44     /* For vertical dithering, the error is spread on the next column (and not next line).  in lv_dither_ordered_hor()
46        The algorithm below is based on few assumptions:  in lv_dither_ordered_hor()
47 …ere would be hard to implement since it means that a pixel on column n depends on the pixel on row…  in lv_dither_ordered_hor()
49 …3. It means that a pixel i,j only depends on the value of a pixel i-7, j-7 to i,j and no other one.  in lv_dither_ordered_hor()
67     /* For vertical dithering, the error is spread on the next column (and not next line).  in lv_dither_ordered_ver()
69        The algorithm below is based on few assumptions:  in lv_dither_ordered_ver()
70 …ere would be hard to implement since it means that a pixel on column n depends on the pixel on row…  in lv_dither_ordered_ver()
72 …3. It means that a pixel i,j only depends on the value of a pixel i-7, j-7 to i,j and no other one.  in lv_dither_ordered_ver()
156 …/* Try to implement error diffusion on a vertical gradient and an horizontal map using those trick…  in lv_dither_err_diff_ver()
162        2 [  ][  ][  ]     Pixel A2 will spread its error on pixel A3 with coefficient 7,  in lv_dither_err_diff_ver()
163 …3 [  ][  ][  ]     Pixel A2 will spread its error on pixel B1 with coefficient 3, B2 with coef 5 a…  in lv_dither_err_diff_ver()
168        This means that the error term depends on pixel W, NW, N and SW.  in lv_dither_err_diff_ver()
171 …W term is painful for us, we can't support it (since to get it, we need its own SW term and so on).  in lv_dither_err_diff_ver()