void fftNormalized(const QVector<int16_t> audioFrame, const uint channel, const uint numChannels, float *freqSpectrum,
const WindowType windowType, const uint windowSize, const float param = 0);
+
+ /** This is linear interpolation with the special property that it preserves peaks, which is required
+ for e.g. showing correct Decibel values (where the peak values are of interest).
+ Consider f = {0, 100, 0}
+ x = {0.5, 1.5}: With default linear interpolation x0 and x1 would both be mapped to 50.
+ This function maps x1 (the first position after the peak) to 100.
+
+ @param in The source vector containing the data
+ @param targetSize Number of interpolation nodes between ...
+ @param left the left array index in the in-vector and ...
+ @param right the right array index (both inclusive).
+ @param fill If right lies outside of the array bounds (which is perfectly fine here) then this value
+ will be used for filling the missing information.
+ */
+ static const QVector<float> interpolatePeakPreserving(const QVector<float> in, const uint targetSize, uint left = 0, uint right = 0, float fill = 0.0);
+
private:
QHash<QString, kiss_fftr_cfg> m_fftCfgs; // FFT cfg cache
QHash<QString, QVector<float> > m_windowFunctions; // Window function cache