6
$\begingroup$

I'm wondering if there is an easy way to get the plot below exported in a small pdf?

The plot below is made in the following steps:

  1. step 1: Create a streamplot
  2. step 2: Create a densityplot, where the densityplot has to have a lot of samplingpoints in some cases !!
  3. step 3: put the two of them together.

VortexAntivortex

Now whenever I save this as a pdf-file it becomes huge due to the large amount of sampling points. So I've always saved it as a pdf-file.

Now I was wondering if it was possible to get this in a single (small !!) pdf-file? For as far as I understand I need to raster the density plot and put in the axes seperately. Now I've seen some answers regarding this for a list density plot, but it doesn't seem to work for a densityplot. Whenever I try to put the plotrangepadding to 0 weird things happen. Next to that I've also not been able to put the legend next to the plot after rasterization. Are there any hints on this ?

Code to reproduce the above example:

vx[x_, y_, d_] := -y*(1/((x - d)^2 + y^2) - 1/((x - 1/d)^2 + y^2)) + 
  y*(1/((x + d)^2 + y^2) - 1/((x + 1/d)^2 + y^2))

vy[x_, y_,d_] := (x - d)/((x - d)^2 + y^2) - (x - 1/d)/((x - 1/d)^2 + 
   y^2) - (x + d)/((x + d)^2 + y^2) + (x + 1/d)/((x + 1/d)^2 + y^2)

v[x_, y_, d_] := Sqrt[( 4 d^2 (-1 + d^2)^2 ((1 + x^2)^2 + 2 (-1 + x^2) y^2 + 
    y^4))/(((d - x)^2 + y^2) ((d + x)^2 + y^2) ((-1 + d x)^2 + 
    d^2 y^2) ((1 + d x)^2 + d^2 y^2))]

part1 = 
 StreamPlot[{vx[x, y, 0.6], vy[x, y, 0.6]}, {x, -2, 2}, {y, -1.4, 
   1.4}, ImageSize -> 650, PlotRangePadding -> None, 
  FrameStyle -> Black, BaseStyle -> FontSize -> 22, 
  PerformanceGoal -> "Quality", StreamStyle -> "PinDart", 
  StreamPoints -> Fine, StreamScale -> .15, 
  AspectRatio -> ((1.4 - (-1.4))/(2 - (-2)))]

legend = 
 BarLegend[{"SunsetColors", {0, 15}}, LegendFunction -> "Panel", 
  LegendLabel -> 
   "|\!\(\*OverscriptBox[\(v\), \(\[RightVector]\)]\)|/(\[Kappa]/2\
\[Pi]R)", LabelStyle -> Directive[Bold, Black, 14], 
  LegendMargins -> 0, LegendMarkerSize -> 400]

part2 = DensityPlot[{v[x, y, 0.6]}, {x, -2, 2}, {y, -1.4, 1.4}, 
  ImageSize -> 650, ColorFunction -> "SunsetColors", 
  PlotRangePadding -> None, FrameStyle -> Black, 
  BaseStyle -> FontSize -> 22, PerformanceGoal -> "Quality", 
  AspectRatio -> ((1.4 - (-1.4))/(2 - (-2))), 
  FrameLabel -> {"x/R", "y/R"}, 
  PlotLabel -> 
   Style["The vortex-antivortex velocity field", Black, 30] , 
  PlotRange -> {0, 15}, PlotPoints -> 200, 
  Epilog -> Style[Circle[{0, 0}, 1], {Thick, Dashed, White}], 
  PlotLegends -> Placed[legende, Right]]

plot = Show[part2, part1]
$\endgroup$
  • 5
    $\begingroup$ Related: Jens's contourDensityPlot with vector contours and raster density. $\endgroup$ – Rahul Jan 16 '15 at 16:34
  • 1
    $\begingroup$ @Rahul, that's for contourplots, there my problem disappears with the fixpolygons. This is not the case for a density plot $\endgroup$ – Nick Jan 17 '15 at 14:38
  • $\begingroup$ What are "fixpolygons"? Perhaps you missed the second part of the answer I linked to, which explicitly deals with a smooth density plot. $\endgroup$ – Rahul Jan 17 '15 at 15:02
  • $\begingroup$ Great question, have you found an answer in newer versions? I have the same problem $\endgroup$ – AimForClarity Jan 29 '18 at 4:41
  • $\begingroup$ @AimForClarity No, I actually just export the figure (without axes) as a (high quality) rastered figure and put a set of axes on top using TikZ. $\endgroup$ – Nick Jan 29 '18 at 13:17
1
$\begingroup$

Try to add the following lines to your code in order to decrease the pdf-file size.

plot = Rasterize[plot, RasterSize -> 1000, ImageSize -> 650,ImageResolution -> 1000];
Export["plot.pdf", plot];

This method give us a file of aproximatelly 200kB instead of the previous 11MB. If you want an eps-file, is better to convert the pdf-file rather than export the eps extension.

$\endgroup$
  • 1
    $\begingroup$ that will indeed work if I wanted to rasterize the complete plot. But it won't keep my axes in a vector-format, that's where my problem is :(. $\endgroup$ – Nick Jan 17 '15 at 14:37

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.