DSP FFT
Posted: Thu Nov 02, 2023 7:03 pm
Hello,
i'm trying to use the DSP Library for FFT computation. I've read the online Espressif documentation but it lacks a lot of information. I also followed the example code but i cannot find any explanation about the following functions:
- dsps_bit_rev_fc32()
- dsps_cplx2reC_fc32()
what do they do? are they needed? the first executes bit reverse operation, what is it? the second converts complex FFT result to real array, but how? After a FFT i don't need any of those functions, if the input array is real i will obtain a simmetric fft COMPLEX array.
Anyway i also tried to compile my project with ESP-32 optimization enabled, the DSP library uses all functions with _ae32 postfix and they always produce NaN output arrays, why? The library works only in ANSI C mode.
Leo
i'm trying to use the DSP Library for FFT computation. I've read the online Espressif documentation but it lacks a lot of information. I also followed the example code but i cannot find any explanation about the following functions:
- dsps_bit_rev_fc32()
- dsps_cplx2reC_fc32()
what do they do? are they needed? the first executes bit reverse operation, what is it? the second converts complex FFT result to real array, but how? After a FFT i don't need any of those functions, if the input array is real i will obtain a simmetric fft COMPLEX array.
Anyway i also tried to compile my project with ESP-32 optimization enabled, the DSP library uses all functions with _ae32 postfix and they always produce NaN output arrays, why? The library works only in ANSI C mode.
Leo