|
From: Philipp K. J. <ja...@ie...> - 2008-02-22 05:20:31
|
Try this (all in one consecutive session): gnuplot> reset gnuplot> unset surface gnuplot> unset hidden3d gnuplot> set xlabel 'x' gnuplot> set ylabel 'y' gnuplot> set style line 1000 lt 1 gnuplot> set pm3d implicit hidden3d 1000 gnuplot> set samples 2 gnuplot> set isosamples 2 gnuplot> splot exp(-x**2-y**2) # totally flat gnuplot> set samples 3 gnuplot> replot # Now there are 3 nodes along the x-axis. # But why? What has set samples to do w/ splot? gnuplot> set isosamples 2,4 gnuplot> replot # Now there are still 3 (not 2!) nodes along the x-axis # and 4 along the y-axis. That's it. Gnuplot version info is attached to my original message on this subject. Best, Ph. On Thursday 21 February 2008 19:47, you wrote: > On Thursday 21 February 2008 19:06, Philipp K. Janert wrote: > > When using > > set pm3d hidden3d > > with splot for plotting a function > > > > it seems as if > > - set samples n > > is used for the number of steps along the x-direction > > - set isosamples n,m > > uses m for the number of steps along the y-direction > > and ignores n. If only a single argument is given, it is > > used for the y-direction. > > I don't think that's right. > "set isosamples XX,YY" causes the surface to be represented by > XX lines with different x, and YY lines with different y. > Try: > set isosamples 10,50 > splot (sin(x)*sin(y))/(x*y) > set isosamples 50,10 > replot > > So isosamples control the number of lines. > For hidden surfaces, that's the end of the story so far as I know. > > For non-hidden surfaces, "set samples" controls how finely each line is > sampled. So "set isosamples 5,5" will give you a small number of lines on > x and y. "set samples 10,10" will make these lines very jagged; "set > samples 100,100" will make these same lines much smoother. > > > I also would expect both arguments to set isosamples to be evaluated. > > For me both arguments work as expected. > Tested both in 4.2 and CVS versions. |