Some parameters passed to the av_dlog can be either float or int, depending on
the mode the file is being compiled as.  Cast those parameters to float and use
appropriate conversion specifiers.