Image data compression device and encoder

Abstract

An image data compression device includes: a quantization section which quantizes image data in unit of a frame; a FIFO buffer section in which quantized data for at least N frames (N is an integer larger than one) quantized by the quantization section is buffered; an encoded data generation section which reads the quantized data from the FIFO buffer section asynchronously from writing into the FIFO buffer section, and generates encoded data; and a rate control section which changes a data size of the encoded data by changing a quantization step of the quantization section by using the data size of the encoded data. The rate control section obtains an average data size by averaging the data sizes of the encoded data for N frames before a frame of the image data quantized by the quantization section, and changes the quantization step based on the average data size.

Description

Japanese Patent Application No. 2003-391087, filed on Nov. 20, 2003, is hereby incorporated by reference in its entirety.

BACKGROUND OF THE INVENTION

The present invention relates to an image data compression device and an encoder.

Moving Picture Experts Group Phase 4 (MPEG-4) has been standardized as a general-purpose coding method for multimedia information such as image data of a still image or a moving image and sound data. A recent portable instrument realizes encoding and decoding of image data conforming to the MPEG-4 standard, and can perform moving image reproduction and transmission/reception through a network.

In the MPEG-4 standard, compressed data obtained by encoding image data of a moving image must be generated at a constant rate. However, when compressing image data of a moving image, compression efficiency changes to a large extent depending on the type of image data. MPEG-4 Visual Part (Recommendation ISO/IEC 14496-2: 1999 (E) Annex L) describes a rate control method for generating compressed data at a constant rate by controlling the amount of code to be generated so that such a change is within a predetermined range.

When performing MPEG-4 encode (compression) processing, a series of processing may be entirely performed by hardware. However, since this increases the circuit scale, it is difficult to achieve a reduction of size when integrating the circuits in an IC (semiconductor device or integrated circuit). In particular, a portable instrument such as a portable telephone cannot satisfy a demand for a reduction of the size of the instrument.

A series of encode processing may be entirely performed by using software.

However, this increases the load imposed on a central processing unit (CPU) which processes the software. Therefore, the time necessary for the CPU to perform another processing is limited, whereby the performance of an instrument provided with the CPU is decreased. Moreover, the processing time of the CPU is increased, whereby power consumption is increased. In particular, a portable instrument such as a portable telephone cannot satisfy a demand for a reduction of power consumption in order to maintain battery life.

Therefore, a series of encode processing may be allocated between hardware and software. However, as a result of studies conducted by the present inventors, it was 10 found that the rate control method disclosed in MPEG-4 Visual Part (Recommendation ISO/IEC 14496-2: 1999 (E) Annex L) cannot be performed when optimizing the allocation between hardware and software for a series of encode processing. Therefore, optimization of allocation of image data compression processing between hardware and software and generation of compressed data at a constant rate cannot achieved in combination.

BRIEF SUMMARY OF THE INVENTION

One aspect of the present invention relates to an image data compression device for compressing image data, the image data compression device including:

• • a quantization section which quantizes image data in unit of a frame;
  • a FIFO buffer section in which quantized data for at least N frames (N is an integer larger than one) quantized by the quantization section is buffered;
  • an encoded data generation section which reads the quantized data from the FIFO buffer section asynchronously from writing into the FIFO buffer section, and generates encoded data by encoding the quantized data; and
  • a rate control section which changes a data size of encoded data to be processed by changing a quantization step of the quantization section by using the data size of the encoded data,
  • wherein the rate control section obtains an average data size by averaging the data sizes of the encoded data for N frames before a frame of the image data quantized by the quantization section, and changes the quantization step based on the average data size.

Another aspect of the present invention relates to an encoder which performs compression processing for image data, the encoder including:

• • an image input interface which performs interface processing for inputting image data;
  • a quantization section which quantizes the image data in unit of a frame;
  • a FIFO buffer section in which quantized data for at least N frames (N is an integer larger than one) quantized by the quantization section is buffered; and
  • a host interface which performs interface processing with a host which reads the quantized data stored in the FIFO buffer section asynchronously from writing into the FIFO buffer section,
  • wherein the host obtains an average data size by averaging data size of encoded data for N frames before a frame of the image data quantized by the quantization section, and
  • wherein the quantization section quantizes the image data by a quantization step changed based on the average data size.

BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWING

FIGS. 1A and 1B are illustrative of MPEG-4 encode processing and decode processing.

FIG. 2 is illustrative of a macroblock.

FIG. 3 is illustrative of an example of DCT coefficients.

FIG. 4 is illustrative of an example of a quantization table.

FIG. 5 is illustrative of an example of quantized DCT coefficients.

FIG. 6 is illustrative of a model equation used in a rate control method.

FIG. 7 is a flow chart of an example of rate control processing using the model equation shown in FIG. 6.

FIG. 8 is a block diagram schematically showing a configuration of an image data compression device in an embodiment of the present invention.

FIG. 9 is illustrative of a rate control method in an embodiment of the present invention.

FIG. 10 is a schematic diagram of operation timing of the rate control method shown in FIG.9.

FIG. 11 is a flow chart of the first half of an example of calculation processing of a quantization parameter.

FIG. 12 is a flow chart of the latter half of an example of calculation processing of a quantization parameter.

FIG. 13 is illustrative of variables used in calculation processing of a quantization parameter.

FIG. 14 is a flow chart of an example of calculation processing of a value of the number of bits used for encoding.

FIG. 15 is illustrative of quantization processing in an embodiment of the present invention.

FIG. 16 is a detailed functional block diagram of an image data compression device in an embodiment of the present invention.

FIG. 17 shows a hardware configuration example of the image data compression device shown in FIG. 16.

FIG. 18 is a flow chart of an example of interrupt reception processing performed by a host.

FIG. 19 is a flow chart of an example of ME interrupt processing.

FIG. 20 is a flow chart of an example of encode completion interrupt processing.

FIG. 21 is a block diagram of a configuration example of a display controller in an embodiment of the present invention.

FIG. 22 is a block diagram of a configuration example of an electronic instrument to which the display controller shown in FIG. 21 is applied.

DETAILED DESCRIPTION OF THE EMBODIMENT

This embodiment has been achieved in view of the above-described technical subjects, and may provide an image data compression device and an encoder which achieve optimization of allocation of image data compression processing between hardware and software and generation of compressed data at a constant rate in combination.

One embodiment of the present invention provides an image data compression device for compressing image data, the image data compression device including:

• • a quantization section which quantizes image data in unit of a frame;
  • a FIFO buffer section in which quantized data for at least N frames (N is an integer larger than one) quantized by the quantization section is buffered;
  • an encoded data generation section which reads the quantized data from the FIFO buffer section asynchronously from writing into the FIFO buffer section, and generates encoded data by encoding the quantized data; and
  • a rate control section which changes a data size of encoded data to be processed by changing a quantization step of the quantization section by using the data size of the encoded data,
  • wherein the rate control section obtains an average data size by averaging the data sizes of the encoded data for N frames before a frame of the image data quantized by the quantization section, and changes the quantization step based on the average data size.

In this embodiment, the FIFO buffer section is provided between the quantization section and the encoded data generation section. This enables the processing of the quantization section and the processing of the encoded data generation section to be performed asynchronously and in parallel. When controlling the generation rate of the encoded data by the encoded data generation section, the rate control section changes the quantization step based on the average data size obtained by averaging the data sizes of the encoded data for a plurality of frames before the frame of the image data quantized by the quantization section.

Therefore, since the image data compression device is configured so that the processing of the quantization section and the processing of the encoded data generation section are performed asynchronously, the generation rate of encoded data can be controlled even if the rate control method disclosed in MPEG-4 Visual Part (Recommendation ISO/IEC 14496-2: 1999 (E) Annex L) cannot be performed, whereby encoded data can be generated by compressing image data at a constant rate.

With this image data compression device, when the quantized data read from the FIFO buffer section is data of an intra-frame-coded macroblock which is encoded within one frame, the rate control section may obtain the data size of the encoded data obtained by encoding the quantized data as the average data size, and may change the quantization step based on the average data size.

Since the image data compression device is configured so that the processing of the quantization section and the processing of the encoded data generation section are performed asynchronously, the generation rate of encoded data can be suitably controlled when there is no continuity with the image data in the previous frame such as an I picture, even if the rate control method disclosed in MPEG-4 Visual Part (Recommendation ISO/IEC 14496-2: 1999 (E) Annex L) cannot be performed, whereby encoded data can be generated by compressing image data at a constant rate.

With this image data compression device, the rate control section may obtain the average data size after the quantization section has quantized the image data for M frames (M≧N, M is an integer) by a predetermined quantization step.

This prevents deterioration of the image quality due to unnecessary rate control performed when the average value of the data size of the encoded data in the previous N frames is not accurate.

This image data compression device may include a quantization table which stores a quantization step value, and

• • the rate control section may change the quantization step by obtaining a quantization parameter based on the average data size and performing quantization using a product of the quantization parameter and the quantization step value.

This image data compression device may include a discrete cosine transform section which supplies the image data subjected to a discrete cosine transform to the quantization section in unit of a frame.

This image data compression device may include:

• • a hardware processing section which processes image data of a moving image by hardware; and
  • a software processing section which generates encoded data by performing encode processing of the quantized data read from the FIFO buffer section by software,
  • the hardware processing section may include the quantization section and the FIFO buffer section, and
  • the software processing section may include the encoded data generation section and the rate control section.

Since most of the quantized moving image data is zero data, the amount of information of data is significantly small in comparison with the data before the quantization in many cases. Moreover, the operation load for encoding is generally small. Therefore, even if the processing in which the amount of information and the operation load are small is performed by the software processing section, the processing load is small. On the contrary, the amount of information is great and the operation is complicated in most quantization processing. Therefore, the processing load is heavy for software processing. Although the above processing is heavy load processing, the necessity of changing the processing is limited when the processing is standardized. Moreover, since most of the processing is repetition, this is suitable for the hardware processing section. Furthermore, since the amount of data processed by the hardware processing section is small, the amount of data transmitted from the hardware processing section to the software processing section is small, whereby the transmission load is reduced. Since the FIFO buffer section is provided between the software processing section and the hardware processing section, the software processing and the hardware processing can be processed in parallel. In addition, a reduction of the size of the device and a reduction of power consumption can be realized by using the software and the hardware for their corresponding processings.

With this image data compression device, the hardware processing section may output a difference between input image data in a current frame and previous image data in a frame one frame before the current frame as motion vector information, may perform a discrete cosine transform for the motion vector information and output the transformed motion vector information to the quantization section as the image data, and may generate the previous image data based on inverse-quantized data obtained by inverse-quantizing the quantized data by the quantization step.

With the image data compression device, the software processing section may encode the quantized data read from the FIFO buffer section into a variable length code.

With this image data compression device, the software processing section may perform scan processing of rearranging the quantized data read from the FIFO buffer section, and may encode a result of the scan processing into a variable length code.

With this image data compression device, the software processing section may obtain a DC component and an AC component from the quantized data read from the FIFO buffer section, may perform scan processing of rearranging the DC component and the AC component, and may encode a result of the scan processing into a variable length code.

Another embodiment of the present invention provides an encoder which performs compression processing for image data, the encoder including:

• • an image input interface which performs interface processing for inputting image data;
  • a quantization section which quantizes the image data in unit of a frame;
  • a FIFO buffer section in which quantized data for at least N frames (N is an integer larger than one) quantized by the quantization section is buffered; and
  • a host interface which performs interface processing with a host which reads the quantized data stored in the FIFO buffer section asynchronously from writing into the FIFO buffer section,
  • wherein the host obtains an average data size by averaging data size of encoded data for N frames before a frame of the image data quantized by the quantization section, and
  • wherein the quantization section quantizes the image data by a quantization step changed based on the average data size.

According to this embodiment, encode processing for compressing image data of a moving image from an imaging section can be allocated between the encoder and the host, for example. Therefore, generation of encoded data and quantization can be performed in parallel in the encode processing. Moreover, a reduction of size and power consumption of a device which includes the encoder can be realized by using the software and the hardware for their corresponding processings.

With this encoder, when the quantized data read from the FIFO buffer section is data of an intra-frame-coded macroblock which is encoded within one frame, the host may obtain the data size of the encoded data obtained by encoding the quantized data as the average data size, and the quantization section may quantize the image data by the quantization step changed based on the average data size.

Embodiments of the present invention are described below in detail with reference to the drawings. Note that the embodiments described hereunder do not in any way limit the scope of the invention defined by the claims laid out herein. Note also that not all of the elements of these embodiments should be taken as essential requirements to the means of the present invention.

1. MPEG-4

The MPEG-4 encode processing is briefly described below. The decode processing of decoding compressed data encoded by the encode processing is also described below.

FIGS. 1A and 1B are illustrative of the MPEG-4 encode processing and decode processing, respectively. The details of the processing are described in “JPEG & MPEG: Illustrated Image Compression Technology”, Hiroshi Ochi and Hideo Kuroda, Nippon Jitsugyo Publishing Co., Ltd., for example. In the following description, only the processing relating to the present invention is mainly described.

In the encode processing shown in FIG. 1A, a motion estimation (ME) between two successive images (two frames) is performed (step S1). In more detail, the difference between two images is calculated (obtained) for a single pixel. Since the difference becomes zero in the image region in which no change occurs between two images, the amount of information can be reduced. The zero data in this image region and the difference (positive and negative components) in the image region in which a change occurs between two images make up information after the motion estimation.

A discrete cosine transform (DCT) is performed (step S2). The DCT is performed in units of 8×8 pixel blocks shown in FIG. 2 to calculate DCT coefficients in block units. The DCT coefficients after the discrete cosine transform represent changes in light and shade of the image in one block by average brightness (DC component) and spatial frequency (AC component). FIG. 3 shows an example of the DCT coefficients in one 8×8 pixel block (quotation from FIG. 5-6 on page 116 of the above-mentioned literature). The DCT coefficient at the upper left corner represents a DC component, and the remaining DCT coefficients represent AC components. The influence on image recognition is small even if high-frequency components of the AC components are omitted.

The DCT coefficients are quantized (step S3). The quantization is performed in order to reduce the amount of information by dividing each DCT coefficient in one block by a quantization step value at the corresponding position in a quantization table. For example, FIG. 5 shows the DCT coefficients in one block obtained by quantizing the DCT coefficients shown in FIG. 3 using a quantization table shown in FIG. 4 (quotation from FIGS. 5-9 and 5-10 on page 117 of the above-mentioned literature). As shown in FIG. 5, most of the DCT coefficients of the high-frequency components become zero data by dividing the DCT coefficients by the quantization step values and rounding off to the nearest whole number, whereby the amount of information is significantly reduced.

A feed-back route is necessary for the encode processing in order to perform the above-described motion estimation between the current frame and the frame subsequent to the current frame. As shown in FIG. 1A, an inverse quantization (iQ), an inverse DCT, and a motion compensation (MC) are performed in the feed-back route (steps S4 to S6). Details of the operation of the motion compensation are omitted. This processing is performed in units of 16×16 pixel macroblocks shown in FIG. 2.

In this embodiment, the series of processing in the steps S1 to S6 is performed by hardware.

DC/AC (direct current/alternating current components) prediction processing performed in a step S7 shown in FIG. 1A and scan processing performed in a step S8 are processing necessary to increase efficiency of encoding into a variable length code (VLC) in a step S9. In the variable length coding in the step S9, the difference in the DC component between adjacent blocks must be encoded, and the order of encoding must be determined for the AC components by scanning the AC components in the block from the low-frequency side to the high-frequency side (also called a “zigzag scan”).

The variable length coding in the step S9 is also called entropy encoding, and has a coding principle in which a component with a higher emergence frequency is represented by using a smaller code. The difference between adjacent blocks is encoded for the DC component, and the DCT coefficients are sequentially encoded for the AC components in the scan order from the low-frequency side to the high-frequency side by utilizing the results obtained in the step S7 and the step S8.

The amount of information of image data to be generated changes depending on complexity of the image and intensity of motion. In order to absorb such a change and transfer information at a constant transfer rate, it is necessary to control the amount of code to be generated. This is achieved by rate control in a step S10. A buffer memory is generally provided for the rate control, and the amount of information to be stored is monitored so that the buffer memory does not overflow to reduce the amount of information to be generated. In more detail, the number of bits which represent the DCT coefficient is reduced by roughening the quantization characteristics in the step S3.

In this embodiment, the series of processing in the steps S7 to S10 is performed by software. Specifically, the series of processing in the steps S7 to S10 is realized by hardware which reads the software.

FIG. 1B shows the decode processing of the image data compressed by the encode processing shown in FIG. 1A. The decode processing is achieved by inversely performing the encode processing shown in FIG. 1A in the reverse order. A “postfilter” shown in FIG. 1B is a filter for eliminating block noise. A “YUV/RGB conversion” shown in FIG. 1B means converting the output from the postfilter into an RGB format from a YUV format.

2. Rate Control

The method described in MPEG-4 Visual Part (Recommendation ISO/IEC 14496-2: 1999 (E) Annex L) is briefly described below relating to the rate control performed in the step S10 shown in FIG. 1A. This method changes a quantization parameter at the time of encoding. The quantization characteristics are changed by changing the quantization step of the quantization (step S3) shown in FIG. 1A by changing the quantization parameter, thereby controlling the amount of code (data size) to be generated.

In this method, the amount of code R generated when encoding one frame is controlled by setting a quantization parameter Qc in frame units. In this case, the quantization parameter Qc is calculated according to a model equation shown in FIG. 6.

In FIG. 6, R denotes the amount of code generated when encoding one frame, Qc denotes the quantization parameter, Ec denotes the frame complexity, and X₁ and X₂ denote parameters of this model. The average absolute value of encoding target pixels is used as the frame complexity Ec. The frame complexity Ec of an inter-frame-coded 20 macroblock is calculated after the motion estimation as a value obtained by dividing the sum of the absolute values of the differences x′_ij between the current frame and the previous frame by an area A. The frame complexity Ec of an intra-frame-coded macroblock is calculated as a value obtained by dividing the sum of the absolute values (|x_ij-μ|) of the differences between the current frame and a reference value μ by the area A. The reference value μ may be the average value of all the pixels in the macroblock.

In FIG. 6, the amount of code to be generated is modeled by the quadratic equation of the frame complexity and the reciprocal of the quantization parameter as described above.

FIG. 7 shows an example of a processing flow of the rate control using the model equation shown in FIG. 6.

The initial frame is encoded using a predetermined quantization parameter (step S30). The initial values of the model parameters X₁ and X₂ are set (step S31). The complexity Ec of the current frame is calculated (step S32). The complexity Ec can be calculated by using the equation shown in FIG. 6. The amount of code used for encoding is calculated based on the remaining available amount of code and the amount of code used in the previous frame (step S33).

The model parameters X₁ and X₂ set in the step S31 and the complexity Ec calculated in the step S32 are set in the model equation shown in FIG. 6. A value obtained by subtracting the number of bits other than information such as the header and the motion vector among the number of bits used in the previous frame from the amount of code used for encoding calculated in the step S33 is set in the model equation shown in FIG. 6 as the amount of code R generated when encoding one frame. The quantization parameter Qc is calculated by solving the quadratic equation shown in FIG. 6 in which the parameter is Qc (step S34).

The frame is quantized and encoded using the quantization parameter Qc calculated in the step S34 (step S35), and the model parameters X₁ and X₂ are calculated from the model equation shown in FIG. 6 based on the quantization parameter and the amount of code generated in the frame encoded one frame before the current frame to update the model parameters X₁ and X₂ (step S36).

When the processing flow is completed under a predetermined condition (step S37: Y), the series of processing is terminated (END). When the processing is not completed (step S37: N), the step S32 is performed. The above-described processing is performed in each frame.

As described above, in the rate control method described in MPEG-4 Visual Part (Recommendation ISO/IEC 14496-2: 1999 (E) Annex L), it is necessary to cause the encoding result in the previous frame to be reflected in the encoding of the subsequent frame.

3. Image Data Compression Device

This embodiment provides an image data compression device which allocates the series of encode processing between hardware and software, and optimizes the allocation.

FIG. 8 shows a block diagram schematically showing a configuration of an image data compression device in this embodiment.

An image data compression device 10 in this embodiment includes a quantization section 20. The quantization section 20 performs the processing in the step S3 shown in FIG. 1A. The quantization section 20 quantizes image data in frame units. Specifically, the quantization section 20 quantizes image data in frame units and generates quantized data. The image data may be represented by the DCT coefficients after the DCT processing in the step S2 shown in FIG. 1A, for example. In this case, the DCT coefficients shown in FIG. 3 are quantized as shown in FIG. 5 by being divided by the quantization step values in the quantization table shown in FIG. 4.

The image data compression device 10 includes a FIFO buffer section 30. The quantized data for at least N frames (N is an integer larger than one) quantized by the quantization section 20 is buffered in the FIFO buffer section 30. The quantized data output from the quantization section 20 in frame units is sequentially written into the FIFO buffer section 30. The FIFO buffer section 30 functions as a first-in first-out storage circuit.

The image data compression device 10 includes an encoded data generation section 40. The encoded data generation section 40 reads the quantized data for one frame from the FIFO buffer section 30, and generates encoded data by encoding the quantized data. The encoded data generation section 40 reads the quantized data for one frame from the FIFO buffer section 30 asynchronously from writing into the FIFO buffer section 30.

The heavy load processing of the quantization section 20 is performed by hardware and the small load encode processing of the encoded data generation section 40 is realized by software processing by providing the FIFO buffer section 30 between the quantization section 20 and the encoded data generation section 40. Moreover, the processing of the quantization section 20 and the encode processing of the encoded data generation section 40 can be performed in parallel.

The following description is given on the assumption that the quantization section 20 is realized by high-speed hardware and the encoded data generation section 40 is realized by low-speed software processing, for example. However, this embodiment is not limited thereto. This embodiment is applied to the case where the encoded data generation section 40 reads the quantized data from the FIFO buffer section 30 asynchronously from writing into the FIFO buffer section 30. Therefore, the quantization section 20 may be realized by high-speed hardware, and the encoded data generation section 40 may be realized by low-speed hardware, for example. The quantization section 20 and the encoded data generation section 40 may be realized by hardware which reads software, and perform processing asynchronously.

The image data compression device 10 includes a rate control section 50. The rate control section 50 changes the data size of the encoded data by changing the quantization step of the quantization section 20 using the data size of the encoded data generated by the encoded data generation section 40. As is clear from FIG. 5, the amount of zero data of the quantized DCT coefficients is increased by increasing the quantization step. The amount of zero data of the quantized DCT coefficients is decreased by decreasing the quantization step. The quantized data of which the amount of zero data is changed in this manner is written into the FIFO buffer section 30. As a result, the size of the encoded data obtained by allowing the encoded data generation section 40 to encode the quantized data read from the FIFO buffer section 30 can be changed corresponding to the quantization parameter.

As described above, in the rate control method described in MPEG-4 Visual Part (Recommendation ISO/IEC 14496-2: 1999 (E) Annex L), it is necessary to cause the encoding result in the previous frame to be reflected in the encoding of the subsequent frame. However, if the quantization of the quantization section 20 and the encoding of the encoded data generation section 40 are allocated between hardware and software, the quantization and the encoding are processed asynchronously. Therefore, the quantized data read from the FIFO buffer section 30 may be the data in a frame two or more frames before the frame of the data quantized by the quantization section 20. Therefore, the rate control method described in MPEG-4 Visual Part (Recommendation ISO/IEC 14496-2: 1999 (E) Annex L) in which the encoding result in the previous frame is caused to be reflected in the encoding of the subsequent frame cannot be realized.

In this embodiment, the rate control section 50 calculates the average data size by averaging the data sizes of the encoded data for N frames before the frame of the image data quantized by the quantization section 20, and changes the quantization step based on the average data size. For example, when the image data quantized by the quantization section 20 is in the Lth frame (L is a positive integer), the rate control section 50 changes the quantization step based on the average data size calculated by averaging the data sizes of the encoded data in the (L-P)th frame (L>P, P is a positive integer) to the (L-P-N+1)th frame (L-P>N−1) for previous N frames of the Lth frame.

FIG. 9 is illustrative of the rate control method in this embodiment. In FIG. 9, sections the same as the sections of the image data compression device 10 shown in FIG. 8 are denoted by the same symbols. Description of these sections is appropriately omitted. In FIG. 9, the FIFO buffer section 30 can store quantized data for seven frames.

FIG. 10 shows a schematic diagram of operation timing of the rate control method shown in FIG. 9.

The quantization section 20 quantizes image data in frame units. For example, a quantization table 22 in which the quantization step values shown in FIG. 4 are set is provided. The quantization section 20 quantizes image data in frame units based on the quantization step values set in the quantization table 22 and the quantization parameter from the rate control section 50. In more detail, the quantization parameter is used as the coefficient of the quantization step values, and the quantization section 20 changes the quantization step by quantizing image data using the product of the quantization parameter and the quantization step values.

The quantization section 20 quantizes image data in frame units at times t1, t2, . . . , and writes the quantized data into the FIFO buffer section 30 in the order of the first frame F₁, the second frame F₂, . . . . The encoded data generation section 40 reads the quantized data from the FIFO buffer section 30 in frame units asynchronously from the write timing of the quantized data into the FIFO buffer section 30, and performs the encode processing.

The rate control section 50 changes the quantization step of the quantization section 20 based on the average data size calculated by averaging the data sizes of the encoded data for four (N=4) frames before the frame (current frame) of the image data quantized by the quantization section 20, for example. This causes the size of the quantized data quantized by the quantization section 20 to be changed, whereby the size of the encoded data generated by the encoded data generation section 40 is also changed.

In FIG. 10, the encoded data generation section 40 sequentially reads the quantized data in the first to fourth frames F₁ to F₄ from the FIFO buffer section 30, and generates and outputs the encoded data in each frame. In this case, the quantization section 20 quantizes image data in the seventh frame F₇.

The rate control section 50 stores the sizes of the encoded data in the first to fourth frames F₁ to F₄, and calculates the average value of the sizes of the encoded data in the first to fourth frames F₁ to F₄ as the average data size. As described with reference to FIGS. 6 and 7, the rate control section 50 calculates the amount of code used for encoding of the seventh frame F₇ using the average data size, and calculates the quantization parameter Qc. As a result, the quantization section 20 quantizes the image data in the seventh frame F₇ using the product of the quantization parameter Qc calculated based on the average value of the data sizes of the encoded data in the first to fourth frames F₁ to F₇ and the quantization step values in the quantization table 22. The quantization result is written into the FIFO buffer section 30. This enables the encoded data to be generated at a desired rate.

In this rate control method, a constant rate must be maintained even when the data size rapidly changes. Therefore, if the value N is decreased, the quantization step follows when the data size in only one frame rapidly changes, whereby the image quality of other frames deteriorates. On the other hand, if the value N is increased, the quantization step changes to only a small extent when the data size in only one frame rapidly changes.

As an example in which the data size rapidly changes, the case where image data of an intra-frame-coded (I) picture (frame which has an intra-frame-coded macroblock which is encoded within one frame) is input can be given. In this case, since there is no continuity with the image data in the previous frame, the generation rate of the encoded data is rapidly decreased, thereby making it necessary to change the rate control method.

In this embodiment, when the quantized data read from the FIFO buffer section 30 is data of an intra-frame-coded macroblock which is encoded within one frame, the rate control section 50 calculates the data size of the encoded data obtained by encoding the quantized data as the average data size, and changes the quantization step based on the average data size. This enables the quantization step to appropriately follow even when there is no continuity with the image data in the previous frame such as the I picture.

The rate control section 50 may calculate the average data size after the quantization section 20 quantizes the image data for M frames (M≧N, M is an integer) by a predetermined quantization step (quantization step determined in advance), and change the quantization step based on the average data size. This prevents unnecessary rate control from being performed when the average value of the data size of the encoded data in the subsequent N frames is not accurate, whereby deterioration of the image quality can be prevented.

3.1 Calculation Processing of Quantization Parameter Qc

The calculation processing of the quantization parameter Qc performed by the rate control section 50 is described below in detail.

FIGS. 11 and 12 show an example of a processing flow of the calculation processing of the quantization parameter Qc. The flow shown in FIGS. 11 and 12 is described below with reference to an explanatory diagram of variables used in the calculation processing of the quantization parameter Qc shown in FIG. 13. The flow shown in FIGS. 11 and 12 is performed in each frame.

The number of bits S used in the previous frame is calculated (step S40). The value of the number of bits Rc used for encoding which is calculated in the previous frame (number of bits used for encoding of the current frame) is set as the variable S.

FIG. 14 shows an example of a flow of calculation processing of the value of the number of bits Rc used for encoding. Whether or not the encoding target image data is image data of an intra-frame-coded macroblock is judged (step S60). When the image data is judged to be image data of an intra-frame-coded macroblock (step S60: Y), the data sizes of the encoded data for previous N frames are replaced by the data size of the encoded data in the current frame (step S61).

When it is judged that the encode target image data is not image data of an intra-frame-coded macroblock, specifically, when it is judged that the encode target image data is image data of an inter-frame-coded macroblock in the step S60 (step S60: N), or after the step S61 has been performed, the value of the variable Rc is calculated (step S62).

In the step S62, the data size of the encoded data in the previous kth frame (k is a positive integer) is denoted by fs(k), and the value of the variable Rc is calculated by dividing the sum of the data sizes of the encoded data for the previous N frames by the number of frames N. The value of the variable Rc is the average data size.

Each of the previous N frames is a frame at least two frames before the current frame.

The value of the variable Rc thus calculated is set as the variable S in the next frame.

The description is given with reference to FIG. 11. When the value of the variable S has been calculated, the number of bits T assigned to the current frame is calculated (step S41). In the step S41, the average number of bits assigned per frame (Rr/Nr) is calculated from the remaining available number of bits Rr and the remaining number of encoding frames Nr, and the number of bits T assigned to the current frame is calculated from the ratio of the resulting value to the number of bits S assigned to the previous frame. In the step S41, the ratio is 0.95:0.05, for example. The number of bits T assigned to the current frame is set so as not to be less than Rs/30 which is the lower limit.

The number of bits T assigned to the current frame is adjusted from the ratio of the number of currently occupied bits B of the FIFO buffer section 30 to the number of bits Bs of the FIFO buffer section 30 (step S41). As a result, the value of the variable T is increased when the number of currently occupied bits B of the FIFO buffer section 30 is smaller than half of the number of bits Bs of the FIFO buffer section 30, and the value of the variable T is decreased when the number of currently occupied bits B of the FIFO buffer section 30 is greater than half of the number of bits Bs of the FIFO buffer section 30.

Whether or not the sum of the number of currently occupied bits B of the FIFO buffer section 30 and the variable T exceeds 90 percent of the number of bits Bs of the FIFO buffer section 30 is judged (step S43). When it is judged that the sum exceeds 90 percent of the variable Bs (step S43: Y), the value of the variable T is set (clipped) at a value obtained by subtracting the value of the variable B from 90 percent of the number of bits Bs of the FIFO buffer section 30 (step S44). Specifically, the sum of the number of currently occupied bits B of the FIFO buffer section 30 and the variable T is set so as not to exceed 90 percent of the number of bits Bs of the FIFO buffer section 30. The value of the variable T is set so as not to be less than Rs/30 which is the lower limit in the same manner as in the step S41.

When it is judged that the sum does not exceed 90 percent of the value of the variable Bs (step S43: N), the value of the variable T is set at a value obtained by subtracting the value of the variable B from the average generation number of bits Rp per frame and adding 10 percent of the variable Bs to the resulting value (step S45). Specifically, the value obtained by subtracting the average generation number of bits Rp per frame from the sum of the variable Bs and the variable T is set so as not to be less than 10 percent of the number of bits Bs of the FIFO buffer section 30.

After the step S44 or S45, the value of the variable T is set so as not to exceed the remaining available number of bits Rr (step S46). The value of the variable T is adjusted so that the value of the variable T does not change between frames to a large extent (step S47).

In order to calculate the quantization parameter Qc, the model equation shown in FIG. 6 is solved as the quadratic equation of the variable Qc. Therefore, the value of the variable tmp is calculated as shown in FIG. 12 (step S48).

When the model parameter X₂ is zero, or the value of the variable tmp is a negative value (step S49: Y), the quantization parameter Qc is calculated from the model equation as the linear equation (step S50). Since the variable R becomes a value obtained by subtracting the number of bits Hp other than the information such as the header among the number of bits used in the previous frame from the number of bits T assigned to the current frame, the quantization parameter Qc is calculated as Qc=X₁×Ec/(T−Hp). The value of the variable Ec is the average absolute value of the pixels in the frame as shown in FIG. 6.

When the model parameter X₂ is not zero and the value of the variable tmp is equal to or greater than zero (step S49: N), the solution of the quadratic equation derived from the model equation shown in FIG. 6 is used as the quantization parameter Qc (step S51).

After the step S50 or the step S51, the processing is performed so that the difference between the quantization parameter Qc and the quantization parameter Qp in the previous frame is within 25 percent and the quantization parameter Qc is 1 to 31 (step S52, step S53, step S54, and step S55). In the step S52 and the step S54, ceil (x) means that the value x is rounded off to an integer in the positive direction.

The quantization step of the quantization section 20 is changed by supplying the quantization parameter Qc thus calculated to the quantization section 20.

Specifically, as shown in FIG. 15, a DCT coefficient D_ij of image data represented by the DCT coefficient is divided by the product of a quantization step value Q_ij at a corresponding position in the quantization table and the quantization parameter Qc to calculate a quantized DCT coefficient d_ij. As a result, the amount of zero data of the quantized DCT coefficients can be increased or decreased.

3.2. Configuration Example

FIG. 16 shows a detailed functional block diagram of an image data compression device in this embodiment. In FIG. 16, sections the same as the sections of the image data compression device 10 shown in FIG. 8 are denoted by the same symbols. Description of these sections is appropriately omitted.

An image data compression device 100 shown in FIG. 16 performs compression processing of image data of a moving image conforming to MPEG-4. The image data compression device 100 includes a hardware processing section 110 and a software processing section 150.

The hardware processing section 110 processes image data of a moving image by hardware. The hardware processing section 110 includes the quantization section 20 and the FIFO buffer section 30. The hardware processing section 110 is realized by hardware such as an ASIC and a dedicated circuit without using software.

The software processing section 150 performs encode processing of quantized data read from the FIFO buffer section 30 by software to generate encoded data. The software processing section 150 includes the encoded data generation section 40 and the rate control section 50. The software processing section 150 is a processing section whose function is realized by software (firmware). The function of the software processing section 150 is realized by a CPU (hardware) or the like which reads the software (firmware).

In more detail, the hardware processing section 110 includes a discrete cosine transform (DCT) section 112, a motion estimation section 114, an inverse quantization section 116, an inverse DCT section 118, and a motion compensation section 120. The DCT section 112 performs the processing in the step S2 shown in FIG. 1A. The motion estimation section 114 performs the processing in the step S1 shown in FIG. 1A. The inverse quantization section 116 performs the processing in the step S4 shown in FIG. 1A. The inverse DCT section 118 performs the processing in the step S5 shown in FIG. 1A. The motion estimation section 120 performs the processing in the step S6 shown in FIG. 1A.

Specifically, the hardware processing section 110 outputs the difference between input image data in the current frame and image data in the previous frame as motion vector information, performs a discrete cosine transform of the motion vector information, and outputs the resulting information to the quantization section. The hardware processing section 110 generates the previous image data based on inverse-quantized data calculated by inverse-quantizing the quantized data by the quantization step.

The hardware processing section 110 does not necessarily include all of these sections. The hardware processing section 110 may have a configuration in which at least one of these sections is omitted.

The encoded data generation section 40 of the software processing section 150 includes a DC/AC prediction section 152, a scan section 154, and a VLC coding section 156. The DC/AC prediction section 152 performs the processing in the step S7 shown in FIG. 1A. The scan section 154 performs the processing in the step S8 shown in FIG. 1A. The VLC coding section 156 performs the processing in the step S9 shown in FIG. 1A.

The software processing section 150 does not necessarily include all of these sections. The software processing section 150 may have a configuration in which at least one of these sections is omitted. For example, the software processing section 150 may encode the quantized data read from the FIFO buffer section 30 into a variable length code. The software processing section 150 may perform scan processing of rearranging the quantized data read from the FIFO buffer section 30, and encode the result of the scan processing into a variable length code. The software processing section 150 may calculate the DC component and the AC component from the quantized data read from the FIFO buffer section 30, perform scan processing of rearranging the DC component and the AC component, and encode the result of the scan processing into a variable length code.

In this embodiment, the steps S1 to S6 shown in FIG. 1A are processed by hardware and the steps S7 to S10 are processed by software for the following reasons. Since zero data accounts for the majority of each block as shown in FIG. 5 after the quantization in the step S3 shown in FIG. 1A, the amount of information on the type of data is significantly small in comparison with the data before the quantization (FIG. 3). Moreover, since the operation load in the steps S7 to S10 is small, the processing load is small even if the processing in the steps S7 to S10 shown in FIG. 1A is processed by software. On the contrary, the quantization in the step S3, DCT in the step S2, and inverse DCT in the step S5 shown in FIG. 1A handle a large amount of information and perform a complicated operation. Therefore, the processing load is heavy for software processing. Although the quantization, DCT, inverse DCT, and motion compensation are heavy load processing, the necessity of changing the processing is limited since the processing is standardized. Moreover, since most of the processing in the steps S1 to S6 shown in FIG. 1A is repetition, this is suitable for hardware processing. Furthermore, since the amount of data after the quantization processed by the hardware processing section 110 is small, the amount of data transferred from the hardware processing section 110 to the software processing section 150 is small, whereby the data transfer control load can be reduced.

FIG. 17 shows a hardware configuration example of the image data compression device 100. In this example, the hardware processing section 110 shown in FIG. 16 is integrated and is mounted on a semiconductor device as an encoder IC (integrated circuit) (encoder in a broad sense) 200. The function of the software processing section 150 shown in FIG. 16 is realized by a host 210. In FIG. 17, sections the same as the sections of the hardware processing section 110 shown in FIG. 16 are denoted by the same symbols. Description of these sections is appropriately omitted.

The host 210 includes a CPU 212 and a memory 214. A program for realizing the functions of the encoded data generation section 40 and the rate control section 50 is stored in the memory 214. The CPU 212 realizes the functions of the encoded data generation section 40 and the rate control section 50 by reading the program stored in the memory 214 and performs processing based on the program.

The encoder IC 200 encodes image data of a moving image obtained by imaging by a camera module (not shown) (imaging section in a broad sense) conforming to the MPEG-4 standard, and generates encoded data at a constant rate. Therefore, the encoder IC 200 includes, in addition to the circuit which realizes the function of each section of the hardware processing section 110 shown in FIG. 16, a host interface (I/F) 202, a camera I/F (image input interface in a broad sense) 204, and a quantization parameter setting register 206.

The encoder IC 200 and the host 210 realize the function of the image data compression device shown in FIG. 8 or 16 by exchanging an interrupt signal and data.

The host I/F 202 performs interface processing with the host 210. In more detail, the host I/F 202 controls generation of an interrupt signal from the encoder IC 200 to the host 210, and transmission/reception of data between the host 210 and the encoder IC 200. The host I/F 202 is connected with the FIFO buffer section 30.

The camera I/F 204 performs interface processing for inputting image data of a moving image input from the camera module (not shown). The camera I/F 204 is connected with the motion estimation section 114.

The camera module (not shown) supplies image data of a moving image obtained by imaging to the encoder IC 200 as input image data. The camera module also supplies a VSYNC signal (vertical synchronization signal) which designates the end of the frame of the input image data to the encoder IC 200. In the encoder IC 200, the camera I/F 204 receives the VSYNC signal from the camera module as a VSYNC interrupt. This causes the encoder IC 200 to start encoding.

The motion estimation section 114 does not perform the motion estimation of the input image data captured for the first time after the encoding is started, and performs the motion estimation after the input image data in the next frame has been captured. Since the details of the motion estimation are the same as described above, description of the operations of the inverse quantization section 116 and the like is omitted. The quantized data for at least one frame is written into the FIFO buffer section 30 in the stage in which the motion estimation is performed. After the motion estimation of the motion estimation section 114 has been completed, the motion estimation section 114 notifies the host 210 of a motion estimation completion interrupt (ME interrupt) through the host I/F 202.

FIG. 18 shows an example of a flow of interrupt reception processing performed by the host 210. A program which realizes the processing shown in FIG. 18 is stored in the memory 214. The CPU 212 reads the program and realizes the processing shown in FIG. 18.

The CPU 212 monitors an interrupt input (step S70: N). When the CPU 212 detects an interrupt (step S70: Y), the CPU 124 judges whether or not the interrupt is an ME interrupt (step S71).

When the CPU 212 judges that the interrupt is an ME interrupt (step S71: Y), ME interrupt processing described later is performed (step S72).

When the CPU 212 judges that the interrupt is not an ME interrupt (step S71: N), the CPU 124 judges whether or not the interrupt is an encode completion interrupt described later (step S73). When the CPU 212 judges that the interrupt is an encode completion interrupt (step S73: Y), encode completion interrupt processing described later is performed (step S74).

When the CPU 212 judges that the interrupt is not an encode completion interrupt in the step S73 (step S73: N), predetermined interrupt processing is performed (step S75).

After the step S72, the step S74, or the step S75, when the processing is not completed (step S76: N), the processing in the step S70 is performed. When the processing is completed (step S76: Y), the series of processing is terminated (END).

FIG. 19 shows an example of a flow of the ME interrupt processing. A program which realizes the processing shown in FIG. 19 is stored in the memory 214. The CPU 212 reads the program and realizes the processing shown in FIG. 19.

The ME interrupt processing is performed in the step S72 shown in FIG. 18.

When the ME interrupt is detected, the CPU 212 reads the complexity Ec generated by the motion estimation section 114 through the host I/F 202 (step S80). The complexity Ec is generated by the motion estimation section 114 according to the equation shown in FIG. 6.

The CPU 212 calculates the quantization parameter Qc (step S81). In more detail, the CPU 212 calculates the value of the quantization parameter Qc as described with reference to FIGS. 11 to 14.

The CPU 212 sets the value of the quantization parameter Qc calculated in the step S81 in a quantization parameter setting register 206 through the host I/F 202 (step S82), and the processing is terminated.

The description is given with reference to FIG. 17. The encoder IC 200 starts processing of the DCT section 112 when the quantization parameter Qc is set in the quantization parameter setting register 206. The quantization section 20 quantizes the DCT coefficients (image data in a broad sense) generated by the DCT section 112 using the quantization parameter set in the quantization parameter setting register 206 and the quantization step values in the quantization table (not shown). The resulting quantized data is written into the FIFO buffer section 30.

When the quantized data is written into the FIFO buffer section 30, the FIFO buffer section 30 notifies the host 210 of the encode completion interrupt which indicates that the encode processing of one frame has been completed through the host I/F 202.

FIG. 20 shows an example of a flow of the encode completion interrupt processing. A program which realizes the processing shown in FIG. 20 is stored in the memory 214. The CPU 212 reads the program and realizes the processing shown in FIG. 20.

The encode completion interrupt processing is performed in the step S74 shown in FIG. 18.

When the CPU 212 detects the encode completion interrupt, the CPU 212 reads the quantized data from the FIFO buffer section 30 in frame units (step S90).

The CPU 212 performs DC/AC prediction processing (step S91), scan processing (step S92), and variable length coding processing (step S93) in macroblock units to generate encoded data.

The CPU 212 adds a macroblock header to the encoded data generated in the step S93. The resulting encoded data is performed for one video object plane (VOP), a GOV header and a VOP header are generated based on the calculated quantization parameter, and, when the encoding of a predetermined number of frames is completed, the resulting encoded data is output as an MPEG-4 file (step S95).

The compression processing of image data is allocated between and performed by the hardware processing section 110 and the software processing section 150 as described above.

4. Display Controller

The function of the above-described encoder IC may be applied to a display controller.

FIG. 21 shows a block diagram of a configuration example of a display controller in this embodiment.

A display controller 300 includes a camera I/F 310, an encode processing section 320, a memory 330, a driver I/F 340, a control section 350, and a host I/F 360.

The camera I/F 310 is connected with a camera module (not shown). The camera module outputs input image data of a moving image obtained by imaging in a YUV format, and outputs a synchronization signal (VSYNC signal, for example) which designates the end of one frame. The camera I/F 310 performs interface processing for receiving the input image data of the moving image generated by the camera module.

The encode processing section 320 is a section in which the functions of the host I/F 202 and the camera I/F 204 in the encoder IC 200 shown in FIG. 17 are omitted. Specifically, the encode processing section 320 has the function of each of the quantization section 20, the FIFO buffer section 30, the DCT section 112, the motion estimation section 114, the inverse quantization section 116, the inverse DCT section 118, the motion compensation section 120, and the quantization parameter setting register 206 shown in FIG. 17.

The memory 330 stores encoded data which is the output from the encode processing section 320. The memory 330 stores image data to be displayed on a display panel. The driver I/F 340 reads the image data from the memory 330 in a predetermined cycle, and supplies the image data to the display driver which drives the display panel. The driver I/F 340 performs interface processing for transmitting the image data to the display driver.

The control section 350 controls the camera I/F 310, the encode processing section 320, the memory 330, and the driver I/F 340. The control section 350 performs reception processing of the input image data from the camera module, encode processing of the input image, write processing of the encoded data into the memory 330, read processing of the display image data from the memory 330, and transmission processing of the image data to the display driver according to instructions from a host (not shown) through the host I/F 360.

FIG. 22 shows a block diagram of a configuration example of an electronic instrument to which the display controller shown in FIG. 21 is applied. FIG. 22 shows the block diagram of a configuration example of a portable telephone as an electronic instrument. In FIG. 22, sections the same as the sections shown in FIG. 21 are denoted by the same symbols. Description of these sections is appropriately omitted.

A portable telephone 400 includes a camera module 410. The camera module 410 includes a charge-coupled device (CCD) camera, and supplies data of an image imaged by the CCD camera to the display controller 300 in a YUV format.

The portable telephone 400 includes a display panel 420. A liquid crystal display panel may be employed as the display panel 420. In this case, the display panel 420 is driven by a display driver 430. The display panel 420 includes a plurality of scan lines, a plurality of data lines, and a plurality of pixels. The display driver 430 has a function of a scan driver which selects the scan line in units of one or more scan lines, and a function of a data driver which supplies voltage corresponding to image data to the data lines.

The display controller 300 is connected with the display driver 430, and supplies image data to the display driver 430.

A host 440 is connected with the display controller 300. The host 440 controls the display controller 300. The host 440 demodulates image data received through an antenna 460 using a modulator-demodulator section 450, and supplies the demodulated image data to the display controller 300. The display controller 300 causes the display driver 430 to display an image on the display panel 420 based on the image data.

The host 440 has the function of the host 210 shown in FIG. 17. The host 440 encodes the image data generated by the camera module 410 using the encode processing section 320, modulates the encoded data using the modulator-demodulator section 450, and directs transmission of the modulated data to other communication device through the antenna 460. The display controller 300 encodes the image data generated by the camera module 410, and outputs the encoded data to the host 440.

The host 440 performs transmission-reception processing of image data, encode processing, imaging of the camera module 410, and display processing of the display panel based on operation information from an operation input section 470.

FIG. 22 illustrates a liquid crystal display panel as an example as the display panel 420. However, the present invention is not limited thereto. The display panel 420 may be an electroluminescence or plasma display device. The present invention may be applied to a display controller which supplies image data to a display driver which drives these display device.

Although only some embodiments of the present invention have been described in detail above, those skilled in the art will readily appreciate that many modifications are possible in the embodiments without materially departing from the novel teachings and advantages of this invention. Accordingly, all such modifications are intended to be included within scope of this invention.

Claims

1. An image data compression device for compressing image data, the image data compression device comprising: a quantization section which quantizes image data in unit of a frame; a FIFO buffer section in which quantized data for at least N frames (N is an integer larger than one) quantized by the quantization section is buffered; an encoded data generation section which reads the quantized data from the FIFO buffer section asynchronously from writing into the FIFO buffer section, and generates encoded data by encoding the quantized data; and a rate control section which changes a data size of encoded data to be processed by changing a quantization step of the quantization section by using the data size of the encoded data, wherein the rate control section obtains an average data size by averaging the data sizes of the encoded data for N frames before a frame of the image data quantized by the quantization section, and changes the quantization step based on the average data size.

2. The image data compression device as defined in claim 1, wherein, when the quantized data read from the FIFO buffer section is data of an intra-frame-coded macroblock which is encoded within one frame, the rate control section obtains the data size of the encoded data obtained by encoding the quantized data as the average data size, and changes the quantization step based on the average data size.

3. The image data compression device as defined in claim 1, wherein the rate control section obtains the average data size after the quantization section has quantized the image data for M frames (M≧N, M is an integer) by a predetermined quantization step.

4. The image data compression device as defined in claim 1, comprising: a quantization table which stores a quantization step value, wherein the rate control section changes the quantization step by obtaining a quantization parameter based on the average data size and performing quantization using a product of the quantization parameter and the quantization step value.

5. The image data compression device as defined in claim 1, comprising a discrete cosine transform section which supplies the image data subjected to a discrete cosine transform to the quantization section in unit of a frame.

6. The image data compression device as defined in claim 1, comprising: a hardware processing section which processes image data of a moving image by hardware; and a software processing section which generates encoded data by performing encode processing of the quantized data read from the FIFO buffer section by software, wherein the hardware processing section includes the quantization section and the FIFO buffer section, and wherein the software processing section includes the encoded data generation section and the rate control section.

7. The image data compression device as defined in claim 6, wherein the hardware processing section outputs a difference between input image data in a current frame and previous image data in a frame one frame before the current frame as motion vector information, performs a discrete cosine transform for the motion vector information and outputs the transformed motion vector information to the quantization section as the image data, and generates the previous image data based on inverse-quantized data obtained by inverse-quantizing the quantized data by the quantization step.

8. The image data compression device as defined in claim 7, wherein the software processing section encodes the quantized data read from the FIFO buffer section into a variable length code.

9. The image data compression device as defined in claim 8, wherein the software processing section performs scan processing of rearranging the quantized data read from the FIFO buffer section, and encodes a result of the scan processing into a variable length code.

10. The image data compression device as defined in claim 9, wherein the software processing section obtains a DC component and an AC component from the quantized data read from the FIFO buffer section, performs scan processing of rearranging the DC component and the AC component, and encodes a result of the scan processing into a variable length code.

11. An encoder which performs compression processing for image data, the encoder comprising: an image input interface which performs interface processing for inputting image data; a quantization section which quantizes the image data in unit of a frame; a FIFO buffer section in which quantized data for at least N frames (N is an integer larger than one) quantized by the quantization section is buffered; and a host interface which performs interface processing with a host which reads the quantized data stored in the FIFO buffer section asynchronously from writing into the FIFO buffer section, wherein the host obtains an average data size by averaging data size of encoded data for N frames before a frame of the image data quantized by the quantization section, and wherein the quantization section quantizes the image data by a quantization step changed based on the average data size.

12. The encoder as defined in claim 11, wherein, when the quantized data read from the FIFO buffer section is data of an intra-frame-coded macroblock which is encoded within one frame, the host obtains the data size of the encoded data obtained by encoding the quantized data as the average data size, and the quantization section quantizes the image data by the quantization step changed based on the average data size.