Oooh, I didn't know you could use map without a function. Thanks!

If I've understood correctly, I've added an assertation to address this. Let me know if I haven't got it right!

I went to fix it, but someone has already. Awesome.

Now specified in the description 🙂

Should be fixed now!

Defined :)

I decided to keep it because typeof NaN === 'number'. I've fixed the name of the tests. I'd prefer to require returning -1, as in other signal processing feature extractors this kind of invalid input would cause feature extraction to fail entirely, and it's better to have a consistent API. A NaN in an FFT would wreck it entirely, same with things like Spectral Centroid, and MFCC.

As I said I would, I've updated the description 🙂

I've added random tests. Let me know if they're not enough, or not as you expect.

Very fun indeed

The correct zcr for [1,0,-1] is 1, because it crosses from positive to negative. I'll update my description. Can't think of any case where filtering zeroes won't work, but would be delighted to be corrected. [1,0,1,0,-1,0,-1,0] would still be 1.

I've removed the input type validation portion of this kata.

Thanks

The tests are good now, please try again.

Solved.