p-MLFD

[rmunderw]

I think there are multiple errors in the paper "a poly reference implementation of the maximum likelihood complex frequency domain estimator and some industrial applicatinos.pdf"

I have programmed up the LSCF, MLFD, and p-LSCF methods and am having trouble with the final piece p-MLFD. The dimensions of the matrices in the early part of the paper don't seem to be right and I think the definition of the X and Y matrices are swapped. The definitions of the R, S, T, V, & W matrices are mostly correct. I could be messing up the un-vec operator of delta vec(D) or delta vec(No). I believe the un-vec of a column array should be a 2 row or column matrix with one row/column being every other value in the vec array.

Has anyone tried to code this up?

Thanks,

 

[IRstuff]

What program are you using? Most definitions of unvec unstacks a column matrix into, often, a square matrix

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[rmunderw]

I've written all of these in Python. The exact implementation of the un-vec is:

alpha[:,0:1] = alpha[:,0:1] + delta_a[::2]
alpha[:,1:2] = alpha[:,1:2] + delta_a[1::2]

the colon's inside the array dimensions are start:stop:step. The delta_a array is the vec'd array, so the length is twice as long as the first dimension of alpha.