Hi need help for overriding system objects in MATLAB

Hi this is my topic in details The thesis focuses on designing and analyzing polynomial sampling rate converters using interpolation techniques. While MATLAB currently supports Lagrange interpolation, the research aims to explore the use of spline interpolation for sampling rate conversion. A MATLAB class will be developed for designing converters based on spline interpolation, and the efficiency of these converters will also be analyzed.

The dsp.FarrowRateConverter class in MATLAB allows for the creation of a polynomial sampling rate converter with user-defined properties, including direct specification of polynomial coefficients, such as for Lagrange or spline interpolation. However, the calculation of spline coefficients must be done externally, as the class does not handle this. The class offers built-in methods for analyzing and comparing converter objects.

The dsp.FarrowRateConverter class can be extended to include a method for calculating spline coefficients within the class.

I have overridden dsp.FarrowRateConverter

classdef SplineFarrowRateConverter < dsp.FarrowRateConverter

% SplineFarrowRateConverter Spline-based sample rate converter with arbitrary conversion factor.

% This class extends dsp.FarrowRateConverter to use Spline interpolation

% instead of Lagrange interpolation.

methods

function obj = SplineFarrowRateConverter(varargin)

% Constructor: Inherit properties from the base class

[e-mail](varargin{:});

% Check if PolynomialOrder is specified and validate it

if strcmp(obj.Specification, 'Polynomial order')

validateattributes(obj.PolynomialOrder, ...

{'numeric'}, ...

{'real', 'integer', 'positive', 'scalar', 'finite'}, ...

'SplineFarrowRateConverter', 'PolynomialOrder');

end

end

function setPolynomialOrder(obj, N)

% Method to set the PolynomialOrder with validation

validateattributes(N, ...

{'numeric'}, ...

{'real', 'integer', 'positive', 'scalar', 'finite'}, ...

'setPolynomialOrder', 'PolynomialOrder');

obj.PolynomialOrder = N;

needToDesignFilter(obj); % Reset filter design to accommodate new order

end

end

methods (Access = protected)

function setFarrowSRCProperties(obj)

% Override the base method to calculate Spline coefficients

if obj.CoefficientsExist

return;

end

isPolynomialOrderSpec = strcmp(obj.Specification, 'Polynomial order');

if isPolynomialOrderSpec

NpOrCf = obj.PolynomialOrder;

else

NpOrCf = obj.Coefficients;

end

coder.extrinsic('dsp.FarrowRateConverter.deriveCoefficients');

if isempty(coder.target)

% Calculate Spline coefficients instead of Lagrange

[L, M, C] = obj.deriveSplineCoefficients(obj.InputSampleRate, ...

obj.OutputSampleRate, obj.OutputRateTolerance, isPolynomialOrderSpec, NpOrCf);

obj.CoefficientsExist = true;

% Update the metadata (fdesign and fmethod)

if isPolynomialOrderSpec

fd = fdesign.polysrc(L, M, 'Fractional Delay', 'Np', NpOrCf, obj.InputSampleRate * L);

if ~isempty(getcurrentspecs(fd))

fmd = fdfmethod.lagrangesrc; % Could be replaced with the equivalent for Spline

else

fmd = [];

end

setMetaData(obj, fd, fmd);

end

else

[L, M, C] = coder.const(@obj.deriveSplineCoefficients, obj.InputSampleRate, ...

obj.OutputSampleRate, obj.OutputRateTolerance, isPolynomialOrderSpec, NpOrCf);

end

obj.InterpolationFactor = L;

obj.DecimationFactor = M;

obj.PolynomialCoefficients = C;

end

function [L, M, C] = deriveSplineCoefficients(obj, FsIn, FsOut, FsTol, isPolynomialOrderSpec, NpOrCf)

% Derive coefficients specific to Spline interpolation

[L, M] = dsp.SampleRateConverter.myrat(FsIn, FsOut, FsTol);

if isPolynomialOrderSpec

% Generate Spline coefficients based on the specified order

C = obj.designSplineCoefficients(NpOrCf);

else

C = (NpOrCf).';

end

end

function C = designSplineCoefficients(~, N)

% Calculate the coefficients for Spline interpolation using B-Spline coefficients

L = N; % Degree of the B-Spline is typically set to N (the polynomial order)

C = computeBSplineCoefficients(L, N); % Call the new function to compute B-Spline coefficients

end

end

end

function a = computeBSplineCoefficients(L, N)

% L: Degree of the B-Spline

% N: Number of control points - 1

% a: Vector containing the B-Spline coefficients

% Initialize the coefficients vector

a = zeros(1, N + 1);

% Define the knot vector t_i = i for i = 0 to N

t = 0:N;

% Loop over each coefficient to calculate a_i

for i = 0:N

prod = 1; % Initialize the product

for k = 0:N

if k ~= i

prod = prod * (t(i+1) - t(k+1)); % MATLAB uses 1-based indexing

end

end

% Calculate the coefficient a_i

a(i+1) = (-1)^(L+1) / prod; % a(i+1) because MATLAB is 1-based

end

end

This is generating error >> SplineFRC3

The class dsp.SampleRateConverter has no Constant property or Static method named 'myrat'.

Error in SplineFarrowRateConverter/deriveSplineCoefficients (line 74)

[L, M] = dsp.SampleRateConverter.myrat(FsIn, FsOut, FsTol);

^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Error in SplineFarrowRateConverter/setFarrowSRCProperties (line 48)

[L, M, C] = obj.deriveSplineCoefficients(obj.InputSampleRate, ...

^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Error in dsp.FarrowRateConverter/validateInputsImpl (line 608)

setFarrowSRCProperties(obj);

^^^^^^^^^^^^^^^^^^^^^^^^^^^

Error in SplineFRC3 (line 12)

response = src(impulse); % Use the object directly to invoke the processing

^^^^^^^^^^^^

This error is occuring when I try to observe its frequency response % Create an instance of your SplineFarrowRateConverter

src = SplineFarrowRateConverter('InputSampleRate', 44100, ...

'OutputSampleRate', 48000, ...

'Specification', 'Polynomial order', ...

'PolynomialOrder', 3); % Adjust the order here

% Generate an impulse signal to analyze the filter response

impulse = [1; zeros(1023, 1)]; % Create an impulse signal

% Process the impulse through the rate converter

response = src(impulse); % Use the object directly to invoke the processing

% Create a digital filter object from the impulse response

% Here, we use dsp.FIRFilter as a representative filter type for visualization

firFilter = dsp.FIRFilter('Numerator', response');

% Visualize the filter's frequency response using fvtool

fvtool(firFilter); Please could you correct the error and also I want to access the computation complexity of this code ,the number of adders and multipliers and also value of coeffiecients and frequency response for polynomial orders more than 3

[Rohitashya]

The functions of Bspline are given in the function

% Function to perform spline interpolation

function y_interp = spline_interp(t, L, knots, x_samples)

m = 1;

N = length(x_samples);

y_interp = 0;

for k = m:m+L

beta = normalized_b_spline(t, L, knots, k);

y_interp = y_interp + beta * x_samples(k);

end

end

% File: spline_interpolation.m

% Function to compute the truncated power function phi_i(t)

function phi = truncated_power(t, ti, L)

if t < ti

phi = 0;

else

phi = (t - ti)^L;

end

end

% Function to compute the normalized L-th order B-spline beta_L_m(t)

function beta = normalized_b_spline(t, L, knots, m)

B_L_m = b_spline(t, L, knots, m);

beta = (knots(end) - knots(1)) * B_L_m;

end

% Function to compute the L-th order B-spline B_L_m(t)

function B = b_spline(t, L, knots, m)

n = length(knots);

if L == 0

if knots(m) <= t && t < knots(m+1)

B = 1;

else

B = 0;

end

else

left_term = (t - knots(m)) / (knots(m+L) - knots(m)) * b_spline(t, L-1, knots, m);

right_term = (knots(m+L+1) - t) / (knots(m+L+1) - knots(m+1)) * b_spline(t, L-1, knots, m+1);

B = left_term + right_term;

end

end

Answer 1

[Divyanshu]

Hello @Rohitashya,

The error 'no constant property or static method named 'func' ' is generally thrown when we try to access some method using 'Dot Indexing' on a class object but the method is not a member of the class.

In your case as well you are using system object 'dsp.SampleRateConverter' but in documentation dsp.SampleRateConverter there is no method 'myRat' for this object.

Therefore you might need to override the class 'dsp.SampleRateConverter' as well and define method 'myRat'.

Additionally, for your second query regarding calculating computation complexity of the code, you can refer the following MATLAB Answer threads:

• https://www.mathworks.com/matlabcentral/answers/66438-how-to-properly-measure-computational-complexity-run-time-in-matlab
• https://www.mathworks.com/matlabcentral/answers/108121-how-to-calculate-the-computational-complexity

Hope it helps!