This is a very personal point of view of the underlying ideas that yields most of the currently reported spectral estimation techniques, procedures and algorithms. After a brief introduction on non-parametric spectral estimation, which includes a shutle aspect about how to use averaging in DFT based methods, the paper describes the potential of the variational approach in deriving already reported estimates and the way out to obtain new ones. Finally, and as the second approach of high interest in spectral estimation, the design and extensions of data-dependent filters for spectral estimation is reported.

Peer Reviewed