doFft function

void doFft(List<int> fft, List<int> waveform)

Performs FFT (Fast Fourier Transform) calculation on the given waveform data.

This method accepts a list of 8-bit unsigned integers representing the waveform data and outputs the FFT result into the provided list of 8-bit unsigned integers.

The input waveform list should contain pairs of bytes representing the waveform data. The output fft list will contain the FFT result, with each pair of bytes representing the real and imaginary components of the FFT result.

The method processes the input waveform data into a workspace, performs the FFT calculation, and then processes the workspace data into the output FFT buffer.

param fft The list of 8-bit unsigned integers to store the FFT result. param waveform The list of 8-bit unsigned integers representing the waveform data.

Implementation

void doFft(List<int> fft, List<int> waveform) {
  final int captureSize = waveform.length;
  // Create workspace with half the capture size
  final Int32List workspace = Int32List(captureSize >> 1);
  int nonzero = 0;

  // Process input waveform data into workspace
  for (int i = 0; i < captureSize; i += 2) {
    // Convert pairs of bytes to 32-bit integers
    // Note: Using bitwise operations ensures 32-bit arithmetic
    workspace[i >> 1] =
        (((waveform[i] ^ 0x80) << 24) | ((waveform[i + 1] ^ 0x80) << 8));
    nonzero |= workspace[i >> 1];
  }

  // Only perform FFT if we have non-zero data
  if (nonzero != 0) {
    _fftReal(captureSize >> 1, workspace);
  }

  // Process workspace data into output FFT buffer
  for (int i = 0; i < captureSize; i += 2) {
    // Process real component
    int tmp = int16(workspace[i >> 1] >> 21);
    // Clamp to 8-bit range
    while (tmp > 127 || tmp < -128) {
      tmp >>= 1;
    }
    fft[i] = uint8(tmp);

    // Process imaginary component
    tmp = int16(workspace[i >> 1]);
    tmp >>= 5;
    // Clamp to 8-bit range
    while (tmp > 127 || tmp < -128) {
      tmp >>= 1;
    }
    fft[i + 1] = uint8(tmp);
  }
}