resolution on plotting for "retina" displays?

Hi,

For all those using newer macs [and I assume other newer computers] the
resolution on the plt.show popup window is a little rough (not sure of the
technical name). Is there a way to double up the resolution? I tried
changing a few things but none of them seemed to make much difference. Does
anyone know specifically what I would need to change?

thanks,

Martin.

···

--
View this message in context: http://matplotlib.1069221.n5.nabble.com/resolution-on-plotting-for-retina-displays-tp39068.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

My understanding is that what gets displayed is always at the native resolution. I would assume you are using the macosx backend, so maybe you are running into a limitation with it? Can you try to see if the problem exists with the Qt4Agg or TkAgg backends?

Cheers!
Ben Root

···

On Tue, Sep 25, 2012 at 8:14 AM, mdekauwe <mdekauwe@…287…> wrote:

Hi,

For all those using newer macs [and I assume other newer computers] the

resolution on the plt.show popup window is a little rough (not sure of the

technical name). Is there a way to double up the resolution? I tried

changing a few things but none of them seemed to make much difference. Does

anyone know specifically what I would need to change?

thanks,

Martin.

I tried your suggestions...the first one (Qt4Agg) resulted in an error
(below) so it looks like I am missing some packages and the second option
(TkAgg) pretty much gave the same quality as the macosx backend.

thanks.

File
"/opt/local/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/matplotlib/pyplot.py",
line 95, in <module>
    new_figure_manager, draw_if_interactive, _show = pylab_setup()
  File
"/opt/local/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/matplotlib/backends/__init__.py",
line 25, in pylab_setup
    globals(),locals(),[backend_name])
  File
"/opt/local/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/matplotlib/backends/backend_qt4agg.py",
line 12, in <module>
    from backend_qt4 import QtCore, QtGui, FigureManagerQT, FigureCanvasQT,\
  File
"/opt/local/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/matplotlib/backends/backend_qt4.py",
line 23, in <module>
    from qt4_compat import QtCore, QtGui, _getSaveFileName, __version__
  File
"/opt/local/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/matplotlib/backends/qt4_compat.py",
line 36, in <module>
    import sip
ImportError: No module named sip

···

--
View this message in context: http://matplotlib.1069221.n5.nabble.com/resolution-on-plotting-for-retina-displays-tp39068p39071.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

I am still doubtful that there is any actual bug here (at least, not without evidence from screen captures), but for something to impact both the TkAgg backend and the macosx backend, it can’t be an Agg-snapping issue. I know it sounds ridiculous, but could it be possible that we are hitting limits of floating point numbers?

What is the resolution of your display and its size?

Ben Root

···

On Tue, Sep 25, 2012 at 8:53 AM, mdekauwe <mdekauwe@…287…> wrote:

I tried your suggestions…the first one (Qt4Agg) resulted in an error

(below) so it looks like I am missing some packages and the second option

(TkAgg) pretty much gave the same quality as the macosx backend.

thanks.

Doesn't Eric have a retina display mac? Eric, do you see this behaviour?

Screenshots would be a real help here as extra information to diagnose
the problem.

···

On Tue, Sep 25, 2012 at 8:09 PM, Benjamin Root <ben.root@...1304...> wrote:

On Tue, Sep 25, 2012 at 8:53 AM, mdekauwe <mdekauwe@...287...> wrote:

I tried your suggestions...the first one (Qt4Agg) resulted in an error
(below) so it looks like I am missing some packages and the second option
(TkAgg) pretty much gave the same quality as the macosx backend.

thanks.

I am still doubtful that there is any actual bug here (at least, not without
evidence from screen captures), but for something to impact both the TkAgg
backend and the macosx backend, it can't be an Agg-snapping issue. I know
it sounds ridiculous, but could it be possible that we are hitting limits of
floating point numbers?

What is the resolution of your display and its size?

Ben Root

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
Matplotlib-users mailing list
Matplotlib-users@lists.sourceforge.net
matplotlib-users List Signup and Options

--
Damon McDougall
http://www.damon-is-a-geek.com
B2.39
Mathematics Institute
University of Warwick
Coventry
West Midlands
CV4 7AL
United Kingdom

Well perhaps it is fine, I wasn't really suggesting it was a bug. Just if I
stretched the plot window a bit the resolution looked a bit fuzzier than i
remembered. Perhaps my memory isn't a good test and I will compare it to the
computer at work. I just did a screen print and it looks fine unless you put
the screen close to your eyes and then the lettering etc is a bit fuzzy, but
you won't see that in the screen dump. Attached anyway.

<http://matplotlib.1069221.n5.nabble.com/file/n39086/Screen_Shot_2012-09-26_at_6.35.55_AM.png>

···

--
View this message in context: http://matplotlib.1069221.n5.nabble.com/resolution-on-plotting-for-retina-displays-tp39068p39086.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

I tried your suggestions...the first one (Qt4Agg) resulted in an error
(below) so it looks like I am missing some packages and the second option
(TkAgg) pretty much gave the same quality as the macosx backend.

thanks.

I am still doubtful that there is any actual bug here (at least, not without
evidence from screen captures), but for something to impact both the TkAgg
backend and the macosx backend, it can't be an Agg-snapping issue. I know
it sounds ridiculous, but could it be possible that we are hitting limits of
floating point numbers?

What is the resolution of your display and its size?

Ben Root

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
Matplotlib-users mailing list
Matplotlib-users@lists.sourceforge.net
matplotlib-users List Signup and Options

Doesn't Eric have a retina display mac? Eric, do you see this behaviour?

I think that what is described is typical for applications that lack specific retina display support. In that case the retina display uses 4 physical pixels for each virtual pixel, so that the program operates as if it were working with a typical dpi. It seems that there are different ways and stages at which this remapping can happen; I don't fully understand it. The way it works with mpl using either the macosx or qt4agg backends is, to me, not bad; maybe a little better than what happens with un-tweaked Firefox or Thunderbird, for example, and better than what happens with "ipython qtconsole". Some degree of fuzziness is seen in all these cases.

My guess is that taking full advantage of the retina's resolution, for text and for lines, is going to require changes to underlying toolkits, and to some of our backends (certainly to the macosx backend). Given that even Mozilla, with its rapid release cycle, still hasn't released versions of Firefox or Thunderbird that properly handle the retina display, I suspect it will be a long time before we can do so in mpl. In the meantime, mpl is perfectly usable with the retina, no worse to my eye than with an ordinary display resolution.

Eric

···

On 2012/09/25 9:23 AM, Damon McDougall wrote:

On Tue, Sep 25, 2012 at 8:09 PM, Benjamin Root <ben.root@...1304...> wrote:

On Tue, Sep 25, 2012 at 8:53 AM, mdekauwe <mdekauwe@...287...> wrote:

Screenshots would be a real help here as extra information to diagnose
the problem.