AAK Schmidt-pair diagnostics for SISO Hankel approximation
==========================================================

.. admonition:: Tutorial goal

   Visualize the first neglected Hankel singular direction that drives the AAK/Nehari approximation barrier.

.. note::

   New to the terminology? See the :doc:`lattice DSP concept map <../../algorithms/concept_map>` and the :doc:`causality/data-use guide <../../theory/causality_and_data_use>` for how online, offline, block, and MIMO examples should be read.

Context
-------

The finite Nehari toy and rational-bridge tutorials show the singular values and a practical
recurrence fit.  This tutorial focuses on the next AAK object: the finite Schmidt pair at
the first neglected singular value.  In the scalar AAK theory, analogous singular-vector
data carries the structure used to build an optimal rational approximant.

This page is still finite-dimensional and diagnostic.  It does not expose a production
``aak_reduce_iir`` routine.  Instead, it shows what such a routine would need to control:
the critical singular direction, the finite rank-r barrier, the Hankelized approximation,
and the resulting rational poles.

Key idea and equations
----------------------

For a finite Hankel matrix ``H`` with singular value decomposition

.. math::

   H = U\Sigma V^T,

the first neglected mode after a rank-``r`` approximation is the finite Schmidt pair
``(u_{r+1}, v_{r+1})`` satisfying

.. math::

   H v_{r+1}=\sigma_{r+1}u_{r+1},\qquad
   H^T u_{r+1}=\sigma_{r+1}v_{r+1}.

The scalar AAK/Nehari theory can be viewed as the infinite-dimensional rational analogue of
this finite singular-direction picture.

How to read the result
----------------------

Look for the highlighted first neglected singular value, small Schmidt-pair residuals, and how the rational fits improve as the rank crosses the critical modes.

Run command
-----------

.. code-block:: bash

   python examples/aak_siso_schmidt_pair_demo.py

Source code
-----------

.. literalinclude:: ../../../examples/aak_siso_schmidt_pair_demo.py
   :language: python
   :linenos: