For the FFT you ALWAYS need a power of 2. Some libraries "pad" the signal with zeros to get the correct length, which slightly affects the frequency spectrum. Unfortunately I don't have specific knowledge of the library you mentioned.

@@MarkNewmanEducation That is a misconception, this is only the case for the first widely known implementation that most textbooks cover. However many modern implementations such as the one in matlab (internally uses fftw) can handle non-power of 2 efficiently without zero padding needed.

How do they do it? What do they pad with? FFT can only work with radix-2 unless they're using some form of the DFT?

From the FFTW documentation: "The current version of FFTW incorporates many good ideas from the past thirty years of FFT literature. In one way or another, FFTW uses the Cooley-Tukey algorithm, the Prime Factor algorithm, Rader’s algorithm for prime sizes, and the split-radix algorithm (with a variation due to Dan Bernstein). Our code generator also produces new algorithms that we do not yet completely understand. [...] FFTW is best at handling sizes of the form 2^a*3^b*5^c*7^d*11^e*13^f, where e+f is either 0 or 1, and the other exponents are arbitrary. Other sizes are computed by means of a slow, general-purpose routine (which nevertheless retains O(n log n) performance, even for prime sizes). (It is possible to customize FFTW for different array sizes. See Section Installation and Customization, for more information.) Transforms whose sizes are powers of 2 are especially fast. " So powers of 2 are still the fastest, but other sizes are much faster than regular DFT. @@MarkNewmanEducation