[johnll]

Software doesn't receive any IR-specific data. All it gets is the intensities at each photosite, and knows what sort of filter (RGB) was at each site. It may conceivably be provided with the IR characteristics of each of these filters. Presumably there might be quite a bit of IR coming through the red filters, probably less or even none coming through the green and blue filters. In the demozaicing, it *might* be possible to isolate the IR components to some extent that way, but it would probably only be a partial fix. I think the IR issue is going to need hardware modifications - probably to the sensor cover glass.