Midpoint prediction error diffusion for display stream compression

A device comprising:

• a memory configured to store content;
• and
• one or more processors in communication with the memory, wherein the one or more processors are configured to:
• receive a first block of samples including at least a first sample, a second sample, and a third sample;
• calculate a predictor value for the first block of samples;
• calculate a residual between the predictor value and the first sample;
• quantize the residual to generate a quantized residual;
• de-quantize the quantized residual to generate a de-quantized residual;
• reconstruct the first sample using the de-quantized residual and the predictor value to generate a first reconstructed sample;
• calculate an error value based on a difference between the first sample and the first reconstructed sample;
• modify the second sample by the error value;
• modify the third sample by the error value;
• calculate a second residual between the predictor value and the modified second sample;
• quantize the second residual to generate a second quantized residual;
• de-quantize the second quantized residual to generate a second de-quantized residual;
• reconstruct the modified second sample using the second de-quantized residual and the predictor value to generate a second reconstructed sample;
• calculate a second error value based on a difference between the modified second sample and the second reconstructed sample;
• and
• modify the modified third sample by the second error value.

An apparatus comprising:

• means for receiving a first block of samples including at least a first sample, a second sample, and a third sample;
• means for calculating a predictor value for the first block of samples;
• means for calculating a residual between the predictor value and the first sample;
• means for quantizing the residual to generate a quantized residual;
• means for de-quantizing the quantized residual to generate a de-quantized residual;
• means for reconstructing the first sample using the de-quantized residual and the predictor value to generate a first reconstructed sample;
• means for calculating an error value based on a difference between the first sample and the first reconstructed sample;
• means for modifying the second sample by the error value;
• means for modifying the third sample by the error value;
• means for calculating a second residual between the predictor value and the modified second sample;
• means for quantizing the second residual to generate a second quantized residual;
• means for de-quantizing the second quantized residual to generate a second de-quantized residual;
• means for reconstructing the modified second sample using the second de-quantized residual and the predictor value to generate a second reconstructed sample;
• means for calculating a second error value based on a difference between the modified second sample and the second reconstructed sample;
• and
• means for modifying the modified third sample by the second error value.

A non-transitory computer-readable storage medium storing instructions that, when executed, cause one or more processors to:

• receive a first block of samples including at least a first sample, a second sample, and a third sample;
• calculate a predictor value for the first block of samples;
• calculate a residual between the predictor value and the first sample;
• quantize the residual to generate a quantized residual;
• de-quantize the quantized residual to generate a de-quantized residual;
• reconstruct the first sample using the de-quantized residual and the predictor value to generate a first reconstructed sample;
• calculate an error value based on a difference between the first sample and the first reconstructed sample;
• modify the second sample by the error value;
• modify the third sample by the error value;
• calculate a second residual between the predictor value and the modified second sample;
• quantize the second residual to generate a second quantized residual;
• de-quantize the second quantized residual to generate a second de-quantized residual;
• reconstruct the modified second sample using the second de-quantized residual and the predictor value to generate a second reconstructed sample;
• calculate a second error value based on a difference between the modified second sample and the second reconstructed sample;
• and
• modify the modified third sample by the second error value.