Frequencies and wavelengths in output spectra refer to bin edges, not centers

[lazygun37]

We create spectra on grids spaced linear in frequency (or log-frequency). Most of the relevant code is in spectra.c. These grids are created via code like:

freqmin = f1;
freqmax = f2;
dfreq = (freqmax - freqmin) / NWAVE_NOW;

When we populate the spectra, we do that by calculating the index of the bin in which to add a photon via

k = (int) ((p[nphot].freq - freqmin) / dfreq);

When we write out the spectra, the frequencies (and associated wavelengths) we print are calculated via

freq = freqmin + i * dfreq;

Here, i is the integer counter of the loop over all bins.

What that means is that we actually print out the lower edge of each frequency bin, not its center.

What I think we should really write out is

freq = freqmin + (i + 0.5) * dfreq;

Not a big deal in general, but it means our spectra are all shifted by half a frequency bin at the moment.