[Date Index] [Thread Index] [Author Index]

LogLinear Plot

• To: mathgroup at christensen.cybernetics.net
• Subject: [mg620] LogLinear Plot
• From: Paul E Howland <PEHOWLAND at taz.dra.hmg.gb>
• Date: Thu, 30 Mar 1995 14:03:40 +0200 (IST)

A colleague of mine is using LogLinearPlot to plot some functions that
show propagation loss (on y-axis) against range (on x-axis).  The range
covers just over three decades on the plot.  The function should appear as
very simple smooth curves.  However, when plotted using LogLinearPlot with
its default number of points, there is a noticeable "kink" in the graph,
where the graph has obviously been undersampled.  The only way my
colleague has found to get around this is to set PlotPoints->1000, which
obviously increases the plot time no end. 

My question is: Does LogLinearPlot sample the function to be plotted in 
the linear domain, before plotting it in the log domain?  This would 
explain the poor sampling of the function for small values of x.  If so, 
why doesn't LogLinearPlot sample the function in the log domain, so that 
the displayed plot has been evenly sampled across its width?  Is there 
any way round this problem, other than increasing PlotPoints to some 
silly value?

Paul E Howland                         

Long Range Ground Radar Systems Section               tel. +44 (0)1684 895767
CSS2 Division, Room BY209                             fax. +44 (0)1684 896315
Defence Research Agency                           email: PEHOWLAND at DRA.HMG.GB 
Malvern, Worcs, WR14 3PS, UK.                      
-----------------------------------------------------------------------------