 2) / 2 2 data and data interpolation by replacing consisting of data _{x x + 2} and median _{x x + 1} immediately after the median, the influence of data interpolation It is possible to interpolate data so that becomes the smallest. This method is effective when an odd number of data is used as a data string.

(2) In the third embodiment, data interpolation, the mean value of the median _{x x + 1} of the difference absolute value the minimum data string, the data _{x x} and median _{x x + 1} immediately before the median _{(x x} + _{x x + 1} ) / 2, and for performing the average value _{(x x + 1 + x x} + 2) / 2 2 data and data interpolation by replacing consisting of data _{x x + 2} and median _{x x + 1} immediately after the median, the influence of data interpolation It is possible to interpolate data so that becomes the smallest. This method is effective when an odd number of data is used as a data string.

  In addition, you may change the said embodiment as follows.

  In the first embodiment or the second embodiment, the minimum value of the sum of absolute differences of two or four adjacent data in packet 1 is obtained, and data insertion or replacement is performed based on the result. Yes, but other methods are possible. That is, not only the sum of absolute differences of 2 data and 4 data, but also a method of obtaining the minimum sum of absolute differences of even data such as 6, 8 and inserting or replacing data based on the result. It may be. As the number of data increases, the data can be increased or decreased at a location where there is less macro change.

  In the third embodiment, the minimum value of the sum of absolute differences of three adjacent data in packet 1 is obtained, and data is inserted or replaced based on the result. good. That is, not only the sum of absolute differences of three data but also a method of obtaining the minimum sum of absolute differences of odd data such as 5, 7 and inserting or replacing data based on the result. Also good. As the number of data increases, the data can be increased or decreased at a location where there is less macro change.

  In the above embodiment, packet compensation is performed when the packet storage number difference deviates by two from the standard, but other methods may be used. That is, it suffices if the packet compensation timing can be adjusted so that packet compensation does not become excessive and packet compensation is not in time.

  In the above embodiment, the DSP 2 includes three types of buffers r_data_buf21, x_data_buf22, and p_data_buf23, but other configurations may be used. For example, the configuration may be such that the p_data_buf 23 is not provided and the data after packet compensation is directly written to the x_data_buf 22 from the r_data_buf 21. That is, any buffer configuration that enables packet compensation by data interpolation or data erasure may be used.

  In the above embodiment, the DSP 2 includes three types of buffers r_data_buf21, x_data_buf22, and p_data_buf23, but other configurations may be used. For example, the configuration may be such that the RP reads r_data_buf21 without providing x_data_buf22. That is, any buffer configuration that enables packet compensation by data interpolation or data erasure may be used.

  In the above embodiment, the three types of buffers r_data_buf21, x_data_buf22, and p_data_buf23 have eight packet storage areas with packet storage numbers from packet storage number 0 to packet storage number 7, but other configurations are also possible. good. For example, the packet storage area of p_data_buf 23 may be smaller than the other two buffers, or the size of each buffer may be different. That is, any structure can be used as long as it can be temporarily stored to the extent that packet compensation is possible.

  In the above embodiment, the packet compensation unit is configured by a DSP, but other configurations may be used. It is only necessary that the above-described packet compensation is possible, and it may be configured by other electronic components (circuits) such as a microcomputer.

  In the above embodiment, a low-pass filter and a high-pass filter are illustrated as the filter 24, but other filters may be used. Further, instead of the filter, for example, signal processing means such as code data combination and error correction may be provided. The filter 24 and the signal processing means may be configured as hardware by a circuit, or may be realized as software such as a program incorporated in the DSP.

  The present invention relates to a streaming data compensation method for compensating for excess and shortage of packets in an asynchronous packet communication in which the sampling frequency of the transmission side and the reception side is not synchronized in the communication protocol, and digital using the streaming data compensation method. Since it is a signal receiving device, it can be widely used particularly in a field where it is necessary to prevent the quality deterioration of streaming data.

It is drawing explaining one Embodiment of the digital signal receiver concerning this invention, Comprising: It is a schematic diagram of a digital signal receiver. It is drawing explaining one Embodiment of the digital signal receiver concerning this invention, Comprising: It is a conceptual diagram of a digital signal receiver. It is drawing explaining one Embodiment of the digital signal receiver concerning this invention, Comprising: It is a conceptual diagram of a digital signal receiver. It is drawing explaining one Embodiment of the digital signal receiver concerning this invention, Comprising: It is a conceptual diagram of a digital signal receiver. It is drawing explaining one Embodiment of the digital signal receiver concerning this invention, Comprising: It is a conceptual diagram of a digital signal receiver. It is drawing explaining one Embodiment of the digital signal receiver concerning this invention, Comprising: It is a conceptual diagram of a digital signal receiver. It is drawing explaining one Embodiment of the digital signal receiver concerning this invention, Comprising: It is a conceptual diagram of a digital signal receiver. FIG. 6 is a diagram for explaining an embodiment of a digital signal receiving apparatus according to the present invention, for explaining that when a moving speed of RP is larger than a moving speed of WP, a packet storage number difference is reduced from a standard value. FIG. It is drawing explaining one Embodiment of the digital signal receiver concerning this invention, Comprising: It is a figure which shows two data with the least change of a data value within 1 packet of data read by RP. It is drawing explaining one Embodiment of the digital signal receiver concerning this invention, Comprising: In 1 packet of the data read by RP, the state which interpolated the data between two data with the least data value change is shown FIG. FIG. 5 is a diagram for explaining an embodiment of a digital signal receiving apparatus according to the present invention, in which data is interpolated between two data with the smallest change in data value, whereby RP moving speed with respect to WP moving speed is illustrated. It is a figure which shows the state by which advance of was eliminated. FIG. 6 is a diagram for explaining an embodiment of a digital signal receiving apparatus according to the present invention, for explaining that a packet storage number difference increases from a standard value when a moving speed of RP is smaller than a moving speed of WP. FIG. It is a figure explaining one Embodiment of the digital signal receiver concerning this invention, Comprising: It is a figure which shows two data with the least change of a data value within 1 packet of data read by RP. It is drawing explaining one Embodiment of the digital signal receiver concerning this invention, Comprising: The state which substituted two adjacent data with the least data value change with the average value in 1 packet of data read by RP. FIG. FIG. 5 is a diagram for explaining an embodiment of a digital signal receiving apparatus according to the present invention, in which RP movement with respect to WP movement speed is performed by interpolating data between two adjacent data with the least change in data value. It is a figure which shows the state by which the delay of speed was eliminated. It is drawing explaining one Embodiment of the digital signal receiver concerning this invention, Comprising: It is a conceptual diagram of a digital signal receiver, and is a figure which shows the state at the time of packet compensation. It is drawing explaining one Embodiment of the digital signal receiver concerning this invention, Comprising: It is a conceptual diagram of a digital signal receiver, and is a figure which shows the state at the time of packet compensation. It is drawing explaining 2nd Embodiment of the digital signal receiver concerning this invention, Comprising: It is a figure which shows 4 continuous data with the least change of a data value within 1 packet of the data read by RP. It is drawing explaining 2nd Embodiment of the digital signal receiver concerning this invention, Comprising: In 1 packet of the data read by RP, between the data of the center of 4 continuous data with the least data value change It is a figure which shows the state which interpolated data. It is drawing explaining 2nd Embodiment of the digital signal receiver concerning this invention, Comprising: It is a figure which shows 4 continuous data with the least change of a data value within 1 packet of the data read by RP. It is drawing explaining 2nd Embodiment of the digital signal receiver concerning this invention, Comprising: In 1 packet of the data which RP reads, two center data of 4 continuous data with the least change of a data value are shown. It is a figure which shows the state replaced with the average value of the center two data. It is drawing explaining 3rd Embodiment of the digital signal receiver concerning this invention, Comprising: It is a figure which shows 3 continuous data with the least change of a data value within 1 packet of data read by RP. It is drawing explaining 3rd Embodiment of the digital signal receiver concerning this invention, Comprising: The state which interpolated the data between 3 continuous data with the least change of a data value within 1 packet of data read by RP FIG. It is drawing explaining 3rd Embodiment of the digital signal receiver concerning this invention, Comprising: It is a figure which shows 3 continuous data with the least change of a data value within 1 packet of data read by RP. It is drawing explaining 3rd Embodiment of the digital signal receiver concerning this invention, Comprising: The center 2 data of 4 continuous data with the least data value change in 1 packet of data read by RP, It is a figure which shows the state replaced with the average value of the center two data.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Receiver, 2 ... DSP (Digital Signal Processor), 3 ... DAC (digital-analog converter), 21 ... r_data_buf, 22 ... x_data_buf, 23 ... p_data_buf, 24 ... filter.

Claims (6)

In a streaming data compensation method for compensating for packet shortage and packet excess caused by asynchronous transmission / reception in streaming data asynchronous packet communication on the receiving side,
The same number of data as the data included in the missing packet or the excess packet is distributed to a plurality of packets , and the sum of the absolute difference values of two adjacent data in the data string is minimized in the continuous data string in the same packet. A streaming data compensation method , comprising: calculating a difference absolute value minimum data string to be and compensating for the insufficient packet or the excess packet by performing data interpolation or data erasure in the difference absolute value minimum data string . The data interpolation is the minimum difference absolute value that minimizes the sum of the absolute difference values of two adjacent data in the data string in a data string consisting of consecutive n data (n is an even number) in the same packet. Calculate the data column,
The streaming data compensation method according to claim 1, wherein the data interpolation is performed by inserting a weighted addition value of two adjacent data in the center of the difference absolute value minimum data string. In the data interpolation, in a data string composed of consecutive n + 1 data (n is an even number) in the same packet, the difference absolute value minimum that minimizes the sum of the difference absolute values of two adjacent data in the data string Calculate the data column,
2 data comprising the median of the differential absolute value minimum data string, the weighted addition value of the data immediately before the median and the median, and the weighted addition of the data immediately after the median and the median The streaming data compensation method according to claim 1, wherein the data interpolation is performed by replacing the data. The data erasure is a minimum difference absolute value that minimizes the sum of absolute differences of two adjacent data in a data string composed of consecutive n data (n is an even number) in the same packet. Calculate the data column,
2. The streaming according to claim 1, wherein the data erasure is performed by replacing two adjacent data in the center of the difference absolute value minimum data sequence with a weighted addition value of the two adjacent data in the center. Data compensation method. In the data erasure, in a data string composed of consecutive n + 1 data (n is an even number) in the same packet, the difference absolute value minimum that minimizes the sum of absolute differences between two adjacent data in the data string Calculate the data column,
Three data consisting of the median of the minimum difference absolute value data string, the data immediately before the median, and the data immediately after the median, a weighted addition value of the data immediately before the median and the median, and 2. The streaming data compensation method according to claim 1, wherein the data erasure is performed by replacing the data immediately after the median with two data consisting of a weighted addition value of the median.   A digital signal receiving apparatus using the streaming data compensation method according to claim 1.

Images