These may be questions for Gordon Richardson, but I thought others might be interested. Why do Canon digicams save JPEG with 4:2:2 chroma subsampling? According to Gordon's JPEG article, "The 4:2:2 option is an old standard from the television industry which is seldom used." It's 2x1x1 horizontally but 1x1x1 vertically. Why does Photoshop SaveForWeb use high quality factors like 98 in conjunction with 2x2 chroma subsampling? Mere stupidity, like the failure to warn about non-sRGB colorspace, perhaps. Is there evidence that different Q factors for qtable 1 and 2 helps? My IJG-based software can't do that; Canon and Photoshop do. At approximately what quality factor (on the IJG) scale should one switch from 2x2 to 1x1 chroma subsampling? Do the new JPEG encoding routines with PaintShopPro work better than IJG libraries, or has nobody evaluated this yet? Comparisons between Photoshop 7 and IJG showed that Photoshop performed better at lower Q factors, while IJG was tighter (smaller) at high Q factors with equal or better quality.