Different hydrogen bonding networks, same principle: hydrogen bonds are the most common fundamental structural driving forces, which determine structural and dynamical properties of numerous functional materials. First-principles calculations of spectroscopic parameters can help to understand local geometric motifs, but also more complex processes such as hydrogen bond lifetimes and ion transport processes in condensed phases. In this feature article, we review the relevance of structure-spectroscopy-relationships, we discuss recent ab initio calculations eludicating the structure of supramolecular assemblies, and highlight the importance of incorporating atomic and molecular mobility by means of molecular dynamics (MD) simulations.Complex hydrogen bonding networks: vinyl-phosphonic acid polymers (left) and aqueous hydrochloric acid (right).1 Sensitivity of spectroscopic signatures to structure Any simulation of a molecular system is based on the computed potential energy surface (PES). This surface, however, is not directly experimentally accessible. One of the first actual observables in a numerical calculation is the atomic structure, be it at equilibrium or as an ensemble average. As a consequence, approximations and errors in the calculation of the PES propagate immediately to structural parameters. Even a highly corrugated landscape is full of local energy minima, at which the atomic forces vanish; hence, the PES is locally quadratic in all coordinates at those points. This in turn results in a strong effect of small perturbative forces on computed geometries near equilibrium.It turns out that many spectroscopic observables exhibit a very different behavior close to the equilibrium geometry: small structural variations can lead to considerable changes in (experimental as well as computed) spectra. As an example from nuclear magnetic resonance spectroscopy (NMR), the 1 H NMR chemical shift calculated for the H-bonding proton in an imidazole dimer is shown in Fig. 1 (top). The slope of d H at equilibrium (indicated by the dashed line) is about 8 ppm/Å ; assuming a conservative estimate for the theoretical/experimental resolution of 0.1 ppm, this corresponds to a spectroscopic distance resolution of almost 0.01 Å [1]. In comparison to this, the energetic resolution of a standard ab initio calculation (assuming a systematic error of 1 kcal/mol and a typical H-bond strength) is roughly 0.1 Å . It is not feasible to reliably compute the energy differences due to small geometric changes more accurately than this, except by means of very resource-intensive quantumchemical reference methods. Hence, the variation of the NMR chemical shift considerably magnifies alterations of Phys. Status Solidi B 249, No. 2, 368-375 (2012)