[QUOTE=timparkin;1020120]Yeah that's fair enough Working with a fixed MTF50 does give a universal reference point but I do honestly think that MTF10 or even MTF5 would be more realistic when comparing the visual resolution of 'system'. However in reality it's probably productive to show both.
You have an interesting point here that I've not fully contemplated. As the spacial frequency of an array of lines and spaces increases The MTF drops (I think more or less linearly). At some point it approaches near zero at some high frequency which means the utility of dynamic range is near zero for that high frequency end point and one can barely tell a space from a line. Full DR is non recoverable from a piece of film with finest lines and spaces. The same holds for actual images. Using MTF10 or MTF5 also will result in reduced dynamic range for fine features. In fact using MTF50 will show a slight reduction in dynamic range for that spacial frequency. I think it is true that the full dynamic range as a function of frequency can only be had when MTF is 100% for a set of line pairs.
So we have a sliding scale of dynamic range starting at 100% at some low frequency (large linewidths) and ending at near 0% for some high frequency (small linewidths). Fine detail is washed out on critically sharp prints at low MTF.
Nate Potter, Austin TX.
Bookmarks