This rebin function resamples a 1D or 2D histogram to new bins.

In the 1D case, we have an array x1 of bin edges (m+1 entries), and counts in each one are recorded in array y1 (m entries). Instead of keeping the data in the x1 bins, we have another set of bins that we want the data sorted into. This new set of bins is represented by x2 (with n+1 entries). The rebin function redistributes the counts in y1 into a new array y2 (n entries).

To do this rebinning, some assumption about the distribution of the counts within each channel is necessary. This script offers the choice between a uniform distribution or a spline fit with specified order.

The function works with array-like objects as determined by Numpy.

Uncertainties in y1 can be propagated through rebin if y1 is a uarray from the Python uncertainties module.

Knoll[1] describes this in Chapter 18.IV.B titled "Spectrum Alignment." He calls this process rebinning, relocating, or spectrum alignment.

[1] Glenn Knoll, Radiation Detection and Measurement, third edition,

Wiley, 2000.