I learned that an optimal fft of a N real input can be done using a full fft on an N/2 input and then a call to dsps_cplx2real_fc32 .
I understand that an fft of a complex input/output can be used as an ifft with some small adjustments.
I am looking to do an efficient ifft where the output is expected to be real and the input is the same format that dsps_cplx2real_fc32 returns.
Is it possible to do it in an efficient way like it is possible with the forward direction? Do we need an inverse of dsps_cplx2real_fc32 and then a call to a full fft with N/2 values?
Please advise.
Request for a feature or assistance.
I learned that an optimal fft of a N real input can be done using a full fft on an N/2 input and then a call to dsps_cplx2real_fc32 . I understand that an fft of a complex input/output can be used as an ifft with some small adjustments.
I am looking to do an efficient ifft where the output is expected to be real and the input is the same format that dsps_cplx2real_fc32 returns. Is it possible to do it in an efficient way like it is possible with the forward direction? Do we need an inverse of dsps_cplx2real_fc32 and then a call to a full fft with N/2 values? Please advise.