Apparatus for encoding moving images to minimize an amount of generated code

A method of encoding moving images, comprising:

• calculating, with circuitry and for each of a plurality of units of processing, amounts of code using a plurality of variable length compression modes, the units of processing being obtained by dividing a predetermined unit of image data, the plurality of variable length compression modes having compression rated different from each other;
• selecting, with the circuitry and for each of the units of processing, a compression mode in which an accumulated amount of code is smaller than or equal to a predetermined amount of code and in which a quantization width is minimum, the compression mode being selected from the variable length compression modes and a fixed length compression mode on the basis of the amounts of code calculated for the variable length compression modes and an amount of code obtained when the image data is compressed by the fixed length compression mode, and the accumulated amount of code being obtained by calculating, among the units of processing included in the predetermined unit, until a unit of processing in which the compression mode is selected;
• and
• compressing, with the circuitry, the units of processing of image data by the compression mode selected to generate compressed data,
• wherein the unit of processing is one of rectangles formed by dividing the image data into rectangles and arranging the rectangles in raster order, and
• the circuitry updates a compression mode selected for one of the rectangles by using information related to a compression rate of the compression mode selected for a rectangle that has already been compressed,
• wherein when the image data includes a plurality of component signals, the amount of code for each compression mode for each component signal of the unit of processing is calculated, and a compression mode on the basis of the sum of the amount of code corresponding to a component signal for each unit of processing is selected,
• wherein compression is performed including quantization and variable length encoding, and the compressed data is caused to include a quantization parameter that is information related to a quantization width corresponding to one of the component signals, a quantization parameter offset that is information related to a difference between the quantization width and a quantization width of another component signal, identification information of a variable length code table for each component signal, and a symbol value including variable length codes of a plurality of component signals, and
• wherein when a variable length code of a component signal of a pixel is represented by a combination of a prefix and a suffix, the symbol value is caused to include a code string in which prefixes of a plurality of component signals continue and a code string in which suffixes of a plurality of component signals continue.

A moving image encoding apparatus comprising:

• circuitry configured to
• calculate, for each of plural units of processing, amounts of code using a plurality of variable length compression modes, the units of processing being obtained by dividing a predetermined unit of image data, the plurality of variable length compression modes having compression rates different from each other;
• select, for each of the units of processing, a compression mode in which an accumulated amount of code is smaller than or equal to a predetermined amount of code and in which a quantization width is minimum, the compression mode being selected from the variable length compression modes and a fixed length compression mode on the basis of the amounts of code calculated for the variable length compression modes and an amount of code obtained when the image data is compressed by the fixed length compression mode, and the accumulated amount of code being obtained by calculating, among the units of processing included in the predetermined unit, until a unit of processing in which the compression mode is selected;
• and
• compress the unit of processing of image data by the compression mode selected to generate compressed data,
• wherein the predetermined amount of code is a value based on an average amount of code per predetermined unit when compressing the predetermined unit for each unit of processing by the fixed length compression mode.