← Algorithms · Size distribution D(R)

Size distribution D(R)

Purpose. For polydisperse populations (especially near-spherical scatterers), a volume size distribution D(R) is more informative than a single p(r). Peaks and widths hint at how many size classes are present and how broad they are — input for parametric mixture fits.

Engine. ATSAS GNOM with system = spheres (volume distribution). Autosaxs again supplies parameter search and quality assessment rather than a new inverse transform.

What autosaxs adds (custom orchestration).

  1. If Rmax is not fixed, searches a plausible Rmax range (time-capped), analogous to the Rg search for monodisperse GNOM.
  2. Runs stability / ensemble checks (including force-zero-style probes) to see whether the chosen maximum size is pathological.
  3. Reports modality and simple parametric PDF hints that help initialize mixture models.
  4. Applies a polydisperse-oriented quality passport (Total Estimate, polydispersity indices, Shannon diagnostics).

In liveview, polydisperse Guinier is shown independently and is not fed into fit_sizes as a hard constraint — polydisperse Guinier quality is often poor by nature.

← p(r) · Mixture → · Polydisperse UI →