Flexible implementations of multiple transforms

A method for video encoding, comprising:

• accessing prediction residuals of a block of a picture in a video;
• transforming said prediction residuals of said block, using a DCT-IV transform type, to form a block of transform coefficients,
• wherein a transform for N-point DCT-IV type is obtained from a transform for 2N-point DCT-II type by performing:
• accessing an array of N prediction residuals,
• padding said array of N prediction residuals with N zeros to form a data array of 2N samples,
• transforming said data array of 2N samples with transform for 2N-point DCT-II type to form an output data array of 2N samples, and
• extracting N odd samples from said output data array of 2N samples as an array of N transform coefficients corresponding to said array of N prediction residuals;
• and
• entropy encoding said transform coefficients.

An apparatus for decoding video data, comprising:

• one or more processors, wherein said one or more processors are configured to:
• access transform coefficients of a block of a picture in a video;
• inverse transform said transform coefficients of said block, using a DCT-IV transform type, to form a block of prediction residuals,
• wherein an inverse transform for N-point DCT-IV type is obtained from an inverse transform for 2N-point DCT-II type by performing:
• accessing an array of N transform coefficients,
• padding said array of N transform coefficients with N zeros to form a data array of 2N samples,
• inverse transforming said data array of 2N samples with said inverse transform for 2N-point DCT-II type to form an output data array of 2N samples, and
• extracting N odd samples from said output data array of 2N samples as an array of N prediction residuals corresponding to said array of N transform coefficients;
• and
• reconstruct said block responsive to said block of prediction residuals.

An apparatus for encoding video data, comprising:

• one or more processors, wherein said one or more processors are configured to:
• access prediction residuals of a block of a picture in a video;
• transform said prediction residuals of said block, using a DCT-IV transform type, to form a block of transform coefficients,
• wherein a transform for N-point DCT-IV type is obtained from a transform for 2N-point DCT-II type by performing:
• accessing an array of N prediction residuals,
• padding said array of N prediction residuals with N zeros to form a data array of 2N samples,
• transforming said data array of 2N samples with said transform for 2N-point DCT-II type to form an output data array of 2N samples, and
• extracting N odd samples from said output data array of 2N samples as an array of N transform coefficients corresponding to said array of N prediction residuals;
• and
• entropy encode said transform coefficients.