JP2007135045A - Compression coding apparatus, compression coding method, and storage medium - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To ensure the total recording time when compressing and coding a dynamic image to record in a storage medium, even when recording time for each recording and total recording frequencies are not predictable. <P>SOLUTION: A means for measuring recording time 107 measures the elapsed time since the recording starts. A means for calculating the maximum amount of assigned codes 108 calculates the maximum amount of assigned codes according to the elapsed time. A means for calculating the amount of assigned codes 110 calculates the variable amount of assigned codes based on the difficulty of coding the input dynamic image. A means for limiting the amount of assigned codes 112 limits the amount of assigned codes to be equal to or less than the maximum amount of assigned codes. A means for determining quantization parameter 113 determines quantization parameter so that the amount of generated codes is equal to the amount of assigned codes. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a compression encoding apparatus, a compression encoding method, and a recording medium that perform compression encoding of a moving image signal, convert it into a variable bit rate encoded stream, and record or transmit the encoded stream.

  In recent years, the MPEG2 system, which is an international standard, has been widely used as a method for compression-coding moving image signals. In the MPEG2 system, each frame of an input moving image signal is divided into 16 × 16 pixel blocks called macroblocks, and interframe prediction by motion compensation, DCT, quantization, and variable length coding are used for each macroblock. Achieves high-efficiency compression.

  Inter-frame prediction is a technique for predicting an encoding target frame from previous and subsequent frames using the strong temporal correlation of moving images. In addition to this, motion compensation finds an image block having a strong correlation with the encoding target macroblock from the previous and subsequent frames, and encodes the difference from the target macroblock, thereby eliminating temporal redundancy of the image, and high The compression ratio is realized.

  DCT is a process of converting image information into frequency components, and each frequency component, that is, a DCT coefficient, is quantized to compress information greatly. Quantization means dividing a DCT coefficient by a value called a quantization parameter, and a frequency component having a small amplitude level becomes zero. In general, a low frequency component occupies most of an image signal, and a high frequency component is small. Therefore, the spatial information of the image is greatly reduced by omitting the high frequency component having a small amplitude level by quantization. Human vision has the characteristic that high-frequency components are insensitive, and even if high-frequency components are removed by quantization, image quality deterioration is not noticeable.

  After that, variable length coding is applied to the quantized DCT coefficients to achieve high compression and high image quality as a whole.

  Now, when a moving image is compressed and recorded on a recording medium such as a DVD, the generated code amount is adjusted by sequentially changing the quantization parameter. If the quantization parameter is increased, the amount of generated codes can be reduced, but the image quality deteriorates. Conversely, if the quantization parameter is reduced, encoding can be performed with high image quality, but the amount of generated code increases accordingly. When recording on a fixed-capacity medium such as a DVD, an increase in the amount of generated code leads to a decrease in recording time. Thus, there is generally a trade-off relationship between image quality and recording time.

  In general, code amount control refers to setting an assigned code amount for each fixed period and adaptively controlling a quantization parameter so that an error between the generated code amount and the assigned code amount becomes small.

  One of the code amount controls is fixed bit rate control. This is to control the quantization parameter so that a fixed code amount is always generated. In case of fixed bit rate control, regardless of the complexity of the input image, encoding is always performed by assigning only a fixed amount of code. If the input image is complex, the Q parameter is set to be large to reduce the amount of generated code. It is necessary to do so, which leads to image quality degradation. On the other hand, if the input image is simple, an amount of code more than necessary is allocated, which is not very good from the viewpoint of coding efficiency.

  On the other hand, there is variable bit rate control as another code amount control. If the variable bit rate control is used, the allocated code amount can be changed according to the degree of difficulty in encoding the image, so that the problem in the fixed bit rate control can be solved. In other words, in the case of an image with a high degree of encoding difficulty, it is possible to encode without increasing the quantization parameter so much by increasing the allocated code amount rather than the target average code amount. It is possible to improve the coding efficiency by making the allocated code amount smaller than the average code amount. If variable bit rate control is used in this way, an appropriate allocated code amount can be set according to the degree of difficulty in encoding an image, and high image quality and high encoding efficiency can be realized.

An example of an apparatus that realizes variable bit rate control is described in Japanese Patent Application Laid-Open No. 2000-174635. In this apparatus, the allocated code amount is controlled to be updated based on the degree of encoding difficulty while preventing the accumulated error between the target average code amount given in advance and the generated code amount from exceeding a preset allowable accumulated error. ing. When the accumulated error is sufficiently small with respect to the allowable accumulated error, a sufficient allocated code amount can be set for an image with a large degree of encoding difficulty, and high image quality can be realized. On the other hand, when the margin for the allowable accumulated error is small, the allocated code amount calculated from the encoding difficulty is corrected so as to approach the target average code amount so as not to exceed the allowable accumulated error. Even when the margin for the allowable accumulated error becomes zero, the target average code amount can be allocated at the minimum. Therefore, it is possible to guarantee the recording time since the set allowable error code amount is not exceeded while realizing high image quality by variable bit rate control.
JP 2000-174635 A

  However, the conventional code amount control method as disclosed in the above Japanese Patent Laid-Open No. 2000-174635 has the following problems.

  When the recording time is set in advance, as in the case of TV program recording, the total code amount can be accurately calculated from the recording time and the target average code amount, and the allowable cumulative error commensurate with the recording media capacity and recording time. Can be set in advance, but if you do not know when the shooting will stop, such as when shooting with a video camera, you cannot know the recording time in advance, so set an appropriate allowable cumulative error in advance. It is difficult.

  The fluctuation amount of the allocated code amount at the variable bit rate can be increased as the margin for the allowable accumulated error is increased, and the effect of the variable bit rate is increased. However, if an image with a high encoding difficulty level continues, an error with respect to the target average code amount is accumulated, and there is no code amount margin. If recording is stopped in such a state, the recording capacity is consumed more than the target average code amount. Furthermore, when such recording is repeated up to the total recording capacity of the recording medium, there is a problem that the actual total recording time becomes extremely shorter than the recording time assumed in advance.

  Conversely, in order to guarantee the total recording time, if only a small amount of accumulated error is allowed for one recording, it is necessary to reduce the fluctuation amount of the assigned code amount. If it shakes greatly, the margin will be consumed immediately, and only the target average code amount can be allocated, and it will approach the fixed bit rate control, so allocation necessary for images with high coding difficulty There was a problem that the code amount could not be given.

  As described above, when the operation of recording a moving image with high encoding difficulty in a relatively short time is repeated to the full capacity of the recording medium, the conventional method guarantees both the image quality and the total recording time. There was a problem that it was not possible.

  In order to solve the above-described problem, the compression encoding apparatus of the present invention compresses and encodes a moving image and outputs an encoded stream, and an assigned code based on the encoding difficulty of the encoded stream. An allocated code amount calculating means for calculating the amount, a maximum allocated code amount calculating means for increasing the maximum allocated code amount in accordance with an elapsed time from the start of compression encoding of the moving image, and the allocated code amount for the maximum allocated code An allocation code amount limiting unit that calculates an effective allocation code amount limited by an amount; and a quantization parameter determination unit that determines a quantization parameter at the time of compression encoding based on the effective allocation code amount.

  As described above, the code amount control method of the present invention increases the maximum allocated code amount according to the recording time, so that no matter how many times recording is stopped and how many times the recording is repeated, Since the accumulated error can be regulated to a predetermined value or less, the recording time can be guaranteed.

  The compression encoding apparatus of the present invention includes compression encoding means for compressing and encoding a moving image and outputting an encoded stream, and allocated code amount calculation means for calculating an allocated code amount based on the encoding difficulty of the encoded stream. A maximum allocated code amount calculating means for increasing the maximum allocated code amount according to an elapsed time from the start of compression encoding of the moving image, and an effective allocated code amount obtained by limiting the allocated code amount by the maximum allocated code amount. Allocation code amount limiting means for calculating, and quantization parameter determining means for determining a quantization parameter at the time of compression encoding based on the effective allocation code amount. As a result, the maximum allocated code amount is increased in accordance with the recording time, so that the accumulated error with respect to the target average code amount is specified to be equal to or less than a predetermined value no matter when recording is stopped or when recording is repeated. Recording time can be guaranteed.

  In the compression encoding apparatus of the present invention according to the above configuration, the maximum allocated code amount calculation means prevents the accumulated error calculated from the target average code amount and the code amount of the encoded stream from exceeding an allowable accumulated error. The maximum allocated code amount may be limited. Thereby, even when recording is performed for a long time, an appropriate maximum allocated code amount according to the accumulated error can be set, so that the recording time can be ensured while realizing high image quality.

  Embodiments of a compression encoding method according to the present invention will be described below in detail with reference to the drawings.

  FIG. 1 is a block diagram showing a configuration of a compression encoding apparatus according to an embodiment of the present invention. In FIG. 1, 101 is a moving image input terminal, 102 is a compression encoding means having a quantization process, 103 is an encoded stream output terminal, 104 is a generated code amount measuring means, 105 is a target average code amount input terminal, 106 Is an accumulated error calculating means, 107 is a recording time measuring means, 108 is a maximum allocated code amount calculating means, 109 is an encoding difficulty level calculating means, 110 is an allocated code amount calculating means, 111 is an allowable accumulated error input terminal, and 112 is an allocation A code amount limiting unit 113 is a quantization parameter determining unit.

  The operation of the compression coding apparatus configured as described above will be described below.

  First, the moving image signal input from the moving image input terminal 101 is encoded by the compression encoding means 102, and the encoded stream is output to the encoded stream output terminal 103.

  The compression encoding unit 102 performs a quantization process for compression, and a quantization parameter Q for quantization is supplied from the quantization parameter determination unit 113.

  The generated code amount measuring means 104 measures the generated code amount B of the encoded stream every predetermined period.

  The accumulated error calculation means 106 calculates an error d between the target average code quantity Ba input from the target average code quantity input terminal 105 and the generated code quantity B obtained from the generated code quantity measurement means 104, and cumulatively adds it. To calculate the cumulative error D. This result is output to the maximum allocated code amount calculation means 108 as accumulated error information.

  On the other hand, the recording time measuring means 107 measures an elapsed time T from the start of recording. The elapsed time T may be calculated based on, for example, the number of frames that have been encoded from the start of recording and the frame rate. This elapsed time information is output to the maximum allocated code amount calculation means 108.

  The maximum allocated code amount calculation unit 108 calculates the maximum allocated code amount Bt_max based on the elapsed time information given from the elapsed time measuring unit 107. The maximum allocated code amount Bt_max defines the upper limit of the allocated code amount Bt. In the present invention, the recording code guarantee is realized by limiting the allocated code amount Bt by the maximum allocated code amount Bt_max so that the generated code amount B does not become too large.

  Hereinafter, a method of calculating the maximum allocated code amount Bt_max in the maximum allocated code amount calculating means 108 will be described with reference to the flowchart of FIG.

The elapsed time T from the start of recording is input from the recording time measuring means 107 (step 201). When the elapsed time T from the start of recording reaches a predetermined time set in advance (step 202), the maximum allocated code amount calculation means 108 first sets the temporary maximum allocated code amount Bt_max ′ using the elapsed time T as a parameter. The calculation is performed using a predetermined function that increases with the elapsed time (step 203). For example,
Bt_max '= a × T + b (a and b are positive real constants set in advance)
Can be calculated as Also, Bt_max ′ can be limited by a certain upper limit value P so that it does not become larger.

  Further, the maximum allocated code amount calculating means 108 uses the accumulated error D input from the accumulated error calculating means 106 (step 204) so as not to exceed the allowable accumulated error Dmax given from the allowable accumulated error input terminal 111. Bt_max ′ is corrected to calculate the maximum allocated code amount Bt_max. That is, when the margin of the accumulated error D with respect to the allowable accumulated error Dmax is small (step 205), correction is performed so as to reduce Bt_max '(step 206). As the margin amount approaches zero, the maximum allocated code amount Bt_max is corrected so as to approach the target average code amount Ba, and when the margin amount becomes zero, Bt_max ′ is set as the target average code amount Ba. Since the allocated code amount Bt is limited by the maximum allocated code amount Bt_max, the accumulated error D does not exceed the allowable accumulated error Dmax. The maximum allocated code amount Bt_max calculated in this way is output to the allocated code amount limiting means 112 (step 207).

  The encoding difficulty level calculation means 109 encodes the code of the current input moving image based on the generated code amount B input from the generated code amount measurement means 104 and the quantization parameter Q input from the quantization parameter determination means 113. The difficulty level X is calculated. First, the difficulty of encoding the current input moving image is obtained from the amount of generated code in a certain section where encoding was completed immediately before and the quantization parameter used in the quantization process at that time. The difficulty of encoding is calculated, for example, as the product of the generated code amount and the quantization parameter. Also, the encoding difficulty level calculation means 109 stores past encoding difficulties, and calculates the average encoding difficulty of the input moving image based on the average of past encoding difficulties. . The current encoding difficulty level is calculated from the ratio between the average encoding difficulty and the current encoding difficulty. And this is output to the allocation code amount calculation means 110 as encoding difficulty level information.

  Based on the encoding difficulty level X input from the encoding difficulty level calculation unit 109, the allocation code amount calculation unit 110 calculates an allocation code amount Bt that is a target code amount for each predetermined period when encoding a moving image. calculate. The allocated code amount calculation means 110 is set larger or smaller than the target average code amount Ba according to the encoding difficulty level information. As described above, the encoding difficulty level X is the ratio of the past encoding difficulty to the current encoding difficulty, so the allocated code amount may be determined by multiplying the target code amount by this ratio. . By calculating the allocated code amount Bt in this way, when a moving image that is more difficult to encode than a previously encoded moving image is input, an allocated code that is larger than the target average code amount Ba so that the image quality does not deteriorate. The amount is set. Conversely, in the case of a moving image that is easy, a code amount smaller than the target average code amount Ba is set, and the code amount can be saved.

  It should be noted that other information may be used for the calculation difficulty and encoding difficulty calculation method as described above, and the calculation element for changing the allocated code amount, or may be calculated by another calculation method. It doesn't matter.

  The allocation code amount limiting unit 112 limits the allocation code amount Bt input from the allocation code amount calculation unit 110 by the maximum allocation code amount Bt_max input from the maximum allocation code amount calculation unit 108. That is, when the allocated code amount Bt is larger than the maximum allocated code amount Bt_max, the maximum allocated code amount Bt_max is set as the allocated code amount, and when the allocated code amount Bt is less than the maximum allocated code amount Bt_max, the allocated code amount Bt is allocated. The code amount is output to the quantization parameter determining unit 113.

  The quantization parameter determining unit 113 changes the quantization parameter Q so that the encoded code amount B of the moving image to be encoded is equal to the allocated code amount Bt.

  Next, the relationship among the elements such as the generated code amount B, the assigned code amount Bt, the maximum assigned code amount Bt_max, the accumulated error D and the like as described above will be described in detail with reference to the drawings.

  FIG. 3 shows changes in allocated code amount and accumulated error for each elapsed time when a moving image having a high encoding difficulty level is input from the recording start time. In FIG. 3A, the assigned code amount when the conventional method is used is indicated by a broken line, and the assigned code amount in the present invention is indicated by a solid line. Since the encoding difficulty level of the input moving image is very high, an allocated code amount that greatly exceeds the target average code amount is required. In the present invention, the maximum allocated code amount Bt_max is set such that the value is suppressed at the beginning of recording and increased every time, so that the allocated code amount actually set in the quantization parameter determining unit 113 is indicated by a solid line. It will be.

  The state of the accumulated error D at this time is shown in FIG. The accumulated error when coding with the conventional allocated code amount is indicated by a broken line, and the accumulated error when encoded with the allocated code amount limited by the maximum allocated code amount Bt_max of the present invention is indicated by a solid line. According to this, in the conventional method in which there is no limitation on the allocated code amount, a large amount of code amount is obviously consumed. When a large amount of code is consumed and recording is stopped and repeated, the actual recording time is extremely large compared to the recording time calculated from the capacity of the recording medium and the target average code amount. It will be short. On the other hand, when the allocated code amount is limited by the maximum allocated code amount of the present invention, since the allocated code amount is kept low by the maximum allocated code amount at the beginning of recording, the cumulative error does not increase so much. Absent. Therefore, even if the recording is stopped in a short time, the accumulated error amount at that time can be suppressed to a predetermined value or less, so that the recording time can be guaranteed even if this operation is repeated.

  In addition, by setting the maximum allocated code amount at the start of recording to be larger than the target average code amount, the allocated code amount can be made larger than the target average code amount, and the code quality is higher than that of the fixed bit rate control. Can be realized.

  Next, FIG. 4 shows a response of variable bit rate control in a normal state when a sufficient time has elapsed from the start of recording in the present invention and the accumulated error D is sufficiently small with respect to the allowable accumulated error Dmax. Yes.

  First, in FIG. 4A, the target average code amount Ba is always set to a constant value regardless of the passage of time. What is indicated by a broken line is the maximum allocated code amount Bt_max. Since a sufficient amount of time has elapsed since the start of recording, the value is sufficiently large and is already fixed at a constant value. In such a case, the allocated code amount Bt can be largely allocated according to the encoding difficulty level X of the image. The allocation code amount Bt at this time is indicated by a solid line. The period in which the assigned code amount Bt is higher than the target average code amount Ba is a period in which a moving image having a high degree of difficulty in encoding is being input, and conversely, the assigned code amount Bt is lower than the target average code amount Ba. The period is a period in which a moving image with a low encoding difficulty level is input. When the allocated code amount Bt as shown in FIG. 4 (a) is set in the quantization parameter determining unit 113, the quantization parameter determining unit 113 sets the Q parameter so that the actual generated code amount B matches the allocated code amount Bt. Therefore, the accumulated error D behaves as shown in FIG. The accumulated error D increases during the period when the allocated code amount Bt is higher than the target average code amount Ba, and conversely, the accumulated error D decreases during the period when the allocated code amount Bt is lower than the target average code amount Ba.

  Next, FIG. 5 shows the response of each element when a moving image having a very high encoding difficulty level is input in the state described with reference to FIG. In FIG. 5A, the allocation code amount when there is no restriction by the maximum allocation code amount Bt_max is indicated by a broken line. This is an allocated code amount that is originally required because the input video is very difficult to encode. However, in practice, correction is performed so that the maximum allocated code amount Bt_max gradually decreases as the accumulated error D approaches the allowable accumulated error Dmax. Since the allocated code amount Bt is clipped at Bt_max by the allocated code amount limiting unit 112, the allocated code amount actually set in the quantization parameter determining unit 113 is Bt_max. For this reason, the cumulative error D exhibits a behavior as shown in FIG. When the cumulative error D reaches the allowable cumulative error Dmax, the maximum allocated code amount Bt_max is corrected so as to be equal to the target average code amount Ba, so that the cumulative error D does not exceed the allowable cumulative error Dmax.

  Thereafter, when a moving image having a low encoding difficulty level is input and the accumulated error D is reduced to allow for a margin, the allocated code amount can be increased by increasing the maximum allocated code amount.

  As described above, in the present invention, when recording is performed for a long time, the maximum allocated code amount is sufficiently large, and the allocated code amount can be sufficiently allocated according to the degree of encoding difficulty, thereby realizing high image quality. .

  As described above, the present invention increases the maximum allocated code amount in accordance with the elapsed time from the recording start time, so that no matter how many times recording is stopped or how many times recording is repeated, Since the accumulated error amount can be suppressed to a predetermined value or less with respect to the recording capacity, the recording time can be guaranteed, which is useful as a compression encoding apparatus or a compression encoding method for a moving image.

1 is a block diagram of a compression coding apparatus according to an embodiment of the present invention. Flowchart for calculating maximum allocated code amount in an embodiment of the present invention The conceptual diagram which shows the allocation code amount at the time of the complicated moving image input from the recording start in one embodiment of this invention, and a cumulative error The conceptual diagram which shows the allocation code amount and accumulation error after progress for a fixed time in one embodiment of this invention The conceptual diagram which shows the allocation code amount and the accumulation error at the time of the complicated moving image being input after progress for a fixed time in one embodiment of this invention

Explanation of symbols

101 moving image input terminal 102 compression encoding means 103 encoded stream output terminal 104 generated code amount measuring means 105 target average code amount input terminal 106 cumulative error calculating means 107 recording time measuring means 108 maximum allocated code amount calculating means 109 encoding difficulty Degree calculation means 110 Allocation code amount calculation means 111 Permissible cumulative error input terminal 112 Allocation code amount restriction means 113 Quantization parameter determination means

Claims (5)

Compression encoding means for compressing and encoding a moving image and outputting an encoded stream;
An allocated code amount calculating means for calculating an allocated code amount based on the encoding difficulty level of the encoded stream;
Maximum allocated code amount calculating means for increasing the maximum allocated code amount according to the elapsed time from the start of compression encoding of the moving image;
An allocated code amount limiting means for calculating an effective allocated code amount obtained by limiting the allocated code amount with the maximum allocated code amount;
A compression coding apparatus comprising: quantization parameter determining means for determining a quantization parameter at the time of compression encoding based on the effective allocation code amount. The maximum allocated code amount calculation unit limits the maximum allocated code amount so that an accumulated error calculated from a target average code amount and a code amount of the encoded stream does not exceed an allowable accumulated error. Compression encoding device. A compression encoding step of compressing and encoding a moving image and outputting an encoded stream;
An allocated code amount calculation step of calculating an allocated code amount based on the encoding difficulty level of the encoded stream;
A maximum allocated code amount calculating step for increasing the maximum allocated code amount according to an elapsed time from the start of compression encoding of the moving image;
An allocated code amount limiting step of calculating an effective allocated code amount obtained by limiting the allocated code amount with the maximum allocated code amount;
A compression encoding method comprising: a quantization parameter determining step for determining a quantization parameter at the time of compression encoding based on the effective allocation code amount. The maximum allocation code amount calculation step limits the maximum allocation code amount so that a cumulative error calculated from a target average code amount and a code amount of the encoded stream does not exceed an allowable cumulative error. Compression encoding method. A compression encoding step of compressing and encoding a moving image and outputting an encoded stream;
An allocated code amount calculation step of calculating an allocated code amount based on the encoding difficulty level of the encoded stream;
A maximum allocated code amount calculating step for increasing the maximum allocated code amount according to an elapsed time from the start of compression encoding of the moving image;
An allocated code amount limiting step of calculating an effective allocated code amount obtained by limiting the allocated code amount with the maximum allocated code amount;
A computer-readable recording medium storing a program for causing a computer to execute a quantization parameter determination step for determining a quantization parameter at the time of compression encoding based on the effective allocation code amount.