[R] spec.pgram() possible bug

[Prof Brian Ripley]

Perhaps you could take the trouble to read the references (or the code)?
Hint: do you know about tapering and how to work out its effect?

On Mon, 26 Jul 2004, David Stoffer wrote:

> > test=spec.pgram(rnorm(128,0,1), kernel("daniell",1))
> > test$df
> [1] 5.374771
> 
> df should be 6= 2*(2*m+1) where m=1 in this case
> 
> ... and so on

> spec.pgram(rnorm(128,0,1), kernel("daniell",1), taper=0)$df
[1] 6

is what you have calculated, inappropriately for the call you 
actually used.

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595