1D Function into 2D Function Interpolation
in [Matlab]
|
From: Matt J on 31 Mar 2010 14:21 "Bruno Luong" <b.luong(a)fogale.findmycountry> wrote in message <hp030o$bic$1(a)fred.mathworks.com>... > "Matt J " <mattjacREMOVE(a)THISieee.spam> wrote in message <hp02pa$7qc$1(a)fred.mathworks.com>... > > > > > > > It does if linear interpolation is used. There's no reason not to do so as far as we've been told. > > Look at the graphic link from OP, he did try both linear and pchip (Hermit) and obviously he is not satisfied. ===================== I think the graph was misinterpreted. The OP claimed to see negative values there, but there are none. Linear interpolation will not, at least without an extrap method specified, produce negative values from positive data. If there was any difficulty with linear interpolation, it is again that the OP did not provide LSF data on a sufficiently wide enough domain. This results in NaNs in certain parts of the graph, were the samples fall outside the upper and lower bounds of the LSF data.
From: Bruno Luong on 31 Mar 2010 14:44 "Matt J " <mattjacREMOVE(a)THISieee.spam> wrote in message <hp03qv$o35$1(a)fred.mathworks.com>... >This results in NaNs in certain parts of the graph, were the samples fall outside the upper and lower bounds of the LSF data. No, in the graphs show the four corner (+/-7,+/-7) fall outside the range (r=7) but the results are extrapolated with finite values; take look again. What OP wanted is monotonic extrapolation to 0 at r="infinity" as I understand. Bruno
From: Bruno Luong on 31 Mar 2010 14:52 "Matt J " <mattjacREMOVE(a)THISieee.spam> wrote in message <hp03qv$o35$1(a)fred.mathworks.com>... > The OP claimed to see negative values there, but there are none. Yes there *are*, because he extrapolates. What he said is *correct* and I have verified it. Bruno
From: Matt J on 31 Mar 2010 14:57 "Bruno Luong" <b.luong(a)fogale.findmycountry> wrote in message <hp055k$fde$1(a)fred.mathworks.com>... > "Matt J " <mattjacREMOVE(a)THISieee.spam> wrote in message <hp03qv$o35$1(a)fred.mathworks.com>... > >This results in NaNs in certain parts of the graph, were the samples fall outside the upper and lower bounds of the LSF data. > > No, in the graphs show the four corner (+/-7,+/-7) fall outside the range (r=7) but the results are extrapolated with finite values; take look again. ======================= Then the OP hasn't shown the code that actually generated the plots. In the posted code, interp1() wasn't called with an extrap option. That makes finite values for r>7 impossible >What OP wanted is monotonic extrapolation to 0 at r="infinity" as I understand. ====== A tail of zeros will also be zero at r=infinity.
From: Royi Avital on 31 Mar 2010 15:03
"Bruno Luong" <b.luong(a)fogale.findmycountry> wrote in message <hp05kl$mu9$1(a)fred.mathworks.com>... > "Matt J " <mattjacREMOVE(a)THISieee.spam> wrote in message <hp03qv$o35$1(a)fred.mathworks.com>... > > The OP claimed to see negative values there, but there are none. > > Yes there *are*, because he extrapolates. What he said is *correct* and I have verified it. > > Bruno Bruno, I try running your code yet it wouldn't work (Matlab 2010a, Downloaded the file you referred me to). Thanks for helping. I would like to try SLM as well if someone more experienced with the syntax would help. Thanks. |