Specify the precision as float to avoid truncating from double. Signed-off-by: Mark Harris <mark.hsj@gmail.com>