JP2005080004A - Dynamic image encoding device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To generate encoded data whose image deterioration can be prevented and which can be decoded by a standard decoding device. <P>SOLUTION: Since a skip code D12 composed of data following an MPEG2 system instead of encoded data D6 can be outputted when an actually encoded code amount increases more than a predicted code amount by providing a VBV (video buffering verifier) occupancy amount calculating part 21 for calculating an occupancy amount in a VBV buffer and a skip control part 23 and skip code generating part 13 for outputting the skip code D12 when the VBV buffer causes underflow, the underflow of the VBV buffer and output encoded data D13 which can be normally recognized without using a special detection circuit or the like can be generated and outputted. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a moving image encoding apparatus, and is suitable for application to, for example, a digital video recording / reproducing apparatus.

  2. Description of the Related Art Conventionally, a moving image encoding apparatus digitizes a signal composed of a plurality of continuous images supplied from the outside (hereinafter referred to as a moving image signal), and then performs an encoding process based on a predetermined image encoding method. By performing the above, for example, encoded data of a predetermined unit such as one frame is generated.

  As such an image encoding method, an image encoding method called MPEG standardized for the purpose of encoding general-purpose images by an MPEG / Moving Picture Expert Group (ISO / IEC) is known.

  In the MPEG2 system, which is standardized as one of the MPEG encoding systems, a moving picture (not shown) is used to monitor the transition of the code amount of the actually encoded moving picture signal (hereinafter referred to as encoded data). Assuming a virtual VBV (Video Buffering Verifier) buffer corresponding to the internal buffer in the decoding apparatus, the code amount of the encoded data in the VBV buffer (hereinafter referred to as the occupation amount) is calculated as needed. The bit rate of the encoded data is set to a set value by adjusting the code amount of the encoded data by controlling the quantization step according to the occupied amount.

  For example, in the moving image encoding apparatus, as shown in FIG. 7, after the encoded data in the frame (n−1) is output, the encoded data in the frame (n) is output in the VBV buffer. The code amount a (n) of the encoded data is calculated.

  Next, the moving image encoding apparatus predicts by calculating the code amount b (n) of the encoded data of the frame (n) based on the encoded data of the frame (n−1), and the frame (n−1) ) Based on the VBV buffer occupancy V (n-1) and the preset bit rate so that the upper and lower limits of the VBV buffer are not exceeded when the encoded data of frame (n) is output. (FIG. 7) After adjusting the quantization scale and the like in the frame (n), the image of the frame (n) is encoded.

  However, when a moving image signal is actually encoded, the actual code amount after encoding may exceed the code amount of the prediction result, for example, in a scene switching scene.

  For this reason, the moving image encoding device monitors the VBV buffer even during the sequential encoding from the upper end of the image of the frame (n). For example, the frame (n) calculated based on the code amount at the upper end of the image. ) When the predicted code amount be (n) (FIG. 7) exceeds the initially predicted code amount b (n), causing an underflow of the VBV buffer. Therefore, the underflow of the VBV buffer is prevented by reducing the generated code amount and suppressing the generated code amount to about the code amount b (n).

  However, when such encoded data is decoded by a decoding device or the like, the image quality of the image of the frame (n) gradually becomes worse from the upper part to the lower part, and the image quality as a moving image is deteriorated. There was a problem.

As a conventional method for solving such a problem, a code indicating that an image of one frame is substituted with an image of the previous frame is set in advance using an MPEG user data area or the like, There has been proposed a moving picture coding apparatus that outputs a code to a user data area instead of coded data when an underflow of a VBV buffer is likely to occur during signal coding (for example, , See Patent Document 1).
Japanese Patent No. 2871316 (page 6, FIG. 1)

  However, in a moving image encoding apparatus to which such a technique is applied, a unique code not defined in the MPEG system is stored in the user data area, and therefore a dedicated code for detecting the code on the decoding apparatus side. It was necessary to provide a detection circuit and the like.

  For this reason, there has been a problem that a general decoding device (not equipped with a detection device) that is widely used cannot correctly decode the encoded data including the skip code described above.

  The present invention has been made in consideration of the above points, and an object of the present invention is to propose a moving picture coding apparatus that can be easily decoded by a decoding apparatus without preventing deterioration in image quality.

  In order to solve such a problem, in the present invention, in a moving image encoding apparatus that performs encoding processing on moving image data according to a predetermined encoding method, and outputs encoded data sequentially generated for each predetermined unit to the outside. Predicted occupancy calculation means for calculating the occupancy of the encoded data when accumulated in the decoding side buffer as the predicted occupancy, and the predicted occupancy calculated by the predicted occupancy calculation means exceeds a predetermined threshold. In place of the encoded data generated at the time, code output means for generating and outputting an instruction code for re-displaying the decoding result of the encoded data before the time according to the encoding method is provided. As a result, when the code amount of the encoded data increases more than the predicted occupancy amount, the code output means uses the data according to the encoding method instead of the encoded data. And to output it generates instruction code consisting of growth.

  As a result, in this moving image encoding apparatus, the encoding amount of encoded data can be suppressed so as not to exceed a predetermined threshold, and encoding that can be normally recognized by the decoding apparatus without using a special detection circuit or the like. Data can be generated.

  According to the present invention, when the code amount actually encoded increases more than expected, an instruction code composed of data according to the encoding method can be output instead of the encoded data. A moving image capable of generating encoded data that can be suppressed so that the code amount does not exceed a predetermined threshold, can be normally recognized by a decoding device without using a special detection circuit, and can prevent deterioration in image quality An encoding device can be realized.

  Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings.

(1) Configuration of Moving Image Encoding Device As shown in FIG. 1, reference numeral 1 denotes a moving image encoding device as a whole. According to the MPEG2 system, image data supplied from the outside (hereinafter referred to as frame data D1). The encoding process is executed sequentially for each call.

  In this case, when the prediction data D2 is not given from the motion compensation circuit 12, the subtracter 2 sends the frame data D1 as it is to the orthogonal transform processing unit 3 as difference data D3.

  The orthogonal transform processing unit 6 generates orthogonal transform coefficient data D4 by performing orthogonal transform processing such as discrete cosine transform on the difference data D3, and sends this to the quantization processing unit 4.

  The quantization processing unit 4 quantizes the orthogonal transform coefficient data D4 by performing a quantization process according to the quantization parameter value of the quantization control signal S1 given according to the feedback control by the encoding control unit 5. Data D5 is generated and sent to the variable length coding control unit 6 and the inverse quantization processing unit 8, respectively.

  The variable length encoding control unit 6 generates encoded data D6 by performing variable length encoding processing on the quantized data D5 given from the quantization processing unit 4, and stores this in the buffer 7.

  On the other hand, the inverse quantization processing unit 8 restores orthogonal transformation coefficient data D8 corresponding to the orthogonal transformation coefficient data D4 by performing inverse quantization processing on the quantized data D5 given from the quantization processing unit 4. This is sent to the inverse orthogonal transform processing unit 9.

  The inverse orthogonal transform processing unit 9 performs inverse orthogonal transform processing on the orthogonal transform coefficient data D8 given from the inverse quantization processing unit 8, thereby restoring the difference data D9 corresponding to the difference data D3 and adding this. To the device 10.

  When the prediction data D2 is given from the motion compensation processing unit 12, the adder 10 adds the prediction data D2 corresponding to the difference data D9 to the difference data D9 given from the inverse orthogonal transform processing unit 9 as necessary. By adding, the frame data D10 corresponding to the frame data D1 is restored and stored in the frame memory 11.

  The motion compensation processing unit 12 sequentially reads frame data D11 for one frame from the frame memory 11, generates prediction data D2 based on a predetermined motion compensation process, and sends this to the subtracter 2 and the adder 10.

  Here, when the prediction data D2 is given as the motion compensation result in the motion compensation circuit 12, the subtracter 2 generates difference data D3 by subtracting the prediction data D2 corresponding to the frame data D1 from the frame data D1. This is sent to the orthogonal transform processing unit 3.

  The difference data D3 is sequentially subjected to orthogonal transform processing, quantization processing, and variable length coding processing as in the case described above, and as a result, is stored in the buffer 7 as encoded data D6.

  In this way, the moving image encoding apparatus 1 executes the encoding process for each input frame data D1, and sequentially outputs the encoded data D6 stored in the buffer 7 via the switch 14 to the output encoded data D13. As output.

(2) Configuration of Encoding Control Unit 5 Here, the encoding control unit 5 of the moving image encoding device 1 includes a VBV occupation amount calculation unit 21, a skip determination unit 22, and rate control as shown in FIG. Part 23.

  The VBV occupation amount calculation unit 21 is configured to calculate the occupation amount in the virtual VBV buffer. For example, when focusing on the frame (n), the encoding of the frame (n−1) which is the previous frame is performed. At the time of sending the data D6, the code amount b (n-1) of the encoded data D6 as shown in FIG. 3 and the VBV buffer occupation amount V (n-1) at this time are recognized. The occupation amount V (n−1) and the code amount b (n−1) are sent to the rate control unit 22 as occupation code amount data D21.

  Then, the VBV occupation amount calculation unit 21 acquires the encoded data D6 from the buffer 7, recognizes the code amount b (n), and outputs the output encoded data D13 of the frame (n) in the VBV buffer. Occupancy V (n)

And is sent as occupancy data D22 to the skip control unit 23.

  When the VBV occupation amount calculation unit 21 receives a notification (details will be described later) that the skip code D12 has been output from the skip control unit 23, the VBV occupation amount calculation unit 21 recognizes the skip code D12 as the encoded data D6 of the frame (n). It is made like that.

  The rate control unit 22 performs variable-length coding processing on the frame data D1 of the frame (n) based on the code amount b (n−1) given as the predicted occupied code amount data D21 and a preset bit rate. A prediction code amount be (n) when converted into encoded data D6 by the unit 6 is calculated.

  Further, the rate control unit 22 outputs the encoded data amount a accumulated in the VBV buffer between the time when the encoded data D6 of the frame (n−1) is output and the time when the encoded data D6 of the frame (n) is output. Using (n) and the predicted code amount be (n) of the frame (n), the predicted occupation amount Ve (n) of the VBV buffer at the time when the encoded data D6 corresponding to the frame (n) is sent is obtained. Next formula

Calculated by

  Here, the rate control unit 22 determines the quantization parameter so that the predicted code amount be (n) satisfies the predicted occupation amount Ve (n)> 0 in order not to cause an underflow of the VBV buffer, The quantization control signal S1 is sent to the quantization processing unit 4.

  As a result, the quantization processing unit 4 sends the data to the orthogonal transform processing unit 6 as quantized data D5 that does not cause an underflow of the VBV buffer.

  The skip control unit 23 determines whether or not the occupation amount V (n) <0 based on the occupation code amount data D22 given from the VBV occupation amount calculation unit 21, that is, whether or not an underflow of the VBV buffer occurs. judge.

  If the occupancy V (n) <0, the skip control unit 23 outputs a skip code generation signal S3 to the skip code generation unit 13 to skip to the skip code generation unit 13 (FIG. 1). By generating the code D12 and outputting the output switching signal S2, the switch 14 (FIG. 1) is switched to the k side (FIG. 1), the skip code D12 is output as the encoded data D13, and the VBV occupation amount is calculated. Notifying the unit 21 that the skip code D12 has been output.

  On the other hand, when the occupation amount V (n) <0 is not satisfied, the skip control unit 23 switches the switch 14 (FIG. 1) to the j side by outputting the output switching signal S2 (FIG. 1), and the encoded data D6 is transferred. The encoded data D13 is output as it is.

  In this way, the encoding control unit 5 calculates the occupation amount V (n) of the VBV buffer using the encoded data D6, determines the quantization parameter of the quantization processing unit 4, and, if necessary, the code A skip code D12 is output as output encoded data D13 instead of the encoded data D6.

(3) Configuration of Skip Code D12 This skip code D12 is configured by data corresponding to an image for one frame, as in the case of normal encoded data (including data indicating an image). In the case of being composed of lines, if one macroblock is composed of 16 × 16 pixels, the one frame is divided into 30 slices in the line direction.

  This slice has a data configuration as shown in FIG. 4, and the (m) th slice 30 is composed of a header portion 31 and a data portion 32.

  The header section 31 includes a slice start code 31A, a quantization scale code 31B, a macroblock type 31C, and a frame motion type 31D, and has a configuration conforming to the MPEG2 system.

  The data section 32 includes a macro block escape 32A, a macro block address addition 32B, a macro block type 32C, and a frame motion type 32D, and has a configuration in which data representing an image is deleted from the data section of a normal slice layer. It has become.

  That is, the slice 30 constitutes a so-called “slice without image data” in which only the data indicating the image is removed while maintaining the header of the standard slice layer, and so on. 30 is interpreted as data indicating that “the same content as the slice at the same position in the previous frame is repeated” in the decoding apparatus based on the MPEG standard.

  Incidentally, the data such as “slice without image data” is originally completely different from the slice (q) in which a certain slice (q) in a certain frame (p) represents the same position in the previous frame (p−1). In order to avoid redundancy and reduce the amount of code when the same code is used, it is defined in the MPEG2 system.

  In addition, since the slice 30 has a configuration in which encoded data representing an image is deleted, the minimum amount of code is required as data for one slice.

  The skip code D12 is configured by combining, for example, 30 slices 40 similar to the slice 30. For this reason, the skip code D12 has a code amount bsc of the code amount b of the frame (n). This is much less than (n) (FIG. 3).

  Further, since the skip code D12 is configured according to the MPEG2 system, the frame (n) is converted by a moving picture decoding apparatus (hereinafter referred to as a standard moving picture decoding apparatus) that does not include a special detection circuit or the like. The same image as the previous frame (n-1) can be displayed again at the timing to be displayed.

(4) Moving Picture Encoding Process Procedure Next, a moving picture encoding process procedure when the skip code D12 is output instead of the encoded data D6 as necessary will be described in detail with reference to the flowchart of FIG. .

  That is, the moving image encoding apparatus 1 enters from the start step SP0 of the routine RT1 and proceeds to step SP1, and based on the frame data D1 of the frame (n) as shown in FIG. The variable-length encoding processing unit 6 or the like waits until the encoded data D6 of the frame (n) as shown in FIG. 6B is generated, and when the encoded data D6 is generated, the process proceeds to the next step SP2. Move.

  The encoding control unit 5 does not calculate the predicted occupation amount Ve (n) of the VBV buffer during the generation of the encoded data D6 of the frame (n) for one frame. 5 to reduce the processing load.

  In step SP2, the encoding control unit 5 determines whether or not the occupation amount V (n) <0 when the output encoded data D13 of the frame (n) is output.

  If a negative result is obtained here, this indicates that the underflow of the VBV buffer does not occur when the encoded data D6 of the frame (n) is output. At this time, the encoding control unit 5 Control goes to step SP3.

  In step SP3, the encoding control unit 5 switches the switch 14 to the j side by outputting the output switching signal S2 (FIG. 1), and outputs the encoded data D6 as it is as the encoded data D13, and the next step SP5. Move on.

  On the other hand, if an affirmative result is obtained in step SP2, this indicates that an underflow of the VBV buffer occurs when the encoded data D6 of the frame (n) is output (FIG. 3). At this time, the encoding control unit 5 proceeds to the next step SP4.

  In step SP4, the encoding control unit 5 outputs the skip code generation signal S3 to the skip code generation unit 13 to generate the skip code D12 and send it to the switch 14, and further switches the output by outputting the output switching signal S2. The device 14 is switched to the k side (FIG. 1), the skip code D12 is output as encoded data D13, and the process proceeds to the next step SP5.

  In step SP5, the encoding control unit 5 determines whether or not all the frame data D1 has been encoded. If a negative result is obtained here, the process returns to step SP1 again and repeats the processing of steps SP1 to SP4 described above. On the other hand, if an affirmative result is obtained here, the process proceeds to the next step SP6, and the moving image encoding processing procedure is terminated.

(5) Decoding Skip Code Next, a case where encoded data including such a skip code D12 is decoded will be described.

  That is, as described above, when the frame data D1 (FIG. 6A) of the frame (n) is simply encoded, the encoded data D6 is generated (FIG. 6B).

  However, when underflow of the VBV buffer occurs in frame (n), output encoded data D13 in which skip code D12 is inserted instead of encoded data (n) is generated (FIG. 6C).

  On the other hand, when such output encoded data D13 is decoded by a general moving picture decoding apparatus (not shown), the moving picture decoding apparatus has a skip code D12 at a position corresponding to the frame (n). Since the skip code D12 does not include data indicating an image, the image of the frame (n−1) is displayed again at the timing of the frame (n) in accordance with the MPEG2 standard (FIG. 6D )).

  At this time, the moving picture decoding apparatus can normally recognize the skip code D12 composed of standard data at the timing of displaying the frame (n), and the image of the frame (n-1) one frame before is again displayed. By displaying it, it is possible to decode a moving image without image quality degradation.

(6) Operation and Effect In the above configuration, the moving image encoding apparatus 1 sequentially encodes the input frame data D1 and outputs it as output encoded data D13, and also the VBV occupation amount calculation unit of the encoding control unit 5 21, the occupancy in the VBV buffer is monitored using the actual code amount b (n) of the encoded data D6, and the frame (n) in which the underflow has occurred in the VBV buffer is encoded. A skip code D12 is inserted in place of the digitized data D6 (FIGS. 6B and 6C).

  At this time, the moving image encoding apparatus 1 is configured such that the skip code D12 is configured by data such as a header other than the data indicating the image, so that the encoded data when the image of the frame (n) is encoded as it is ( Since the code amount of the skip code D12 can be reduced compared to the code amount of n), underflow of the VBV buffer can be prevented, and the amount of code occupied by the VBV buffer is reduced because the amount of code to be transmitted is small. It can be recovered (increased), and the frequency of occurrence of underflow of the VBV buffer in the subsequent encoding can be reduced (FIG. 3).

  Furthermore, the moving image encoding apparatus 1 does not constantly monitor the encoded data D6, but calculates the occupation amount of the VBV buffer at the time when one frame is accumulated, so that the processing of the encoding control unit 5 is performed. Since the load can be reduced and processing such as highly accurate calculation of the quantization parameter can be performed instead, the image quality of the moving image can be improved.

  In addition, the generated skip code D12 does not set a unique code in the user data area of the MPEG2 system, but is configured by standard data in accordance with the MPEG2 system, so that a special detection circuit, etc. The skip code D12 can be recognized by a standard moving picture decoding apparatus that does not include the frame (n-1) at the timing when the image of the frame (n) should be displayed by the standard moving picture decoding apparatus. The image can be displayed again, and the image quality of the moving image is not deteriorated.

  According to the above configuration, the moving image encoding apparatus 1 calculates the occupation amount of the VBV buffer using the actual code amount of the encoded data D6, and the frame in which occurrence of underflow is detected in the VBV buffer is calculated. In place of the encoded data D6, by inserting a skip code D12 which is composed of data according to the MPEG2 system and has a small code amount, underflow of the VBV buffer can be prevented and at the same time the occupation amount can be increased. In addition, since the skip code D12 can be normally recognized by a more general moving image decoding apparatus, it is possible to prevent a deterioration in the image quality of the moving image when decoded.

(7) Other Embodiments In the above embodiment, the case where the skip code D12 having the data configuration shown in FIG. 4 is generated has been described. However, the present invention is not limited to this, and MPEG2 You may make it produce | generate the skip code which consists of the other various data structure which consists of the syntax according to a system as an instruction code.

  In the above-described embodiment, the case where the skip code D12 for redisplaying the image one frame before is output as the instruction code has been described. However, the present invention is not limited to this. Various other codes such as a code for redisplaying the screen may be output as the instruction code.

  Furthermore, in the above-described embodiment, the case where the predicted occupation amount Ve of the VBV buffer is calculated based on the encoded data D6 has been described. However, the present invention is not limited to this, and other various calculation methods. For example, the predicted occupation amount Ve may be calculated by referring to a predetermined table showing the relationship between the code amount and the quantization parameter.

  Furthermore, in the above-described embodiment, the case where the present invention is applied to the moving picture encoding apparatus 1 conforming to the MPEG2 system has been described. However, the present invention is not limited to this, and other examples such as MPEG4 are used. You may make it apply to the encoding apparatus based on a various encoding system.

  The present invention can also be applied to a personal computer or the like having a moving image encoding function.

It is a block diagram which shows the structure of a moving image encoder. It is a functional block diagram which shows the process of an encoding control part. It is a basic diagram with which it uses for description of the occupation amount of the VBV buffer by embodiment. It is a basic diagram which shows the data structure of the skip code | cord | chord by embodiment. It is a flowchart which shows a moving image encoding process procedure. It is a basic diagram which shows the mode of the encoding and decoding of an image. It is a basic diagram with which it uses for description of the occupation amount of the conventional VBV buffer.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Moving image encoder, 5 ... Encoding control part, 13 ... Skip code generation part, 14 ... Switch, 21 ... VBV occupation amount calculation part, 22 ... Rate control part, 23 ... Skip control unit.

Claims (4)

In a moving image encoding apparatus that performs encoding processing on moving image data according to a predetermined encoding method, and outputs encoded data sequentially generated for each predetermined unit to the outside.
Predicted occupancy calculation means for calculating the occupancy of the encoded data when accumulated in the decoding-side buffer as a predicted occupancy;
Instead of the encoded data generated when the predicted occupancy calculated by the predicted occupancy calculation means exceeds a predetermined threshold, the decoding result of the encoded data before the time is redisplayed. And a code output means for generating and outputting an instruction code according to the encoding method. The moving picture encoding apparatus according to claim 1, wherein the instruction code is configured by a part (or all) of a header defined in accordance with the encoding method. The moving image encoding apparatus according to claim 1, wherein the instruction code is a skip code defined in the encoding method. The moving picture encoding apparatus according to claim 1, wherein the predicted occupation amount calculation means is calculated based on an actual code amount of the encoded data in units of one frame or one field.

Images