The scale of "broadened intensities" should be independent of calculator (`SpinWaveTheory` or `dynamical_correlations`)

[kbarros]

The scale of Sunny-calculated dynamical structure factor intensities is should be consistent across calculation methods, following this precise specification.

In Sunny 0.5.11, the result of spin wave calculations (using SpinWaveTheory and intensities_broadened) followed this convention, whereas classical dynamics calculations (using dynamical_sampler and intensities_interpolated) did not.

Concretely, SpinWaveTheory reports intensities as $S(q,\omega)$, while classical dynamics reports intensities as $S(q,\omega) \Delta\omega$. In other words, only the SWT calculation is truly a "density" in energy space.

In Sunny 0.6, this bug is fixed. All calculation methods properly report the structure factor as a density in energy space.

Additional notes:

• SpinW uses the $S(q,\omega) \Delta \omega$ convention for its broadened intensities, so to obtain quantitative agreement with Sunny 0.6, one must divide the SpinW result by a $\Delta \omega$ factor.
• The results of "binned" intensities are already integrated, and this issue is not relevant to that case.

[kbarros]

Fixed in #272.