Hi all, I am trying to implement (if possible ?) a lossless transform from a YCbCr 4:2:2 raw input into JPEG.
I've written a quick sample application: * https://github.com/malaterre/PublicRep/blob/master/dicom/yuv2jpg/yuv2jpg.cxx Basically: ``` cinfo->in_color_space = JCS_YCbCr; /* colorspace of input image, 24bits */ jpeg_set_defaults(cinfo); jpeg_set_quality(cinfo, 100, TRUE /* limit to baseline-JPEG values */); ``` I also made sure that chroma subsampling was setup to reflect input: ``` cinfo->comp_info[0].h_samp_factor = 2; cinfo->comp_info[0].v_samp_factor = 1; cinfo->comp_info[1].h_samp_factor = 1; cinfo->comp_info[1].v_samp_factor = 1; cinfo->comp_info[2].h_samp_factor = 1; cinfo->comp_info[2].v_samp_factor = 1; ``` The above code seems to be giving the correct output for simple inputs (*), but I see difference at the bit-level when re-decompressing back to YCbCr 4:2:2 raw for real-life examples. Could someone confirm that: 1. My code properly skips color transform step, 2. Discrete cosine transform + quantization step are stable (feeding the output of this step back as input should produce the same output). Thanks -Mathieu (*) % convert -size 512x512 -depth 8 xc:#000000 black.ppm % convert -size 512x512 -depth 8 xc:#ffffff white.ppm % convert -size 512x512 -depth 8 xc:#ff0000 red.ppm % convert -size 512x512 -depth 8 xc:#00ff00 green.ppm % convert -size 512x512 -depth 8 xc:#0000ff blue.ppm % for f in *.ppm; do cjpeg -sample 2x1,1x1,1x1 -optimize -quality 100 -outfile $f.jpg $f; done -- You received this message because you are subscribed to the Google Groups "libjpeg-turbo User Discussion/Support" group. To unsubscribe from this group and stop receiving emails from it, send an email to libjpeg-turbo-users+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/libjpeg-turbo-users/d2cffd89-1bc2-4790-9e9f-24faf62af87dn%40googlegroups.com.
