Michiel de Hoon provided a patch for this which I just applied to the trunk.
As I don't have a Mac, I can't test it -- any feedback is welcome. Mike Jae-Joon Lee wrote: > Thomas, > As John suggested before, please check if the size differences go away > if you use the same dpi, actually dpi=72. > > After some quick look, it seems that the osx backend does not scale > the font size correctly respecting the dpi. > At line 124 of bacend_macosx.py, > > size = prop.get_size_in_points() > weight = prop.get_weight() > style = prop.get_style() > gc.draw_text(x, y, unicode(s), family, size, weight, style, angle) > > My quick guess is that replacing > > size => size*self.dpi/72. > > in the last call may solve the problem. Since I don't use the osx > backend, I wonder if others can test this. > > -JJ > > > On Thu, Apr 30, 2009 at 9:32 AM, Michael Droettboom <md...@stsci.edu> wrote: > >> I forwarded this message to Michiel de Hoon, the author of the Mac OS-X >> backend, in case he has any thoughts. >> >> Mike >> >> Thomas Robitaille wrote: >> >>> Hi Jae-Jong and John, >>> >>> Thanks for your replies! While experimenting with this to send >>> screenshots, I realized that my default backend was set to MacOSX, not >>> WXAgg. The WXAgg output to the screen actually agrees with the PNG output in >>> terms of font sizes. But the font sizes differ between the MacOSX and WXAgg >>> backends. Attached are screenshots using the MacOSX and WXAgg backends. You >>> can see the font size is different. Is this a bug? >>> >>> Thanks, >>> >>> Thomas >>> >>> >>> >>> ------------------------------------------------------------------------ >>> >>> >>> ------------------------------------------------------------------------ >>> >>> >>> On 29 Apr 2009, at 00:24, Jae-Joon Lee wrote: >>> >>> >>>> On Tue, Apr 28, 2009 at 11:09 PM, John Hunter <jdh2...@gmail.com> wrote: >>>> >>>>> If you want the relative fontsizes in the figure window and saved figure >>>>> to >>>>> agree, pass the same "dpi" to the figure command and savefig command. >>>>> >>>> John, >>>> I thought the font size (which is specified in points) is independent >>>> of dpi, i.e., font size in "pixel" actually scales with the dpi. I >>>> think it should be filed as a bug if the relative font size depends on >>>> the dpi. >>>> >>>> Anyhow, I just did a quick test and the (relative) font size does not >>>> seem to vary with dpi. >>>> >>>> Thomas, >>>> What version of mpl are you using? >>>> With the mpl from the svn trunk, I don't see any significant change as >>>> you described. >>>> The WxAgg figure and the png output are actually drawn by an identical >>>> backend, so there should be no significant difference. There can be >>>> some subtle difference due to different dpi, but I don't see a >>>> difference as large as 30%. Can you post a some sample images? i.e., a >>>> screenshot of WxAgg figure and the png output. >>>> >>>> I can see that the text in pdf output occupies a bit larger area than >>>> png (when usetex=False), but, to me, this seems to be due to different >>>> amount of kernings (it seems that no kerning is applied for pdf text) >>>> instead of different font size. >>>> >>>> So, can you first check if the difference goes away when you use same >>>> dpi as John suggested? And if that is the case, can you try the latest >>>> svn and check if the relative font size still depends on the dpi? >>>> >>>> Regards, >>>> >>>> -JJ >>>> >>> ------------------------------------------------------------------------ >>> >>> >>> ------------------------------------------------------------------------------ >>> Register Now & Save for Velocity, the Web Performance & Operations >>> Conference from O'Reilly Media. Velocity features a full day of expert-led, >>> hands-on workshops and two days of sessions from industry leaders in >>> dedicated Performance & Operations tracks. Use code vel09scf and Save an >>> extra 15% before 5/3. http://p.sf.net/sfu/velocityconf >>> ------------------------------------------------------------------------ >>> >>> _______________________________________________ >>> Matplotlib-users mailing list >>> Matplotlib-users@lists.sourceforge.net >>> https://lists.sourceforge.net/lists/listinfo/matplotlib-users >>> >>> >> -- >> Michael Droettboom >> Science Software Branch >> Operations and Engineering Division >> Space Telescope Science Institute >> Operated by AURA for NASA >> >> >> ------------------------------------------------------------------------------ Register Now & Save for Velocity, the Web Performance & Operations Conference from O'Reilly Media. Velocity features a full day of expert-led, hands-on workshops and two days of sessions from industry leaders in dedicated Performance & Operations tracks. Use code vel09scf and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf _______________________________________________ Matplotlib-users mailing list Matplotlib-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/matplotlib-users