FFTInterface Manual

Table of Contents

• About
• Classes
• FFTInterface
• Further Reading

About

The fast Fourier transform (FFT) is used to analyze a given signal in order to transform it into the frequency domain. The maxon::FFTInterface gives access to implementations of various algorithms.

Classes

Various implementations of the FFT are registered at maxon::FFTClasses:

• maxon::FFTClasses::Generic: Generic Cinema 4D implementation.
• maxon::FFTClasses::Kiss: Fast FFT algorithm using the KISS library.
• maxon::FFTClasses::Cooley: Cooley Tukey algorithm.

FFTInterface

The maxon::FFTInterface provides the functions to perform the FFT:

• maxon::FFTInterface::Transform1D(): Computes a one-dimensional FFT transformation.
• maxon::FFTInterface::Transform2D(): Computes a two-dimensional FFT transformation.
• maxon::FFTInterface::GetSupportOptions(): Returns the supported features of the implementation. See maxon::FFT_SUPPORT.

  // This example shows how to use FFT to get the spectrum of a signal.
 
  // synthesize a signal of one second
 
  // signal parameters
  const maxon::Int   sampleCount = 500;
  const maxon::Float sampleTime  = 1.0; // seconds
  const maxon::Float samplingFrequency = maxon::Float(sampleCount) / sampleTime;
 
  const maxon::Float step = (2 * maxon::PI) / maxon::Float(sampleCount);
 
  // create signal
  maxon::BaseArray<maxon::Float> signal;
  signal.EnsureCapacity(sampleCount) iferr_return;
 
  for (maxon::Int i = 0; i < sampleCount; ++i)
  {
    const maxon::Float pos = maxon::Float(i) * step;
    maxon::Float       sample = 0.0;
 
    // 1Hz component
    sample += maxon::Sin(pos);
    // 2Hz component
    sample += maxon::Sin(pos * 2);
    // 4Hz component
    sample += maxon::Sin(pos * 4) * 0.5;
    // 8Hz component
    sample += maxon::Sin(pos * 8) * 0.5;
 
    signal.Append(sample) iferr_return;
  }
 
  // perform FFT
 
  // create FFTRef
  const maxon::FFTRef genericFFT = maxon::FFTClasses::Generic().Create() iferr_return;
 
  // check if this FFT implementation supports Transform1D()
  const maxon::FFT_SUPPORT options = genericFFT.GetSupportOptions();
  if (!(options & maxon::FFT_SUPPORT::TRANSFORM_1D))
    return maxon::UnexpectedError(MAXON_SOURCE_LOCATION);
 
  // transform signal
  maxon::BaseArray<maxon::Complex64> complexDFT;
  genericFFT.Transform1D(signal.ToArray(), complexDFT.ToArray()) iferr_return;
 
  // convert to amplitude spectrum
 
  maxon::BaseArray<maxon::Float> spectrum;
  spectrum.EnsureCapacity(complexDFT.GetCount()) iferr_return;
 
  for (const auto DFT : complexDFT)
  {
    const maxon::Float amplitude = DFT.GetLength();
    spectrum.Append(amplitude) iferr_return;
  }
 
  // display spectrum
 
  // spectrum resolution
  const maxon::Float resolution = samplingFrequency / sampleCount;
 
  const maxon::Float spectrumRange = maxon::Float(spectrum.GetCount());
  const maxon::Int   maxFrequencyIndex = maxon::Int(spectrumRange * 0.5);
 
  for (maxon::Int i = 0; i < maxFrequencyIndex; ++i)
  {
    const maxon::Float frequency = (maxon::Float)i * resolution;
    const maxon::Float amplitude = spectrum[i];
    const maxon::Float normalizedAmplitude = (amplitude / maxon::Float(sampleCount));
    DiagnosticOutput("@ Hz: @", frequency, normalizedAmplitude);
  }

Further Reading

• Algorithms
• Arrays Manual
• Published Objects Usage