Bug in mode 8 of bands init

[Higginbottom]

Mode 8 is the automatic log spaced bands for photon generation.
The lines which work out the depicting of the bands are

f1_log = log10 (f1);
f2_log = log10 (f2);
df = (f2_log - f1_log) / (band->nbands + 1);

but df should be divided by nbands not nbands+1

e.g. if we want 5 bands to run from 0 to 10, clearly each band must be 2 wide, but (10-0)/5+1 does not equal 2! This was causing the final band to be missed off - a lot of frequencies in log space!

I will correct, and submit a pull request.