| Re: Quasi-freehand curve fitting... [message #11193] |
Mon, 09 March 1998 00:00 |
ez95
Messages: 1
Registered: March 1998
|
Junior Member |
|
|
It looks like
y=a Arctan (bx+c) +d
In article <6e0mdb$4s0@scotsman.ed.ac.uk>,
R.Brockie@roe.ac.uk wrote:
>
> Hello folks,
>
> This is something which has been exercising my mind for a little while
> now. My problem is the following:
>
> I have pairs of numbers (x,y) which rescribe the performance of a
> system, usually with a fairly abrupt change of gradient. eg:
>
> | x x x
> | x
> y| x
> | x
> | x x x x
> ---------------------
> x
>
> Now, this preformance does not have any particular functional form,
> but I would like to draw a curve through the points to draw the
> eye. The closest I have got to what I wish is to use INTERPOL to
> linearly interpolate between the points and then SMOOTH to round the
> corners. However, getting a nice amount of smoothing drags the curve
> away from the points. I would like a nice smooth curve which does go
> through all the points, like one would draw freehand.
>
> --
> R.
>
> -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= -=-=-=-=-=-=-
> Richard Brockie B.Sc.(Hons), The tall blond one.
> Adaptive Optics Chap. Email: R.Brockie@roe.ac.uk
> -=-=-=-=-=-=-=-=-=-=-= http://www.roe.ac.uk/rmbwww =-=-=-=-=-=-=-=-=-=-=-
>
-----== Posted via Deja News, The Leader in Internet Discussion ==-----
http://www.dejanews.com/ Now offering spam-free web-based newsreading
|
|
|
|
| Re: Quasi-freehand curve fitting... [message #11196 is a reply to message #11193] |
Mon, 09 March 1998 00:00  |
steinhh
Messages: 260
Registered: June 1994
|
Senior Member |
|
|
Richard Brockie (rmb@muckleroe.roe.ac.uk) wrote:
> Thanks to Jim Pendleton <jimp@rsinc.com>, who suggested the SPLINE
> routine. I have ended up using SPL_INIT and SPL_INTERP as SPLINE has a
> few limitations, the main being that it requires monotonically
> increasing x-values which I do not necessarily have.
IDL> ix = sort(x)
IDL> x = x(ix)
IDL> y = y(ix)
Now you do :-)
Stein Vidar
|
|
|
|
| Re: Quasi-freehand curve fitting... [message #11197 is a reply to message #11193] |
Mon, 09 March 1998 00:00 |
rmb
Messages: 5
Registered: October 1994
|
Junior Member |
|
|
I, Richard Brockie (rmb@muckleroe.roe.ac.uk) wrote:
: eye. The closest I have got to what I wish is to use INTERPOL to
: linearly interpolate between the points and then SMOOTH to round the
: corners. However, getting a nice amount of smoothing drags the curve
: away from the points. I would like a nice smooth curve which does go
: through all the points, like one would draw freehand.
Thanks to Jim Pendleton <jimp@rsinc.com>, who suggested the SPLINE
routine. I have ended up using SPL_INIT and SPL_INTERP as SPLINE has a
few limitations, the main being that it requires monotonically
increasing x-values which I do not necessarily have.
--
R.
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= -=-=-=-=-=-=-
Richard Brockie B.Sc.(Hons), The tall blond one.
Adaptive Optics Chap. Email: R.Brockie@roe.ac.uk
-=-=-=-=-=-=-=-=-=-=-= http://www.roe.ac.uk/rmbwww =-=-=-=-=-=-=-=-=-=-=-
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
