Transform hardware architecture for video coding

A device for transforming video data, the device comprising:

• a pre-transform buffer configured to store sample values of a block of the video data, wherein the sample values comprise a plurality of frequency domain sample values;
• a transform circuit comprising N computation units, wherein the transform circuit is coupled to the pre-transform buffer and is configured to:
• receive M sample values of the stored sample values of the block of the video data from the pre-transform buffer, wherein M represents a non-zero integer value;
• process the M sample values with the N computation units of the transform circuit to generate intermediate values, wherein processing the M sample values to generate the intermediate values includes feeding back temporary values from output of one or more of the N computation units to input of one or more of the N computation units for transforming the same block of the video data that includes the M sample values, and wherein N represents a non-zero integer value;
• and
• output a first set of the intermediate values for storage;
• and
• a transpose buffer coupled to the pre-transform buffer and the transform circuit, wherein the transpose buffer is configured to receive the first set of intermediate values and store the first set of intermediate values,
• wherein the pre-transform buffer is configured to store a second set of the intermediate values that are to be later retrieved for storage in the transpose buffer,
• wherein the transform circuit comprises a first transform circuit, the device further comprising a second transform circuit, wherein the second transform circuit is configured to process the intermediate values generated by the first transform unit to generate final transform values, the final transform values comprising a plurality of residual sample values representing differences between a current video block and a predictive block,
• wherein to process the intermediate values, the second transform circuit is configured to:
• receive a plurality of the intermediate values from the transpose buffer;
• and
• process the plurality of intermediate values with computation units of the second transform circuit to generate the final transform values, wherein processing the plurality of intermediate values to generate the final transform values includes feeding back temporary values from output of one or more of the computation units of the second transform circuit to input of one or more of the computation units of the second transform circuit.

A device for transforming video data, the device comprising:

• a pre-transform buffer;
• a transpose buffer;
• and
• a transform circuit comprising:
• N computation units;
• means for receiving M sample values of a block of the video data from the pre-transform buffer, wherein M represents a non-zero integer value, and wherein the sample values comprise a plurality of frequency domain sample values;
• means for processing the M sample values with the N computation units of the transform circuit to generate intermediate values, wherein the means for processing the M sample values to generate the intermediate values includes means for feeding back temporary values from output of one or more of the N computation units to input of one or more of the N computation units for transforming the same block of the video data that includes the M sample values, and wherein N represents a non-zero integer value;
• and
• means for storing a first set of the intermediate values in the transpose buffer;
• wherein the pre-transform buffer is configured to store a second set of the intermediate values that are to be later retrieved for storage in the transpose buffer,
• wherein the transform circuit comprises a first transform circuit, the device further comprising a second transform circuit comprising means for processing the intermediate values generated by the first transform unit to generate final transform values, the final transform values comprising a plurality of residual sample values representing differences between a current video block and a predictive block, and
• wherein the means for processing the intermediate values comprises:
• means for receiving a plurality of the intermediate values from the transpose buffer;
• and
• means for processing the plurality of intermediate values with computation units of the second transform circuit to generate the final transform values, wherein the means for processing the plurality of intermediate values to generate the final transform values includes means for feeding back temporary values from output of one or more of the computation units of the second transform circuit to input of one or more of the computation units of the second transform circuit.

A computer-readable storage medium storing instruction that when executed cause a first transform circuit to:

• receive M sample values of a block of the video data from a pre-transform buffer, wherein M represents a non-zero integer value, and wherein the sample values comprise a plurality of frequency domain sample values;
• process the M sample values with N computation units of the first transform circuit to generate intermediate values, wherein the instructions that cause the first transform circuit to process the M sample values to generate the intermediate values include instructions that cause the first transform circuit to feed back temporary values from output of one or more of the N computation units to input of one or more of the N computation units for transforming the same block of the video data that includes the M sample values, and wherein N represents a non-zero integer value;
• store a first set of the intermediate values in a transpose buffer;
• and
• store a second set of the intermediate values in the pre-transform buffer that are to be later retrieved for storage in the transpose buffer,
• wherein the computer-readable storage medium further comprises instruction that when executed cause a second transform circuit to process the intermediate values generated by the first transform unit to generate final transform values, the final transform values comprising a plurality of residual sample values representing differences between a current video block and a predictive block, and
• wherein the instructions that causes the second transform circuit to process the intermediate values comprise instructions that cause the second transform circuit to:
• receive a plurality of the intermediate values from the transpose buffer;
• and
• process the plurality of intermediate values with computation units of the second transform circuit to generate the final transform values, wherein the instructions that cause the second transform circuit to process the plurality of intermediate values to generate the final transform values include instructions that cause the second transform circuit to feed back temporary values from output of one or more of the computation units of the second transform circuit to input of one or more of the computation units of the second transform circuit.