Line-based compression for digital image data

1. A method comprising:

computing a minimum absolute sample difference (MASD) for a pixel in a line of pixels based on differences between different combinations of neighboring pixels of the pixel, wherein the neighboring pixels comprise a left neighboring pixel in the line of pixels, a top left neighboring pixel in a previous line of pixels, and a top neighboring pixel in the previous line of pixels, and wherein the smallest of the differences between the different combinations of the neighboring pixels of the pixel is selected as the MASD for the pixel; and

in response to the MASD for the pixel being below a threshold, encoding the pixel using variable length coding.

2. The method of claim 1 , further comprising in response to the MASD for the pixel being above a threshold, encoding the pixel using fixed length coding.

3. The method of claim 1 , comprising:

determining the threshold using a quantization parameter; and

computing a prediction error based on the MASD;

quantizing the prediction error based on the quantization parameter.

4. The method of claim 1 , further comprising:

counting a number of bits required for encoding a previous line of pixels using each of a plurality of variable length entropy codes as the previous line is encoded; and

selecting a variable length entropy code for the variable length coding that requires a least number of bits.

5. The method of claim 3 , wherein quantizing the prediction error based on the quantization parameter comprises quantizing the prediction error using a linear quantization based on the quantization parameter.

6. The method of claim 1 , wherein a variable length code used for the variable length coding is a signed exponential-Golomb code.

7. The method of claim 1 , wherein the computing step and the encoding step are performed independently for each color component of the pixel.

8. One or more computer-readable non-transitory storage media embodying software that is operable when executed to:

compute a minimum absolute sample difference (MASD) for a pixel in a line of pixels based on differences between different combinations of neighboring pixels of the pixel, wherein the neighboring pixels comprise a left neighboring pixel in the line of pixels, a top left neighboring pixel in a previous line of pixels, and a top neighboring pixel in the previous line of pixels, and wherein the smallest of the differences between the different combinations of the neighboring pixels of the pixel is selected as the MASD for the pixel; and

in response to the MASD for the pixel being below a threshold, encode the pixel using variable length coding.

9. The media of claim 8 , wherein the software is further operable when executed to:

in response to the MASD being above a threshold, encoding the pixel using fixed length coding.

10. The media of claim 8 , wherein the software is further operable when executed to:

count a number of bits required for encoding the line of pixels using each of a plurality of variable length entropy codes; and

select a variable length entropy code for the variable length coding that requires a least number of bits.

11. The media of claim 8 , wherein the software is further operable when executed to:

count a number of bits required for encoding a previous line of pixels using each of a plurality of variable length entropy codes as the previous line is encoded; and

select a variable length entropy code for the variable length coding that requires a least number of bits.

12. The media of claim 8 , wherein the software is further operable when executed to:

determine the threshold using a quantization parameter;

compute a prediction error based on the MASD; and

quantize the prediction error based on the quantization parameter.

13. The media of claim 12 , wherein quantizing the prediction error based on the quantization parameter comprises quantizing the prediction error using a linear quantization based on the quantization parameter.

14. A system comprising: one or more processors; and a memory coupled to the processors comprising instructions executable by the processors, the processors being operable when executing the instructions to:

compute a minimum absolute sample difference (MASD) for a pixel in a line of pixels based on differences between different combinations of neighboring pixels of the pixel, wherein the neighboring pixels comprise a left neighboring pixel in the line of pixels, a top left neighboring pixel in a previous line of pixels, and a top neighboring pixel in the previous line of pixels, and wherein the smallest of the differences between the different combinations of the neighboring pixels of the pixel is selected as the MASD for the pixel; and

in response to the MASD for the pixel being below a threshold, encode the pixel using variable length coding.

15. The system of claim 14 , wherein the processors are further operable when executing the instructions to:

in response to the MASD for the pixel being above a threshold, encode the pixel using fixed length coding.

16. The system of claim 14 , wherein the processors are further operable when executing the instructions to:

count a number of bits required for encoding the line of pixels using each of a plurality of variable length entropy codes; and

select a variable length entropy code for the variable length coding that requires a least number of bits.

17. The system of claim 14 , wherein the processors are further operable when executing the instructions to:

count a number of bits required for encoding a previous line of pixels using each of a plurality of variable length entropy codes as the previous line is encoded; and

select a variable length entropy code for the variable length coding that requires a least number of bits.

18. The system of claim 14 , wherein the processors are further operable when executing the instructions to:

determining the threshold using a quantization parameter; and

computing a prediction error based on the MASD;

quantizing the prediction error based on the quantization parameter.

19. The system of claim 18 , wherein quantizing the prediction error based on the quantization parameter comprises quantizing the prediction error using a linear quantization based on the quantization parameter.