Visible to Intel only — GUID: GUID-6864C4A1-1A43-42CC-8AF6-E86D2561729E
Visible to Intel only — GUID: GUID-6864C4A1-1A43-42CC-8AF6-E86D2561729E
Special Arguments
This section describes the flag and hint arguments used by the Fourier transform functions.
The Fourier transform functions require you to specify the flag and hint arguments.
The flag argument specifies the result normalization method. The following table lists the possible values for the flag argument. Specify one and only one of the represented values in the flag argument. The A and B factors are multipliers used in the DFT computation.
Value | A | B | Description |
---|---|---|---|
IPP_FFT_DIV_FWD_BY_N | 1/N | 1 | Forward transform is done with the 1/N normalization. |
IPP_FFT_DIV_INV_BY_N | 1 | 1/N | Inverse transform is done with the 1/N normalization. |
IPP_FFT_DIV_BY_SQRTN | 1/N1/2 | 1/N1/2 | Forward and inverse transform is done with the 1/ N1/2 normalization. |
IPP_FFT_NODIV_BY_ANY | 1 | 1 | Forward or inverse transform is done without the 1/ N or 1/N1/2 normalization. |
You can also select the algorithmic implementation of the function. The following values are supported:
- ippAlgHintAccurate: All output samples are exact; accuracy takes precedence over performance.
- ippAlgHintFast, ippAlgHintNone: Function performance takes precedence over accuracy, and some output samples may differ slightly from the exact result.