You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
line 194: vDSP_ctoz((COMPLEX*) _A, 2, &(_compl_buf), 1, self.spectrumResolution);
The vDSP examples by apple seem to suggest that "self.spectrumResolution" should either be divided by 2, or the FFT should increase its log2 count by 1 (depending on the actual size of the buffer obv).
This also affects subsequent calls like vDSP_zaspec.
The documentation mentions that vDSP_DFT_Execute is preferred over the FFT variants, whenever possible.
The text was updated successfully, but these errors were encountered:
type "vdsp" in the search field and you should end up with the latest and greatest:
The referencedoc entry for the fft routines will mention the DFT alternative, the vDSP example code has a lot of commenting that should clarify its use.
Note above the N/2 and LOG2N parameters.
For my own purposes I seem to get the best results using the DCT in lieu of either the DFT, or FFT.
line 194: vDSP_ctoz((COMPLEX*) _A, 2, &(_compl_buf), 1, self.spectrumResolution);
The vDSP examples by apple seem to suggest that "self.spectrumResolution" should either be divided by 2, or the FFT should increase its log2 count by 1 (depending on the actual size of the buffer obv).
This also affects subsequent calls like vDSP_zaspec.
The documentation mentions that vDSP_DFT_Execute is preferred over the FFT variants, whenever possible.
The text was updated successfully, but these errors were encountered: