scribeGriff / ConvoLab

ConvoLab: A signal processing library for Dart

Convolab

A signal processing library for Dart

Note: This library is not currently being maintained.

Overview

The ability to create 2-dimensional, canvas based plots of signal waveforms is available through the simplot library.

The graph-centric algorithms (ie, scc, 2-sat, Prim's MST, Kruskal's MST, all pairs shortest path and the knapsack algorithm) have been moved to the graphlab library.

The library is currently moving its 1-dimensional data structure from List to Sequence, which extends the ListBase class and adds a number of methods and capabilities.

Some features of the library include:

• Flexible data structures:
  Sequences: generate and easily manipulate impulse, step, position or arbitrary sequences
  Complex numbers: create complex numbers and handle complex calculations
• Hyperbolic and logarithmic functions: sinh(), cosh(), log2(), log10() etc.
• Simple waveform generator: square, pulse, triangular, etc.
• Polynomial string construction in text, HTML or latex format
• Common signal processing algorithms including (more are in development):
  fft(): fast and discrete Fourier transform
  ifft(): fast and discrete inverse Fourier transform
  fsps(): partial sums of Fourier series
  conv(): convolution
  deconv(): deconvolution
  corr(): cross and auto correlation
  filter(): a 1D transposed direct form II digital filter structure

Most library access is through top level function calls.

Library Usage:

Add the following to your pubspec.yaml:

convolab:
  git: git://github.com/scribeGriff/ConvoLab.git

Then import the library to your app:

import 'package:convolab/convolab.dart';

Sequences

The library is currently moving to using Sequences as the main data structure. A sequence extends from the ListBase class and therefore shares many of the same methods of the List data type. Some examples of using a Sequence data structure:

// Creates a List object.
var list = [1, 2, 3, 4];
// Converts the List into a Sequence.
var seq1 = sequence(list);
// Check if seq1 is a Sequence.
print(seq1 is Sequence);
// Print the sequence.
print(seq1);
// Prints:
// true
// [1, 2, 3, 4]

// Any Iterable, including another sequence, can be passed to the
// sequence function.
var seq2 = sequence(list.map((element) => element * 2));

// Sequences can be added, subtracted, multiplied, and divided element
// by element:
var seq3 = seq1 + seq2;
print(seq3);
// Prints:
// [3, 6, 9, 12]

// and element by number:
var seq4 = seq3 * 2;
print(seq4);
// Prints: 
// [6, 12, 18, 24]

// Sequences can be made periodic:
print(sequence(seq1, 4));
// Prints:
// [1, 2, 3, 4, 1, 2, 3, 4, 1, 2, 3, 4, 1, 2, 3, 4]

Sequences allow us to perform a variety of tasks on sampled data. For example, suppose we had a sample sequence, x(n), and we wanted to know what x1(n) = 2x(n - 5) - 3x(n + 4) was.

// x(n)
Sequence x = sequence([1, 2, 3, 4, 5, 6, 7, 6, 5, 4, 3, 2, 1]);
// n, zeroth position at element 2.
Sequence n = x.position(2);
// n - 5
Sequence nm5 = shiftseq(n, 5);
// n + 4
Sequence np4 = shiftseq(n, -4);
// x1(n) = 2x(n - 5) - 3x(n + 4)
var x1 = addseqs(x * 2, x * -3, nm5, np4);
print(x1.x);
print(x1.n);
// Prints:
// [-3, -6, -9, -12, -15, -18, -21, -18, -15, -10, -5, 0, 5, 10, 12, 14, 12, 10, 8, 6, 4, 2]
// [-6, -5, -4, -3, -2, -1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15]

The function, addseqs(), allows us to add sequences that have position information. In this example, the n = 0 sample point was the 3rd element in the sequence x(n). The function shiftseq() (as well as foldseq()), are also methods of the Sequence class. Therefore, we could do the following:

var x2 = position(11, 5);
print(x2);
print(x2.shiftseq(2).foldseq(negate:true));
// Prints:
// [-5, -4, -3, -2, -1, 0, 1, 2, 3, 4, 5]
// [-7, -6, -5, -4, -3, -2, -1, 0, 1, 2, 3] 

The position() function creates a sample position sequence 11 samples long and places the 0 position at the 6th position in the sequence (sequences, like all Iterables in the Dart programming language, are based on the first element being element 0).

To convert a sequence to a list, simply use the toList() method of the ListBase class.

Sequence methods (other than those inherited from ListBase)

• + operator - adds a sequence to another sequence of the same length or to a number
• - operator - subtracts a sequence from another sequence of the same length or from a number
• * operator - multiplies a sequence by another sequence of the same length or by a number
• / operator - divides a sequence by another sequence of the same length or by a number
• position() - creates a position sequence
• shiftseq() - shifts a sequence
• foldseq() - folds(ie, reverses) a sequence
• toReal() - converts a complex sequence to a real valued sequence
• toComplex() - converts a real sequence to a complex valued sequence
• equals() - returns a boolean indicating element by element equivalence between two sequences
• abs() - returns a sequence with the absolute value of each element
• min() - returns the minimum of a sequence
• max() - returns the maximum of a sequence
• sum() - returns the sum of a sequence
• prod() - returns the product of a sequence
• energy() - returns the energy of a sequence
• power() - returns the power of a sequence
• iterator - returns an iterator to the sequence (allows for in and forEach)
• middle - returns the index of the middle element of the sequence (truncates)

Sequence functions

• sequence() - create a new sequence from an Iterable
• position() - creates a position sequence
• zeros() - creates a sequence of zeros
• ones() - creates a sequence of ones
• impseq() - creates an impulse sequence
• stepseq() - creates a step sequence
• shiftseq() - shifts a sequence
• foldseq() - folds (ie, reverses) a sequence
• addseqs() - adds sequences that are of differing lengths and/or have position information
• multseqs() - multiplies sequences that are of differing lengths and/or have position information
• evenodd() - decompose a sequence into its even and odd components.

---

Complex Numbers

The library contains support for complex numbers. A complex number can be defined as follows:

var c = complex(1, 2);
print(c.string);
// Prints:
// 1.00 + 2.00j

The complex class implements the following methods:

• string - returns the complex number as a string
• magnitude - returns the magnitude of a complex number
• phase - returns the phase of a complex number
• croud2 - rounds fractional part of complex number to two significant digits
• conj - returns the complex conjugate of the complex number
• recip - returns the reciprocal of the complex number
• cexp - returns the complex exponential of the complex number
• csin - returns the complex sine of the complex number
• ccos - returns the complex cosine of the complex number
• ctan - returns the complex tangent of the complex number
• scale() - scales the complex number by a number of type int or double
• + operator adds a complex number to another complex number
• - operator subtracts a complex number from another complex number
• * operator multiplies a complex number by another complex number
• / operator divides a complex number by another complex number
• == operator checks if two complex numbers are equal (hashCode is implemented)

---

Poly Strings

To convert a sequence to a readable polynomial string, the library contains a function called pstring(). The function can generate strings in text, HTML, or Latex format. In its simplest form, pstring() requires only a sequence of polynomial coefficients:

// Simple case - defaults
Sequence coefficients = sequence([2, 5, -3, 7]);
String polyString = pstring(coefficients);
print(polyString);
// prints:
// $$f(x) = 2 + 5x^{-1} - 3x^{-2} + 7x^{-3}$$

The Latex format, for use with MathJax in a browser, is the default. The pstring() function also accepts several named optional parameters:

• index: for causal signals, this is the n = 0 position index (default = 0)
• type: String representing desired format text, html, latex (default = 'latex')
• variable: String representing the variable name (default = 'x')
• name: String representing the function name (default = 'f')

Some additional examples:

// Setting options - html format.
polyString = pstring(coefficients, type: 'html', name: 'y', variable: 'n');
print(polyString);
// prints:
// y(n) = 2 + 5n<sup>-1</sup> - 3n<sup>-2</sup> + 7n<sup>-3</sup>

// Send the output to element on a webpage
query("#myDiv").appendHtml(pstring(test, type:'html'));

// Working with causal signals.
zeroIndex = 2;
polyString = pstring(coefficients, index: zeroIndex, type: 'html', name: 'y', variable: 'n');
print(polyString);
// prints:
// y(n) = 2n<sup>2</sup> + 5n - 3 + 7n<sup>-1</sup>

// Another example working with causal signals.
coefficients = sequence([1, 1, 1, 1, 1, 1]);
polyString = pstring(coefficients, index: zeroIndex, type: 'text');
print(polyString);
// prints:
// f(x) = x^2 + x + 1 + x^-1 + x^-2 + x^-3

The library functions conv() and deconv() both implement the PolyString class as a format() method to the function results. In the case of deconv(), the format() method handles remainders.

---