On 29 May 2012 10:19, octave <oct...@duncans.org.uk> wrote: > "fracshift" in signal-1.1.3 does not work on row vectors when the shift > parameter is an integer. It calls "circshift" when the shift parameter > is an integer, but that function only shifts in the column direction. > > Hence the following command: > > fracshift([1 0 0 0],1) > > gives the output: > > [1 0 0 0]
I'm not sure what's the best way to act here and I'm CC'ing the function author Eric Chassande-Mottin. Using a non-integer for shift value works whether x is column or vector. fracshift([1 0 0 0],1.5) ans = 0.12584 0.63632 0.63632 -0.21131 This is definetily inconsistent. However, the idea behind this function seems to be a circshift that accepts integers. If so, then it should behave like circshift in which case the bug would be that it acts on x even when it's a row vector and not that it doesn't act on it if n is an integer. Anyone has any comment? > Also, "circshift" itself appears to be redundant, as it duplicates the > functionality of the built-in command "shift". circshift exists in matlab so it must be there for compatibility. However, looking at their code, they do differ a bit and maybe it would be easier to maintain if one was a wrapper of the other. That however belongs to octave core, maybe you should implement a bug report at https://savannah.gnu.org/bugs/?group=octave (or even better, a patch). Carnë ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ Octave-dev mailing list Octave-dev@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/octave-dev
