TWI453733B - Device and method for audio quantization codec - Google Patents

Description

Audio quantization codec device and method thereof

The present invention relates to a quantization apparatus, and more particularly to an audio quantization codec apparatus and method thereof.

The speech signal is originally an analog signal. After digitization and compression, distortion is generated. Generally, the compression ratio is high and the signal distortion is large, but the required transmission code rate is low. Therefore, in the case of insufficient transmission bandwidth, a protocol with a high compression ratio is usually selected under the condition that the content of the call can be recognized. If there is no problem with the transmission bandwidth, the G.711 protocol with a small signal distortion is generally a better choice.

Please refer to FIG. 1 , which is a prior art speech encoding and decoding system diagram, including: a voice input signal 100 , a speech encoder 200 , a memory 300 , a speech decoder 400 , and a voice output signal 500 . The voice input signal 100 is a real sound, which is an analog signal. For example, if the speech encoder 200 is a 16-bit mono, if the sampling is performed at a frequency of 8 KHz per second, the amount of data is 128 kbits per second. When the voice input signal 110 is input to the voice encoder 200, the voice input signal 100 is sampled as a mono-channel data of 128 kbits per second, and then compressed and encoded, and stored in the memory 300. The practical application of the speech encoder 200 is a compressor. In practical applications, in order to reduce the usage of the memory 300, 16-bit voice data is generally compressed into lower resolution data (such as 5 bit or 4 bit) and stored in the memory 300, which can effectively reduce The amount of memory 300 used. Finally, the speech decoder 400 interprets the compressed lower resolution data stored in the memory 300, converts it into 16-bit mono speech data, and converts it into a speech output signal 500.

Next, please refer to FIG. 2A, which is a detailed block diagram of the prior art speech encoder 200. The speech encoder 200 includes an analog-to-digital converter 210, a down-converter 220, a quantizer 230, and a data encoder 240. The analog digital converter 210 receives the analog voice input signal and converts it into a digital first voice material. The down-converter 220 is coupled to the analog-to-digital converter 210 to down-code the first speech material. The quantizer 230 is coupled to the down-converter 220, receives the first speech data and quantizes to generate a digital code comprising symbol data and digital data. The data encoder 240 is coupled to the quantizer 230 and receives at least one digital code to generate a series of speech data.

In another previous embodiment, please refer to FIG. 2B. The first first memory 110 has stored a digital first audio data, and the down converter 220 performs a subtraction on the first voice data. The quantizer 230 is coupled to the down-converter 220, receives the first speech data and quantizes to generate a digital code comprising symbol data and digital data. The data encoder 240 is coupled to the quantizer 230 and receives at least one digital code to generate a series of speech data.

An example is given below: Next, please refer to FIG. 3, the 16-bit voice data converted by the analog-to-digital converter 210, which includes the first voice data of 7 digits: 111110110001000, 1111001100001000, 1111111100001000, 0000000100001000, 0000010100001000, 0000100000001000, 0000111100001000. The down-converter 220 then converts the seventeen 16-bit first voice data into an 8-bit first voice material with positive and negative signs. The down-converter 220 directly removes the first bit to the eighth bit of the first bit of the 16-bit voice data, and retains only the data of the 9th bit to the 16th bit of the original first voice, and the data left by the original voice is New first voice material. Therefore, the first voice data only leaves 8 bits of data with positive and negative signs, and the data range is -128 to 127. Among them, the first bit to the seventh bit represent digital data (the amount of voice signals), and the eighth bit represents symbol data (positive and negative values of voice signals). Therefore, after the first speech data is down-coded by the down-converter 220, the new seven first speech data are obtained as 11111011, 11110011, 11111111, 00000010, 00000101, 00001000, and 00001111 (-5, -13, -1, 2, 5, 8, 15).

Next, the quantizer 230 quantizes the first speech data to generate a digital code, and the manner of quantization can be performed using a look-up table method. Although the voice data is positive or negative, in order to save memory usage, usually only a semi-circular quantization table is established. Before performing the quantification procedure, the symbolic bits representing the speech data belonging to the positive half cycle or the negative half cycle are recorded, and then the speech data is taken to an absolute value, and the speech data is quantized using the positive half-cycle quantization table. The following is a list of 5 bit tables, and the first speech data (-5, -13, -1, 2, 5, 8, 15) is quantized using the quantization table 1. For example, the sign bit of the recorded voice data -5, -13, -1 is 1 and the sign bit of 2, 5, 8, and 15 is 0, and then all the data are taken as absolute values (5, 13, 1, 2, 5, 8, 15), wherein 5 according to the quantization table 1, the best index code is 3, and the corresponding quantized digit code is 00011, 13 according to the quantization table 1, the best index code is 7, and the corresponding The quantized digit code is 00111, and the digit data after the fifth bit plus the sign bit is 10011, 10111, respectively.

Therefore, the absolute value of the first voice data (-5, -13, -1, 2, 5, 8, 15) is obtained according to Table 1 as (3, 7, 1, 2, 3, 4, 7). The corresponding binary digits are (00011, 00111, 00001, 00010, 00011, 00100, 00111), and the binary digits after the fifth digit replaces the sign bit are (10011, 10111, 10001, 00010, 00011) , 00100, 00111). Among them, the table code of the 13 pairs is the closest to 15, so it is selected as the corresponding table value.

Table 1

In practical applications, in order to reduce the quantization error, multiple quantization tables are often used to correspond to different dynamic range speech data. Referring to FIG. 4, the digit code 600 is composed of symbol data 612 plus digital data 614. The serial voice data is composed of a plurality of digit codes 600 and the data amount of the seven digit code 600 is 35 bits. The concatenated voice data usually also includes a frame header data 606 (Frame Header), and the frame tag data 606 records the index of the optimal quantization table corresponding to the current frame, and the frame switching control code (generally using the speech coding data). Special code with the same number of bits but not repeated). The length of the data of the frame mark data 606 depends on the number of quantization tables. For example, when eight quantization tables are used, the frame mark data 606 needs 3 bits to correspond to the index of the optimal quantization table, and 32 quantization tables are used. The box tag data 606 requires 5 bits to correspond to the index of the best quantization table. Taking the data length of the 10-bit frame as an example (the index of the 5-bit frame switching control code plus the 5-bit optimal quantization table), the size of the serialized speech data is 35+10=45 bits. Therefore, the original 7-bit 16-bit first voice data has a total of 112 bits, and the 16-bit data is converted into a first voice data of only 7 8 bits and a total of 56 bits through the down-converter 220. The quantized result of the quantizer 230 is used to convert each 8-bit data (table code) into 5-bit index code data. Finally, the data amount of the seven 5-bit data is 35 bits. From this, we can see that the original 112-bit data amount passes through the down-converter 220 and the quantizer 230, and finally becomes a data amount of only 35 bits. After that, plus 10bit frame mark data 606, the total amount of data is now 45bit.

It can be known from the above prior art that when performing speech quantization coding, each quantized digital code contains a symbol code and a digital code, and each digital code contains a symbol code, which inevitably wastes the stored data amount. Therefore, in order to reduce the amount of waste of stored data, it is necessary to propose a new architecture to reduce the amount of stored data.

The invention provides an audio quantization codec device, which uses a memory for encoding and decoding signals, and the memory records a plurality of digital first speech data and a second encoded data string, including: audio quantization coding module and audio quantization Decoding module. The audio quantization coding module comprises: a quantizer, a signal divider and a data encoder. The signal splitter reads the digital first speech data and performs a plurality of zero crossing point judgments to sequentially generate a plurality of first symbol data, and cuts all digits of the first speech data into a plurality of frames. The quantizer is connected to the signal splitter, and receives the first digit of the first voice data and the first symbol data corresponding to each frame, and quantizes the plurality of digits of the first voice data corresponding to each frame received by the corresponding frame. a plurality of first digital data, and correspondingly generating a first frame mark data according to the quantized result of each frame. The data encoder is connected to the quantizer and the signal divider, and receives the first digital data, the first symbol data and the first frame label data and encodes the plurality of first encoded data strings. The audio quantization decoding module comprises: a data decoder and an inverse quantizer. The data decoder is connected to the memory, reads the second encoded data string and performs decoding to generate a plurality of second decoded data strings, each second decoded data string includes: a second frame mark data, a second symbol data, A plurality of second digital materials. The inverse quantizer is connected to the data decoder, receives the second decoded data string, and performs inverse quantization of the second digital data according to the values of the second frame mark data and the second symbol data to sequentially generate a plurality of digital second voices data.

The invention further provides an audio quantization coding method, which is applied to the encoding of a voice digital signal, comprising: reading a plurality of digital first speech data and performing a plurality of zero crossing point judgments to sequentially generate a plurality of first symbol data, and Cutting all digits of the first voice data into a plurality of frames; receiving a plurality of digits of the first voice data and the first symbol data corresponding to each frame, and corresponding to each of the received frames The digital first data is quantized to generate a plurality of first digital data, and corresponding to the quantized result of each frame, corresponding to the first frame marking data; and receiving the first digital data, the first symbol data and the first The frame marks the data and encodes into a plurality of first encoded data strings.

The present invention further provides an audio quantization decoding method, which is applied to digital voice data decoding, comprising: reading a second encoded data string and performing decoding to generate a plurality of second decoded data strings, each second decoded data string comprising: a second frame mark data, a second symbol data, and a plurality of second digital data; and receiving the second decoded data string, and performing the second number according to the value of the second frame mark data and the second symbol data The inverse quantization of the data sequentially generates a plurality of digital second speech data.

The present invention further provides an audio quantization and decoding method, comprising: reading the digital first speech data and performing a plurality of zero crossing point judgments to sequentially generate a plurality of first symbol data, and the digital first speech Cutting the data into a plurality of frames; receiving the plurality of digits of the first voice data and the first symbol data corresponding to each frame, and quantizing the digital first voice data corresponding to the frame received each time a plurality of first digital data, and correspondingly generating a first frame mark data according to the quantized result of the frame; receiving the first digital data, the first symbol data and the first frame mark data and encoding the first code Data string; reading a second encoded data string and decoding to generate a plurality of second decoded data strings, each second decoded data string comprising: a second frame mark data, a second symbol data, and a plurality of And receiving the second decoded data string, and performing inverse quantization of the second digital data according to the values of the second frame mark data and the second symbol data to sequentially generate the plurality of numbers The second voice data.

The present invention proposes a more efficient encoding device. In the prior art encoding device, each quantized digital code contains a symbol code, which inevitably wastes the stored data amount. The invention only proposes that only one symbol code is used, and a plurality of digital data are connected in series, and the amount of stored data can be reduced while maintaining the resolution, and in the pursuit of sound quality, the amount of data can be increased by a slight increase to achieve a significant improvement in sound quality. The effect.

The above and other objects, features and advantages of the present invention will become more apparent and understood.

Please refer to FIG. 5A , which is an embodiment of the audio quantization coding module 200A of the present invention, including: a quantizer 230, a signal divider 250 and a data encoder 240. The signal divider 250 includes a register 251. The signal divider 250 reads the stored digital first voice data from the first memory 110 and performs a zero crossing point judgment to sequentially generate a plurality of first symbol data. 612, and cutting the digital first voice data into a plurality of frames. The quantizer 230 is connected to the signal divider 250, and sequentially according to the plurality of digits of the first voice data corresponding to the frame cut by the signal divider 250 and the corresponding first symbol data, and corresponding to the frame. The digitized first voice data is quantized and the first digital data is generated in a one-to-one correspondence, and the first frame marker data 606 is generated according to the result quantized by the frame. The data encoder 240 is connected to the quantizer 230 and the signal divider 250, and receives the first digital data, the first symbol data and the frame marker data and encodes the first encoded data string into a first encoded data string. A frame mark data, a first symbol data, and a plurality of first digital materials. Then, the first encoded data string is stored in the second memory 300.

In practice, the first memory 110 and the second memory 300 may be different blocks in the same memory.

Next, please refer to FIG. 5B, which is an embodiment of the audio quantization coding module 200B of the present invention. The main difference between FIG. 5B and FIG. 5A is that the plurality of digits of the first voice data corresponding to the frame read by the quantizer 230 in FIG. 5B are directly read from the first memory 110. And quantify. In FIG. 5A, the plurality of digits of the first voice data corresponding to the frame read by the quantizer 230 are first read by the first memory 110 and then placed in the register 251 of the signal divider 250. Then, the digital first data is read by the register 251 of the signal divider 250 and quantized.

Next, please refer to FIG. 6A, which is an embodiment of the audio quantization coding module 200C of the present invention. In the embodiment of FIG. 5A, a down code converter 220 is added. The down code converter 220 is connected between the first memory 110 and the signal divider 250, and performs a code reduction operation on all the first voice data temporarily stored in one frame of the first memory 110, and then stored in the temporary register. 251.

Next, FIG. 6B is an embodiment of the audio quantization coding module 200D of the present invention. In the embodiment of FIG. 5B, a down code converter 220 is added. The down code converter 220 is connected between the first memory 110 and the signal divider 250 and the quantizer 230, and performs the action of reducing the first voice data stored in the first memory 110 to be transmitted to the quantizer 230. .

The quantizer 230 includes: a control unit and a vector unit, and the control unit calculates a quantization error or the like according to all the first voice data or the subtracted first voice data corresponding to each frame for the selection of the quantization table, And generate frame marker data. The vector unit receives the first voice data and performs a lookup of the quantization table to generate digital data.

The down-converter 220 of the present invention is not limited to only 16 bits reduced to 8 bits, and can be changed from 16 bits to 10 bits, or 24 bits to 12 bits. The present invention does not have a format, and is selected according to the design of the system.

The first voice data may be a subtracted data, for example, 16bit down to 8bit or 10bit. In some embodiments of the present invention, when the plurality of first voice data is negative, the symbol data representing the positive and negative symbols in the digit code 600 is omitted and integrated into the frame marker material, and a new number is formed. The data 614, which only contains the content representing the sound material, does not contain positive and negative information, as shown in FIG. In this way, the number of bits of the digit code 600 in the prior art can be reduced, for example, the original 5 bits is reduced to 4 bits to reduce the amount of data. Or, the number of bits of the digital code 600 is increased. For example, only 4 bits of the 5 bits of the original digital code are used. In the present invention, 1 bit, that is, 5 bits of digital data is added to improve the analysis of the encoded data. degree.

Please refer to FIG. 7, which is a schematic structural diagram of the data of the serial voice data of the present invention. The individual serial voice data 620, 622 includes frame mark data 606, symbol data 612 and multiple digital data 614. In other words, each concatenated speech data begins with frame tag data 606 and symbol data 612, and finally before the next frame tag data 606. The length of the concatenated speech data 620, 622 depends on the number of points of the digital data among the lengths of the two zero crossing points, that is, the frame mark data 606 + the symbol data 612 + the number of points of the two zero crossing points × The length of the digital material (for example, 4bit or 5bit).

In other words, the positive or negative sign of the symbol data 612 is determined by the signal divider to determine the zero crossing point, and the zero crossing point is judged according to the data of the two consecutive first speech data. And set. When the signal divider 250 of the present invention receives a positive first speech data and a negative first speech data (whichever occurs first), it can be determined that there is a zero crossing point, and the signal divider 250 generates a representative zero. The symbol data 612 generated by the crossover point is provided to the data encoder 240. The positive sign of the symbol data 612 can be represented by 0, and the negative sign of the symbol data 612 can be represented by 1. For example, the first first voice data is A, and the second first voice data is B. When A<0 and B>=0, the signal divider 250 generates the symbol data 612 as “0” (positive sign) ), that is, the first voice data is negatively positive, that is, the subsequent first voice data will be positive; when A>=0, and B<0, the signal divider 250 generates the symbol data 612. It is "1" (negative sign), that is, the first voice data is changed from positive to negative, that is, the subsequent first voice data will be negative. The above is only one embodiment of the zero crossing point judgment of the present invention, and the present invention is not limited to this manner.

Example 1:

This embodiment illustrates the case where the digital data encoding length representing the voice signal is fixed to 4 bits. There may be two or more quantization tables corresponding to the frame mark data 606. At this time, the frame mark data 606 must use at least one bit to indicate which table to use. For example, when only two quantization tables are used, the frame mark data 606 can set the table value corresponding to Table 2 (the first table in this embodiment) to "0", and Table 3 (this embodiment) The table value of the second table) can be set to "1". When five quantization tables are used, at this time, the frame mark data 606 needs 3 bits, and the table values respectively mapped are 000, 001, 010, 011, 100, and the like. The number of quantifiable tables that can be mapped in the present invention may be a single table or a plurality of tables.

In the case of multiple quantization tables, the most suitable table is generally determined based on the size of the quantization error of the data. Next, please refer back to Figure 3, which is a sequence of multiple first speech data (-6, -12, -1, 3, 5, 8, 15, 8, 5) with two zero crossing points. occur. Assuming a total of 8 tables are available for use, after comparison screening, Tables 2 and 3 are the most suitable tables for the (-6, -12, -1) sequence and (3, 5, 8, 15 respectively). , 8, 5) sequence. Suitable quantification table selections are well known to those skilled in the art and will not be described again.

First, after the first zero crossing point, when (-6, -13, -1) is encountered first, the negative symbol data 612 is taken as 1 in the front, and then (-6, -13, -1) The absolute value becomes (6, 13, 1). First, the frame mark data 606 of Table 2 is first set to 000, and finally, after looking up the table 2, the index code is (4, 8, 1), and finally The sequence obtained by mapping the digital data of Table 2 is (0100, 1000, 0001). Thereafter, the value 000 of the frame mark material 606 and the value 1 of the symbol data 612 are added. The resulting concatenated speech material 620 is (000, 1, 0100, 1000, 0001).

Then, after the second zero crossing point, the frame mark data 606 of Table 3 is first set to 01, and finally (3, 5, 8, 15, 8, 5), and the symbol data 612 is 0 for positive Value, then map 3, find the closest form code, which is well known in the art, you can get the index code (1, 2, 3, 5, 3, 2), and finally map 3 The code of the digital data is (0001, 0010, 0101, 0101, 0011, 0010). Then, the concatenation voice data obtained by adding the value 001 of the frame mark data 606 and the value 0 of the symbol data 612 is (001, 0, 0001, 0010, 0011, 0101, 0011, 0010).

Next, please refer to FIG. 7 , and finally combine the positive and negative frame coded data and add the frame switch control code 1111 to obtain (1111, 000, 1, 0100, 1000, 0001, 1111, 001, 0). , 0001, 0010, 0101, 0101, 0011, 0010) complete speech data sequence string.

In the foregoing embodiment, by adding a symbol data of one bit to the start code, the symbol bit of the subsequent signed digital code can be omitted. When the number of points per two zero crossing points is increased, the amount of data that can be omitted is increased.

From another point of view, taking 4 bit quantization coding as an example, the coded quantization result of the prior art includes one symbol data for each 4-bit digital code, and the numerical data is only 3 bits. In the present invention, in the 4 bit architecture of the cum, the original 4 bits can be used as the numerical data. In the application of such a look-up table method, the result of the look-up table, that is, the resolution of the audio signal coding can be nearly doubled under the same amount of data, and the quality of the voice signal coding is greatly improved.

The foregoing disclosure is part of an encoding device. Next, please refer to FIG. 8, which illustrates the audio quantization decoding module of the present invention, which can decode the serial voice data encoded by the present invention. The audio quantization decoding module includes a data decoder 410 and an inverse quantizer 420. The data decoder 410 reads the second encoded data string in the second memory 300 and performs decoding to generate a plurality of second decoded data strings, each of the second decoded data strings comprising: a second frame marking material, a second symbol data, a plurality of second digital data; and an inverse quantizer 420, connected to the data decoder, receiving the second decoded data strings, and according to the second frame marking data, the second symbol data The values are inverse quantized by the second digital data to sequentially generate a plurality of digital second speech data, and then stored in the third memory 510.

In practice, the second memory 300 and the third memory 510 may be different blocks in the same memory.

For example, after decoding by the data decoder 410, the speech data sequence string of the foregoing example 1 can be obtained (1111, 000, 1, 0100, 1000, 0001, 1111, 001, 0, 0001, 0010, 0011, 0101, 0011, 0010). ).

Then, the frame switching control code 1111 of the voice data sequence string is first removed, and then the inverse quantizer 420 takes the table 2 of the matching data, and takes the symbol data as 1, indicating that the next digital data will be a negative value, and then Table 2 obtains the index code (4, 8, 1), and finally the mapping table 2 can get the table code (6, 12, 1). Since the symbol data 612 is 1 and is negative, the obtained multiple voice data is (-6, -12, -1). In the same, in the second sequence, the frame switching control code 1111 of the speech data sequence string is first removed, and 001 represents the second quantization table (Table 3), the symbol data 612 is 0, and the index code is (2). 3, 4, 8, 4, 2), and finally corresponding to Table 3, the form code (5, 8, 12, 25, 12, 5) can be obtained individually. Finally, the inverse meter will restore the data to a plurality of first voice data (-5, -12, -1, 5, 8, 12, 25, 12, 5). Finally, a plurality of first voice materials are stored in the third memory 510.

Referring to FIG. 9, the flowchart of the audio quantization coding of the present invention includes the following steps: Step 110: reading a plurality of digits of the first voice data and performing a plurality of zero-crossing point judgments to sequentially generate a plurality of firsts. Symbol data, and cutting the digital first speech data into a plurality of frames.

In step 110, the digitizing the first voice data may be further coded.

Step 120: sequentially receiving the frames and the first symbol data, and quantizing the digital first speech data included in the received frame each time to generate a plurality of first digital data, and according to The quantized results of the frames correspond to generating a plurality of first frame tag data.

Step 130: Receive the first digital data, the first symbol data, and the first frame label data, and encode the plurality of first encoded data strings, each of the first encoded data strings corresponding to the frame The first frame tag data, the first symbol data, and the first digital data corresponding to the frame are included.

The first voice data is obtained by means of a mapping method. The calculation of the zero crossing point is judged by multiplying two consecutive first speech data, and when the two consecutive first speech data are multiplied, the value is negative, indicating that the zero crossing point is established.

Referring to FIG. 10, a flowchart of audio quantization decoding according to the present invention includes the following steps: Step 210: Read a second encoded data string and perform decoding to generate a plurality of second decoded data strings, each of the second The decoded data string includes: a second frame mark data, a second symbol data, and a plurality of second digital materials.

Step 220: Receive the second decoded data strings, and perform inverse quantization of the second digital data according to the second frame marking data and the second symbol data to sequentially generate a plurality of digital second voice data. .

Please refer to FIG. 11 , the flow chart of the audio quantization codec of the present invention includes the following steps: Step 310: Read the digital first speech data and perform a plurality of zero crossing point judgments to sequentially generate a plurality of numbers. a symbol data, and cutting the digital first speech data into a plurality of frames.

In step 310, the digitizing the first voice data may be further coded.

Step 320: sequentially reading the frames and the first symbol data, and quantizing the digital first speech data included in the received frame to generate a plurality of first digital data, and A plurality of first frame tag data are generated corresponding to the quantized results of the frames.

Step 330: Receive the first digital data, the first symbol data, and the first frame label data, and encode the plurality of first encoded data strings, each of the first encoded data strings corresponding to the frame The first frame tag data, the first symbol data, and the first digital data corresponding to the frame are included.

Step 340: Read a second encoded data string and perform decoding to generate a plurality of second decoded data strings, each of the second decoded data strings: a second frame mark data, a second symbol data, and a plurality of Second digital data.

Step 350: Receive the second decoded data strings, and perform inverse quantization of the second digital data according to the second frame mark data and the second symbol data to sequentially generate a plurality of digital second voice data. .

The first digital data and the first frame mark data are generated by using the one or more quantization tables to perform lookup by using the first voice data. The second voice data is generated by using the one or more quantization tables to perform the lookup table by using the second frame mark data, the second symbol data, and the second digital data. The zero crossing point is calculated by multiplying two consecutive first speech data to a negative value, and determining that the zero crossing point is established. The first symbol data and the second symbol data represent a positive value or a negative value.

Next, please refer to the speech codec system diagram of the present invention, as shown in FIGS. 12A and 12B, which are respectively codec embodiments of the speech encoder using the 5A and 5B diagrams. In the embodiment of FIGS. 12A and 12B, the audio quantization codec device uses the memory (including the first memory 110, the second memory 300, and the third memory 510) for signal encoding and decoding, the first memory. The 110 record has a plurality of digits of the first voice data, and the second memory 300 records the second coded data string. The codec includes a signal divider 250, a quantizer 230, a data encoder 240, a data decoder 410, and an inverse quantizer 420.

The signal divider 250 reads the plurality of digits of the first voice data and performs a plurality of zero-crossing point determinations to sequentially generate the plurality of first symbol data, and cuts the plurality of digits of the first voice data into a plurality of frames. . The quantizer 230 is connected to the signal divider 250, and receives a plurality of digital first voice data and a first symbol data corresponding to each frame, and receives a plurality of digital first voice data corresponding to each received frame. After the quantization, a plurality of first digital data are generated, and a first frame mark data is generated according to the result of the frame quantization. The data encoder 240 is connected to the quantizer 230 and the signal divider 250. The receiving quantizer 230 generates a plurality of first digital data, first symbol data and first frame label data for each frame and encodes the first encoded data into a first encoded data. string. The data decoder 410 is connected to the second memory 300, and reads the second encoded data string and decodes it to generate a plurality of second decoded data strings. Each second decoded data string includes: a second frame mark data, a second symbol data, and a plurality of second digital materials. The inverse quantizer 420 is connected to the data decoder 410, receives the second decoded data string, and performs inverse quantization of the plurality of second digital data according to the values of the second frame marking data and the second symbol data to sequentially generate a plurality of digits. After the two voice data, the data is stored in the third memory 510.

In practice, the first memory 110, the second memory 300, and the third memory 510 may be different blocks among the same memory.

While the preferred embodiment of the invention has been described above, it is not intended to limit the invention, and it is obvious to those skilled in the art that the invention may be modified and modified without departing from the spirit and scope of the invention. The patent protection scope of the invention is subject to the definition of the scope of the patent application attached to the specification.

100. . . Voice input signal

110. . . First memory

200. . . Speech encoder

210. . . Analog digital converter

220. . . Minus converter

230. . . Quantizer

240. . . Data encoder

250. . . Signal splitter

251. . . Register

300. . . Second memory

400. . . Speech decoder

410. . . Data decoder

420. . . Inverse quantizer

500. . . Voice output signal

510. . . Third memory

600. . . Digital code

606. . . Frame marker data

612. . . Symbolic data

614. . . Digital data

624. . . First encoded data string

626. . . First encoded data string

700. . . Audio codec

Figure 1 is a diagram of a prior speech coding and decoding system (prior art);

Figure 2A is a first embodiment of a functional block diagram of a prior speech coder (prior art);

Figure 2B is a second embodiment of the functional block diagram of the prior speech coder (prior art);

Figure 3 is a previous analog digital sampling map (prior art);

Figure 4 is a diagram of a previously concatenated speech data (previous technique);

Figure 5A is a first embodiment of a functional block diagram of a speech coder of the present invention;

Figure 5B is a second embodiment of the functional block diagram of the speech coder of the present invention;

Figure 6A is a third embodiment of the functional block diagram of the speech coder of the present invention;

Figure 6B is a fourth embodiment of the functional block diagram of the speech coder of the present invention;

Figure 7 is a diagram showing an embodiment of the serial voice data of the present invention;

Figure 8 is a functional block diagram of the speech decoder of the present invention;

Figure 9 is a flow chart of the audio quantization coding of the present invention;

Figure 10 is a flow chart of the audio quantization decoding of the present invention;

Figure 11 is a flow chart of the audio quantization codec of the present invention;

Figure 12A is a first embodiment of a functional block diagram of a speech codec of the present invention; and

Figure 12B is a second embodiment of the functional block diagram of the speech codec of the present invention.

110. . . First memory

200. . . Speech encoder

230. . . Quantizer

240. . . Data encoder

250. . . Signal splitter

300. . . Second memory

Claims (24)

An audio quantization codec device uses a memory for signal encoding and decoding. The memory records a plurality of digital first speech data and a second encoded data string, and includes: an audio quantization coding module, including: a signal a splitter, connected to the memory, reading the digital first speech data and performing a plurality of zero crossing point judgments to sequentially generate a plurality of first symbol data, and according to the distribution of the first symbol data The digital first voice data is cut into a plurality of frames, and each of the digital first voice data contained in the frame belongs to the same positive half cycle or negative half cycle, and has the same first symbol data; a quantizer And connecting the signal divider to receive the digital first voice data corresponding to each of the frames and the first symbol data, and each of the digits of the first voice data corresponding to the received frame Quantizing and generating a plurality of first digital data, and calculating the quantization error according to the digital first speech data belonging to the same positive half cycle or negative half cycle corresponding to the frame a first frame marking data; and a data encoder connecting the quantizer and the signal divider, and receiving the first digital data, the first symbol data generated by the quantizer for each frame The first frame marking data is encoded into a first encoded data string; and an audio quantization decoding module includes: a data decoder connected to the memory, reading the second encoded data string and decoding to generate a plurality of a second decoded data string, each of the second decoded data strings includes: a second frame mark data, a second symbol data, and a plurality of second digital data; and an inverse quantizer connected to the data decoder to receive The second decoded data string is subjected to inverse quantization of the second digital data according to the second frame mark data and the second symbol data to sequentially generate a plurality of digital second voice data. The audio quantization codec device of claim 1, further comprising: a down code converter, coupled between the memory and the signal divider, and performing the code reduction operation on the first voice data. The audio quantization codec device of claim 1, further comprising: a down code converter connected between the memory and the signal splitter, and connected to the quantizer to subtract the first voice data The action of the code. The audio quantization and codec device of claim 1, wherein the first encoded data string corresponds to the frame and includes the first frame mark data, the first symbol data, and the first corresponding to the frame Digital data. The audio quantization and codec device of claim 1, wherein the digits of the first voice data corresponding to the frame read by the quantizer are read from the memory. The audio-quantity codec device of claim 1, wherein the signal divider comprises a register, the register is configured to store the digitized first voice data corresponding to the frame, and the quantizer is The digitized first voice data corresponding to the read frame is read from the register of the signal divider. The audio quantization and codec device of claim 1, wherein the quantizer uses one or more quantization tables to find the first speech data corresponding to each frame by looking up the table and calculating the quantization error. The quantization table corresponding to the minimum quantization error is obtained, and the first digital data is obtained according to the quantization table. The audio quantization and codec device of claim 1, wherein the inverse quantizer uses the one or more quantization tables to include the second frame marker data, the second symbol data, and the second symbol data included in each of the decoded data strings. The second digital data corresponding to the quantization table corresponding to the second frame mark data corresponding to the second voice data. The audio quantization codec device of claim 1, wherein each of the zero crossing points is calculated The two consecutive first speech data are multiplied, and when they are negative, the zero crossing point is established. The audio quantization and codec device of claim 1, wherein the first symbol data and the second symbol data represent a positive value or a negative value. An audio quantization coding method includes: reading a plurality of digital first speech data and performing a plurality of zero crossing point judgments to sequentially generate a plurality of first symbol data, and according to the distribution of the first symbol data The digital first voice data is cut into a plurality of frames, so that the digital first voice data included in each frame belongs to the same positive half cycle or negative half cycle, and has the same first symbol data; The digitized first voice data and the first symbol data corresponding to the frame, and quantizing the digitized first voice data corresponding to the frame received each time to generate a plurality of first digital data And correspondingly generating a first frame mark data according to the calculation of the quantization error corresponding to the digital first voice data belonging to the same positive half cycle or negative half cycle corresponding to the frame; and receiving corresponding to each of the The first digital data, the first symbol data and the first frame label data are generated by the frame and encoded into a first encoded data string. The method as claimed in claim 11, wherein the first encoded data string corresponds to the frame and includes the first frame mark data, the first symbol data, and the first numbers corresponding to the frame. data. The audio quantization coding method of claim 11, further comprising: performing a code reduction operation on the digital first speech data. The audio quantization coding method of claim 11, wherein the digital data and the frame mark data are generated by using the one or more quantization tables to perform a lookup table using the first voice data. The audio quantization coding method according to claim 11, wherein the zero-crossing point judgment determines that the zero-crossing point is established by multiplying two or more first speech data by a negative value. An audio quantization decoding method includes: reading a second encoded data string and decoding to generate a plurality of second decoded data strings, each of the second decoded data strings comprising: a second frame mark data, a second a symbol data, a plurality of second digital data, each of the second decoded data strings belonging to the same positive half cycle or negative half cycle; and receiving the second decoded data string, and according to the second frame marking data, the first The value of the two symbol data is subjected to inverse quantization of the second digital data to sequentially generate a plurality of digits of the second voice data. The audio quantization decoding method of claim 16, wherein the second voice data is generated by using one or more quantization tables to perform table lookup by using the frame mark data, the symbol data, and the digital data. An audio quantization codec method includes: reading the digital first speech data and performing a plurality of zero crossing point judgments to sequentially generate a plurality of first symbol data, and according to the distribution of the first symbol data The digital first speech data is cut into a plurality of frames, such that the digital first speech data included in each frame belongs to the same positive half cycle or negative half cycle, and has the same first symbol data; Each of the digitized first voice data corresponding to the frame and the first symbol data, and each of the digitized first voice data corresponding to the received frame is quantized to generate a plurality of first Digital data, and corresponding to the calculation of quantization error corresponding to the digital first data of the same positive half cycle or negative half cycle corresponding to the frame, corresponding to generating a first frame mark data; receiving corresponding to each The first digital data generated by the frame, the first symbol data and the first frame label data are encoded into a first encoded data string; a second encoded data string is read and decoded A plurality of second decoding data a string, each of the second decoded data strings comprising: a second frame mark data, a second symbol data, a plurality of second digital data; and receiving the second decoded data strings, and according to the second frame The value of the mark data and the second symbol data is subjected to inverse quantization of the second digital data to sequentially generate a plurality of digital second voice data. The method for encoding and decoding an audio code according to claim 18, wherein the first encoded data string corresponds to the frame and includes the first frame mark data, the first symbol data, and the first corresponding to the frame. Digital data. The audio quantization and coding and decoding method of claim 18, further comprising: performing a code reduction operation on the digital first speech data. The method of claim 18, wherein the first digital data and the first frame marking data use the one or more quantization tables to include the first voices included in each of the frames. The data is obtained by looking up the table and calculating the quantization error, and the quantization table corresponding to the minimum quantization error is obtained, and the first digital data is obtained according to the quantization table. The audio quantization and coding and decoding method of claim 18, wherein the second voice data uses the one or more quantization tables to include the second frame marker data and the second symbol included in each of the decoded data strings. The data and the second digital data are generated according to a corresponding one of the quantization tables corresponding to the second frame mark data. The audio quantization codec method of claim 18, wherein each of the zero crossing points is calculated by multiplying two consecutive first speech data, and when it is a negative value, the zero crossing point is established. The audio quantization decoding method of claim 18, wherein the first symbol data and the second symbol data represent a positive value or a negative value.