%% LMS Homework Assignment
%
% Introduction
%
% This homework assignment requires you to experiment with a simple LMS
% filter architecture. Slide 15 of the most recent lecture
% (http://www.echelonembedded.com/dsphwlab/notes/class11.pdf) gives the LMS
% architecture for Adaptive Noise Cancellation.
%
% For this homework assignment, implement an LMS filter on your DSP board.
% The easiest way to do this is likely to start with one of the Chassaing
% examples. You should feel free to use as much pre-written code from the
% book (or any other source except another student). The goal is simply to
% experiment with an LMS filter; not to spend much time implementing code.
%
% To experiment, inject the "signal" through one side of the stereo audio
% input, and the "noise reference" into the other side or the stereo input
% (so, for example, the right channel could be dedicated to the signal, and
% the left channel to the noise reference).
%
% The adaptively filtered signal "e" should be played real-time out the
% audio output.
%
% Demonstration
%
% Three .mat files are available on the class website (and are referenced
% by this .m file). Each includes the following signals:
%   * input: The input signal corrupted by a noise source. * noiseRef: The
%   reference noise. This will be adaptively filtered, and removed from the
%   input to produce the error signal, e. * e: The input - noiseRef; i.e.,
%   the "filtered" signal. This is the result we are observing on the codec
%   output. * sig: The vector used to form the signal portion of input. *
%   noiseInFilt: A filtered version of the noise added to the sig. This is
%   the corrupting noise source. For the lawnmower example, this input has
%   undergone two transformations (filtering and summation with a delayed
%   copy). * fs: The sample rate of the vectors.
%
% For your demonstration, you simply need to inject each of the three input
% and noiseRef signals and play the output signal. Note that the error
% signal examples provided were formed at different beta values.
%
% In addition to demonstrating the operation of your LMS filter, you will
% need to be prepared to intelligently discuss your architecture and the
% operation of the filter.

%% Sin Example
load sin.mat
% creates fs, sig, noiseInFilt, input, noiseRef, e
soundsc(sig(7*fs:fs*10),fs);
soundsc(input(7*fs:fs*10),fs);
soundsc(e(7*fs:fs*10),fs);f

%% Lawnmower Example
% noiseInFilt = noiseInFilt1 + [zeros(1, nDelay), .1 * noiseInFilt2(nDelay+1:end)];
load lawnmower.mat
soundsc(sig(1:fs*10),fs);
soundsc(input(1:fs*10),fs);
soundsc(e(1:fs*10),fs);

%% White Noise Example
% noiseInFilt = noiseInFilt1 + [zeros(1, nDelay), .1 * noiseInFilt2(nDelay+1:end)];
load noise.mat
soundsc(sig(1:fs*10),fs);
soundsc(input(1:fs*10),fs);
soundsc(e(1:fs*10),fs);