I am developing a pedometer algorithm on a RFduino for my Bachelor thesis and I wanted to use FFT/FHT for filtering noise and simultaneously extracting frequency information for the feature detection. Firstly I wanted to use the Arduino FFT library, but my data consists of real values so I figured out that the FHT algorithm would be a lot better, because it is faster and more memory efficient. I found the Arduino FHT Library for my project. I only have one problem at the moment. I don't know if it is possible to make an inverse FHT in order to go back in the time domain after filtering the unwanted frequencies. I found a similar question about the arduino fft library and the inverse FFT with the given library is possible. So I am wandering if the same will be possible for the FHT one?
Asked
Active
Viewed 391 times
1
-
[FHT is not necessarily faster/better](https://en.wikipedia.org/wiki/Discrete_Hartley_transform#Fast_algorithms) - just use a real-to-complex FFT for the forward transform (and *vice versa*). – Paul R Jun 14 '17 at 08:54
-
Thank you, @PaulR! I've researched a little bit more and it seems this libraries do not work with the RFduino. So I am going to use the CMSIS FFT library instead. – ap213 Jun 15 '17 at 09:43