From: Ian S. <ian...@st...> - 2003-12-05 18:14:51
|
Marc, I've been talking to some other users of non 8*n bit images here, and the consensus is that vil file loaders should default to keeping data in the least significant bits of the C++ pixel type. So get_view() on an NITF image would return pixels valued at 0 to 2047. If you think NITF and other non 8*n bit images should have their bits shifted right, let me know and I'll open the debate to the reflector. I've added a vil_property_quantisation_depth which you may find useful. Ian. |