Re: Not able to get the matlab equivalent output from tfestimate function (transfer function estimate)

[Mike Miller]

On Thu, Feb 14, 2013 at 10:58 PM, Sarbjit singh wrote:
> I could see difference in data itself when tfestimate is calculated from
> Octave using the same inputs (x and y)
>
> octave-3.6.2.exe:17> txy_matlab = txy;
> octave-3.6.2.exe:18> W_matlab = W;
> octave-3.6.2.exe:19> [txy W]=tfestimate(x,y,10,[],[]);
> octave-3.6.2.exe:20> size(txy)
> ans =
>
>    6   1

Well you already mentioned that you noticed the arguments are in a
different order. 10 here means FFT size to Octave, what you want is

octave:19> [txy, W] = tfestimate (x, y, [], [], 10);
octave:20> size (txy)
ans =

   129     1

octave:21> size (W)
ans =

   129     1


> octave-3.6.2.exe:26> max(W)
> ans =  0.50000
> octave-3.6.2.exe:27> max(W_matlab)
> ans =  3.1416

Yes, just another small incompatibility, so simply multiply Octave's W
by 2*pi to get the equivalent x-axis scale until we can fix this.
Regardless, it doesn't affect the value of txy.

-- 
mike