A dither pattern is formed in such a manner that threshold values 1 to 16 are dispersed in regard to the order of magnitudes of the threshold values. First, there is made a calculation of multiplying each value of the image data by a ratio of 1/2. Thereby, there is produced the divided 17-valued data where every value of all pixels is “2”. Next, binarization is carried out for the divided data by using the dither pattern. As a result, binary data at the first pass is obtained. Next, there is obtained the data by a product of a ratio (1/2+1/2) obtained by adding 1/2 to the above ratio and each value of the image data. Binarization is carried out for the image data by using the dither pattern. This produces binary data. Finally, dot data for the second pass is obtained by subtracting data from the binary data.