CN1741392A - The method and apparatus that data are encoded and deciphered - Google Patents

Abstract

Methods and apparatus for encoding and decoding data are provided. The means for encoding data includes: a quantizer for quantizing data input to it into a predetermined bit quantization; a DPCM operator for generating differential data by performing a DPCM operation on the quantized data; an entropy encoding A device for entropy coding the differential data input from the DPCM operator and outputting the entropy coded differential data into a bit stream; a quantization error minimizer for receiving input data and quantized data, and adjusting the input data and a header encoder for encoding the maximum and minimum values to be included in the bitstream.

Description

Method and apparatus for encoding and decoding data

This application is a divisional application of an invention patent application with a filing date of November 27, 2002, an application number of 02154294.5, and an invention title of "Method and Device for Encoding and Decoding Data".

technical field

The present invention relates to an apparatus and method for encoding and decoding data, and more particularly, to a differential pulse code modulation (DPCM) operation, a method of normalizing data, and a method using the DPCM operation and The method for standardizing data is a method and device for encoding and decoding data.

Background technique

FIG. 1 is a block diagram showing the structure of a conventional device for encoding data and a conventional device for decoding data. Referring to FIG. 1 , a conventional apparatus for encoding data includes a quantizer 100 , a DPCM operator 110 and an entropy encoder 120 . According to a conventional method for encoding data, in the quantizer 100 , input data is quantized with a predetermined number of bits and input to the DPCM operator 110 . The DPCM operator 110 generates differential data and outputs the differential data to the entropy encoder 120 by subtracting the current quantized data input from the quantizer 100 from the previously quantized data. The entropy encoder 120 encodes the differential data into a bit stream using a predetermined entropy encoding method.

Referring to FIG. 1 , a conventional apparatus for decoding data includes an entropy decoder 130 , an inverse DPCM operator 140 and an inverse quantizer 150 . The bit stream in which data is encoded is input to the entropy decoder 130 . The entropy decoder 130 reversely performs the processing performed by the entropy encoder to output differential data, and then the inverse DPCM operator 140 converts the differential data input from the entropy decoder 130 into quantized data and outputs the quantized data to inverse quantization device 150. The inverse quantizer 150 inverse quantizes the quantized data input from the inverse DPCM operator 140, and then outputs decoded data.

However, since the aforementioned conventional method for encoding data only performs simple DPCM operations, the amount of data to be encoded is reduced, and this method encodes data having a plurality of components without considering the The characteristics of each of the x, y, and z components, especially when the data is fed in continuously, it cannot be encoded efficiently enough.

Contents of the invention

SUMMARY OF THE INVENTION In order to solve the above and other problems, it is an object of the present invention to provide a DPCM operation and a DPCM operator capable of significantly reducing the size of data to be encoded.

Another aspect of the present invention is to provide a method and apparatus for encoding and decoding data that minimize quantization errors generated during the quantization process and that encode the sign of the data.

Another aspect of the present invention is to provide a method and apparatus for encoding and decoding data using a method of normalizing data by normalizing the x, y and z components according to one component Normalizing data per component can reduce the size of data with multiple components.

Another aspect of the present invention is to provide a method and apparatus for encoding data, wherein the DPCM operation, error minimization method and data normalization method according to the present invention are applied for encoding and decoding position interpolation A method for key value data of a device, the key value data is used to represent the position of an object appearing in a three-dimensional animation, and a method and device for decoding a bit stream, which has been decoded Encoding is performed according to methods and apparatus for encoding data.

Therefore, in order to achieve the above and other aspects of the present invention, there is provided an apparatus for generating differential data, the apparatus comprising: a cyclic DPCM operator for performing a DPCM operation on quantized data to generate differential data, for differential performing a cyclic quantization operation on the data in order to reduce their range, and then outputting cyclically quantized differential data; a predictive cyclic DPCM operator for performing a predictive DPCM operation on the quantized data in order to generate predicted differential data, for which the predicted differential data undergoes a cyclic quantization operation in order to reduce their range, and then outputs cyclic quantized predicted differential data; and a selector for selecting the cyclic quantized differential data and the predicted differential data according to the value of the differential data input to it one.

In order to achieve the above and other aspects of the present invention, a method for generating differential data is provided. The method includes: (a) generating differential data by performing a DPCM operation on the quantized data, generating predicted differential data by performing a predictive DPCM operation on the quantized data, (b) performing a cyclic quantization operation on the differential data and the predicted differential data, The cyclic quantization difference data and the cyclic quantization prediction difference data are generated so as to reduce their ranges, and (c) one of the cyclic quantization difference data and the cyclic quantization prediction difference data is selected according to their magnitudes.

In order to achieve the above and other aspects of the present invention, there is provided an apparatus for generating quantized data using differential data. The apparatus comprises: an inverse cyclic DPCM operator for performing an inverse cyclic quantization operation on differential data input thereto in order to extend their range, performing an inverse DPCM operation on the result of the inverse cyclic quantization operation, and then outputting quantized data; an inverse predictive cyclic DPCM operator for performing inverse cyclic quantization operations on differential data input thereto to extend their range, performing inverse predictive DPCM operations on the results of the inverse cyclic quantization operations, and then outputting quantized data; and a determination unit, Used to output the differential data to the inverse loop DPCM operator or the inverse predictive loop DPCM operator according to the type of DPCM that has been performed on the differential data.

In order to achieve the above and other aspects of the present invention, there is provided a method of generating quantized data using differential data. The method includes: (a) identifying the kind of DPCM that has been performed on the input differential data, (b) performing an inverse loop quantization operation on the input differential data in order to extend their range, and (c) if the input differential data has been performed If the DPCM operation is performed, the quantized data is generated by performing the reverse cycle DPCM operation on the reverse cycle quantized differential data, and if the input differential data has been subjected to the predictive DPCM operation, the reverse cycle DPCM is performed on the reverse cycle quantized differential data operations to generate quantitative data.

In order to achieve the above and other aspects of the present invention, an apparatus for encoding data is provided. The apparatus includes: a quantizer for quantizing data input to it into predetermined quantization bits; a DPCM operator for generating differential data by performing a DPCM operation on the quantized data; an entropy encoder for Entropy encoding is performed on the differential data input to it from the DPCM operator, and the entropy-encoded differential data is output as a bit stream; a quantization error minimizer is used to receive input data and quantized data, and adjust the maximum of the input data and minimum values, thereby minimizing quantization errors and outputting the maximum and minimum values; and a header encoder for encoding the maximum and minimum values to be included in the bitstream.

In order to achieve the above and other aspects of the present invention, a method for encoding data is provided. The method comprises: (a) quantizing input data by using predetermined quantization bits; (b) generating differential data by performing a DPCM operation on the quantized data; (c) entropy encoding the differential data generated in step (b) , generating a bitstream; (d) adjusting the maximum and minimum values in the input data using the input data and the quantization data, thereby minimizing the quantization error; and (e) encoding the maximum and minimum values to be included in the bitstream .

In order to achieve the above and other aspects of the present invention, there is provided an apparatus for encoding data having a plurality of components. The device includes: a normalizer for calculating the maximum range in the data range of the components, and normalizing the data of each component according to the maximum range; a quantizer for normalizing the normalized data with predetermined quantization bits performing quantization; a DPCM operator for performing a DPCM operation on the quantized data and then outputting differential data; and an entropy encoder for performing entropy encoding on the differential data and outputting a bitstream into which the differential data has been encoded.

To achieve the above and other aspects of the present invention, there is provided a method for encoding data for encoding data having a plurality of components. The method includes: (a) calculating the maximum range among the data ranges of the components, and normalizing the data of each component according to the maximum range; (b) quantizing the normalized data by using predetermined quantization bits; (c) quantizing performing a DPCM operation on the data to generate differential data; and (d) generating a bitstream into which the data has been encoded by performing entropy encoding on the differential data.

To achieve the above and other aspects of the present invention, there is provided an apparatus for decoding data that decodes a bit stream into which data having a plurality of components is encoded. The device includes: an entropy decoder for entropy decoding the bit stream input to it and outputting differential data; an inverse DPCM operator for performing inverse DPCM operations on differential data and outputting quantized data; an inverse a quantizer configured to perform inverse quantization on the quantized data and output normalized data; and a denormalizer configured to receive a minimum value among each component data and a maximum value among each component data from the bitstream to obtain normalized data The maximum range of , and denormalize the canonical data based on the maximum range and minimum value.

To achieve the above and other aspects of the present invention, there is provided a method for decoding data that decodes a bit stream into which data having a plurality of components is encoded. The method includes: (a) generating differential data by performing entropy decoding on an input bit stream; (b) generating quantized data by performing an inverse DPCM operation on the differential data; (c) generating a specification by performing inverse quantization on the quantized data data; and (d) using the minimum value among each component data and the maximum value among each component data decoded from the bitstream to obtain the maximum range of the normalized data, and based on the maximum range and the minimum value, the normalized data Execute the reverse norm.

To achieve the above and other aspects of the invention, there is provided means for encoding data for key value data of a position interpolator representing the position of an object having x, y and z components to encode. The apparatus includes: a normalizer for calculating a maximum range among data ranges of the components, according to the maximum range, normalizing the key value data of each of the x, y and z components and outputting the normalized key value data; a quantizer to quantize the canonical key value data with predetermined quantization bits; a floating-point encoder to receive the minimum and maximum ranges used in the canonical operation and convert the input value to a decimal number ; a DPCM processor for obtaining differential data and predicted differential data of the quantized key value data, and performing cyclic quantization operations on the differential data and predicted differential data, thereby reducing their range; an entropy encoder for differential the data performs entropy encoding and outputs a bitstream into which the key value data is encoded; and a key value header encoder for encoding the information required to decode the bitstream to be included in the bitstream to encode.

To achieve the above and other aspects of the invention, there is provided a method for encoding data for key value data of a position interpolator representing the position of an object having x, y and z components to encode. . The method comprises: (a) calculating a maximum range among the data ranges of the components, normalizing key value data for each of the x, y, and z components based on the maximum range, and then generating normalized key value data ; (b) quantize the canonical key value data using predetermined quantization bits; (c) convert the minimum and maximum ranges used in the canonical operation into decimal numbers; (d) obtain the difference data of the quantized key value data and predicted differential data, and perform a cyclic quantization operation on the differential data and the predicted differential data, thereby reducing their range; (e) generate a bitstream into which the key value data is encoded by performing entropy encoding on the differential data; and ( f) Encoding the information to be included in the bitstream required for decoding the bitstream.

In order to achieve the above and other aspects of the present invention, there is provided an apparatus for decoding data for a key value of a position interpolator that will represent the position of an object with x, y and z components The bit stream into which the data is encoded is decoded. The device includes: a key value header decoder for decoding information required for decoding from the bit stream; an entropy decoder for entropy decoding the bit stream and outputting differential data ; an inverse DPCM processor for performing an inverse cyclic DPCM operation or an inverse predictive cyclic DPCM operation on differential data according to the type of DPCM performed on the differential data, and thereby outputting quantized key value data; an inverse quantizer, Used to dequantize quantized key value data and output canonical key value data; a floating point decoder for receiving each of the x, y, and z components from the key value header decoder The minimum value among the key value data of the component and the maximum range among the data ranges of the x, y, and z components, convert the minimum value and maximum range into a binary number, and output the binary number to the denormalizer; and an inverse A normalizer for receiving a minimum value among key value data of each of the x, y, and z components and a maximum range among data ranges of the x, y, and z components from the floating-point number decoder, and The key value data for the x, y, and z components are denormalized.

To achieve the above and other aspects of the present invention, there is provided a method for decoding data for a key value of a position interpolator that will represent the position of an object with x, y and z components The bit stream into which the data is encoded is decoded. The method comprises: (a) decoding header information required for decoding from the bit stream; (b) generating differential data by entropy decoding the bit stream; (c) generating differential data by performing The type of DPCM, perform the reverse cycle DPCM operation or the reverse prediction cycle DPCM operation on the differential data, generate quantized key value data; (d) convert the x, y and z components that have been decoded in step (a) Convert the minimum value among the key value data of each component of and the maximum range among the data ranges of the x, y, and z components into binary numbers, and convert the key value data of each of the x, y, and z components The minimum value among them and the maximum value among all key value data of the x, y and z components are converted into binary numbers; (e) by performing inverse quantization on the quantized key value data according to predetermined quantization bits, generating a canonical key value data; and (f) the data of the x, y and z components based on the minimum value among the key value data of each of the x, y and z components that have been converted in step (d) The largest range of ranges to denormalize the key value data.

Description of drawings

By describing in detail preferred embodiments of the present invention in conjunction with the accompanying drawings, the above-mentioned purposes and advantages of the present invention will become clearer, wherein:

The block diagram of Fig. 1 shows the structure that is used for encoding the conventional apparatus of data and is used for the structure of the conventional apparatus of decoding data;

The block diagram of Fig. 2A shows the apparatus for performing DPCM operation according to the preferred embodiment of the present invention, and Fig. 2B shows the flowchart of the DPCM operation according to the preferred embodiment of the present invention;

The block diagram of Fig. 3 A shows, according to the preferred embodiment of the present invention, is used to utilize DPCM operator to encode the device of data, and Fig. 3 B shows, is used for the device of encoding data shown in Fig. 3 A Operation flow chart;

Figure 4A is a block diagram illustrating an apparatus for encoding data that minimizes quantization errors, according to a preferred embodiment of the present invention. Figure 4B shows a flowchart of a method of minimizing quantization error;

Figure 5A is a block diagram illustrating an apparatus for encoding data according to a preferred embodiment of the present invention, which normalizes input data having multiple components and then encodes the normalized data, and Figure 5B A flowchart illustrating a method of normalizing data;

Figure 6A is a block diagram showing an apparatus for encoding data according to a preferred embodiment of the present invention, which is used to encode key value data for a position interpolator, and Figure 6B shows an apparatus for encoding key value data using A flowchart of a method for encoding key value data;

7A and 7B show a flowchart of the operation of the floating-point number encoder shown in FIG. 6A;

7C and 7D show a flowchart of the operation of an entropy decoder according to a preferred embodiment of the present invention;

FIG. 8 shows step S683 shown in FIG. 7C;

Figures 9A and 9B illustrate the steps shown in Figure 7D;

Figure 10A is a block diagram showing an inverse DPCM operator according to a preferred embodiment of the present invention; and Figure 10B shows a flow chart of inverse DPCM operations;

The block diagram of Fig. 11 shows the apparatus for decoding data using an inverse DPCM operator according to a preferred embodiment of the present invention;

The block diagram of Fig. 12A shows an apparatus for decoding data using a denormalizer according to a preferred embodiment of the present invention, and Fig. 12B shows an operation flowchart of the denormalizer;

The block diagram of Fig. 13A shows the apparatus for decoding key value data according to a preferred embodiment of the present invention, and Fig. 13B shows the flow of the method for decoding key value data picture;

FIG. 14A shows a detailed flowchart of step S1320 shown in FIG. 13B;

FIG. 14B shows a detailed flowchart of step S1330 shown in FIG. 14A;

FIG. 14C shows a detailed flowchart of step S1380 shown in FIG. 13B;

Fig. 15 shows the component order of the bit stream that is input to the entropy decoder according to the preferred embodiment of the present invention;

Figures 16A and 16B show the results of predictive DPCM operations according to a preferred embodiment of the present invention;

Figures 17A to 17C show the results of cyclic DPCM operations according to a preferred embodiment of the present invention;

Figures 18A and 18B show the key and key value data for the position interpolator; and

19 to 27 show examples of program codes by which the process of reading a bit stream in the method for decoding data according to a preferred embodiment of the present invention can be realized.

Detailed ways

The invention will be more fully described below with reference to the accompanying drawings.

FIG. 2A is a block diagram showing an apparatus for performing DPCM operations according to a preferred embodiment of the present invention, and FIG. 2B shows a flowchart of DPCM operations according to a preferred embodiment of the present invention.

With reference to Fig. 2 A, according to the preferred embodiment of the present invention, be used for carrying out the apparatus of DPCM operation and comprise: a cyclic DPCM manipulator 200, it comprises the DPCM manipulator 210 that is used to carry out general DPCM operation, output differential data then and input A cyclic quantization operator 220 that performs a cyclic quantization operation on its differential data in order to reduce their range; a predictive cyclic DPCM operator 230 that includes a predictive DPCM operation for performing a predictive DPCM operation on the quantized data input to it, and then outputs a predictive differential data A predictive DPCM operator and a cyclic quantizer 250 for performing a cyclic quantization operation on the differential data input to it in order to reduce their range; One differential data is selected between the differential data input from the DPCM operator 200 and the differential data input from the prediction cycle DPCM operator 230 .

The operation of DPCM according to the preferred embodiment of the present invention will be described below with reference to FIG. 2B.

In step S210 , the data quantized by any one of the quantization methods is input to the loop DPCM operator 200 and the predictive loop DPCM operator 230 .

In the loop DPCM operator 200 , the DPCM operator 210 generates differential data by performing a general DPCM operation on the quantized data input thereto, and outputs the generated differential data to the loop quantizer 220 . Then, the loop quantizer 220 performs loop quantization on the differential data input from the DPCM operator 210 in step S220a.

Specifically, the DPCM operator 210 calculates difference data by subtracting the previous quantization data from the current quantization data input in step S210. Thereafter, the loop quantizer 220 performs loop quantization on the calculated differential data. The operation of the cyclic DPCM operator 200 is represented by the following equation.

${\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; p</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; Circular\; Quantization</span>}<span\; class="notranslate">\; (</span>{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; p</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; -</span>{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; p</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1,2</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

In Equation (1), n represents the number of data. Loop quantization will be described later.

The predicted DPCM operator 240 in the predicted loop DPCM operator 230 calculates predicted differential data by performing a predicted DPCM operation on the differential data input thereto, and then the loop quantizer 250 performs loop quantization on the predicted differential data in step S220b.

Specifically, the predictive DPCM operator 240 generates and outputs predicted difference data by subtracting the predicted data from the current quantized data. In order to obtain prediction data for the current quantized data, a difference value is obtained by subtracting the quantized data before the previous quantized data from the previous quantized data, and then adding the difference value to the previous difference data, thereby Calculate forecast data for current data.

The predicted data calculated by the predicted DPCM operator 240 may not exceed the maximum range of the quantized data input to it. In other words, if the predicted data exceeds the maximum value in the quantization range of the input quantized data, then the predicted DPCM operator 240 will set the predicted data to the maximum value in the quantized range of the input quantized data and subtract Current data, generate differential data for the current data. If the predicted data is smaller than the minimum value in the quantization range of the input data, the predicted DPCM operator 240 determines the current data as the predicted difference data. Hereinafter, the operation of the predictive DPCM operator 240 will be referred to as "modified predictive DPCM operation".

The predicted difference data calculated by the predicted DPCM operator 230 is input to the loop quantizer 250, and a loop quantization operation is performed.

The operation of the predictive cyclic DPCM operator 230 can be represented by the following equation.

$\stackrel{\widehat{~}}{p}=(2^{\mathrm{nQuantBit}}-1)-{\tilde{p}}_i$ (if $2\&times;{\tilde{p}}_{i-1}-{\tilde{p}}_{i-2}>2^{\mathrm{nQuantBit}}-1$ ) …(2)

${\stackrel{\widehat{~}}{p}}_i={\tilde{p}}_i$ (if $2\&times;{\tilde{p}}_{i-1}-{\tilde{p}}_{i-2}<0$ )

${\stackrel{\widehat{~}}{p}}_i={\tilde{p}}_i-(2\&times;{\tilde{p}}_{i-1}-{\tilde{p}}_{i-2})$ (otherwise)

${\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; p</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; Circular\; Quantization</span>}<span\; class="notranslate">\; (</span>{\stackrel{\stackrel{<span\; class="notranslate">\; ^</span>}{<span\; class="notranslate">\; ~</span>}}{\mathrm{<span\; class="notranslate">\; p</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; )</span>$

In Equation (2), i is an integer between 2 and n-1, and n represents the number of data.

Figure 16A shows the results of performing a simple predictive DPCM operation on 50 quantized data with a maximum value of 1024 and a minimum value of 0, and Figure 16B shows the modified The result of the DPCM operation is predicted so that the differential data for the data is adjusted according to equation (2) in case the predicted data for the data exceeds the maximum value among the data or falls below the minimum value among the data.

As a result of the simple predictive DPCM operation, the predictive differential data has a range approximately 3,000 wide, as shown in FIG. 16A. As a result of the modified predictive DPCM operation, the predictive difference data has a narrow range of not more than 2,000, as shown in FIG. 16B. This means that it is easier to obtain predicted differential data with a narrower range in modified predictive DPCM operation than in simple predictive DPCM operation.

The loop quantization will be described below.

FIG. 17A shows quantized data, and FIG. 17B shows the result of performing a DPCM operation on the quantized data shown in FIG. 17A. As shown in FIG. 17B, the range of differential data after the DPCM operation can be increased to twice as wide as before the DPCM operation is performed. Therefore, it can be said that using the cyclic quantization operation makes it possible to perform the DPCM operation while keeping the range of differential data on which the DPCM operation is to be performed within the range of the input data.

The cyclic quantization operation is performed on the assumption that the maximum value and the minimum value within the quantization range are cyclically connected to each other. Therefore, if differential data as a result of performing linear DPCM on two consecutive quantized data is larger than half of the maximum value in the quantization range, their values can be reduced by subtracting the maximum value from each of the differential data.

If the differential data is less than half of the minimum value in the quantization range, their values can be further reduced by adding the maximum value in the quantization range to each of the differential data.

分别表示输入差分数据和循环量化的差分数据，则循环量化可以由下列等式表示。Suppose X and

denote the input differential data and the differential data of cyclic quantization respectively, then the cyclic quantization can be expressed by the following equation.

$\mathrm{<span\; class="notranslate">\; Circular\; Quantization</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; Min</span>}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; ,</span><span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span>$

X′ _i ＝X _i -(2 ^nQBits -1) (if X≥0)

X′ _i ＝X _i +(2 ^nOBits -1) (otherwise)

In Equation (3), nQBits represents a bit size for quantization. Fig. 17C shows the result of loop quantization performed on the differential data shown in Fig. 17B.

Referring again to FIG. 2B, the circularly quantized differential data and the circularly quantized predicted differential data are output to the selector 260, and then the selector 260 performs a sum of absolute difference (SAD) operation on the input differential data in step S230. Here, the SAD operation is performed so that the absolute values of the input data are all added. In step S230, the selector adds the absolute values of all cyclically quantized differential data and adds the absolute values of all cyclically quantized predicted differential data.

Thereafter, the selector 260 compares the sum of the absolute values of the quantized differential data with the sum of the absolute values of the predicted differential data, and selects differential data having a smaller sum of absolute values from between the quantized differential data and the predicted differential data And output the selected differential data in step S240. The reason why the selector 260 selects differential data having a smaller absolute value sum is that they presumably have a narrower range of values. For example, in the process of entropy coding the predicted differential data, it can be expected that the number of bits required to encode the differential data with a smaller absolute value sum is less than the number of bits required to encode the differential data with a larger absolute value sum number of bits.

It is obvious to those skilled in the art that the SAD operation is just one of many methods for selecting a certain set of differential data, thus, the selector 260 can select differential data using different methods.

Fig. 3A is a block diagram showing a device for encoding data according to a first embodiment of the present invention, and Fig. 3B shows a device for encoding data according to a first embodiment of the present invention Flow chart of the operation of the device. An apparatus for encoding data according to a first embodiment of the present invention includes the aforementioned DPCM operator of the present invention.

The apparatus for encoding data shown in FIG. 3A includes: a quantizer 310 for quantizing input data with predetermined quantization bits; a DPCM processor 340 for performing cyclic DPCM operations and prediction on the quantized data and an entropy encoder 350 for performing a predetermined entropy encoding operation on the differential data input from the DPCM processor 340, and then outputting a bit stream. The DPCM processor 340 includes the aforementioned DPCM operator of the present invention.

3A and 3B, in step S310, the data to be encoded is input to a device for encoding data, and in step S320, is quantized in a quantizer 310 using a predetermined number of quantization bits and is Output to DPCM processor 340.

The loop DPCM operator 200 in the DPCM processor 340 performs a DPCM operation and loop quantization operation on the quantized data input from the quantizer 310 and then outputs the result to the selector 260 . In addition, the predictive loop DPCM operator performs a predictive DPCM operation and loop quantization on the quantized data input from the quantizer 310 in step S330.

The selector 260 performs a SAD operation on the differential data input from the cyclic DPCM operator 200 and the differential data input from the predictive cyclic DPCM operator 230, selects the differential data to be output to the entropy decoder 350, and converts Selected differential data output.

The entropy encoder 350 performs a predetermined entropy encoding operation on the differential data input from the selector 260, thereby entropy encoding them and generating a bit stream in step S350.

FIG. 4A is a block diagram showing an apparatus for encoding data according to a second embodiment of the present invention, and FIG. 4B is a flowchart showing the operation of the quantization error minimizer shown in FIG. 4A.

The apparatus for encoding data according to the second embodiment of the present invention includes maximum and minimum values for minimizing quantization errors in a bit stream, and encodes the bit stream so that when encoding data The quantization error can be minimized during decoding.

Referring to FIG. 4A, the apparatus for encoding data according to a preferred embodiment of the present invention includes a quantizer 310, a DPCM operator 110/340, an entropy encoder 350, a quantization error minimizer 320, and a header encoder 370, wherein , the quantization error minimizer 320 is used to receive data and quantized data, adjust the maximum and minimum values in the data so as to minimize the quantization error, and output the adjusted maximum and minimum values, and the header encoder 370 is used to convert the quantized The maximum and minimum values input by the error minimizer 320 are encoded into header information, and this information is included in the bitstream.

The quantization error minimizer 320 includes: an initial value setting unit 321 for setting a minimum error value e _min , an adjusted minimum value min' and a quantization error minimum value min _min ; an adjusted minimum value update unit 323 for for updating the adjusted minimum value min' by performing a predetermined operation; a determination unit 325 for determining the quantization error minimum value min _min to be used for The minimum value of inverse quantization; an error value update unit 327, which is used to calculate the quantization error value by using the updated and adjusted minimum value, and if the calculated error value is less than the minimum error value e _min , the adjusted minimum value min' and the calculated error values are respectively updated to a quantization error minimum value min _min and a minimum error value e _min , and the updated results are output to the adjusted minimum value update unit 323 .

The quantization error minimizer 320 controls the quantization range by performing quantization and inverse quantization on the input data so as to minimize the quantization error.

逆量化值

以及误差e_i进行计算。Specifically, when Max represents the fixed maximum value to be used for quantization, Min represents the adjusted minimum value to be used for quantization, _Xi represents the input value, and nQuanBit represents the number of bits used for quantization, then using etc. Equation (4) quantifies the input value

inverse quantization value

And the error e _i is calculated.

${\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; floor</span>}<span\; class="notranslate">\; (</span>\frac{{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; Min</span>}}{\mathrm{<span\; class="notranslate">\; Max</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; Min</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; *</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; 2</span>}}^{\mathrm{<span\; class="notranslate">\; nQuantBit</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 0.5</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span>$

${\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; =</span>\frac{{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; *</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; Max</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; Min</span>}<span\; class="notranslate">\; )</span>}{{\mathrm{<span\; class="notranslate">\; 2</span>}}^{\mathrm{<span\; class="notranslate">\; nQuantBit</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; Min</span>}$

${\mathrm{<span\; class="notranslate">\; e</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; -</span>{\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}}$

In order to reduce the error sum Σe _i , the quantization error minimizer 320 controls Min so that the error sum Σe _i can be minimized and the quantized data can be quantized using the minimum value that can minimize the quantization error in the process of decoding the data. to decode.

The quantizer 310, DPCM operator 110/340, and entropy encoder 350 are almost the same as the corresponding elements of the aforementioned means for encoding data, so they will not be repeated here. The operation of quantization error minimizer 320 is described more fully below in connection with FIG. 4B.

The initial value setting unit 321 receives the same data as the data input to the quantizer 310 and the quantized data output from the quantizer 310 in step S402.

The initial value setting unit 321 obtains the maximum value max and the minimum value min from among the input data and calculates the number of digits of the minimum value min. Next, the initial value setting unit 321 calculates and sets a quantization error e as an initial minimum error value e _min by inversely quantizing the quantized data using the maximum value max and the minimum value min in step S404 .

The initial value setting unit 321 subtracts the result of dividing the quantization step size QuantSpace by 2 from the minimum value among the input data and sets the result of the subtraction as the initially adjusted minimum value min′. Then, the initial value setting unit 321 sets the initially adjusted minimum value min' as the quantization error minimum value m _min and outputs the adjusted minimum value min' to the adjusted minimum value updating unit 323 in step S406.

The adjusted minimum value updating unit 323 updates the adjusted minimum value min' input from the initial value setting unit 321 by performing a predetermined operation. According to a preferred embodiment of the present invention, the adjusted minimum value update unit 323 updates the adjusted minimum value min' input from the initial value setting unit 321 according to equation (5) and in step S408 the updated and adjusted minimum The value min′ is output to the determination unit 325 .

${\mathrm{<span\; class="notranslate">\; min</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; Mantissa</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; min</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{{\mathrm{<span\; class="notranslate">\; 10</span>}}^{\mathrm{<span\; class="notranslate">\; Exponent</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; min</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; )</span>}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 5</span>}<span\; class="notranslate">\; )</span>$

In step S410, the determination unit 325 calculates the number of digits of the updated and adjusted minimum value min', compares the calculated result with the number of digits of the minimum value min among the input data calculated by the initial value setting unit 321, and The updated and adjusted minimum value min' is compared with the result of adding the minimum value min among the input data to the result obtained by dividing the quantization bit size QuantSpace by 2.

则确定单元325将当前所存储的量化误差最小值min_min确定为将被用于逆量化的最小值，并将该量化误差最小值min_min输出给输出单元329。如果经过更新和调节的最小值min’的数字数不大于输入数据当中的最小值min的数字数并且经过更新和调节的的最小值min’不大于

则确定单元325将经过更新和调节的最小值min’输出给误差值更新单元327。If the number of digits of the updated and adjusted minimum min' is greater than the number of digits of the minimum min' in the input data, or if the updated and adjusted minimum min' is greater than

Then the determining unit 325 determines the currently stored minimum value of the quantization error min _min as the minimum value to be used for inverse quantization, and outputs the minimum value of the quantization error min _min to the output unit 329 . If the number of digits of the updated and adjusted minimum value min' is not greater than the number of digits of the minimum value min in the input data and the updated and adjusted minimum value min' is not greater than

Then the determining unit 325 outputs the updated and adjusted minimum value min′ to the error value updating unit 327 .

In step S414, the error value updating unit 327 dequantizes the quantized data using the updated and adjusted minimum value min' and the maximum value max obtained by the initial value setting unit 321, and calculates a quantization error e.

In step S446, the error value update unit 327 compares the newly calculated error value e with the minimum error value e _min . As a comparison result, if the newly calculated error value e is smaller than the minimum error value e _min , then in step S418, the error value update unit 327 updates the minimum error value e _min with the newly calculated error value e, and uses the updated and adjusted The minimum value min' updates the quantization error minimum value min _min . On the other hand, if the newly calculated error value e is greater than the minimum error value e _min , the error value updating unit 327 will execute step S408 again without updating the minimum error value e _min and the minimum quantization error value min _min .

The maximum and minimum values for minimizing the quantization error obtained through the above processing are output to the header encoder 370 and encoded as header information. Then, the header information is included in the bitstream generated by the entropy encoder 350 .

Fig. 5A is a block diagram showing an apparatus for encoding data according to a third embodiment of the present invention, and Fig. 5B is a flowchart showing the operation of the normalizer shown in Fig. 5A. The apparatus for encoding data according to the third embodiment of the present invention improves the efficiency of encoding data by normalizing input data having a plurality of components.

As shown in FIG. 5A, the apparatus for encoding data according to the third embodiment of the present invention includes: a normalizer 300 for calculating the maximum value among the data ranges of the components input to the apparatus for encoding data range, and normalize the data of each of the x, y and z components according to the maximum range; the quantizer 310 is used to quantize the normalized data with a predetermined number of quantization bits; the DPCM operator 110/340, for performing a DPCM operation on the quantized data and outputting differential data; and an entropy encoder 350 for performing entropy encoding on the differential data and outputting an encoded bit stream.

The quantizer 310, DPCM operator 110/340 and entropy encoder 350 are the same as the corresponding elements of the apparatus for encoding data according to the first embodiment of the present invention, so detailed descriptions about them will not be repeated here. The operation of the normalizer 300 is described below only in conjunction with FIG. 5B.

In the case where the data to be encoded includes multiple components, for example, if the input data includes three components x, y, and z representing the position of the object in three-dimensional space, each of the x, y, and z components The components are input to a normalizer 300 of the means for encoding data. Then, the normalizer 300 calculates the maximum range among the ranges of the x, y, and z components, and normalizes the data of each of the x, y, and z components according to the maximum range.

In step S512, the normalizer 300 receives the data of each component and calculates the data ranges of the x, y and z components according to equation (6). Then, in step S514, the normalizer 300 calculates the maximum range among the data ranges of the x, y, and z components by comparison.

${\mathrm{<span\; class="notranslate">\; Max</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; Max</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}^{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; (</span>{\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{\mathrm{<span\; class="notranslate">\; p</span>}}}_{\mathrm{<span\; class="notranslate">\; ij</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 6</span>}<span\; class="notranslate">\; )</span>$

${\mathrm{<span\; class="notranslate">\; Min</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; Min</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}^{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; (</span>{\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{\mathrm{<span\; class="notranslate">\; p</span>}}}_{\mathrm{<span\; class="notranslate">\; ij</span>}}<span\; class="notranslate">\; )</span>$

Range _max =Max(Max _x -Minx, Max _y -Min _y , Max _z -Min _z )

Thereafter, in step S516, the normalizer 300 performs normalization using the data of each of the x, y, and z components of the following equations.

In Equations (6) and (7), n represents the number of data, i=0, 1, . . . , n−1, and j represents each of x, y, and z components. The x, y, and z components are normalized using the largest range, Range _max , among their data ranges, as shown in equation (7). As a result of this specification, the redundancy of data concerning components not having the maximum range Range _max increases, so that the efficiency of encoding the data is improved.

The normalizer 300 normalizes each of the input components, and outputs the normalized result to the quantizer 310, and then the quantizer 320 quantizes the normalized data input from the normalizer 300 using a predetermined number of quantization bits, and The quantized results are output to the DPCM operator 110/340. Thereafter, the DPCM operator 110/340 generates differential data by performing a DPCM operation on the quantized data input from the quantizer 310, and outputs the differential data to the entropy encoder 350, thereby being able to encode them.

The following will describe an apparatus for encoding data according to a fourth embodiment of the present invention, in which the aforementioned DPCM operator of the present invention, an apparatus capable of minimizing quantization errors for encoding data according to the present invention are integrated The aforementioned means and the aforementioned means for encoding data according to the invention are capable of encoding data having a plurality of components.

Although the apparatus for encoding data according to the fourth embodiment of the present invention is specifically described as encoding key value data of position interpolators among three-dimensional animation data, it is obvious that it can also encode other general type of data to encode.

A position interpolator is information about representing an animation path in keyframe-based animation, which is one of the basic techniques for representing computer-synthesized 3D animation. Keyframe-based animations are used to specify the sequence of animations based on when the key data occurs. A frame corresponding to the instant at which key data appears is called a key frame.

In keyframe-based animation, multiple keyframes and multiple frames between keyframes constitute an animation, and these frames are interpolated between keyframes (or use interpolation to generate frame).

International multimedia standards, such as MPEG-4 Binary Format for Scenes (BIFS) and Virtual Reality Modeling Language (VRML), support keyframe-based animation using interpolator nodes. In MPEG-4BIFS and VRML, there are various interpolators including scalar interpolator, positional interpolator, coordinate interpolator, directional interpolator, standard interpolator and color interpolator. These interposers and their functions and characteristics are shown in Table 1.

Table 1 Type of interposer feature Function scalar interpolator Linear Interpolation of Scalar Variables Capable of representing area, diameter, and density position interpolator Linear interpolation on 3D coordinates Parallel motion in 3D space directional interposer Linear interpolation of 3D coordinate axes and rotation Rotation in 3D space coordinate interpolator Linear interpolation of variables in 3D coordinates 3D variant Standard Interposer Linear interpolation of standard 3D coordinates Ability to represent variables in standard 3D vectors color interpolator Linear interpolation of color information Variables that can be represented in colors

The position interpolator shown in Table 1 is used to represent the position information of the keyframe-based animation and includes key and key value fields. The key field represents the position of each keyframe on the time axis with a discrete number ranging from -∞ to ∞. Each of the key value fields specifies information on the position of the target at the time indicated by each key, and includes three components of x, y, and z. Each of the key value fields includes the same number of key values as each of the key fields has.

Examples of position interpolators are shown in Figures 18A and 18B. Specifically, FIG. 18A shows key data, and FIG. 18B shows key value data.

Linear interpolation has been used in MPEG-4 BIFS and VRML. However, in order to express animation smoothly and naturally using linear interpolation, a considerable amount of key data and key value data is required. In addition, in order to store and transmit such animations, a large-capacity memory and a large amount of time are required. Therefore, it is preferable to compress the interpolators for storage and transmission of these interpolators.

In MPEG4-BIFS, a method of encoding and decoding an interpolator node called predictive MF coding (PMFC) is used. In the PMFC method, similar to the operation of the conventional apparatus for encoding data shown in FIG. 1, the key value data of the position interpolator is encoded using a quantizer, a DPCM operator, and an entropy encoder. The quantizer and DPCM operator reduce redundant key value data, and the output of the DPCM operator is input to the entropy encoder. However, in the PMFC method, entropy coding is performed on differential data obtained by the DPCM operator, so its coding efficiency is not high enough. Also, due to the limitations entropy encoders have, it is nearly impossible to provide high-quality animations.

Therefore, the present invention provides a device for encoding a key value, in which device is integrated the aforementioned DPCM operator according to the invention, the aforementioned DPCM operator for encoding data according to the invention enabling quantization error minimization The apparatus and the aforementioned apparatus for encoding data according to the invention are capable of encoding data having a plurality of components, thereby increasing the efficiency of encoding key value data.

FIG. 6A is a block diagram showing an apparatus for encoding key value data according to a fourth embodiment of the present invention, and FIG. 6B is an operation flowchart of the apparatus for encoding key value data.

Referring to FIG. 6A, the apparatus for encoding key value data includes: a normalizer 300 for normalizing the input key value data for each component according to a maximum range among the data ranges of the x, y and z components Quantizer 310 is used to quantize the key value data of the norm using predetermined quantization bits; quantization error minimizer 320 is used to receive the minimum value and the maximum range from the normalizer 300, and adjust and output the maximum and minimum value, so that the quantization error can be minimized; the floating-point number encoder 330 is used to receive the minimum value and the maximum range from the quantization error minimizer 320, and convert the maximum range and the minimum value for minimizing the quantization error into a decimal number; DPCM processor 340, is used for obtaining the difference data of quantization key value data and prediction difference data and carries out circular quantization operation so that reduce the range of difference data; a bit stream into which the key value data is encoded; and a key value header encoder 360 for encoding information required to decode the bit stream and enabling the information to be included in the bit stream.

The operation of the apparatus for encoding key value data according to the present invention will be described below with reference to FIG. 6B.

In step S600, the key value data of each of the x, y, and z components is input to the normalizer 300 of the apparatus for encoding key value data. Then, in step S610, the normalizer 300 calculates the maximum range among the data ranges of the x, y, and z components, and normalizes the input key value data of the x, y, and z components according to the maximum range. The process of normalizing the key value data input to the normalizer 300 has been described above, and will not be repeated here.

The key value data of the standardized x, y and z components are input to the quantizer 310, and then in step S620, the quantizer 310 utilizes a predetermined number of quantization bits nQuanBit, according to equation (8) to the normalized key value data carried out.

${\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; p</span>}}}_{\mathrm{<span\; class="notranslate">\; ij</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; floor</span>}<span\; class="notranslate">\; (</span>{\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{\mathrm{<span\; class="notranslate">\; p</span>}}}_{\mathrm{<span\; class="notranslate">\; ij</span>}}<span\; class="notranslate">\; \&times;</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; 2</span>}}^{\mathrm{<span\; class="notranslate">\; nQuantBit</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 0.5</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 8</span>}<span\; class="notranslate">\; )</span>$

In Equation (8), floor() is a function for converting an input floating-point number into the largest integer not larger than the input floating-point number. The quantizer 310 outputs quantized key value data to the DPCM processor 340 and the quantization error minimizer 320 .

In step S630, the quantization error minimizer 320 receives the minimum value and the maximum range from the normalizer 300, calculates the minimum value and the maximum value using the minimum value and the maximum range input from the normalizer 300, and determines a minimum value and a maximum value so that The quantization error is minimized, a minimum value and a maximum range are calculated using the determined minimum and maximum values, and output to the floating point converter 330, thereby encoding them into a key value header. Here, the minimum and maximum values determined by the quantization error minimizer 320 are used to minimize the quantization error.

Step S650 performed by the floating-point number encoder 330 is described below in conjunction with FIG. 7A .

In step S651, the floating-point number encoder 330 receives the minimum value x_min, y_min, z_min, the maximum value max of the component having the largest range, the maximum value max about the maximum value max, among the key value data of each of the x, y, and z components. The information nWhichAxis belonging to which component and the number of digits nKeyValueDigit of the original key value data.

In order to improve the efficiency of encoding key value data by reducing the number of bits required for encoding, the floating point encoder 330 converts x_min, y_min, z_min, and max represented by binary numbers into decimal numbers in step S652.

Computers store floating-point numbers as 32-bit binary numbers. When a floating point number of the binary system is input, the floating point number encoder 330 converts the floating point number into a mantissa of the decimal system and its exponent according to equation (9).

For example, the floating point number 12.34 in the decimal system can be converted by the computer into binary as shown below.

$\frac{\mathrm{<span\; class="notranslate">\; 0</span>}}{\mathrm{<span\; class="notranslate">\; 1</span>}}\frac{\mathrm{<span\; class="notranslate">\; 10001010111000010100011</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}\frac{\mathrm{<span\; class="notranslate">\; 10000010</span>}}{\mathrm{<span\; class="notranslate">\; 3</span>}}$

1: sign

2: mantissa in binary system

3: Exponent in binary system

This binary number can be converted back to the original decimal number according to equation (9) as shown below.

$\frac{\mathrm{<span\; class="notranslate">\; 0</span>}}{\mathrm{<span\; class="notranslate">\; 1</span>}}\frac{\mathrm{<span\; class="notranslate">\; 1234</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}\frac{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; 3</span>}}$

1: sign

2: mantissa in decimal system

3: The exponent in the decimal system

In order to include the mantissa and exponent of the decimal system in the bitstream, the number of bits required to represent the mantissa and the number of bits required to represent the exponent must be calculated. The exponent has a value between -38 and 38, whereby the exponent can be represented using 7 bits together with a sign. The number of bits required to represent a mantissa depends on the number of digits in that mantissa. The following table shows various ranges of values and the number of bits required to represent the mantissas of the various ranges of values.

Table 2 range of values mantissa number of digits the number of bits required to represent the mantissa 0 0 0 1-9 1 4

10-99 2 7 100-999 3 10 1000-9999 4 14 10000-99999 5 17 100000-999999 6 20 1000000-9999999 7 twenty four

The floating point number encoder 330 checks whether the number of digits of x_min, y_min and z_min are the same in step S453. If they are not the same, they are output to the key value header encoder 360 and encoded into key value headers in step S654.

If the number of digits of x_min, y_min and z_min are the same, then in step S655, the floating point number encoder 330 checks whether they are the same number of digits as the original key value data. If the number of digits of x_min, y_min and z_min is different from the number of digits of the original key value data, then, in step S656, one of the digits of x_min, y_min and z_min is output to the key value header encoder 360 and encoded is the key title.

Thereafter, the floating-point number encoder 330 checks in step S657 from which component the maximum value max among all the key value data of the x, y, and z components is selected, and determines whether the number of digits of max is the same as The number of digits of the minimum value among the key value data of the components for which the maximum value max is obtained is the same.

If the number of digits of max is different from the number of digits of the minimum value among the key value data of the components that obtain the maximum value max, the number of digits of max is output to the key value header encoder 360 and encoded as a key value header . On the other hand, if the number of digits of max is the same as the number of digits of the minimum value among the key value data of the component to which the maximum value max belongs, then in step S658, state information describing that they are identical is output to the key value header encoder 360.

The aforementioned operations of the floating-point number encoder 330 are used to encode the mantissa of the converted floating-point number. The following describes the process of encoding information related to the exponent of the floating-point number into the key value header.

In step S659, the floating-point number encoder 330 identifies which one is the maximum value among the absolute values of the exponents of x_min, y_min, z_min, and max, stores the identified maximum value as nMaxExp, and encodes nMaxExp into several bits of encryption in the key-value header.

Thereafter, the floating-point number encoder 330 checks whether the signs of the exponents of x_min, y_min, z_min, and max are the same in step S660. If they are the same, bSameSignExp is set to 1 and their signs are output to the key value header encoder 360 in step S661. On the other hand, if they are not the same, bSameSignExp is set to 0 in step S662, and then the encoding of the exponent-related information of x_min, y_min, z_min, and max is completed.

Referring to FIG. 7B , the floating-point number encoder 330 encodes the sign of the converted floating-point number in step S663 before encoding x_min, y_min, z_min, and max into the key value header.

Next, in step S664, the floating-point number encoder 330 refers to Table 2 to calculate the number of bits required to encode the floating-point number, and outputs the mantissa of the floating-point number with the same number of bits as the calculation result to the key value header encoder 360 .

In step S665 , the floating-point number encoder 330 calculates the number of bits required to encode nMaxExp with reference to Table 2, and outputs the exponent of the floating-point number having as many bits as the calculation result to the key value header encoder 360 .

The floating point encoder 330 checks whether x_min, y_min, z_min and max have the same sign according to bSameSignExp. If they do not have the same sign, their signs are output to the key value header encoder 360 and encoded in step S667, whereby the x, y and z components input from the normalizer 300 are completed The encoding process of the maximum and minimum values among the key data of each component in .

The floating point number encoder 330 receives the maximum and minimum values required to minimize the quantization error from the quantization error minimizer 320, and encodes them into a key header through the above-mentioned steps.

Referring again to FIG. 6B , the DPCM processor 340 performs a cyclic DPCM operation and a prediction cyclic DPCM operation on the quantized key value data and outputs differential data to the entropy encoder 350 in step S670 . The DPCM processor 340 is composed of the aforementioned DPCM operators whose operation and structure have been described above. The difference between the DPCM processor 340 and the aforementioned DPCM operators is only that the DPCM processor 340 performs round-robin DPCM operations and predictive round-robin DPCM operations on key value data for each of the x, y, and z components, respectively.

In step S680 , the differential data output from the DPCM processor 340 is encoded in the entropy encoder 350 .

Referring to FIG. 7C , the entropy encoder 350 entropy-encodes differential data of key value data for each of x, y, and z components.

The entropy encoder 350 checks whether quantization values of the x, y, and z components are identical with reference to differential data of each of the x, y, and z components, and if they are identical, entropy encoding is completed in step S681. For example, in the case of a three-dimensional animation, a train moves in the same horizontal direction as the direction of the x component, and the key value data of the x component is changed. However, the y and z components are almost unchanged. Therefore, it is estimated that the y and z components will have the same value if they are quantized. Therefore, it is not necessary to encode all the key value data that are actually the same, whereby the efficiency of encoding can be improved by encoding only one of the key value data into the key value header.

If the quantization key value data of each of the x, y, and z components is different, in step S682, the entropy encoder 350 reads out a predetermined value that has been set in advance. Thereafter, the entropy encoder 350 encodes the differential data of the quantized key value data of each of the x, y, and z components with the unary AAC function in step S683, or encodes them with the continuous quantized AAC function in step S685. coding.

The unary AAC function is described below in conjunction with FIG. 8 . The unary AAC function converts the symbol to be encoded into multiple bits consisting of a series of 0s, a flag bit 1 indicating the end of the series of 0s, and a bit representing the sign of the symbol. Here, multiple 0s correspond to the magnitude of the sign. For example, use the unary AAC function to encode 256 into a series of bits consisting of 256 0s, a marker bit 1 indicating the end of the series of 0s, and a 0 indicating the sign of 256, that is, the positive sign. Due to the unary AAC function, the redundancy of the bits representing the symbols to be encoded is increased, improving the efficiency of encoding the symbols.

A method of encoding symbols using the SQ AAC function will be described below with reference to FIGS. 7D, 9A, and 9B. The SQ AAC function encodes symbols continuously updating the quantization range.

FIG. 9A shows a method for encoding symbols using the SQ AAC function. As shown in FIG. 9A, the input symbol to be encoded is 1, and the minimum and maximum values in the encoding range are 0 and 9, respectively.

Referring to Fig. 9A, in the first step of encoding the symbol, the encoding range is divided into two subfields, i.e. upper bound and lower bound, and then, it is checked whether the symbol to be encoded, i.e. 1, belongs to the upper bound or the lower bound . Since 1 belongs to the lower bound, 0 is encoded and the lower bound is updated to a new encoding range. Therefore, the new encoding range for the second step is from 0 to 4.

In the second step, the new coded range 0 to 4 is divided into two subfields, an upper bound and a lower bound, and then, it is checked whether 1 belongs to the upper bound or the lower bound. Since 1 belongs to the lower bound, 0 is encoded, and the maximum value among the encoding ranges is updated to 1 which is the maximum value in the lower bound. Therefore, the new encoding range for the third step is from 0 to 1 that was originally used as a lower bound.

In the third step, the coding range of 0 to 1 is divided into an upper bound and a lower bound, and then it is checked whether 1 belongs to an upper bound. Since 1 is equal to the upper limit value, 1 is encoded, and then the minimum value in the corresponding encoding range is updated to 1. Therefore, a new encoding range is [1, 1], whereby its maximum and minimum values are identical to each other. When the minimum and maximum values in the encoding range are the same, encoding processing using the SQ AAC function is completed.

The flowchart of FIG. 7D shows the encoding process of the differential data of the key value data of the components using the SQ AAC function, which is performed in the entropy encoder 350 according to the present invention.

7D, in step S692, the entropy encoder 350 receives differential data of the key value data (hereinafter referred to as symbols), the number of bits nQP required for entropy encoding of the symbols, and the symbols that will be encoded as bit streams. The first subscript of nStartIndex and the number (n) of symbols to be encoded.

The entropy encoder 350 encodes the sign of the sign and converts the sign into a positive number in step S694. The reason for converting the sign to a positive number is that negative numbers are not allowed in subsequent processing.

The entropy encoder 350 identifies the maximum value among the positive numbers in step S696, stores the maximum value as nMax, and encodes nMax into as many bits as nQP.

和最大值 分别设置为0和nMax。另外，熵编码器350将一个比特的标记

分派到符号中的每一个符号。标记 被用于在对符号编码进行期间改变概率模型。开始时将标记 设置为‘假’并且当下一个将被编码的值是1时将其转换成‘真’，然后，到目前为止已经被用于对符号编码进行的概率模型被其他的模型代替。The entropy encoder 350 initializes the range in which the symbols will be encoded, and the minimum value in the encoded range

and the maximum Set to 0 and nMax respectively. In addition, the entropy encoder 350 marks a bit

Dispatched to each of the symbols. mark is used to change the probability model during encoding of symbols. mark at the beginning Set to 'false' and convert it to 'true' when the next value to be encoded is 1, then the probability model that has been used to encode symbols so far is replaced by another model.

Next, the entropy encoder 350 identifies a differential reference point (i=nStartIndex) for each of the x, y, and z components in step S700, assuming it is encoded first, and sets the flag bDone to "true", which indicates that Whether the encoding of all symbols is complete.

The entropy encoder 350 repeatedly performs steps S702 to S718, which will be described below, until the maximum and minimum values in the encoding range for each of the symbols become the same, that is, until there are no remaining symbols to be encoded. Here, when the maximum and minimum values in the encoding range for one symbol are the same, it can be considered that the encoding of the symbol is completed.

如果 等于

方法移动到用于对下一个符号进行编码的步骤。如果不等，则在步骤S704中将bDone设置为‘假’，这意味着对当前符号(当前正在被编码的符号)的编码还没有完成。Entropy encoder 350 checks in step S702 Is it equal to

if equal

The method moves to the step for encoding the next symbol. If not, set bDone to 'false' in step S704, which means that the encoding of the current symbol (the symbol currently being encoded) has not been completed.

The entropy encoder 350 calculates an intermediate value nMid in the encoding range for the current symbol at step S706 and compares the intermediate value nMid with the value of the current symbol at step S708. In a preferred embodiment of the invention, the intermediate value nMid is calculated according to the following equation:

$\mathrm{<span\; class="notranslate">\; Mid</span>}<span\; class="notranslate">\; =</span>\frac{{\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; P</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; \&Center\; Dot;</span>\mathrm{<span\; class="notranslate">\; max</span>}}<span\; class="notranslate">\; +</span>{\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; P</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; \&CenterDot;</span>\mathrm{<span\; class="notranslate">\; <figure-callout\; id="2"\; label="min"\; filenames="A200510092099.300701.png,A200510092099.300711.png"\; state="\{\{state\}\}">min</figure-callout></span>}}}{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; +</span>{\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; P</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; \&Center\; Dot;</span>\mathrm{<span\; class="notranslate">\; min</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 10</span>}<span\; class="notranslate">\; )</span>$

As a result of the comparison, if the current symbol value is not greater than the middle value nMid, which means that the current symbol belongs to the lower bound, then in step S710, 0 is sent to the bitstream, and the middle value nMid is used to replace the maximum range

当当前正被编码的符号的值第一次超过的中间值nmid时，换言之，当第一次将1发给比特流时，在步骤S712中将 设置为‘真’，从而改变用于对符号进行编码的概率模型。On the other hand, if the current symbol value is greater than the middle value nMid, which means that the current symbol belongs to the upper bound, a 1 is emitted to the bitstream and the middle value nMid is substituted for the smallest value in the coding range for the current symbol

When the value of the symbol currently being encoded exceeds the median value nmid for the first time, in other words, when 1 is sent to the bitstream for the first time, in step S712 Set to 'true' to alter the probability model used to encode symbols.

There are two probabilistic models used in the SQ AAC function. One is notFoundContext for entropy encoding the bits output before the first time a 1 is emitted from each symbol, and the other is for entropy encoding the bits output just after the first time a 1 is emitted from each symbol Encoded FoundContext. The purpose of using two different probability models is to increase the probability of generating 0 during the use of noFoundContext. Since the probability of generating 0 increases, the entropy coding efficiency of symbols is improved.

是真还是假。如果 是真，则熵编码器350在步骤S716中使用FoundContext对符号进行编码，否则，熵编码器350在步骤S718中使用notFoundContext对符号进行编码。Thereafter, the entropy encoder 350 checks in step S714

True or false. if If it is true, the entropy encoder 350 uses FoundContext to encode the symbol in step S716, otherwise, the entropy encoder 350 uses notFoundContext to encode the symbol in step S718.

When the encoding process from step S702 to step S718 is completed for the current symbol, it means that the entropy encoder has only completed the first stage of encoding the current symbol. In step S720, the entropy encoder 350 encodes the subscript i is incremented by 1. Next, the entropy encoder 350 checks in step S722 whether encoding of all symbols is completed in the current stage. If there are still symbols to be encoded in the current stage, the entropy encoder 350 performs steps S702 to S722 again.

When the encoding of all symbols in one stage is completed, in step S724, the entropy encoder 350 checks whether the encoding of all stages has been completed according to the flag nDone. If bDone is true, the entropy encoder 350 completes all encoding processes in the current stage and starts encoding the next component, if not, the entropy encoder 350 performs steps S700 to S722 again to encode the remaining symbols.

Figure 9B shows the process of encoding multiple symbols with the SQ AAC function. Specifically, FIG. 9B shows the process of encoding 0, 1, 2, 3, 4, and 9 with the SQ AAC function.

Referring again to FIG. 6B, if the apparatus for encoding data according to the present invention generates a bit stream by entropy encoding all input key value data, then the key value header encoder 360 will encode the encoded key value in step S730. Information necessary for decoding the key value data is encoded into key value header information, the key value header information is added to the bit stream, and the synthesized bit stream is output.

The key value header encoder 360 encodes the number of input key value data to be encoded and the number of digits. Next, the key value header encoder 360 identifies whether each of the x, y, and z components has the same quantization value (for example, even if the key value data of each of the x, y, and z components varies, but Since the key value data of the x component varies little, the quantized key value data of the x component have the same value) and the result is coded into a token.

For example, if the quantization value of the x component is not the same, then whether the key value data of the x component has been encoded using the cyclic DPCM operation or the predictive cyclic DPCM operation, and whether the unary AAC function will be used or the SQ AAC will be used The function encodes the key value data of the x component are encoded into tokens. If it is not desired to encode the key value data of the x component with the unary AAC function, then both the encoded bit size of the x component to be entropy encoded and the starting subscript of the x component are encoded into the key value header.

Also, if the key value data of each of the y and z components are different, then the header information of y and z exactly corresponding to the aforementioned header information of the x component is encoded as the header.

If the quantized value of the x component is different, the key value data of the x component is not encoded with the unary AAC function, and the start subscript of the x component to be entropy encoded is 1, and then, the first quantized key value reference point Coded as an internal fiducial. In the same way, the first quantization key value reference points for each of the y and z components are encoded as internal reference points.

Key value header encoder 360 identifies which of the x, y, and z components has the largest range. The key value header encoder 360 sets the variable nWhichAxis to 0 if the x component has the largest range. The key value header encoder 360 sets nWhichAxis to 1 if component y has the largest range. Key value header encoder 360 sets nWhichAxis to 2 if component z has the largest range. Thereafter, the key value header encoder 360 encodes nWhichAxis, x_min, y_min, z_min, and max into a key header using the floating point number encoder 330 .

A computer program for implementing a method of encoding data according to a fourth embodiment of the present invention will be described below.

The device for encoding data according to the invention encodes the position interpolator into a bit stream, which will be described below. For a better understanding, the traditional program code and variables will be described in the form of SDL language below.

Fig. 19 shows an example of program code for implementing a compressed position interpolator. In Figure 19, the first class for reading a compressed position interpolator is shown. The first class includes PoslKeyValueHeader and PoslKeyValue. PoslKeyValueHeader contains header information for decoding PoslKeyValue, class PoslKeyValue reads key value data about compressed position interpolators from said bitstream. The function qf_start() is used to start an arithmetic decoder before reading out the ACC-encoded part of the bitstream.

Fig. 20 shows the program code of PoslKeyValueHeader. Referring to FIG. 20, nNumKeyValueCodingBit represents the bit size of nNumberOfKeyValue among variables stored in PoslKeyValueHeader, nNumberOfKeyValue represents the number of key value data, and nKVQBit represents the quantization bit size of key value data.

x_keyvalue_flag, y_keyvalue-flag, and z_keyvalueflag state whether x, y, and z components have the same quantization value, and nKVDigit represents the maximum number of most significant bits of each data in key value data. NKVDPCMOrder_X, nKVDPCMOrder_Y, nKVDPCMOrder_Z respectively correspond to the orders of DPCM used for the key value data of each of the x, y, and z components. This flag is set to 0 if DPCM has been performed, and is set to 1 if predictive DPCM has been performed.

Each of blsUnaryAAC_X, blsUnaryAAC_Y, blsUnaryAAC_Z indicates that unary AAC functions have been used during entropy encoding. nKVCodingBit_X, nKVCodingBit_Y and nKVCodingBit_Z specify the coding bit sizes for the x, y and z components, respectively. nStartIndex_X nStartIndex_Y and nStartIndex_Z indicate the start index of each component axis to be encoded. firstKV_X, firstKV_Y, and firstKV_Z represent first quantization key value data of each of the x, y, and z components, respectively, which are encoded as internal data.

Fig. 21 shows the program code of the class KeyValueMinMax. Referring to FIG. 21 , the class KeyValueMin restores the maximum and minimum values that have been used to normalize key value data. bUse32float indicates whether 32-bit floating-point numbers have been used to store the maximum and minimum values. If bUse32Float is 0, the maximum and minimum values used to normalize the key value data are encoded in the IEEE Standard 754 floating point format. Otherwise, encode the maximum and minimum values used to canonicalize the key value data with a floating point encoder.

nWhichAxis represents a component having the largest range among x, y, and z. In this embodiment, if nWhichAxis is 0, the X axis has the maximum range, if nWhichAxis is 1, the Y axis has the maximum range, and if nWhichAxis is 2, the Z axis has the maximum range.

BAllSameMantissaDigitFlag indicates whether the mantissa of the minimum value has the same number of digits among key value data of each of the x, y, and z components, and bSameKVDigitFlag indicates that the mantissa of the minimum value has the same number of digits as nKVDigit. nMantissaDigit_X, nMantissaDigit_Y, and nMantissaDigit_Z indicate the number of digits of the mantissa of the minimum value among the key value data of each of the x, y, and z components, respectively.

BMaxDigitFlag states whether the mantissa of the maximum value has the same number of digits as the minimum value of the component from which the maximum value max is obtained. If the number of digits of the mantissa of the maximum value is different from the number of digits of the mantissa of the minimum value, the number of digits of the mantissa of the maximum value is read from the bitstream.

nMantissaDigit_M represents the number of digits of the mantissa of the maximum value, and nExponentBits represents the number of bits required to encode the maximum absolute exponent among the exponents of the maximum and minimum values. BAllSameExponentSign indicates whether the signs of the x, y and z components are the same. When bAllSameExponentSign is true, nExponentSign indicates the sign of the exponent.

fpnMin_X, fpnMin_Y, fpnMin_Z, and fpnMax represent floating-point numbers decoded in the decimal system. The method of decoding fpnMin_X, fpnMin_Y, fpnMin_Z and fpnMax will be described below using FloatingPointNumber. fMin_X, fMin_Y, and fMin_Z represent the minimum value among the key values of each of the x, y, and z components, and fMax represents the maximum value among the key value data of the component having the largest range.

The program code of the class FloatingPointNumber is described below in conjunction with FIG. 22 . Class FloatingPointNumber represents floating point numbers using the decimal system. nMantissa represents the mantissa of the floating-point number in the decimal system, and nExponent represents the exponent of the floating-point number.

nSign represents the sign of the floating-point number, and nExponentSign represents the sign of the exponent of the floating-point number.

The program code of class PoslKeyValue is described below in conjunction with FIG. 23 . Referring to FIG. 23, among the variables stored in the class PoslKeyValue, keyValue_X, keyValue_Y, and keyValue_Z represent an array of key value data for each of the x, y, and z components in the position interpolator, respectively. If nStartIndex_X is set to 0, keyValue_X[0] is decoded from the bitstream using an arithmetic decoder. If nStartIndex_X is set to 1, keyValue_X[0] is decoded from the key value header decoder. In the same manner, keyValue_Y[0] and keyValue_Z[0] are determined. When arithmetically decoding keyValue_X[0], keyValue_Y[0], and keyValue_Z[0] from the bitstream, use the decodeUnaryAAC or decodeSQAAC function.

The context models kVXSignContext, kVYSignContext and kVZSignContext are used to decode the sign of keyValue_X, keyValue_Y and keyValue_Z, respectively. The context models kVXSignContext, kVYSignContext and kVZSignContext are sent to the decodeUnaryAAC or decodeSQAAC functions.

MaxValueContext, FoundContext and NotFoundContext are used to entropy decode the absolute value of the key value. For example, kVXMaxValueContexr, kVXFoundContext and kVXNotFoundContext are used to decode keyValue_X. MaxValueContext, FoundContext and NotFoundContext are sent to decodeUnaryACC or decodeSQAAC functions.

The context models kVXUCotext, kVYUContext and kVZUContext are used to decode keyValue_X, keyValue_Y and keyValue_Z and sent to the decodeUnaryAAC function.

An inverse DPCM operator that outputs quantized data by restoring differential data generated by the aforementioned DPCM operator will be described below with reference to FIGS. 10A and 10B.

FIG. 10A is a block diagram showing an inverse DPCM operator according to the present invention, and FIG. 10B is a flowchart of an inverse DPCM operation.

Referring to FIG. 10A, the inverse DPCM operator according to the present invention includes: an inverse cyclic DPCM operator 1020 for performing an inverse cyclic quantization operation on the differential data input to it so as to expand their range, and performing an inverse DPCM operation on the input differential data, Quantized data is then output; an inverse prediction loop DPCM operator 1030 for performing an inverse loop quantization operation on the differential data input to it so as to extend their range, performing an inverse prediction DPCM operation on the input differential data, and then outputting quantized data; and The determining unit 1010 is configured to output the differential data to the inverse loop DPCM operator 1020 or the inverse prediction loop DPCM operator 1030 according to the type of DPCM that has been performed on the differential data.

The inverse DPCM operation is described below with reference to FIG. 10B. Referring to FIGS. 10A and 10B , differential data subjected to an inverse DPCM operation is input to a determination unit 1010 . Then, in step S1010, the determination unit 1010 identifies which DPCM has been performed on the input differential data, determines which inverse DPCM will be performed on the input differential data according to the result, and outputs the differential data to the inverse loop DPCM operator 1020 or the inverse prediction Cyclic DPCM operator 1030 .

执行按照下述等式的逆循环量化获得量化的差分数据 If the differential data is input to the reverse loop DPCM operator 1020 and nMax is the maximum value in the quantization range of the input differential data, the reverse loop DPCM operator 1020 passes the input differential data in step S1020

Perform inverse loop quantization according to the following equation to obtain quantized differential data

和将从等式(11)获得的逆循环量化的差分数据

代入等式(12)获得逆DPCM的差分数据的值A和逆DPCM的差分数据的值B。Thereafter, the reverse cycle DOCM operator 1020 inputs the differential data by respectively

and will be the difference data obtained from the inverse loop quantization of equation (11)

Substituting into equation (12) obtains the value A of the differential data of the inverse DPCM and the value B of the differential data of the inverse DPCM.

In Equation (12), n represents the number of data, and i represents an integer between 1 and n-1.

If A is not less than 0 and not greater than nMax, the reverse cycle DPCM operator 1020 outputs A as the data of the reverse cycle DPCM If A is less than 0 or greater than nMax, the inverse loop DPCM operator 1020 outputs B in step S1040.

The inverse prediction loop DPCM operator 1030 uses equation (11) to predict the differential data in step S1020 Perform reverse cycle quantization to obtain predicted difference data for reverse cycle quantization

代入等式(14)计算B。if $P=2\&times;{\tilde{P}}_{i-1}+{\tilde{P}}_{i-2},$ Where P represents the prediction data used for decoding, then the inverse loop DPCM operator 1030 passes Substituting into equation (13) to calculate A and will

Substitute into equation (14) to calculate B.

In equations (13) and (14), n represents the number of data, and i represents an integer between 1 and n-1.

If A is not less than 0 and not greater than nMax, then the reverse cycle DPCM operator 1020 outputs A as the data of the reverse prediction cycle DPCM If A is less than 0 or greater than nMax, the inverse predictive loop DPCM operator 1020 outputs B in step S1040 as

The apparatus for decoding data according to the first embodiment of the present invention will be described below with reference to FIG. 11 . The apparatus for decoding data according to the first embodiment of the invention uses a DPCM operator according to the invention to decode a bit stream in which input data has been used according to the first embodiment of the invention , the device encoding used to encode the data.

11, the apparatus for decoding data according to the first embodiment of the present invention includes an entropy decoder 1120, an inverse DPCM processor 1130 and an inverse quantizer 1140 composed of the aforementioned inverse DPCM operators.

The entropy decoder 1120 generates differential data by entropy decoding the input bit stream and outputs the differential data to the inverse DPCM processor 1130 .

As described above, the inverse DPCM processor 1130 identifies which DPCM is performed on the input differential data, performs an inverse cyclic DPCM operation or an inverse predictive cyclic DPCM operation on the differential data, and outputs quantized data to the inverse quantizer 1140 .

The inverse quantizer 1140 inverse quantizes the quantized data input from the inverse DPCM processor 1130 using predetermined quantization bits and outputs the retrieved data.

The apparatus for decoding data according to the second embodiment of the present invention will be described below with reference to FIGS. 12A and 12B. An apparatus for decoding data according to a second embodiment of the present invention decodes a bit stream in which data having a plurality of components has been normalized using a normalizer according to the present invention.

Referring to FIG. 12A, according to the second embodiment of the present invention, the device for decoding data includes: an entropy decoder 1120 for entropy decoding an input bit stream and outputting differential data; an inverse DPCM operator 140 /1130 is used to perform inverse DPCM operation on the differential data input from the entropy decoder 1120 and output quantized data; the inverse quantizer 1140 is used to inverse quantize and output the quantized data input from the inverse DPCM operator 140/1130 normative data; and a denormalizer 1150 configured to receive the minimum value among the key value data of each component and the maximum value of the component with the largest range, calculate the maximum range of the normative data, and dequantize from the maximum range according to the maximum range The quantized data input by the unit 1140 is denormalized.

The operation and structure of the entropy decoder 1120, the inverse DPCM operator 140/1130 and the inverse quantizer 1140 are the same as those of the corresponding elements of the apparatus for decoding data according to the first embodiment of the present invention, The description will not be repeated here. The operation of the denormalizer 1150 will be described below with reference to FIG. 12B.

The denormalizer 1150 receives in step S1210, decoded from the input bit stream, the maximum value fMax for normalization, and receives the minimum values fMin_X, fMin_Y and fMin_Z of the x, y and z components, respectively, and which component axis has Information about the maximum value nWhichAxis and information about the encoding type bUse32Float.

The denormalizer 1150 identifies from bUse32Float whether fMax, fMin_X, fMin_Y, and fMin_Z are encoded using the IEEE Standard 754 floating-point number format, and if not, the denormalizer 1150 calculates the maximum value Range _max in the maximum range using the following equation. If fMax, fMin_X, fMin_Y, and fMin_Z are coded in 32 bits, the denormalizer 1150 determines fMax as Range _max in step S1220 .

Range _max = fMax-fMin_X (if nWhichAxis = 0)....(15)

Range _max = fMax-fMin_Y (if nWhichAxis = 1)

Range _max = fMax-fMin_Z (if nWhichAxis = 2)

If the Range _max is determined, the denormalizer 1150 normalizes the data of each of the x, y, and z components according to the Range _max in step S1230 using the following equation.

$\begin{array}{c}{\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; P</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; x</span>}}<span\; class="notranslate">\; =</span>{\stackrel{\stackrel{<span\; class="notranslate">\; ^</span>}{<span\; class="notranslate">\; ~</span>}}{\mathrm{<span\; class="notranslate">\; P</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; x</span>}}<span\; class="notranslate">\; \&times;</span>{\mathrm{<span\; class="notranslate">\; Range</span>}}_{\mathrm{<span\; class="notranslate">\; max</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; f\; Min</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; x</span>}\\ {\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; P</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}}<span\; class="notranslate">\; =</span>{\stackrel{\stackrel{<span\; class="notranslate">\; ^</span>}{<span\; class="notranslate">\; ~</span>}}{\mathrm{<span\; class="notranslate">\; P</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}}<span\; class="notranslate">\; \&times;</span>{\mathrm{<span\; class="notranslate">\; Range</span>}}_{\mathrm{<span\; class="notranslate">\; max</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; f\; Min</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; Y</span>}\\ {\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; P</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; z</span>}}<span\; class="notranslate">\; =</span>{\stackrel{\stackrel{<span\; class="notranslate">\; ^</span>}{<span\; class="notranslate">\; ~</span>}}{\mathrm{<span\; class="notranslate">\; P</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; z</span>}}<span\; class="notranslate">\; \&times;</span>{\mathrm{<span\; class="notranslate">\; Range</span>}}_{\mathrm{<span\; class="notranslate">\; max</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; f\; Min</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; Z</span>}\end{array}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0,1</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 16</span>}<span\; class="notranslate">\; )</span>$

In Equation (16), n represents the number of data.

The apparatus and method for decoding data according to the third embodiment of the present invention will be described below with reference to FIGS. 13A and 13B. According to the third embodiment of the present invention, the means for decoding the data decodes the bit stream, and has utilized the method for encoding the data according to the present invention to interpolate the positions representing the objects in the three-dimensional animation The key value data of is encoded into this bitstream.

Referring to FIG. 13A, according to the third embodiment of the present invention, the device for decoding data includes: a key value header decoder 1110, used for decoding information required for decoding from the bit stream; An entropy decoder 1120 is used to entropy decode the bit stream input to it, and then output differential data; an inverse DPCM processor 1130 is used to perform reverse cycle DPCM on the differential data input to it from the entropy decoder 1120 Operation or inverse prediction cycle DPCM operation, then output quantized key value data; inverse quantizer 1140, used to inversely quantize the quantized key value data input from inverse DPCM processor 1130 and output standardized key value data; Floating point number decoder 1160 for receiving the minimum and maximum values from key value header decoder 1110 and converting the minimum value and the component with the maximum range among the key value data of each of the x, y and z components and the denormalizer 1150 for receiving from the floating-point number decoder 1160 the minimum value among the key value data of each component in the x, y and z components and the component with the maximum range The maximum value of , according to the minimum value and maximum value from the floating-point number decoder 1160, calculate the maximum range, and then denormalize the key value data input from the inverse quantizer 1140.

The method for decoding data according to the third embodiment of the present invention will be described below with reference to FIG. 13B . Referring to FIG. 13B , a bit stream of encoded key value data is input to an entropy decoder 1120 and a key value header decoder 1110 . Then, in step S1300, the key value header decoder 1110 decodes the header information required for decoding from the bit stream and outputs the decoded header information to the entropy decoder 1120 and the inverse DPCM processor 1130 and inverse quantizer 1140 .

In step S1320 , the entropy decoder 1120 performs entropy decoding on the bit stream and outputs the differential data to the inverse DPCM processor 1130 .

FIG. 14A is a detailed flowchart of step S1320. Referring to FIG. 14A, the bit stream Pi is input to the entropy decoder 1120 in step S1321. FIG. 15 shows the sequence of encoded key value data input to the entropy decoder 1120 and to be entropy-decoded.

The entropy decoder 1120 checks in step S1322 whether key value data of components, such as x, have the same quantization value. If the key value data of a component have the same quantization value, all symbols of the component are decoded in step S1323, and they are set to the minimum value input from the key value header decoder 1110, such as fMin_X.

If the key value data of the components have different quantization values, the entropy decoder 1120 checks in step S1324 whether the key value data has been encoded using the unary AAC function. If the key value data have been encoded using the unary AAC function, they are decoded using the unary AAC function in step S1325.

The unary AAC function reads 0s from the bit stream until the bit "1" appears, converts the number of 0s into an absolute value, and reads the bit following "1" as the sign of the value, if the bit is "0 ", the sign of the value is positive, and if the bit is "1", the sign of the value is negative, and the decoded value is output.

If the key value data of the component has not been encoded using the unary AAC function, the entropy decoder 1120 decodes the bit stream using the SQ AAC function to encode the key value data of the component into One of the bitstreams, which will be described below.

初始化为“假”。In step S1331, the entropy decoder 1120 decodes the signs of all symbols from the bit stream, decodes the maximum value nMax, and converts the decoding range of all symbols (range from 0 to nMax) and mark

Initialized to "false".

Thereafter, in step S1332, the entropy decoder 1120 determines the key value reference point (i=nStartlndex) to be decoded first and sets bDone to 'true'.

The entropy decoder 1120 performs steps S1333 to S1343 again to decode the symbols.

是否相同。Specifically, in step S1333, the entropy decoder 1120 checks the maximum value in the decoding range and minimum

Is it the same.

和最小值

相同，则在步骤S1334中，将被编码的符号被确定为译码范围内的最大值(或最小值)，如果不是，则在步骤S1335中，熵译码器1120将bDone设置为‘假’，并利用下述等式(17)更新译码范围内的中间值nMid。if max

and minimum

Same, then in step S1334, the symbol to be encoded is determined to be the maximum value (or minimum value) in the decoding range, if not, then in step S1335, the entropy decoder 1120 sets bDone to 'false' , and use the following equation (17) to update the intermediate value nMid in the decoding range.

是否为真，从而确定用于译码的概率模型。如果 是真，则熵译码器1120在步骤S1337中利用FoundContext对比特流进行译码，否则，熵译码器1120在步骤S1338中利用notFoundContext对比特流进行译码。The entropy decoder 1120 checks the context flag in step S1336

True to determine the probability model used for decoding. if If it is true, then the entropy decoder 1120 uses FoundContext to decode the bit stream in step S1337, otherwise, the entropy decoder 1120 uses notFoundContext to decode the bit stream in step S1338.

如果从比特流译码的比特是1，则熵译码器1120在步骤S1341中利用中间值nMid代替译码范围中的最小值

并将上下文标记 设置为‘真’。The entropy decoder 1120 that has decoded a bit from the bit stream checks in step S1339 whether this bit read from the bit stream is 1, if not, the entropy decoder 1120 uses intermediate The value nMid replaces the maximum value in the decoding range

If the bit decoded from the bitstream is 1, the entropy decoder 1120 replaces the minimum value in the decoding range with the intermediate value nMid in step S1341

and mark the context Set to 'true'.

The entropy decoder 1120 adds 1 to the decoding subscript i in step S1342, and checks in step S1343 whether all the key value data of the components in the current stage have been completely decoded, if there are still remaining keys to be decoded in the current stage , execute steps S1333 to S1342.

If all the key value data of the current stage have been decoded, the entropy decoder 1120 checks in step S1344 whether an additional stage is needed to decode the key value data of the current component, and if so, then execute steps S1332 to S1343. The entropy decoder checks that if all key values of the current component have been decoded, the entropy decoder starts decoding the key value data of the next component.

13A and 13B, the differential data of the key value data decoded by the entropy decoder 1120 is input to the inverse DPCM processor 130, and then the inverse DPCM processor 130 performs a reverse cycle DPCM operation or Inverse prediction for cyclic DPCM operations. The inverse cyclic DPCM operation and the predictive cyclic DPCM operation are identical to those of the previously described inverse DPCM operator, except that they are performed separately on the key value data for each of the x, y, and z components.

Inverse DPCM processor 1130 converts the quantized key value data generated by the inverse DPCM operation output to the inverse quantizer 1140, and then, in step S1370, the inverse quantizer 1140 uses the quantized bit nKeyValueQBits input from the key value header decoder 1110 to Perform inverse quantization. when n means

${\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; P</span>}}}_{\mathrm{<span\; class="notranslate">\; ij</span>}}<span\; class="notranslate">\; =</span>\frac{{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; P</span>}}}_{\mathrm{<span\; class="notranslate">\; ij</span>}}}{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; 2</span>}}^{\mathrm{<span\; class="notranslate">\; nKeyValueQBits</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; nStartIndex</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; z</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 18</span>}<span\; class="notranslate">\; )</span>$

The inverse quantization performed in the inverse quantizer 1140 when the number of key value data input from the key value header decoder 1110 can be represented by the following equation.

The inverse quantizer 1140 generates normalized data by performing inverse quantization on the key value data of each of the x, y, and z components, and outputs the normalized data of each of the x, y, and z components to the denormalizer 1150. The key value data.

The denormalizer 1150 denormalizes the normalized key value data input from the dequantizer 1140 using the information input from the floating-point number decoder 1160 .

Inverse normalization of the normalized key value data for each of the x, y, and z components is described below with reference to FIG. 14C.

The floating point decoder 1160 receives the message bUse32Float and checks whether bUse32Float is true. If bUse32Float is true, the floating-point number decoder receives the maximum value fMax for denormalization and the minimum value among the key value data for each of the x, y, and z components in the floating-point number format of IEEE Standard 754 fMin_X, fMin_Y, and fMin_Z. If not, the floating-point number decoder receives the maximum value fMax for denormalization and the minimum values fMin_X, fMin_Y, and fMin_Z among the key value data for each of the x, y, and z components in step S1382, and The information nWhichAxis about which component the maximum value fMax belongs to is received.

The floating-point number decoder 1160 converts fMax, fMin_X, fMin_Y, and fMin_Z expressed in the decimal system into binary numbers according to Equation (9) in step S1384, and outputs the converted values to the denormalizer 1150.

The denormalizer 1150 checks whether fMax, fMin_X, fMin_Y and fMin_Z have been encoded as 32 bits according to bUse32Float. If fMax, fMin_X, fMin_Y and fMin_Z have been coded as 32 bits, then the denormalizer 1150 determines fMax as Range _max , if not, the denormalizer calculates the maximum in the maximum range according to equation (15) in step S1386 Value Range _max .

If Range _max is determined, the denormalizer 1150 denormalizes the key value data of each of the x, y, and z components according to Equation (16) according to Range _max in step S1388.

24 to 27 show examples of program codes executed by the apparatus for decoding data according to the third embodiment of the present invention.

FIG. 24 shows an example of a C++ program code for realizing an inverse cyclic DPCM operation, and FIG. 25 shows an example of a C++ program code for realizing an inverse predictive cyclic DPCM operation.

Both the inverse cyclic DPCM operation and the inverse predictive cyclic DPCM operation include an inverse cyclic quantization procedure for selecting a value in the quantization range.

Examples of C++ program code for implementing entropy decoding (Adaptive Arithmetic Decoder) operations are shown in FIGS. 26 to 31. Here, qf_decode() denotes a function for reading out one bit from a bit stream.

The present invention can be realized by computer readable codes written on a computer readable recording medium. Here, the computer-readable recording medium includes any recording medium that can be read by a computer system. For example, the computer-readable recording medium may be ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical data storage, and carrier wave (transmission via Internet). The computer-readable recording medium can be transferred to a computer system connected through a network, and the computer can read the recording medium in a decentralized manner.

As described above, it can be considered that the DPCM operator and its operations according to the present invention can greatly reduce the size of differential data by performing predictive DPCMA operations and cyclic quantization operations as well as conventional DPCM operations. In addition, by encoding data using the DPCM operation according to the present invention, the efficiency of encoding data can be significantly improved.

In addition, by normalizing data having a plurality of components according to one of x, y, and z components using the normalizer according to the present invention, it is possible to increase the efficiency of encoding data, thereby reducing the size of the data.

In addition, by performing the cyclic quantization operation after performing the DPCM operation and predicting the DPCM operation, the key value data of the position interpolator can be encoded and decoded at a higher compression ratio, thereby reducing the range of differential data to be encoded .

Although the present invention has been described with reference to its preferred embodiments, it will be understood by those skilled in the art that changes may be made in form and detail without departing from the spirit and scope of the invention as defined by the appended claims. various modifications.

Claims (17)

1. one kind is used for data are carried out apparatus for encoding, and this device comprises:

A quantizer, the data-measuring that is used for inputing to it is predetermined bit quantization;

A DPCM operator is used for generating differential data by quantized data being carried out the DPCM operation;

An entropy coder is used for the differential data from the input of DPCM operator is carried out entropy coding, and will outputs to through the differential data of entropy coding in the bit stream;

A quantization error minimizes device, is used for receiving input data and quantized data, regulates the minimum and maximum value in the input data, thereby makes the quantization error minimum, and export minimum and maximum value; And

A heading code device is used for the minimum and maximum value that will be included in bit stream is encoded.

2. device as claimed in claim 1, wherein, described quantization error minimizes device and comprises:

An initial value is provided with the unit, is used for receiving input data and quantized data, by the counting the number of words of minimum value in the middle of calculating input data and the quantization error minimum error values e is set _MinInitial value, and minimum value min ' and quantization error minimum value min through overregulating are set _MinInitial value;

Minimum value updating block through overregulating is used for by the minimum value min ' through overregulating is carried out a predetermined operation minimum value min ' through overregulating being upgraded;

A determining unit is used for according to through the counting the number of words and be worth of the minimum value upgrading and regulate, with quantization error minimum value min _MinBe defined as to be used to the minimum value of re-quantization; And

An error amount updating block is used for according to calculating quantization error value through the minimum value of upgrading and regulating, if the quantization error value of being calculated is less than minimum error values e _Min, then the minimum value through overregulating is updated to quantization error minimum value min _Min, the quantization error value of being calculated is updated to minimum error values e _Min, and with minimum error values e _MinExport to minimum value updating block through overregulating.

3. device as claimed in claim 2, wherein, described initial value is provided with the unit according to the minimum value in result who removes the predetermined quantitative step-length with predetermined constant and the input data, and the initial value of the minimum value through overregulating is set.

4. the device of counting as claim 2, wherein, counting the number of words and importing counting the number of words of data and compare in the minimum value by will inputing to described determining unit through overregulating, and after the value of minimum value compared in the middle of the value of the minimum value through overregulating that will import and the input data, described determining unit was output as current quantization error minimum value will be used to the minimum value of contrary standard.

5. one kind is used for the data with a plurality of components are carried out apparatus for encoding, and this device comprises:

A normalizer is used to calculate the maximum magnitude in the middle of the data area of component, and according to this maximum magnitude the data of each component in the component is carried out standard;

A quantizer is used to utilize predetermined quantization bit to quantizing through specification data;

A DPCM operator is used for quantized data is carried out DPCM operation and output differential data; And

An entropy coder is used for differential data is carried out entropy coding and output with encode into wherein bit stream of differential data.

6. device as claimed in claim 5, wherein, described normalizer calculates the central minimum and maximum value of data of each component in the component, utilize the central maximum magnitude of data area of the minimum and maximum value acquisition component of component, and the data of each component in the component are carried out standard according to this maximum magnitude.

7. one kind is used for device that data are deciphered, is used for the bit stream that the digital coding that will have a plurality of components is advanced is wherein deciphered, and described device comprises:

An entropy encoder is used for the bit stream that inputs to it is carried out entropy decoding and output differential data;

A contrary DPCM operator is used for differential data is carried out contrary DPCM operation and output quantized data;

An inverse quantizer is used for quantized data is carried out re-quantization and exports authority data; And

A contrary normalizer is used to receive the central minimum value of each component data and has the maximum of the component of maximum magnitude, obtain the maximum magnitude of authority data, and this maximum magnitude carries out contrary standard to authority data.

8. device as claimed in claim 7, wherein, described contrary normalizer utilizes the maximum magnitude and the minimum value of the affiliated component of authority data, and authority data is carried out contrary standard.

9. device as claimed in claim 7, wherein, described contrary normalizer carries out contrary standard by authority data being multiply by maximum magnitude and the minimum value of component under the authority data being added on the result of multiplication to authority data.

10. one kind is used for data are carried out Methods for Coding, and this method comprises the steps:

(a) utilize the predetermined quantitative bit that the input data are quantized;

(b) by quantized data being carried out the DPCM operation, generate differential data;

(c) by the differential data that generates in step (b) is carried out entropy coding, generate bit stream;

(d) utilization input data and quantized data are adjusted in the minimum and maximum value in the quantized data, thus making amount error minimum; And

(e) the minimum and maximum value that will be included in the bit stream is encoded.

11. method as claimed in claim 10, wherein, step (d) comprises the steps:

(d1) by utilize input data and quantized data calculate the minimum value of input data and quantization error count the number of words, minimum error values e is set _MinInitial value, and minimum value min ' and quantization error minimum value min through overregulating are set _MinInitial value;

(d2) by the minimum value min ' through overregulating is carried out scheduled operation, upgrade minimum value min ' through overregulating;

(d3) according to through the counting the number of words and size of the minimum value upgrading and regulate, with quantization error minimum value min _MinBe defined as to be used to the minimum value of re-quantization; And

(d4) if in step (d3), also not with quantization error minimum value min _MinBe defined as to be used to the minimum value of re-quantization, then according to minimum value, calculate quantization error value, and the minimum value through overregulating is updated to quantization error minimum value min through upgrading and regulating _Min, the quantization error of calculating is updated to minimum error values, if the quantization error value of being calculated is less than minimum error values e _Min, then move to step (d2).

12. method as claimed in claim 11 wherein, in step (d1), according to the minimum value in the middle of result who removes the predetermined quantitative step-length with predetermined constant and the input data, is provided with the initial value of the minimum value through overregulating.

13. method as claimed in claim 11, wherein, in step (d3), counting the number of words and importing counting the number of words of data of minimum value by will be through overregulating compares, and the value of the minimum value in the middle of the minimum value through overregulating that will import and the input data compares, and current quantization error minimum value is defined as and will be used to the minimum value of re-quantization.

14. one kind is used for data are carried out Methods for Coding, this method is encoded to the data with a plurality of components, and described method comprises:

(a) calculate the maximum magnitude in the data area of component and the data of each component are carried out standard according to this maximum magnitude;

(b) utilize the predetermined quantitative bit that authority data is quantized;

(c) by quantized data being carried out the DPCM operation, generate differential data; And

(d), generate wherein bit stream is advanced in digital coding by differential data being carried out entropy coding.

15. method as claimed in claim 14, wherein, step (a) comprises the steps:

(a1) calculate minimum and maximum value in the middle of the data of each component;

(a2) utilize minimum and maximum value to calculate maximum magnitude in the data area of the data area of component and component; And

(a3) according to maximum magnitude the data of each component are carried out standard.

16. the method that data are deciphered, this method are used for the bit stream that the digital coding that will have a plurality of components is advanced is wherein deciphered, described method comprises the steps:

(a) carry out entropy coding by bit stream, generate differential data input;

(b) by differential data being carried out contrary DPCM operation, generating quantification data;

(c) by quantized data is carried out re-quantization, generate authority data; And

(d) utilize and to decipher from bit stream, the minimum value in the middle of each component data and have the maximum of the component of maximum magnitude obtains the maximum magnitude of authority data, and according to this maximum magnitude authority data is carried out contrary standard.

17. method as claimed in claim 16 wherein, in step (d), is used the contrary standard specification data of minimum value of the affiliated component of described maximum magnitude and authority data.

Images