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

Abstract

The method and apparatus that data are encoded and deciphered is provided.Being used for that data are carried out apparatus for encoding comprises: a quantizer, and 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, be used for the minimum and maximum value that will be included in bit stream is encoded.

Description

The method and apparatus that data are encoded and deciphered

The application be that November 27, application number in 2002 are 02154294.5 the applying date, denomination of invention divides an application for the application for a patent for invention of " method and apparatus that data are encoded and deciphered ".

Technical field

The present invention relates to a kind of apparatus and method that data are encoded and deciphered of being used for, more specifically say, relate to the method for a kind of differential pulse-code modulation (DPCM) operation, a kind of authority data and the method and apparatus that a kind of method of using described DPCM operation and described authority data is encoded and deciphered data.

Background technology

The block diagram of Fig. 1 shows conventional being used for data is carried out apparatus for encoding and the conventional structure that is used for device that data are deciphered.With reference to Fig. 1, conventional being used for carried out apparatus for encoding to data and comprised quantizer 100, DPCM operator 110 and entropy coder 120.According to routine be used for data are carried out Methods for Coding, in quantizer 100, utilize the bit quantization input data of predetermined quantity and it inputed to DPCM operator 110.DPCM operator 110 generates differential data and this differential data is exported to entropy coder 120 by cutting from the data that formerly quantize from the current quantized data of quantizer 100 inputs.Entropy coder 120 uses predetermined entropy coding method, and this differential data is encoded into bit stream.

With reference to Fig. 1, traditional being used for comprises entropy encoder 130, contrary DPCM operator 140 and inverse quantizer 150 to the device that data are deciphered.The bit stream that encodes the data in wherein is transfused to entropy encoder 130.The processing that entropy encoder 130 reverse execution are carried out by entropy coder, the output differential data, then, contrary DPCM operator 140 will be converted to quantized data from the differential data of entropy encoder 130 inputs and this quantized data will be exported to inverse quantizer 150.150 pairs of quantized datas from contrary DPCM operator 140 inputs of inverse quantizer carry out re-quantization, then the data of output decoding.

But, because being used for of aforementioned routine carries out Methods for Coding to data and only carries out simple DPCM operation, to such an extent as to reduced the data volume that to be encoded, and this method is encoded to the data with a plurality of components and is not considered the characteristic of each component in x, the y of data and the z component, particularly when data were imported continuously, its code efficiency was not high enough.

Summary of the invention

In order to solve above-mentioned and other problems, one object of the present invention is exactly that a kind of DPCM operation and DPCM operator that can reduce the size of the data that will be encoded significantly will be provided.

Another aspect of the present invention provides a kind of method and apparatus that is used to encode and data are deciphered, and they can make that the quantization error that generates is minimum and can encode to the sign of data during quantification treatment.

Another aspect of the present invention provides standard is carried out in a kind of utilization to data method, the method and apparatus that data are encoded and deciphered, the method of described authority data can reduce the size of the data with a plurality of components by coming that according to one-component the data of each component in x, y and the z component are carried out standard.

Another aspect of the present invention provides a kind of being used for data is carried out Methods for Coding and device, wherein, be applied to being used to encoding and decipher the method for the key value data of position interpolater according to DPCM of the present invention operation, error minimize method and data standard method, described key value data are used for expressing the position of the target of present three-dimensional animation, a kind of method and apparatus that is used for bit is flow to row decoding also is provided, and this bit stream is by according to being used for that data are carried out Methods for Coding and device is encoded.

Therefore, in order to realize above-mentioned and other aspects of the present invention, a kind of device that is used to generate differential data is provided, this device comprises: a circulation DPCM operator, be used for quantized data is carried out the DPCM operation,, differential data carried out the circulation quantization operation so that reduce their scope so that generate differential data, then, the differential data of output circulation quantification; A prediction loop DPCM operator, be used for quantized data is carried out prediction DPCM operation, so that the differential data of generation forecast is carried out the circulation quantization operation to the differential data of described prediction, so that reduce their scope, export the differential data of circulation quantitative prediction then; And a selector, be used for according to inputing to the value of its differential data, select in differential data that circulation quantizes and the predicted difference divided data.

In order to realize above-mentioned and other aspects of the present invention, provide a kind of method that is used to generate differential data.This method comprises: (a) by quantized data being carried out the DPCM operation, generate differential data, by quantized data being carried out prediction DPCM operation, the generation forecast differential data, (b) carry out the circulation quantization operation by differential data to differential data and prediction, generation circulates and quantizes differential data and circulation quantification predicted difference divided data, so that reduce their scope, and (c) according to one in their the amplitude selection circulation quantification differential data and the quantification predicted difference divided data that circulates.

In order to realize above-mentioned and other aspects of the present invention, provide a kind of device that uses differential data generating quantification data.This device comprises: a contrary circulation DPCM operator, be used for the differential data that inputs to it is carried out contrary circulation quantization operation, and so that expand their scope, the result against the circulation quantization operation is carried out contrary DPCM operation, export quantized data then; An inverse prediction circulation DPCM operator is used for that the differential data that inputs to it is carried out contrary circulation quantization operation and just expands their scope, and the result against the circulation quantization operation is carried out inverse prediction DPCM operation, exports quantized data then; And a determining unit, be used for kind according to the DPCM that differential data has been carried out, differential data is exported to contrary circulation DPCM operator or inverse prediction circulation DPCM operator.

In order to realize above-mentioned and other aspects of the present invention, provide a kind of method of utilizing differential data generating quantification data.This method comprises: (a) kind of the DPCM that differential data of input carried out of identification, (b) the input difference data are carried out contrary circulation quantization operation, so that expand their scope, if and (c) the input difference data have been performed the DPCM operation, then carry out contrary circulation DPCM operation and generate quantized data by the differential data that contrary circulation is quantized, if and the input difference data have been performed prediction DPCM operation, then carry out inverse prediction circulation DPCM operation and generate quantized data by the differential data that contrary circulation is quantized.

In order to realize above-mentioned and other aspects of the present invention, provide a kind of being used for that data are carried out apparatus for encoding.Described device comprises: a quantizer, and the data-measuring that is used for inputing to it is predetermined quantization bit; 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 that inputs to it from the DPCM operator is carried out entropy coding, and will be output as bit stream through the differential data of entropy coding; A quantization error minimization 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, be used for the minimum and maximum value that will be included in bit stream is encoded.

In order to realize above-mentioned and other aspects of the present invention, provide a kind of being used for that data are carried out Methods for Coding.This method comprises: (a) utilize predetermined quantization bit, quantize the input data; (b) by quantized data being carried out the DPCM operation, generate differential data; (c) by the differential data that generates is carried out entropy coding, generate bit stream in step (b); (d) utilize input data and quantized data to be adjusted in the minimum and maximum value of importing in the data, thereby make the quantization error minimum; And (e) the minimum and maximum value that will be included in the bit stream is encoded.

In order to realize above-mentioned and other aspects of the present invention, provide a kind of being used for that the data with a plurality of components are carried out apparatus for encoding.This device comprises: a normalizer, be used for calculating the maximum magnitude of the data area of component, and according to this maximum magnitude, the data of each component of standard; A quantizer is used for predetermined quantization bit, to quantizing through specification data; A DPCM operator is used for quantized data is carried out the DPCM operation, exports differential data then; With an entropy coder, be used for differential data is carried out entropy coding, and be encoded into wherein bit stream of output differential data.

In order to realize above-mentioned and other aspects of the present invention, provide a kind of being used for that data are carried out Methods for Coding, this method is used for the data with a plurality of components are encoded.This method comprises: (a) calculate the maximum magnitude in the middle of the data area of component, and according to the data of this each component of maximum magnitude standard; (b) utilize predetermined quantization bit, quantize through specification data; (c) generate differential data by quantized data being carried out the DPCM operation; And, generate be encoded into wherein bit stream of data (d) by differential data being carried out entropy coding.

In order to realize above-mentioned and other aspects of the present invention, a kind of device that data are deciphered of being used for is provided, the bit stream that this device advances wherein the digital coding that will have a plurality of components is deciphered.This 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 exported authority data; And a contrary normalizer, be used for receiving minimum value in the middle of each component data and the maximum in the middle of each component data from bit stream, obtain the maximum magnitude of authority data, and, authority data is carried out contrary standard according to this maximum magnitude and minimum value.

In order to realize above-mentioned and other aspects of the present invention, a kind of method that data are deciphered of being used for is provided, the bit stream that this method is advanced wherein the digital coding that will have a plurality of components is deciphered.This method comprises: (a) by incoming bit stream being carried out entropy decoding, generate differential data; (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 minimum value in the middle of each component data from bit stream, decipher and the maximum in the middle of each component data, obtain the maximum magnitude of authority data, and, authority data is carried out contrary standard according to maximum magnitude and minimum value.

In order to realize above-mentioned and other aspects of the present invention, provide a kind of being used for that data are carried out apparatus for encoding, this device is used for expression is had the key value data of position interpolater of position of the target of x, y and z component and encodes.Described device comprises: a normalizer, be used to calculate the maximum magnitude in the middle of the data area of component, and according to this maximum magnitude, the key value data of each component in standard x, y and the z component are also exported the key value data of standard; A quantizer is used to utilize predetermined quantization bit, quantizes the key value data of standard; A floating-point encoding device is used for being received in minimum value that standard operation uses and maximum magnitude and converts input value to decimal number; A DPCM processor is used to obtain to quantize the differential data and the predicted difference divided data of key value data, and differential data and predicted difference divided data are carried out the circulation quantization operation, thereby reduces their scope; An entropy coder is used for differential data is carried out entropy coding, and output is advanced wherein bit stream with the key value digital coding; And a key value heading code device, be used for encoding to bit being flow to row decoding information required, that will be included in bit stream.

In order to realize above-mentioned and other aspects of the present invention, provide a kind of being used for that data are carried out Methods for Coding, this method is used for expression is had the key value data of position interpolater of position of the target of x, y and z component and encodes.。Described method comprises: (a) calculate a maximum magnitude in the middle of the data area of component, according to this maximum magnitude, the key value data of each component in standard x, y and the z component generate the key value data of standard then; (b) utilize predetermined quantization bit, quantize the key value data of standard; (c) minimum value that will use in standard operation and maximum magnitude convert decimal number to; (d) acquisition quantizes the differential data and the predicted difference divided data of key value data, and differential data and predicted difference divided data are carried out the circulation quantization operation, thereby reduces their scope; (e), generate wherein bit stream is advanced in the key value digital coding by differential data being carried out entropy coding; And (f) encode to bit being flow to row decoding information required, that will be included in the bit stream.

In order to realize above-mentioned and other aspects of the present invention, a kind of device that data are deciphered of being used for is provided, this device is used for deciphering with the bit stream that the key value digital coding of the position interpolater of the position of x, y and z representation in components target is advanced wherein.This device comprises: a key value title decoder, be used for to from bit stream, decoding required information decipher; An entropy encoder is used for bit stream is carried out entropy decoding and output differential data; A contrary DPCM processor is used for the type of basis to the DPCM of differential data execution, differential data is carried out against circulation DPCM operation or inverse prediction circulation DPCM operated, and export the key value data that quantize thus; An inverse quantizer is used for the key value data that quantize are carried out re-quantization and exported the key value data of standard; A floating number decoder, be used for the maximum magnitude in the middle of the data area of minimum value in the middle of key value title decoder receives the key value data of each component x, y and the z component and x, y and z component, convert minimum value and maximum magnitude to binary number, and this binary number is exported to contrary normalizer; And contrary normalizer, be used for the maximum magnitude in the middle of the data area of minimum value in the middle of the floating number decoder receives the key value data of each component x, y and the z component and x, y and z component, and the key value data of x, y and z component are carried out contrary standard.

In order to realize above-mentioned and other aspects of the present invention, a kind of method that data are deciphered of being used for is provided, this method is used for deciphering with the bit stream that the key value digital coding of the position interpolater of the position of x, y and z representation in components target is advanced wherein.Described method comprises: (a) to from bit stream, decipher required heading message and decipher; (b) generate differential data by bit stream being carried out entropy decoding; (c), differential data is carried out contrary circulation DPCM operation or inverse prediction circulation DPCM operation, the key value data of generating quantification by type according to the DPCM that differential data is carried out; (d) will be in step (a) the minimum value in the middle of the key value data of each component decoding, in x, y and the z component and the central maximum magnitude of data area of x, y and z component convert binary number to, and convert the maximum in the middle of all key value data of the minimum value in the middle of the key value data of each component in x, y and the z component and x, y and z component to binary number; (e) by according to predetermined quantization bit, the key value data that quantize are carried out re-quantization, generate the key value data of standard; And, the key value data are carried out contrary standard (f) according to the maximum magnitude in the middle of the data area of the minimum value in the middle of the key value data of each component that in step (d), be converted, in x, y and the z component and x, y and z component.

Description of drawings

In conjunction with the drawings the preferred embodiments of the present invention are described in detail, it is clearer that above-mentioned purpose of the present invention and advantage will become, wherein:

The block diagram of Fig. 1 shows the structure of the conventional equipment that is used for the conventional equipment that data are encoded and is used for data are deciphered;

The block diagram of Fig. 2 A shows according to device the preferred embodiments of the present invention, that be used to carry out the DPCM operation, and Fig. 2 B shows the flow chart according to the DPCM operation of the preferred embodiments of the present invention;

The block diagram of Fig. 3 A show according to the preferred embodiments of the present invention, be used to utilize the DPCM operator that data are carried out apparatus for encoding, and Fig. 3 B shows operational flowchart shown in Fig. 3 A, that be used for data are carried out code device;

The block diagram of Fig. 4 A show according to the preferred embodiments of the present invention, be used for data are carried out apparatus for encoding, this device can make the quantization error minimum.Fig. 4 B shows the flow chart of the method that makes the quantization error minimum;

The block diagram of Fig. 5 A show according to the preferred embodiments of the present invention, be used for data are carried out apparatus for encoding, this device carries out standard to the input data with a plurality of components, then specification data is encoded, and Fig. 5 B shows the flow chart that data is carried out the method for standard;

The block diagram of Fig. 6 A show according to the preferred embodiments of the present invention, be used for data are carried out apparatus for encoding, this device is used for the key value data of position interpolater are encoded, and Fig. 6 B shows the flow chart that is used for the key value data are carried out Methods for Coding;

Fig. 7 A and 7B show the operational flowchart of the floating-point encoding device shown in Fig. 6 A;

Fig. 7 C and 7D show the operational flowchart according to the entropy encoder of the preferred embodiments of the present invention;

Fig. 8 shows the step S683 shown in Fig. 7 C;

Fig. 9 A and 9B show the step shown in Fig. 7 D;

The block diagram of Figure 10 A shows the contrary DPCM operator according to the preferred embodiments of the present invention; And Figure 10 B shows the flow chart of contrary DPCM operation;

The block diagram of Figure 11 shows according to device the preferred embodiments of the present invention, that be used to utilize contrary DPCM operator that data are deciphered;

The block diagram of Figure 12 A shows according to device the preferred embodiments of the present invention, that be used to utilize contrary normalizer that data are deciphered, and Figure 12 B shows the operational flowchart of contrary normalizer;

The block diagram of Figure 13 A show according to the preferred embodiments of the present invention, be used for device that the key value data are deciphered and Figure 13 B and show the flow chart that is used for method that the key value data are deciphered;

Figure 14 A shows the detail flowchart of the step S1320 shown in Figure 13 B;

Figure 14 B shows the detail flowchart of the step S1330 shown in Figure 14 A;

Figure 14 C shows the detail flowchart of the step S1380 shown in Figure 13 B;

Figure 15 show according to the preferred embodiments of the present invention, be transfused to component order to the bit stream of entropy encoder;

Figure 16 A and 16B show the result according to the prediction DPCM operation of the preferred embodiments of the present invention;

Figure 17 A shows result according to the circulation DPCM of the preferred embodiments of the present invention operation to 17C;

Figure 18 A and 18B show the key and the key value data of position interpolater; And

Figure 19 to 27 shows the example of program code, by these program codes, can be implemented in according to the process that reads bit stream the preferred embodiments of the present invention, that be used for method that data are deciphered.

Embodiment

Below in conjunction with accompanying drawing the present invention is described more fully.

The block diagram of Fig. 2 A shows according to device the preferred embodiments of the present invention, that be used to carry out the DPCM operation, and Fig. 2 B shows the flow chart according to the DPCM operation of the preferred embodiments of the present invention.

With reference to Fig. 2 A, comprise according to device the preferred embodiments of the present invention, that be used to carry out the DPCM operation: a circulation DPCM operator 200, it comprises and is used to carry out general DPCM operation, exports the DPCM operator 210 of differential data and the differential data that inputs to it is carried out the circulation quantization operation quantize operator 220 so that reduce the circulation of their scope then; A prediction loop DPCM operator 230, it comprise be used for to the quantized data that inputs to it carry out prediction DPCM operation, then prediction of output differential data prediction DPCM operator and be used for the differential data that inputs to it is carried out the circulation quantization operation so that reduce the circulation quantizer 250 of their scope; And a selector 260, be used for according to the value that inputs to its differential data, selecting a differential data from the differential data of circulation DPCM operator 200 inputs with between the differential data of prediction loop DPCM operator 230 inputs.

With reference to Fig. 2 B, the DPCM operation according to the preferred embodiments of the present invention is described below.

In step S210, the data that quantized according to any quantization method are transfused to circulation DPCM operator 200 and prediction loop DPCM operator 230.

In circulation DPCM operator 200, DPCM operator 210 generates differential data, and the differential data that is generated is exported to circulation quantizer 220 by the quantized data that inputs to it being carried out general DPCM operation.Then, circulation quantizer 220 quantizes carrying out circulation from the differential data of DPCM operator 210 inputs in step S220a.

Specifically, DPCM operator 210 is by deducting at preceding quantized data the differential data of calculating from the current quantized data of importing among step S210.After this, 220 pairs of differential datas that calculated of circulation quantizer are carried out the circulation quantification.The operation of circulation DPCM operator 200 is represented by following equation.

${\hat{p}}_i=\mathrm{CircularQuantization}({\tilde{p}}_i-{\tilde{p}}_i)(i=\mathrm{1,2},...,n)---\left(1\right)$

In equation (1), n represents the quantity of data.To describe circulation subsequently quantizes.

Prediction DPCM operator 240 in prediction loop DPCM operator 230 is by carrying out prediction DPCM operational computations predicted difference divided data to the differential data that inputs to it, then, circulation quantizer 250 is carried out the circulation quantification to the predicted difference divided data in step S220b.

Specifically, prediction DPCM operator 240 is by deducting prediction data from current quantized data, the differential data of generation forecast and with its output.In order to obtain to be used for the prediction data of current quantized data, obtain difference value by the quantized data that from quantized data the preceding, deducted before quantized data the preceding, then this difference value is added to the preceding on the differential data, thereby calculates the prediction data that is used for current data.

Cannot be by the prediction data that prediction DPCM operator 240 calculates above the maximum magnitude of the quantized data that inputs to it.In other words, if prediction data surpasses the maximum in the quantizing range of importing quantized data, so, prediction DPCM operator 240 will this prediction data be arranged to import in the quantizing range of quantized data maximum and by from this prediction data, deducting current data, generate the differential data that is used for current data.If prediction data is less than the minimum value in the quantizing range of input data, so, prediction DPCM operator 240 is defined as the predicted difference divided data with current data.After this, the operation of prediction DPCM operator 240 will be called as " through the prediction DPCM operation of revising ".

The predicted difference divided data of being calculated by prediction DPCM operator 230 is transfused to circulation quantizer 250, and is performed the circulation quantization operation.

The operation of prediction loop DPCM operator 230 can be represented by following equation.

$\stackrel{\widehat{~}}{p}=(2^{\mathrm{nQuantBit}}-1)-{\tilde{p}}_i$ If ( $2\×{\tilde{p}}_{i-1}-{\tilde{p}}_{i-2}>2^{\mathrm{nQuantBit}}-1$ ) …(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)

${\hat{p}}_i=\mathrm{CircularQuantization}\left({\stackrel{\widehat{~}}{p}}_i\right)$

In equation (2), i be one 2 and n-1 between integer, n represents the quantity of data.

Figure 16 A show to 50 have maximum be 1024 and minimum value be the result that 0 quantized data is carried out simple forecast DPCM operation, and Figure 16 B shows the data identical with the data of the situation of Figure 16 A is carried out result through the prediction DPCM operation of revising, thereby, maximum in the middle of the prediction data that is used for data surpasses data or be lower than under the situation of the minimum value in the middle of the data, the differential data that will be used for data is regulated according to equation (2).

As the result of simple forecast DPCM operation, the predicted difference divided data has about 3,000 wide scopes, shown in Figure 16 A.As the result through the prediction DPCM operation revised, the predicted difference divided data has and is not more than 2,000 close limit, shown in Figure 16 B.This means in than easier acquisition in simple forecast DPCM operation and have predicted difference divided data than close limit through the prediction DPCM operation of revising.

To describe circulation below quantizes.

Figure 17 A shows quantized data, and Figure 17 B shows the result who the quantized data shown in Figure 17 A is carried out the DPCM operation.Shown in Figure 17 B, the twice that the scope of the differential data after the DPCM operation can be increased to before the execution DPCM operation is wide.Therefore, we can say, use the circulation quantization operation can be so that the scope that remains on the input data in the scope of the differential data that will be performed the DPCM operation with in interior, is carried out the DPCM operation.

Circulate each other on the connection basis in the maximum and the minimum value of hypothesis in the quantizing range, carry out the circulation quantization operation.Therefore, if as the result's who two continuous quantized datas is carried out linear DPCM differential data greater than in the quantizing range peaked half, then can reduce their value by in the middle of each of differential data, deducting maximum.

If differential data then by the maximum in the quantizing range is added on each of differential data, can further reduce their value less than half of the minimum value in the quantizing range.

Suppose X and

Represent the differential data that input difference data and circulation quantize respectively, then circulation quantizes and can be represented by following equation.

$\mathrm{CircularQuantization}\left(X_i\right)={\hat{X}}_i=\mathrm{Min}\left(\right|X_i|,|X_i^{\&prime;}\left|\right)---\left(3\right)$

X ' _i=X _i-(2 ^NQBits-1) (if X 〉=0)

X ' _i=X _i+ (2 ^NOBits-1) (otherwise)

In equation (3), nQBits represents the bit size that is used to quantize.Figure 17 C shows the differential data shown in Figure 17 B is carried out the result that circulation quantizes.

Referring again to Fig. 2 B, the predicted difference divided data that differential data that circulation quantizes and circulation quantize is exported to selector 260, and then, selector 260 is carried out difference absolute value summation (SAD) operation to the differential data of input in step S230.Here, thus carrying out the SAD operation all is added the absolute value of input data.In step S230, the absolute value addition of the predicted difference divided data that the absolute value phase adduction of the differential data that selector quantizes all circulations quantizes all circulations.

After this, selector 260 compares the absolute value sum of the differential data that the quantizes absolute value sum with the differential data of prediction, between the differential data of the differential data that quantizes and prediction, select to have less absolute value and differential data and among step S240, export selected differential data.Selector 260 select to have less absolute value and the reason of differential data be according to inferring that they have value than close limit.For example, the predicted difference divided data is being carried out in the process of entropy coding, can expect to have less absolute value and differential data encode required bit number be less than to have big absolute value and the differential data required bit number of encoding.

For those skilled in the art clearly, SAD operates only multiple a kind of of certain method of organizing differential data that be used for selecting, and thus, selector 260 can adopt diverse ways to select differential data.

The block diagram of Fig. 3 A show according to the first embodiment of the present invention, be used for data are carried out apparatus for encoding and Fig. 3 B shows according to the flow chart first embodiment of the present invention, that be used for data are carried out the operation of apparatus for encoding.According to the first embodiment of the present invention, be used for that data are carried out apparatus for encoding and comprise aforesaid DPCM operator of the present invention.

Being used for shown in Fig. 3 A carries out apparatus for encoding to data and comprise: a quantizer 310 is used to utilize predetermined quantization bit that the input data are quantized; A DPCM processor 340 is used for quantized data is carried out circulation DPCM operation and predicted that DPCM operates, exports then differential data; And an entropy coder 350, be used for carry out predetermined entropy coding operation, output bit flow then from the differential data of DPCM processor 340 inputs.DPCM processor 340 comprises aforesaid DPCM operator of the present invention.

With reference to Fig. 3 A and 3B, in step S310, the data that are encoded are transfused to being used for carries out apparatus for encoding to data, and in step S320, is utilized the quantization bit of predetermined quantity to quantize in quantizer 310 and is exported to DPCM processor 340.

200 pairs of quantized datas from quantizer 310 inputs of circulation DPCM operator in DPCM processor 340 are carried out DPCM operation and circulation quantization operation, then the result are outputed to selector 260.In addition, described prediction loop DPCM operator quantizes carrying out prediction DPCM operation from the quantized data of quantizer 310 inputs and circulating in step S330.

260 pairs of selectors are carried out the SAD operation from the differential data of circulation DPCM operator 200 inputs with from the differential data of prediction loop DPCM operator 230 inputs, selection will be exported to the differential data of entropy encoder 350, and in step S340 selected differential data be exported.

350 pairs of differential datas from selector 260 inputs of entropy coder are carried out predetermined entropy coding operation, thereby they are carried out entropy coding and generate bit stream in step S350.

The block diagram of Fig. 4 A shows the flow chart that is used for data are carried out apparatus for encoding and Fig. 4 B according to second embodiment of the invention and shows the operation that quantization error shown in Fig. 4 A minimizes device.

Be used for that according to second embodiment of the invention data are carried out apparatus for encoding and comprise minimized maximum of quantization error and the minimum value that is used to make a bit stream, thereby and this bit stream encoded when to data are encoded described quantization error can being minimized.

With reference to Fig. 4 A, according to the preferred embodiment of the present invention, be used for that data are carried out apparatus for encoding and comprise quantizer 310, DPCM operator 110/340, entropy coder 350, quantization error minimizes device 320 and heading code device 370, wherein, quantization error minimizes device 320 and is used to receive data and quantized data, maximum in the adjusting data and minimum value are so that can make the quantization error minimum, and the minimum and maximum value of output through overregulating, maximum and minimum value that heading code device 370 is used for minimizing device 320 inputs from quantization error are encoded into heading message, and this information is included in the bit stream.

Quantization error minimizes device 32O and comprises: an initial value is provided with unit 321, is used to be provided with minimum error values e _Min, the minimum value min ' through overregulating and quantization error minimum value min _MinMinimum value updating block 323 through overregulating is used for by carrying out the predetermined minimum value min ' of operation renewal through overregulating; A determining unit 325 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; An error amount updating block 327 is used to utilize the minimum value through upgrading and regulating, and calculates quantization error value, if the error amount that is calculated is less than minimum error values e _Min, minimum value min ' that will be through overregulating and the error amount that is calculated are updated to quantization error minimum value min respectively _MinWith minimum error values e _Min, and the result that will upgrade exports to the minimum value updating block 323 through overregulating.

Quantization error minimizes device 320 by quantizing range is controlled in execution quantification of input data and re-quantization, so that make the quantization error minimum.

Specifically, when representing fixed maximum values, the Min that will be used to quantize, Max represents the minimum value through overregulating, the X that will be used to quantize _iRepresent that input value and nQuanBit represent the bit number that is used to quantize, then utilize equation (4) quantizing input value

The re-quantization value

And error e _iCalculate.

${\tilde{x}}_i=\mathrm{floor}\left(\frac{X_i-\mathrm{Min}}{\mathrm{Max}-\mathrm{Min}}\right)*(2^{\mathrm{nQuantBit}}-1)+0.5---\left(4\right)$

${\hat{x}}_i=\frac{{\tilde{X}}_i*(\mathrm{Max}-\mathrm{Min})}{2^{\mathrm{nQuantBit}}-1}+\mathrm{Min}$

$e_i=X_i-{\hat{X}}_i$

In order to reduce sum of errors ∑ e _i, quantization error minimizes device 320 control Min, thereby can make sum of errors ∑ e _iMinimum and use can make the minimum value of the quantization error minimum in the process that data are deciphered, and quantized data is deciphered.

Therefore the counter element of the aforementioned means that quantizer 310, DPCM operator 110/340 and entropy coder 350 and being used for encoded to data no longer repeats much at one here.Below in conjunction with Fig. 4 B the operation that quantization error minimizes device 320 is described more fully.

Initial value is provided with unit 321 receives the data identical with the data that input to quantizer 310 and exports from quantizer 310 in step S402 quantized data.

Initial value is provided with unit 321 and obtains counting the number of words of maximum max and minimum value min and calculated minimum min in the middle of the input data.Then, initial value is provided with unit 321 and calculates and quantize error e and this quantization error e is set to initial minimum error values e by utilizing maximum max and minimum value min that quantized data is carried out re-quantization in step S404 _Min

Initial value is provided with the minimum value of unit 321 in the middle of the input data and deducts the minimum value min ' that quantization step QuantSpace is set to initial adjustment divided by 2 resulting results and this subtraction result.Then, initial value is provided with unit 321 the minimum value min ' of described initial adjustment is set to quantization error minimum value m _MinAnd will minimum value min ' in step S406 export to minimum value updating block 323 through overregulating through overregulating.

Minimum value updating block 323 through overregulating upgrades the minimum value min ' through overregulating that unit 321 inputs are set from initial value by carrying out predetermined operation.According to the preferred embodiments of the present invention, the minimum value updating block 323 through overregulating upgrades from initial value according to equation (5) and the minimum value min ' through overregulating of unit 321 inputs is set and will exports to determining unit 325 through the minimum value min ' that upgrades and regulate among step S408.

${\min }^{\&prime;}=\frac{\mathrm{Mantissa}\left({\min }^{\&prime;}\right)+1}{{10}^{\mathrm{Exponent}\left({\min }^{\&prime;}\right)}}---\left(5\right)$

In step S410, determining unit 325 is calculated counting the number of words through the minimum value min ' that upgrades and regulate, this result calculated and the counting the number of words of minimum value min that is provided with by initial value in the middle of the input data of calculating unit 321 are compared, and will be added to through minimum value min ' the minimum value min central that upgrades and regulate quantization bit size QuantSpace is compared divided by the result on the 2 gained results with importing data.

If through counting the number of words of the minimum value min ' that upgrades and regulate greater than the counting the number of words of the minimum value min in the middle of the input data, if perhaps through the minimum value min ' that renewal and regulate greater than

Then determining unit 325 is with the current quantization error minimum value min that stores _MinBe defined as being used to the minimum value of re-quantization, and with this quantization error minimum value min _MinExport to output unit 329.If through counting the number of words of the minimum value min ' that renewal and regulate be not more than counting the number of words of minimum value min in the middle of the input data and through renewal and regulate minimum value min ' be not more than

Then determining unit 325 will be exported to error amount updating block 327 through the minimum value min ' that upgrades and regulate.

In step S414, error amount updating block 327 utilizes process renewal and the minimum value min ' that regulates and by initial value unit 321 acquisition maximum max is set quantized data is carried out re-quantization, and calculates the quantification error e.

In step S446, error amount e that error amount updating block 327 will newly calculate and minimum error values e _MinCompare.Result as a comparison is if the new error amount e that calculates is less than minimum error values e _Min, then in step S418, error amount updating block 327 utilizes the new error amount e that calculates to upgrade minimum error values e _Min, and utilize through the minimum value min ' that upgrades and regulate and upgrade quantization error minimum value min _MinOn the other hand, if newly the error amount e that calculates is greater than minimum error values e _Min, error amount updating block 327 execution in step S408 and not to minimum error values e once more then _MinWith quantization error minimum value min _MinUpgrade.

Make the minimum and maximum value of quantization error minimum be exported to heading code device 370 and be encoded as heading message through above-mentioned processing being used to of obtaining.Then, this heading message is included in the bit stream that is generated by entropy coder 350.

The block diagram of Fig. 5 A show according to third embodiment of the invention, be used for data are carried out apparatus for encoding the operational flowchart that shows normalizer shown in Fig. 5 A of Fig. 5 B.According to third embodiment of the invention, be used for data are carried out apparatus for encoding by the input data with a plurality of components are carried out standard, improved the efficient that data are encoded.

Shown in Fig. 5 A, according to third embodiment of the invention, be used for that data are carried out apparatus for encoding and comprise: normalizer 300, be used to calculate input to and be used for data are carried out maximum magnitude in the middle of the data area of component of apparatus for encoding, and the data of each component in x, y and the z component are carried out standard according to this maximum magnitude; Quantizer 310 is used to utilize the quantization bit of predetermined quantity to quantizing through specification data; DPCM operator 110/340 is used for quantized data is carried out DPCM operation and output differential data; And entropy coder 350, be used for differential data is carried out entropy coding and output process bitstream encoded.

Quantizer 310, DPCM operator 110/340 and entropy coder 350 are with identical according to counter element first embodiment of the invention, that be used for data are carried out apparatus for encoding, and therefore the detailed description about them no longer repeats here.The operation of normalizer 300 is only described in conjunction with Fig. 5 B below.

Comprise in the data that will be encoded under the situation of a plurality of components, for example, if the input data comprise three component x, y and the z of the expression position of target in three dimensions, then each component in x, y and the z component is inputed to the normalizer 300 that is used for data are carried out apparatus for encoding.Then, normalizer 300 calculates the central maximum magnitude of scope of x, y and z component, and according to this maximum magnitude the data of each component in x, y and the z component is carried out standard.

In step S512, normalizer 300 receives the data of each component and calculates the data area of x, y and z component according to equation (6).Then, in step S514, normalizer 300 calculates the central maximum magnitude of data area of x, y and z component by relatively.

${\mathrm{Max}}_j={\mathrm{Max}}_{i=0}^n\left({\stackrel{\&OverBar;}{p}}_{\mathrm{ij}}\right)---\left(6\right)$

${\mathrm{Min}}_j={\mathrm{Min}}_{i=0}^n\left({\stackrel{\&OverBar;}{p}}_{\mathrm{ij}}\right)$

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

After this, in step S516, normalizer 300 utilizes the data of each component in following equation x, y and the z component to carry out standard.

In equation (6) and (7), n represents the quantity of data, i=0, and 1 ..., n-1, and j represents each component in x, y and the z component.Shown in equation (7), use the central maximum magnitude Range of their data areas _MaxX, y and z component are carried out standard.As the result of this standard, about not having maximum magnitude Range _MaxThe redundancy of data of component increase, so, improved the efficient that data are encoded.

Each component in 300 pairs of inputs of normalizer component carries out standard, and the result of standard exported to quantizer 310, then, quantizer 320 utilizes the quantization bit of predetermined quantity, authority data from normalizer 300 inputs is quantized, and the result that will quantize exports to DPCM operator 110/340.After this, DPCM operator 110/340 generates differential data by the quantized data execution DPCM from quantizer 310 inputs is operated, and this differential data is exported to entropy coder 350, thereby can encode to them.

To describe below according to fourth embodiment of the invention be used for data are carried out apparatus for encoding, the integrated aforesaid DPCM operator of the present invention, the aforementioned means that can encode to data that is used for the aforementioned means that can make the quantization error minimum of data being encoded and is used for data being encoded in this device with a plurality of components according to the present invention according to the present invention.

Although according to fourth embodiment of the invention be used for data carried out the key value data that apparatus for encoding is described as specially to the position interpolater in the middle of the three-dimensional animation data encode, but clearly, it also can be encoded to the data of other general type.

The position interpolater is about being illustrated in the information based on the animation path in the animation of key frame, is one of basic fundamental of the synthetic 3D animation of expression computer based on the animation of key frame.Be used to moment of occurring according to key data, the order of regulation animation based on the animation of key frame.The corresponding frame of moment that occurs with key data is referred to as key frame.

In the animation based on key frame, a plurality of key frames and a plurality of frames between key frame have constituted an animation, and these frames are interpolated (or the use interpolation method generates the frame between the key frame) between the key frame.

Such as MPEG-4 binary format (BIFS) that is used for scenery and virtual reality modeling language international multimedia standards such as (VRML), support to use the animation based on key frame of interpolater node.In MPEG-4BIFS and VRML, the various interpolaters that comprise scalar interpolater, position interpolater, coordinate built device, directed interpolater, standard interpolater and color interpolator are arranged.These interpolaters and their function and feature are shown in Table 1.

Table 1

The type of interpolater	Feature	Function
			The scalar interpolater	The linear interpolation of scalar variable	Can represent area, diameter and density
The position interpolater	Linear interpolation on the 3D coordinate	Parallel motion in 3d space
			Directed interpolater	The linear interpolation of 3D reference axis and rotation amount	Rotation in the 3d space
The coordinate built device	The linear interpolation of the variable in the 3D coordinate	The 3D variant
			The standard interpolater	The linear interpolation of standard 3D coordinate	Can represent the variable in the standard 3D vector
Color interpolator	The linear interpolation of colouring information	Can be illustrated in the variable in the color

Position interpolater shown in the table 1 is used to indicate based on the positional information of the animation of key frame and comprises key and the key value field.Cipher key field is shown the position of each key frame on time shaft with the discontinuous numerical table of scope between-∞ and ∞.Each define objective in the key value field is in the information of the position in the moment of being represented by each key, and comprises x, y and three components of z.In the key value field each all comprise with cipher key field in the identical key value of quantity of the key value that each had.

Example about the position interpolater is shown among Figure 18 A and the 18B.Specifically, Figure 18 A shows key data, and Figure 18 B shows the key value data.

In MPEG-4BIFS and VRML, used linear interpolation.But, smoothly and naturally represent animation in order to utilize linear interpolation, need the key data and the key value data of quite big quantity.In addition, in order to store and to transmit this animation, need have jumbo memory and a large amount of time.Therefore, preferably compress interpolater so that storage and transmit these interpolaters.

In MPEG4-BIFS, used the method that the interpolater node that is referred to as prediction MF coding (PMFC) is encoded and deciphered.In the PMFC method, with class of operation shown in Figure 1, as to be used for conventional equipment that data are encoded seemingly, utilize quantizer, DPCM operator and entropy coder that the key value data of position interpolater are encoded.Quantizer and DPCM operator have reduced redundant key value data, and the output of DPCM operator is transfused to entropy coder.But, in the PMFC method, the differential data that is obtained by the DPCM operator is carried out entropy coding, so its code efficiency is not high enough.In addition, because the restriction that entropy coder had, it may provide high quality animation hardly.

Therefore, the invention provides a kind of being used for carries out apparatus for encoding to key value, integrated according to aforementioned DPCM operator of the present invention, the aforementioned means that can encode to data that is used for the aforementioned means that can make the quantization error minimum of data being encoded and is used for data being encoded in this device with a plurality of components according to the present invention according to the present invention, increased the efficient that the key value data are encoded thus.

The block diagram of Fig. 6 A show according to fourth embodiment of the invention, be used for the key value data are carried out apparatus for encoding, Fig. 6 B is the operational flowchart that is used for the key value data are carried out apparatus for encoding.

With reference to Fig. 6 A, be used for that the key value data are carried out apparatus for encoding and comprise: normalizer 300 is used for according to maximum magnitude in the middle of the data area of x, y and z component the input key value data of each component being carried out standard; Quantizer 310 is used to utilize predetermined quantization bit, and the key value data of standard are quantized; Quantization error minimizes device 320, is used for receiving minimum value and maximum magnitudes from normalizer 300, and regulates and export minimum and maximum value, so that can make the quantization error minimum; Floating-point encoding device 330 is used for minimizing device 320 from quantization error and receives minimum value and maximum magnitude, and will be used to make the maximum magnitude of quantization error minimum and minimum value to convert decimal number to; DPCM processor 340 is used to obtain to quantize the differential data and the predicted difference divided data of key value data and carries out the circulation quantization operation so that reduce the scope of differential data; Entropy coder 350 is used for differential data is carried out entropy coding, exports then wherein bit stream is advanced in the key value digital coding; And key value heading code device 360, be used for decoding bit stream required information is encoded and made this information can be included in bit stream.

Describe according to the operation that is used for the key value data are carried out apparatus for encoding of the present invention below in conjunction with Fig. 6 B.

In step S600, the key value data of each component are transfused to the normalizer 300 that is used for the key value data are carried out apparatus for encoding in x, y and the z component.Then, in step S610, normalizer 300 calculates the maximum magnitude in the middle of the data area of x, y and z component, and according to this maximum magnitude the x, the y that are imported and the key value data of z component is carried out standard.Described above the key value data that input to normalizer 300 are carried out the processing of standard, will no longer repeat here.

Be transfused to quantizer 310 by the key value data of the x of standard, y and z component, in step S620, quantizer 310 utilizes the quantization bit nQuanBit of predetermined quantity then, according to equation (8) the key value data of standard is carried out.

${\tilde{p}}_{\mathrm{ij}}=\mathrm{floor}({\stackrel{\&OverBar;}{p}}_{\mathrm{ij}}\&times;(2^{\mathrm{nQuantBit}}-1)+0.5---\left(8\right)$

In equation (8), floor () is a function that is used for the floating number of input is converted to the maximum integer of the floating number that is not more than this input.Quantizer 310 minimizes the key value data that device 320 outputs quantize to DPCM processor 340 and quantization error.

In step S630, quantization error minimizes device 320 and receives minimum value and maximum magnitude from normalizer 300, utilization is from minimum value and the maximum magnitude calculated minimum and the maximum of normalizer 300 inputs, determine that a minimum value and a maximum are so that minimize quantization error, utilize determined minimum and maximum value calculation minimum value and maximum magnitude, and, thus they are encoded into the key value title to floating number transducer 330 this minimum value of output and maximum magnitudes.Here, minimize device 320 definite minimum and maximums by quantization error and be used to make the quantization error minimum.

The step S650 that is carried out by floating-point encoding device 330 is described below in conjunction with Fig. 7 A.

In step S651, floating-point encoding device 330 receives minimum value x_min, y_min, z_min in the middle of the key value data of each component of x, y and z component, has the maximum max of the component of maximum magnitude, belongs to the information nWhichAxis of which component and the nKeyValueDigit that counts the number of words of primary key Value Data about maximum max.

In order to improve the efficient that the key value data are encoded by reducing the required bit number of coding, floating-point encoding device 330 will convert decimal number to by x_min, y_min, z_min and the max of binary number representation in step S652.

Computer is with the form storage floating number of 32 bits.When importing the floating number of binary system, floating-point encoding device 330 converts this floating number to the mantissa of decimal system and its index according to equation (9).

For example, the floating number 12.34 of decimal system can convert binary number as follows to by computer.

$\frac{0}{1}\frac{10001010111000010100011}{2}\frac{10000010}{3}$

1: sign

2: the mantissa in the binary system

3: the index in the binary system

As follows, this binary number can be converted back to original decimal number according to equation (9).

$\frac{0}{1}\frac{1234}{2}\frac{2}{3}$

1: sign

2: the mantissa in the decimal system

3: the index in the decimal system

For mantissa and index with decimal system are included in the bit stream, must calculate required bit number of expression mantissa and the required bit number of expression index.Index has the value between-38 and 38, can utilize 7 and sign to come together to represent index thus.The required bit number of expression mantissa depends on counting the number of words of this mantissa.Following table shows the value of various scopes and represents the required bit number of mantissa of the value of various scopes.

Table 2

The scope of value	Mantissa counts the number of words	The required bit number of expression 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	24

Floating-point encoding device 330 checks in step S453 whether counting the number of words of x_min, y_min and z_min be identical.If they are inequality, then in step S654, export to key value heading code device 360 and they are encoded to the key value title.

If counting the number of words of x_min, y_min and z_min is identical, then in step S655, floating-point encoding device 330 checks whether they are identical with counting the number of words of primary key Value Data.If counting the number of words of x_min, y_min and z_min is different with counting the number of words of primary key Value Data, so, at step S656, one in the counting the number of words of x_min, y_min and z_min is exported to key value heading code device 360 and is encoded as the key title.

After this, floating-point encoding device 330 checks that in step S657 the maximum max in the middle of all key value data of x, y and z component selects in the middle of the key value data of which component, and the minimum value in the middle of the key value data with the component that obtains maximum max of whether counting the number of words of definite max count the number of words identical.

If counting the number of words of the minimum value in the middle of the key value data with the component that obtains maximum max of counting the number of words of max is different, then counting the number of words of max is exported to key value heading code device 360 and is encoded as a key value title.On the other hand, if max count the number of words with maximum max under the key value data of component in the middle of minimum value count the number of words identically, then in step S658, describe their identical state informations and be exported to key value heading code device 360.

The aforementioned operation of floating-point encoding device 330 is to be used for the mantissa of passing through the floating number of changing is encoded.Describe below and to be encoded into the processing of key value title with the information of the correlation of indices of floating number.

In step S659, which was a maximum in the middle of floating-point encoding device 330 was identified in the absolute value of index of x_min, y_min, z_min and max, the maximum of being discerned is stored as nMaxExp, and nMaxExp is encoded in the key value title of some bits.

After this, floating-point encoding device 330 checks in step S660 whether the sign of index of x_min, y_min, z_min and max is identical.If they are identical, then bSameSignExp is set to 1 and their sign exported to key value heading code device 360 in step S661.On the other hand, if they are inequality, then bSameSignExp is set to 0 in step S662, finish then to the coding of the correlation of indices information of x_min, y_min, z_min and max.

With reference to Fig. 7 B, floating-point encoding device 330 is with x_min, and y_min, z_min and max are encoded into before the key value title, in step S663 the sign through the floating number of conversion is encoded.

Then, in step S664, floating-point encoding device 330 calculates the floating number required bit number of encoding with reference to table 2, and the mantissa that will have with the floating number of result of calculation as much bit exports to key value heading code device 360.

In step S665, floating-point encoding device 330 calculates the required bit number of encoding to nMaxExp with reference to table 2, and will have with the index of the floating number of result of calculation as much bit and export to key value heading code device 360.

Floating-point encoding device 330 checks according to bSameSignExp whether x_min, y_min, z_min and max have identical sign.If they do not have identical sign, then their sign is exported to key value heading code device 360 and is encoded in step S667, thus, finish encoding process to the minimum and maximum value in the middle of the key data of each component from x, the y of normalizer 300 input and z component.

Floating-point encoding device 330 minimizes device 320 receptions from quantization error and makes the minimum required minimum and maximum value of quantization error, and through above-mentioned steps they is encoded into the key title.

Referring again to Fig. 6 B, in step S670, the key value data of 340 pairs of quantifications of DPCM processor are carried out circulation DPCM operation and prediction loop DPCM operates and export differential datas to entropy coder 350.DPCM processor 340 is made up of aforesaid DPCM operator, and the operation and the structure of DPCM operator are described in the above.Difference between DPCM processor 340 and the aforesaid DPCM operator only is that DPCM processor 340 is carried out circulation DPCM operation to the key value data of each component in x, y and the z component respectively and prediction loop DPCM operates.

In step S680, in entropy coder 350, the differential data of exporting from DPCM processor 340 is encoded.

With reference to Fig. 7 C, the differential data of the key value data of each component of entropy coder 350 couples of x, y and z component carries out entropy coding.

Entropy coder 350 checks with reference to the differential data of each component of x, y and z component whether the quantized value of x, y and z component is identical, if they are identical, then finishes entropy coding in step S681.For example, under the situation of three-dimensional animation, a row train is with the horizontal motion identical with the x component direction, and the key value data of x component change.But y and z component are almost constant.Therefore, if y and z component are quantized, estimate that they will have identical value.Therefore, do not need all practically identical key value data are encoded, thus can be by only being encoded into the efficient that the key value title improves coding with one in the middle of the key value data.

If the quantification key value data of each component are all different in x, y and the z component, then in step S682, entropy coder 350 is read the predetermined value that has been set in advance.After this, entropy coder 350 is encoded to the differential data of the quantification key value data of each component in x, y and the z component with unary AAC function in step S683, perhaps with quantizing the AAC function continuously they is encoded in step S685.

Below in conjunction with Fig. 8 unary AAC function is described.Unary AAC function becomes the symbol transition that will be encoded by a series of 0, one to be used to indicate the marker bit 1 of this series 0 end and the position of a sign of representing symbol to form a plurality of bits.Here, a plurality of 0 amplitude corresponding to symbol.For example, utilize unary AAC function with 256 be encoded into by 256 0, be used to indicate marker bit 1 that this series 0 finishes and represent 256 sign, i.e. positive sign, 0 a series of bits of forming.Because unary AAC function, increased the redundancy of the bit of the symbol that expression will be encoded, improved efficient to encoding symbols.

With reference to Fig. 7 D, 9A and 9B the method for SQ AAC function to encoding symbols of utilizing described below.SQ AAC function upgrades quantizing range ground continuously to encoding symbols.

Fig. 9 A shows the method for SQ AAC function to encoding symbols of utilizing.Shown in Fig. 9 A, be 1 with the incoming symbol that is encoded, minimum value in coding range and maximum are respectively 0 and 9.

With reference to Fig. 9 A, in first step, coding range is divided into two subdomains to this encoding symbols, i.e. the upper bound and lower bound then, are checked the symbol that is encoded, promptly 1, belong to the upper bound or belong to lower bound.Because 1 belong to described lower bound, so 0 is encoded and lower bound is updated to a new coding range.Therefore, the new coding range that is used for second step is from 0 to 4.

In second step, new coding range 0 to 4 is divided into two subdomains, and promptly the upper bound and lower bound then, check that 1 belongs to the upper bound or belongs to lower bound.Because 1 belong to lower bound,, and the maximum in the middle of the coding range is updated to as peaked 1 in the lower bound so 0 is encoded.Therefore, the new coding range that is used for third step is former from 0 to 1 of the lower bound that is used as.

In third step, 0 to 1 coding range is divided into a higher limit and lower limit, checks then whether 1 belong to a upper bound.Because 1 equals higher limit, so 1 is encoded, the minimum value in corresponding coding range is updated to 1 then.Therefore, a new coding range is [1,1], and thus, its minimum and maximum value is mutually the same.When the minimum in coding range is identical with maximum, utilize the encoding process of SQ AAC function to finish.

The flow chart of Fig. 7 D shows and utilizes the encoding process of SQ AAC function to the differential data of the key value data of component, according to the present invention, this processing be entropy coder 350 in carry out.

With reference to Fig. 7 D, in step S692, entropy coder 350 receive the key value data differential data (below be referred to as symbol), to symbol carry out the required bit number nQP of entropy coding, expression will be encoded as bit stream symbol the first subscript n StartIndex and with the quantity (n) of the symbol that is encoded.

Entropy coder 350 is encoded to the sign of symbol in step S694 and symbol transition is become positive number.Become the reason of positive number to be in follow-up processing, not allow negative symbol transition.

Entropy coder 350 is discerned the maximum in the middle of the positive number in step S696, this maximum is stored as nMax, and nMax is encoded to the as many bit with nQP.

350 pairs of entropy coders will carry out initialization to the scope of encoding symbols therein, minimum value that will be in coding range

And maximum Be set to 0 and nMax respectively.In addition, entropy coder 350 is with the mark of a bit

Be assigned to each symbol in the symbol.Mark Be used to during symbolic coding is carried out, change probabilistic model.During beginning with mark Be set to value of false and convert thereof into value of true when the next value that will be encoded is 1, then, the probabilistic model that up to the present has been used to symbolic coding is carried out is replaced by other model.

Then, entropy coder 350 is discerned the difference datum mark (i=nStartIndex) of each component in x, y and the z component in step S700, suppose that it at first is encoded, and mark bDone is set to " very ", whether this description of symbols is finished the coding of all symbols.

Entropy coder 350 repeats the step S702 that will be described below to S718, the minimum and maximum value in the coding range of each that is used for symbol become identical till, promptly till not having remaining symbol to be encoded.Here, when when the minimum and maximum value of the coding range that is used for a symbol is identical, can thinks and finish coding symbol.

Entropy coder 350 is checked in step S702 Whether equal

If Equal

Method moves to the step that is used for next encoding symbols.If not etc., then bDone is set to value of false in step S704, this means the coding of current sign (the current symbol that is being encoded) is not also finished.

Entropy coder 350 calculates the median nMid of the coding range be used for current sign and in step S708 the value of median nMid and current sign is compared at step S706.In a preferred embodiment of the invention, calculate median nMid according to following equation:

$\mathrm{nMid}=\frac{{\hat{P}}_{i\&CenterDot;\max }+{\hat{P}}_{i\&CenterDot;\mathrm{<figure-callout\; id="2"\; label="min"\; filenames="A200510092099.300701.png,A200510092099.300711.png"\; state="\{\{state\}\}">min</figure-callout>}}}{2}+{\hat{P}}_{i\&CenterDot;\min }+1---\left(10\right)$

Result as a comparison if the current sign value is not more than median nMid, this means that current sign belongs to lower bound, then sends 0 to bit stream in step S710, and replaces maximum magnitude in the coding range that is used for current sign with median nMid

On the other hand,, then send 1, and be used for the minimum value of the coding range of current sign with median nMid replacement to bit stream if the current sign value, this means that current sign belongs to the upper bound greater than median nMid

When median nmid that the value of the current symbol that just is being encoded surpasses for the first time, in other words, when issuing bit stream with 1 for the first time, will in step S712 Be set to value of true, thereby change the probabilistic model that is used for encoding symbols.

Two kinds of probabilistic models that use in SQ AAC function are arranged.A kind of is to be used for sending the notFoundContext that the bit of exporting before 1 carries out entropy coding from each symbol for the first time, and another kind is to be used for sending the FoundContext that the bit of exporting after 1 carries out entropy coding from each symbol for the first time just.The purpose of using two kinds of different probability models is to generate 0 probability during being increased in use noFoundContext.Owing to generate 0 probability increase, therefore improved symbol is carried out entropy-coding efficiency.

After this, entropy coder 350 is checked in step S714

Be very or vacation.If Be true, then entropy coder 350 uses FoundContext to encoding symbols in step S716, otherwise entropy coder 350 uses notFoundContext to encoding symbols in step S718.

When finishing when current sign carried out encoding process from step S702 to step S718, this means that this entropy coder only finished the phase I that current sign is encoded, in step S720, the entropy coder 350 subscript i that will encode increases by 1.Then, entropy coder 350 is checked the coding of whether finishing all symbols in the current generation in step S722.If the symbol that in the current generation, will be encoded in addition, then entropy coder 350 once more execution in step S702 to step S722.

When the coding finished all symbols in a stage, in step S724, entropy coder 350 checks that according to mark nDone whether the coding in all stages is all finished.If bDone is true, then entropy coder 350 is finished the whole encoding process in the current generation and is begun next component is encoded, if not, then entropy coder 350 once more execution in step S700 to S722, to remaining encoding symbols.

Fig. 9 B shows with the processing of SQ AAC function to a plurality of encoding symbols.Specifically, Fig. 9 B shows with SQ AAC function 0,1,2,3,4 and 9 processing of encoding.

Referring again to Fig. 6 B, if be used for that data are carried out apparatus for encoding and generate bit stream by all input key value data are carried out entropy coding according to of the present invention, then key value heading code device 360 will be deciphered required information to the key value data of coding and be encoded into the key value heading message in step S730, this key value heading message is added in the bit stream, and the synthetic bit stream of output.

The quantity of the input key value data that 360 pairs of key value heading code devices will be encoded is encoded with counting the number of words.Then, whether key value heading code device 360 identification x, y (for example have identical quantized value with each component in the z component, even the key value data variation of each component in x, y and the z component, but because the degree of the key value data variation of x component is very little, so the quantification key value data of x component have identical value) and the result is encoded into a mark.

For example, if the quantized value of x component is inequality, so, be to have utilized circulation DPCM operation or utilized prediction loop DPCM operation that the key value data of x component are encoded, and be all to be encoded into mark with unary AAC function or with the key value data of x component being encoded with SQ AAC function.If do not expect the key value data of x component to be encoded, so, will all be encoded into the key value title by the beginning subscript of the coded-bit size of the x component of entropy coding and x component with unary AAC function.

Equally, if y is different with the key value data of each component in the z component, so, be encoded into title with complete corresponding y of the heading message of aforesaid x component and the heading message of z.

If the quantized value difference of x component is then encoded to the key value data of x component without unary AAC function, and will be 1, and then, quantize the key value datum mark with first and be encoded to the internal reference point by the beginning subscript of the x component of entropy coding.According to identical mode, first of each component in y and the z component is quantized the key value datum mark be encoded to the internal reference point.

Key value heading code device 360 is identified in which component has maximum magnitude in x, y and the z component.If the x component has maximum magnitude, then key value heading code device 360 variable nWhichAxis are set to 0.If component y has maximum magnitude, then key value heading code device 360 nWhichAxis are set to 1.If component z has maximum magnitude, then key value heading code device 360 nWhichAxis are set to 2.After this, key value heading code device 360 utilizes floating-point encoding device 330 that nWhichAxis, x_min, y_min, z_min and max are encoded into the key title.

Use description to realize according to fourth embodiment of the invention below, data are carried out the computer program of Methods for Coding.

Be used for that data are carried out apparatus for encoding the position interpolater is encoded into bit stream according to of the present invention, this will be described below.In order to understand better, below will be with the formal description of SDL language traditional program code and variable.

Figure 19 shows the example of the program code of realizing the compression position interpolater.In Figure 19, show first class that is used to read a compression position interpolater.First class comprises PoslKeyValueHeader and PoslKeyValue.PoslKeyValueHeader comprises the heading message that is used to decipher PoslKeyValue, and class PoslKeyValue reads the key value data of relevant compression position interpolater from described bit stream.Function qf_start () is used to start an arithmetic decoder before reading the partial bit stream of being encoded by ACC.

Figure 20 shows the program code of PoslKeyValueHeader.With reference to Figure 20, nNumKeyValueCodingBit represents the bit size of nNumberOfKeyValue in the middle of the variable in being stored in PoslKeyValueHeader, nNumberOfKeyValue represents the quantity of key value data and the quantization bit size that nKVQBit represents the key value data.

X_keyvalue_flag, whether y_keyvalue-flag has identical quantized value with z_keyvalueflag explanation x, y with the z component, and nKVDigit represents the maximum number of the highest significant position of each data in the key value data.NKVDPCMOrder_X, nKVDPCMOrder_Y, nKVDPCMOrder_Z correspond respectively to the exponent number of the DPCM of the key value data that are used to each component in x, y and the z component.If carried out DPCM, then this mark is set to 0, if carried out prediction DPCM, then this mark is set to 1.

BlsUnaryAAC_X, blsUnaryAAC_Y, unary AAC function has been used in each explanation among the blsUnaryAAC_Z during entropy coding.NKVCodingBit_X, nKVCodingBit_Y and nKVCodingBit_Z illustrate the coded-bit size that is used for x, y and z component respectively.NStartIndex_X nStartIndex_Y and nStartIndex_Z represent the beginning subscript of each component axle that will be encoded.FirstKV_X, firstKV_Y and firstKV_Z represent the first quantification key value data of each component in x, y and the z component respectively, and they all are encoded into internal data.

Figure 21 shows the program code of class KeyValueMinMax.With reference to Figure 21, class KeyValueMin recovers to be used to the minimum and maximum value of standard key value data.BUse32float illustrates whether 32 bit floating numbers have been used to store minimum and maximum value.If bUse32Float is 0, with the floating number form of ieee standard 754 the minimum and maximum value that is used for standard key value data is encoded so.Otherwise, the minimum and maximum value that is used for standard key value data is encoded with the floating-point encoding device.

NWhichAxis is illustrated in the component that has maximum magnitude in the middle of x, y and the z.In the present embodiment, if nWhichAxis is 0, then X-axis has maximum magnitude, if nWhichAxis is 1, then Y-axis has maximum magnitude, and if nWhichAxis is 2, then the Z axle has maximum magnitude.

Whether the mantissa of the minimum value in the middle of the key value data of BAllSameMantissaDigitFlag explanation each component in x, y and z component has identical counting the number of words, and bSameKVDigitFlag represent minimum value mantissa count the number of words identical with nKVDigit.NMantissaDigit_X, nMantissaDigit_Y and nMantissaDigit_Z are illustrated respectively in the counting the number of words of mantissa of the minimum value in the middle of the key value data of each component of x, y and z component.

BMaxDigitFlag illustrates whether peaked mantissa has identical the counting the number of words of minimum value with the component that therefrom obtains maximum max.If the mantissa that the counting the number of words of peaked mantissa is different from minimum value counts the number of words, then from bit stream, read counting the number of words of peaked mantissa.

NMantissaDigit_M represents counting the number of words of peaked mantissa, and nExponentBits represents the required bit number of encoding of the maximum adiabatic index in the middle of the index of minimum and maximum value.Whether the sign of BAllSameExponentSign explanation x, y and z component is identical, and when bAllSameExponentSign is a true time, nExponentSign represents the sign of index.

FpnMin_X, fpnMin_Y, fpnMin_Z and fpnMax are illustrated in floating number decoded in the decimal system.To the method that fpnMin_X, fpnMin_Y, fpnMin_Z and fpnMax are deciphered be described with FloatingPointNumber below.FMin_X, fMin_Y and fMin_Z are illustrated in the central minimum value of key value of each component in x, y and the z component, and fMax is illustrated in the central maximum of key value data of the component with maximum magnitude.

The program code of class FloatingPointNumber is described below in conjunction with Figure 22.Class FloatingPointNumber uses decimal system to represent floating number.NMantissa represents the mantissa of the floating number in the decimal system, and nExponent represents the index of this floating number.

NSign represents the sign of floating number, and nExponentSign represents the sign of the index of this floating number.

The program code of class PoslKeyValue is described below in conjunction with Figure 23.Referring to Figure 23, in the variable in being stored in class PoslKeyValue, keyValue_X, keyValue_Y and keyValue_Z are illustrated respectively in the array of the key value data of x, y in the interpolater of position and each component in the z component.If nStartIndex_X is set to 0, then use arithmetic decoder from bit stream, to decipher keyValue_X[0].If nStartIndex_X is set to 1, then from key value title decoder, decipher keyValue_X[0].In an identical manner, determine keyValue_Y[0] and keyValue_Z[0].As arithmetically decoding keyValue_X[0 from bit stream], keyValue_Y[0] and keyValue_Z[0] time, use decodeUnaryAAC or decodeSQAAC function.

Context model kVXSignContext, kVYSignContext and kVZSignContext are respectively applied for the sign of keyValue_X, keyValue_Y and keyValue_Z are deciphered.Context model kVXSignContext, kVYSignContext and kVZSignContext are sent to decodeUnaryAAC or decodeSQAAC function.

MaxValueContext, FoundContext and NotFoundContext are used to the absolute value of key value is carried out entropy decoding.For example, kVXMaxValueContexr, kVXFoundContext and kVXNotFoundContext are used to keyValue_X is deciphered.MaxValueContext, FoundContext and NotFoundContext are sent to decodeUnaryACC or decodeSQAAC function.

Context model kVXUCotext, kVYUContext and kVZUContext are used to keyValue_X, keyValue_Y and keyValue_Z are deciphered, and are sent to the decodeUnaryAAC function.

Below in conjunction with Figure 10 A and 10B the differential data that is generated by aforementioned DPCM operator by recovering is described, the contrary DPCM operator of output quantized data.

The block diagram of Figure 10 A shows according to contrary DPCM operator of the present invention, and Figure 10 B is the flow chart of contrary DPCM operation.

Referring to Figure 10 A, comprise according to contrary DPCM operator of the present invention: contrary circulation DPCM operator 1020, be used for the differential data that inputs to it is carried out contrary circulation quantization operation so that expand their scope, the input difference data are carried out contrary DPCM operation, export quantized data then; Inverse prediction circulation DPCM operator 1030 is used for the differential data that inputs to it is carried out contrary circulation quantization operation so that expand their scope, and the input difference data are carried out inverse prediction DPCM operation, exports the data that quantize then; And determining unit 1010, be used for the DPCM kind that basis has been carried out differential data, differential data is exported to contrary circulation DPCM operator 1020 or inverse prediction circulation DPCM operator 1030.

Below in conjunction with Figure 10 B contrary DPCM operation is described.Referring to Figure 10 A and 10B, the differential data of the contrary DPCM operation of experience is transfused to determining unit 1010.Then, in step S1010, which kind of DPCM determining unit 1010 identifications have carried out to the input difference data, determine to carry out which kind of contrary DPCM to the input difference data according to the result, and differential data is exported to contrary circulation DPCM operator 1020 or inverse prediction circulation DPCM operator 1030.

If it is maximum in the quantizing range of input difference data that described differential data is transfused to contrary circulation DPCM operator 1020 and nMax, contrary circulation DPCM operator 1020 by in step S1020 to the input difference data

Execution quantizes to obtain the differential data of quantification according to the contrary circulation of following equation

After this, contrary circulation DOCM operator 1020 is by difference input difference data

With the differential data that will quantize from the contrary circulation that equation (11) obtains

Substitution equation (12) obtains against the value B of the differential data of the value A of the differential data of DPCM and contrary DPCM.

In equation (12), n represents the quantity of data, and i represent 1 and n-1 between integer.

If A is not less than 0 and be not more than nMax, then contrary circulation DPCM operator 1020 output A are as the data of contrary circulation DPCM If A is less than 0 or greater than nMax, then contrary circulation DPCM operator 1020 is exported B in step S1040.

Inverse prediction circulation DPCM operator 1030 passes through to use the differential data of equation (11) to prediction at step S1020 Carry out contrary circulation and quantize to obtain the predicted difference divided data that contrary circulation quantizes

If $P=2\&times;{\tilde{P}}_{i-1}+{\tilde{P}}_{i-2},$ Wherein P represents the prediction data that is used to decipher, and then contrary circulation DPCM operator 1030 will be by will Substitution equation (13) calculates A and incites somebody to action

Substitution equation (14) calculates B.

In equation (13) and (14), n represents the quantity of data, i represent 1 and n-1 between integer.

If A is not less than 0 and be not more than nMax, so, contrary circulation DPCM operator 1020 output A are as the data of inverse prediction circulation DPCM If A is less than 0 or greater than nMax, then inverse prediction circulation DPCM operator 1020 is exported the B conduct in step S1040

Describe according to first embodiment of the invention below in conjunction with Figure 11, be used for the device that data are deciphered.The device use that is used for data are deciphered according to first embodiment of the invention flows to row decoding according to DPCM operator of the present invention to bit, in this bit stream, the input data have been utilized according to first embodiment of the invention, are used for data are carried out the apparatus for encoding coding.

Referring to Figure 11, according to the contrary DPCM processor 1130 and the inverse quantizer 1140 that the device that data are deciphered is comprised entropy encoder 1120, is made up of aforementioned contrary DPCM operator of being used for of first embodiment of the invention.

Entropy encoder 1120 generates differential data and this differential data is exported to contrary DPCM processor 1130 by incoming bit stream being carried out entropy decoding.

As mentioned above, which kind of DPCM contrary DPCM processor 1130 identifications have carried out to the input difference data, and differential data is carried out contrary circulation DPCM operation or inverse prediction circulation DPCM operation, and quantized data is exported to inverse quantizer 1140.

Inverse quantizer 1140 utilizes predetermined quantization bit, to carry out the data that re-quantization and output obtain again from the quantized data against 1130 inputs of DPCM processor.

Describe according to second embodiment of the invention below in conjunction with Figure 12 A and 12B, be used for the device that data are deciphered.According to second embodiment of the invention be used for the device that data are deciphered is flow to row decoding to bit, in this bit stream, the data with a plurality of components have been utilized according to normalizer standard of the present invention.

Referring to Figure 12 A, according to second embodiment of the invention, be used for the device that data are deciphered is comprised: entropy encoder 1120 is used for incoming bit stream is carried out entropy decoding and output differential data; Contrary DPCM operator 140/1130 is used for carry out contrary DPCM operation and output quantized data from the differential data of entropy encoder 1120 inputs; Inverse quantizer 1140 is used for the quantized data from contrary DPCM operator 140/1130 input is carried out re-quantization and exports authority data; And contrary normalizer 1150, be used to receive the minimum value in the middle of the key value data of each component and have the maximum of the component of maximum magnitude, calculate the maximum magnitude of authority data, and the quantized data from inverse quantizer 1140 inputs is carried out contrary standard according to this maximum magnitude.

The operation of entropy encoder 1120, contrary DPCM operator 140/1130 and inverse quantizer 1140 and structure and according to first embodiment of the invention, the operation of respective element that is used for device that data are deciphered is identical with structure, no longer is repeated in this description here.Below in conjunction with Figure 12 B the operation of contrary normalizer 1150 is described.

Contrary normalizer 1150 receives in step S1210 to be deciphered from incoming bit stream, the maximum fMax that is used for standard, and receive minimum value fMin_X, fMin_Y and the fMin_Z of x, y and z component and the information bUse32Float which component axle to have peaked information nWhichAxis and relevant type of coding about respectively.

Contrary normalizer 1150 is encoded according to the floating number form whether bUse32Float identification fMax, fMin_X, fMin_Y and fMin_Z are utilized ieee standard 754, if not, the following equation of then contrary normalizer 1150 uses calculates the maximum Range in the maximum magnitude _MaxIf fMax, fMin_X, fMin_Y and fMin_Z are by with 32 codings, then contrary normalizer 1150 is defined as Range with fMax in step S1220 _Max

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 Range _MaxBe determined, contrary normalizer 1150 in step S1230 according to Range _Max, utilize following equation that the data of each component of x, y and z component are carried out standard.

$\begin{array}{c}{\hat{P}}_{i,x}={\stackrel{\widehat{~}}{P}}_{i,x}\&times;{\mathrm{Range}}_{\max }+\mathrm{fMin}\_X\\ {\hat{P}}_{i,y}={\stackrel{\widehat{~}}{P}}_{i,y}\&times;{\mathrm{Range}}_{\max }+\mathrm{fMin}\_Y\\ {\hat{P}}_{i,z}={\stackrel{\widehat{~}}{P}}_{i,z}\&times;{\mathrm{Range}}_{\max }+\mathrm{fMin}\_Z\end{array}(i=\mathrm{0,1},...,n)---\left(16\right)$

In equation (16), n represents the quantity of data.

Describe according to third embodiment of the invention below in conjunction with Figure 13 A and 13B, be used for the apparatus and method that data are deciphered.According to third embodiment of the invention, be used for the device that data are deciphered is flow to row decoding to bit, utilized according to of the present invention data have been carried out Methods for Coding, the key value digital coding of representing the position interpolater of the target in the three-dimensional animation has been advanced in this bit stream.

Referring to Figure 13 A, according to third embodiment of the invention, be used for the device that data are deciphered is comprised: key value title decoder 1110 is used for deciphering from the required information of the decoding of bit stream; Entropy encoder 1120 is used for the bit stream that inputs to it is carried out entropy decoding, exports differential data then; Contrary DPCM processor 1130 is used for the contrary circulation DPCM operation of the differential data execution that inputs to it from entropy encoder 1120 or inverse prediction circulation DPCM operation, the key value data of output quantification then; Inverse quantizer 1140 is used for carrying out re-quantization from the quantification key value data of contrary DPCM processor 1130 inputs and exporting the key value data of standard; Floating number decoder 1160 is used for receiving minimum and maximum and converting minimum value in the middle of the key value data of each component of x, y and z component and maximum with component of maximum magnitude to binary number from key value title decoder 1110; And contrary normalizer 1150, be used for receiving the minimum value in the middle of the key value data of each component of x, y and z component and having the maximum of the component of maximum magnitude from floating number decoder 1160, according to minimum value and maximum from floating number decoder 1160, calculate maximum magnitude, then to carrying out contrary standard from the key value data of inverse quantizer 1140 inputs.

Describe according to third embodiment of the invention below in conjunction with Figure 13 B, be used for the method that data are deciphered.With reference to Figure 13 B, will input to entropy encoder 1120 and key value title decoder 1110 through the bit stream of key value data of coding.Then, in step S1300, decipher and the heading message of decoding is exported to entropy encoder 1120, contrary DPCM processor 1130 and inverse quantizer 1140 to deciphering by required heading message from bit stream for key value title decoder 1110.

In step S1320,1120 pairs of bit streams of entropy encoder carry out entropy decoding and differential data are exported to contrary DPCM processor 1130.

Figure 14 A is the detail flowchart of step S1320.With reference to Figure 14 A, in step S1321, bit stream Pi is inputed to entropy encoder 1120.Figure 15 shows the order of the key value data that input to process coding entropy encoder 1120 and that will carry out entropy decoding.

Entropy encoder 1120 is checked component in step S1322, x for example, the key value data whether have identical quantized value.If the key value data of component have identical quantized value, then in step S1323 with all symbol substitutions of this component, and they are set to from the minimum value of key value title decoder 1110 inputs, for example fMin_X.

If the key value data of component have different quantized values, then entropy encoder 1120 checks in step S1324 whether these key value data have been utilized unary AAC function coding.If these key value data have been utilized unary AAC function coding, then utilize unary AAC function that they are deciphered at step S1325.

Unary AAC function reads 0 from bit stream, till bit " 1 " occurs, convert 0 quantity to absolute value, read then a sign of " 1 " as value, if this position is " 0 ", then Zhi sign is for just, if this position is " 1 ", then Zhi sign is negative, and output is through the value of decoding.

If the key value data of component are not utilized unary AAC function coding, then entropy encoder 1120 utilizes SQ AAC function that bit is flow to row decoding, this bit stream is meant and utilizes SQ AAC function that wherein bit stream is advanced in the key value digital coding of component that this will be described below.

In step S1331, entropy encoder 1120 sign to all symbols from bit stream is deciphered, and nMax deciphers to maximum, and with the decoding range (scope from 0 to nMax) and the mark of all symbols

Be initialized as " vacation ".

After this, in step S1332, entropy encoder 1120 determines that the key value datum mark (i=nStartlndex) and the bDone that will at first be deciphered are set to value of true.

Entropy encoder 1120 execution in step S1333 is once more deciphered symbol to S1343.

Specifically, in step S1333, the maximum that entropy encoder 1120 is checked in decoding range And minimum value

Whether identical.

If maximum

And minimum value

Identical, then in step S1334, the symbol that is encoded is confirmed as maximum (or minimum value) in the decoding range, if not, then in step S1335, entropy encoder 1120 bDone are set to value of false, and utilize following equation (17) to upgrade the interior median nMid of decoding range.

Entropy encoder 1120 is checked contextual tagging in step S1336

Whether be true, thus the probabilistic model that is identified for deciphering.If Be true, then entropy encoder 1120 utilizes FoundContext that bit is flow to row decoding in step S1337, otherwise entropy encoder 1120 utilizes notFoundContext that bit is flow to row decoding in step S1338.

The entropy encoder of deciphering from a bit of bit stream 1120 is checked in step S1339 whether this bit of reading is 1 from bit stream, if not, then entropy encoder 1120 is used the maximum in the median nMid replacement decoding range in step S1340

If from the bit of bit stream decoding is 1, then entropy encoder 1120 utilizes the minimum value in the median nMid replacement decoding range in step S1341

And with contextual tagging Be set to value of true.

Entropy encoder 1120 will be deciphered subscript i and add 1 in step S1342, whether all key value data of checking component in the current generation in step S1343 are deciphered fully, if also residue is wanted decoded symbol in the current generation, then execution in step S1333 is to S1342.

If with all key value data decodings of current generation, whether entropy encoder 1120 is checked in step S1344 needs additional phase to come the key value data of current component are deciphered, if, then once more execution in step S1332 to S1343.If entropy encoder checks all key values of current component is deciphered that then entropy encoder begins the key value data of next component are deciphered.

Referring to Figure 13 A and 13B, differential data by the key value data of entropy encoder 1120 decoding is transfused to contrary DPCM processor 130, and contrary then DPCM processor 130 is carried out contrary circulation DPCM operation or inverse prediction circulation DPCM operation according to the type of performed DPCM.Except the key value data of each component of x, y and z component were carried out separately, contrary circulation DPCM operation and prediction loop DPCM operation were identical with aforementioned operation against the DPCM operator.

Contrary DPCM processor 1130 will be by the quantification key value data of contrary DPCM operation generation Export to inverse quantizer 1140, then, in step S1370, it is right that inverse quantizer 1140 utilizes from the quantization bit nKeyValueQBits of key value title decoder 1110 inputs Carry out re-quantization.When n represents

${\hat{P}}_{\mathrm{ij}}=\frac{{\tilde{P}}_{\mathrm{ij}}}{(2^{\mathrm{nKeyValueQBits}}-1)}(i=\mathrm{nStartIndex},...,n,j=x,y,z)---\left(18\right)$

During from the quantity of the key value data of key value title decoder 1110 input, the re-quantization of carrying out in inverse quantizer 1140 can utilize following equation to represent.

Inverse quantizer 1140 generates specification data by the key value data of each component in x, y and the z component are carried out re-quantization, and the key value data of the standard of each component in contrary normalizer 1150 output x, y and z component.

Contrary normalizer 1150 utilizes from the information of floating number decoder 1160 inputs, and the key value data from the standard of inverse quantizer 1140 inputs are carried out contrary standard.

The contrary standard that the standard key value data of each component in x, y and the z component are carried out is described below in conjunction with Figure 14 C.

Floating number decoder 1160 receives information bUse32Float and checks whether bUse32Float is true.If bUse32Float is true, then the floating number decoder receive the maximum fMax that is used for contrary standard and have x, the y of floating number form of ieee standard 754 and the key value data of each component of z component in the middle of minimum value fMin_X, fMin_Y and fMin_Z.If not, then the floating number decoder receives minimum value fMin_X, fMin_Y and the fMin_Z in the middle of the key value data of each component of the maximum fMax that is used for contrary standard and x, y and z component in step S1382, and receives the information nWhichAxis that belongs to which component about maximum fMax.

Floating number decoder 1160 converts binary number to according to fMax, fMin_X, fMin_Y and fMin_Z that equation (9) will be represented in decimal system in step S1384, and the value after will changing is exported to contrary normalizer 1150.

Contrary normalizer 1150 checks according to bUse32Float whether fMax, fMin_X, fMin_Y and fMin_Z have been encoded as 32 bits.If fMax, fMin_X, fMin_Y and fMin_Z have been encoded as 32 bits, then contrary normalizer 1150 is defined as Range with fMax _Max, in step S1386, calculate maximum Range in the maximum magnitude according to equation (15) if not, then contrary normalizer _Max

If Range _MaxBe determined, then contrary normalizer 1150 in step S1388 according to Range _Max, the key value data of each component in x, y and the z component are carried out contrary standard according to equation (16).

Figure 24 to 27 shows according to third embodiment of the invention, is used for the example of the program code of device operation that data are deciphered.

Figure 24 shows the example of the C++ program code that is used to realize contrary circulation DPCM operation, and Figure 25 shows the example of the C++ program code that is used to realize inverse prediction circulation DPCM operation.

Contrary circulation DPCM operation and inverse prediction circulation DPCM operation all comprise the contrary circulation quantification program that is used for selecting at quantizing range certain value.

The example that is used for realizing the C++ program code of entropy decoding (self adaptation arithmetic decoder) operation is shown in Figure 26 to 31.Here, qf_decode () expression is used for reading from bit stream the function of a bit.

The computer-readable code that utilization is written on the computer readable recording medium storing program for performing can be realized the present invention.Here, computer readable recording medium storing program for performing comprises any recording medium that can be read by computer system.For example, computer readable recording medium storing program for performing can be ROM, RAM, CD-ROM, tape, floppy disk, optical data storage and carrier wave (passing through internet transmission).Computer readable recording medium storing program for performing can be transferred to the computer system that connects by network, and computer can be with the dispersing mode reading ﹠ recording medium.

As mentioned above, can think, operate and circulation quantization operation and traditional DPCM operation by carrying out prediction DPCMA, can significantly reduce the scale of differential data according to DPCM operator of the present invention and operation thereof.In addition, data are encoded, can improve the efficient that data are encoded significantly by utilizing according to DPCM operation of the present invention.

In addition,, the data with a plurality of components are carried out standard, can improve the efficient that data are encoded, thereby reduce the scale of data according to the one-component in x, y and the z component by utilizing according to normalizer of the present invention.

In addition, by after carrying out DPCM operation and prediction DPCM operation, carrying out the circulation quantization operation, can encode to the key value data of position interpolater and decipher with higher compression ratio, thus the scope of the differential data that minimizing will be encoded.

Although the present invention describes with reference to its preferred embodiment, it should be appreciated by those skilled in the art, can be under the situation that does not break away from the spirit and scope of the present invention that are defined by the following claims, it is carried out various modifications on form and the details.

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