Commentary on the denominator recursion for Burg's block algorithm

An order-recursive form for the denominator in the equation for the reflection coefficients in Burg's block algorithm was originally presented by N. Andersen in 1978, although its derivation was not given. This recursion was later repeated in a tutorial paper by Kay and Marple, also without derivation, but with data and error indexing defined differently. This letter shows a derivation of the denominator recursion using the notation defined in the latter work and demonstrates that the recursion given there is incorrect. A computer program written in the C language is also presented which implements Burg's algorithm, and includes references to specific equations from the latter work.