Frame segmentation and grouping for audio encoding

A method to encode blocks of audio in frames, the method comprising:

• receiving an input signal that includes a set of block groups in a frame, where each block includes content;
• in a first loop, selecting each block group from the set of block groups, in turn, as a selected first block group for a potential merge;
• in a second loop, selecting each block group from the set of block groups different from the first block group, in turn, as a selected second block group for potential merge;
• obtaining a first measure of quality associated with a merge of the selected first block group with the selected second block group, the first measure of quality being a measure of side information cost associated with the merge of the selected first block group with the selected second block group;
• obtaining a second measure of quality associated with the merge of the selected first block group with the selected second block group, the second measure of quality being a measure of distortion due to spectral coefficient quantization associated with the merge of the selected first block group with the selected second block group;
• comparing the first measure of quality and the second measure of quality of the second block group to prior iterations of the second loop to selectively identify the second block group as a merge candidate block;
• after the second loop has completed all iterations, selectively merging the selected first block group with the identified merge candidate block;
• and
• after the second loop has completed all iterations, assembling the frame as an encoded signal and outputting the encoded signal.

An apparatus to encode blocks of audio information arranged in frames, where each frame includes a set of block groups, and each block includes content, the apparatus comprising:

• an electronic processor configured to:
• receive an input signal that includes the blocks of audio information, wherein the blocks of audio information includes a set of block groups for a respective frame,
• obtain, for each respective block group, a first measure of quality, wherein the first measure of quality indicates a side information cost associated with a merge of two or more blocks of audio information to form the respective block group,
• obtain, for each respective block group, a second measure of quality, wherein the second measure of quality indicates an estimated distortion due to spectral coefficient quantization associated with a merge of the two or more blocks of audio information to form the respective block group,
• merge, based on the first measure of quality and the second measure of quality, at least two block groups of the set of block groups to generate an encoded signal that represents contents of the input signal and associated control parameters for each block group in the set, and
• output the encoded signal.

A non-transitory computer-readable storage medium recording a program of instructions that is executable by a device to perform a method to process blocks of audio information arranged in frames, the method comprising:

• receiving an input signal including the blocks of audio information, wherein the blocks of audio information includes a set of block groups for a respective frame;
• obtaining, for each respective block group, a first measure of quality, wherein the first measure of quality indicates a side information cost associated with a merge of two or more blocks of audio information to form the respective block group;
• obtaining, for each respective block group, a second measure of quality, wherein the second measure of quality indicates an estimated distortion due to spectral coefficient quantization associated with a merge the two or more blocks of audio information to form the respective block group;
• merging, based on the first measure of quality and the second measure of quality, at least two block groups of the set of block groups to generate an encoded signal representing contents of the input signal and associated control parameters for each block group in the set;
• and
• outputting the encoded signal.

A method to encode blocks of audio in frames, the method comprising:

• receiving an input signal that includes a set of two or more audio block groups in a frame, where each block includes content;
• selecting, as merge candidates, one or more pairs of audio block groups from the frame;
• obtaining, for each merge candidate, a first measure of quality associated with a merge of the audio block groups, wherein the first measure of quality indicates a side information cost associated with the merge of the audio block groups;
• obtaining, for each merge candidate, a second measure of quality associated with a merge of the audio block groups, wherein the second measure of quality indicates a distortion due to spectral coefficient quantization associated with the merge of the audio block groups;
• selecting, based on the first and second measures of quality, one pair of audio block groups from the merge candidates that yields the largest improvement in coding accuracy;
• merging the selected pair of audio block groups;
• and
• assembling the frame as an encoded signal and outputting the encoded signal.