Bill - Thank you for finding this! But I still don’t understand what was wrong with the way that the cx array was sized.
Surely the size of the original x array (NMAX real elements) is equivalent to the size of the original cx array 2*(NMAX/2) complex elements... I just did this test (using the original code): call four2a(x,NMAX,1,-1,0) !r2c FFT to freq domain 205 write(*,*) x(1:4) 206 write(*,*) cx(0:1) 207 write(*,*) x(NMAX-3:NMAX) 208 write(*,*) cx(NMAX/2-2:NMAX/2-1) with these results: -21077.0000 0.00000000 32318.1348 -31433.4023 ( -21077.0000 , 0.00000000 ) ( 32318.1348 , -31433.4023 ) 36066.5547 41465.5625 34216.6055 26828.1055 ( 36066.5547 , 41465.5625 ) ( 34216.6055 , 26828.1055 ) which seem to confirm that elements x(1:4) line up with cx(0:1) and that x(NMAX-3:NMAX) line up with cx(NMAX/2-2:NMAX/2-1), as expected. So why did you need to add the extra element at the end of cx? Steve > On Jul 2, 2017, at 10:12 AM, Bill Somerville <g4...@classdesign.com> wrote: > > On 02/07/2017 16:06, Bill Somerville wrote: >> Steve, I didn't fully analyse the issue. Does the c1 vector in >> ft8_downsample() need to be one longer as well? > > Hi Steve, > > ok, I see not as it is complex to complex for the return to time domain. > > 73 > Bill > G4WJS. > > > ------------------------------------------------------------------------------ > Check out the vibrant tech community on one of the world's most > engaging tech sites, Slashdot.org! http://sdm.link/slashdot > _______________________________________________ > wsjt-devel mailing list > wsjt-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/wsjt-devel ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot _______________________________________________ wsjt-devel mailing list wsjt-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/wsjt-devel
