Prev: splitting a data set for n-fold testing - best practices?
Next: Why is MOVEFILE so slow?
From: Ryan Webb on 16 Feb 2010 18:33
Same problem here, in addition the figure is cropped on the left and right. This has been an ongoing problem dating back to at least 2008a. Any word on this??
From: Jonas on 16 Feb 2010 21:41
"Ryan Webb" <webbr(a)econ.donotinclude.queensu.ca> wrote in message <hlf9ve$euq$1(a)fred.mathworks.com>...
> Same problem here, in addition the figure is cropped on the left and right. This has been an ongoing problem dating back to at least 2008a. Any word on this??

I get the same problem on Snow Leopard with 2010a - saving to eps looks horrible.

However, export_fig solves the issue for me (note that it crops the figure tightly, but I like that a lot).
From: Jonas on 16 Feb 2010 22:12
"Jonas" <remove_this_iop(a)gmx.at> wrote in message <hlfl02$bv1$1(a)fred.mathworks.com>...
> "Ryan Webb" <webbr(a)econ.donotinclude.queensu.ca> wrote in message <hlf9ve$euq$1(a)fred.mathworks.com>...
> > Same problem here, in addition the figure is cropped on the left and right. This has been an ongoing problem dating back to at least 2008a. Any word on this??
>
> I get the same problem on Snow Leopard with 2010a - saving to eps looks horrible.
>
> However, export_fig solves the issue for me (note that it crops the figure tightly, but I like that a lot).

Oops, I wrote too soon. Export_fig works only if saving to pdf, not when saving to eps. However, pdf is fine if you want to edit the figure later, e.g. in Illustrator.
From: Oliver Woodford on 17 Feb 2010 11:58
"Chris" wrote:
> No, unfortunately that doesn't work either... My .pdfs still have the blurriness with imagesc and the white diagonal lines with pcolor.

Bluriness of images when viewing pdfs (or eps files, for that matter) is generally caused by the viewing (or printing) software bilinearly interpolating the colors of an image when magnifying it to screen resolution. I've had 4 prints of my thesis ruined because of this.

There are two ways to avoid it:
1. Change the interpolation settings in your viewing/printing software, or change to a software that doesn't interpolate.
2. Increase the resolution of your image by, say, 8 (for good jpeg compression) before generating the figure, using nearest neighbor interpolation.

Method 2 is failsafe because it doesn't matter what viewing/printing software anyone else has, the figure will still look right to them too.

HTH,
Oliver
From: nick.steinmetz on 22 Feb 2010 17:28
On Feb 16, 7:12 pm, "Jonas" <remove_this_...(a)gmx.at> wrote:
> "Jonas" <remove_this_...(a)gmx.at> wrote in message <hlfl02$bv...(a)fred.mathworks.com>...
> > "Ryan Webb" <we...(a)econ.donotinclude.queensu.ca> wrote in message <hlf9ve$eu...(a)fred.mathworks.com>...
> > > Same problem here, in addition the figure is cropped on the left and right. This has been an ongoing problem dating back to at least 2008a. Any word on this??
>
> > I get the same problem on Snow Leopard with 2010a - saving to eps looks horrible.
>
> > However, export_fig solves the issue for me (note that it crops the figure tightly, but I like that a lot).
>
> Oops, I wrote too soon. Export_fig works only if saving to pdf, not when saving to eps. However, pdf is fine if you want to edit the figure later, e..g. in Illustrator.

For me, export_fig didn't even work saving to .pdf (to get rid of
lines), but on a whim I wondered whether the white/gray lines would
actually be there in the printed version (or whether they were just an
artifact of the display) and lo and behold the printed version looks
beautiful. If your application requires a nice electronic document
then this doesn't help you at all, but if you only need a printed
version (I am putting the pcolor image on a poster) then just try
printing it!

To be clear, I am using Matlab R2009b on Windows 7 (I know this post
was about macs, but I don't think the problem was specific to mac. but
sorry if this fails entirely there.). I export the pcolor figure using
-painters renderer; nothing else fancy.

Hope that helps,

Nick
 |  Next  |  Last
Pages: 1 2
Prev: splitting a data set for n-fold testing - best practices?
Next: Why is MOVEFILE so slow?