> Date: Sat, 15 Jun 2013 16:12:06 -0400
> From: etter...@gmail.com
> To: tomd...@comcast.net
> CC: firstname.lastname@example.org
> Subject: Re: [Gimp-user] A sad case of regression ?
> Not exactly, no, the edited image that is now on my screen, the xcf, is
> probably a resolution of 300 x 300 and may be a print size of 12 x 16;
> But the exported
> image is a resolution of 72 and is not meant for printing. *That* is the
> one that I have to re-open (because I can't force it not to close when I
> export.) *That's* the one I have to mail, and if I decided to make a
> tweak, I can't just save and mail -- I have to export, re-open ... I
> don't see any way around the repeated reopenings except to make sure
> everything I do is perfect the first time, and that's even less likely than
> the developers reconsidering this. Thank you Tom.
I'm not making any sense of this at all. Image resolution is a piece of
metadata and does not in any way dictate the size of the image as measured in
actual image pixels. If you open an image whose tags say "300 dpi", when you
save (or export) it the output file will contain that 300dpi setting. Now if
the image is 12"x16" and tagged as 300dpi this means that the image's physical
PIXEL dimensions are 3600x4800. And when you export this image to a JPG, that
JPG will still be be 3600x4800 pixels large (and tagged as 300dpi) unless you
specifically dictated to GIMP otherwise. Going to the Image menu and selecting
"Resize image..." rescales the image to a different size in pixels (but doesn't
necessarily change the dpi metadata); selecting "Print Size..." lets you set
the dpi metadata directly, but doesn't change the pixel content of the image.
Also, when you export the image to a JPEG, if suddenly your open image window
"disappears", well that is not supposed to happen at all and sounds something
like a GIMP program crash, but we don't have enough information as is to
determine that. And when GIMP crashes, you at least get a message telling you
in no ambiguous terms that something crashed.
Numbers may not lie, but neither do they tell the whole truth.
gimp-user-list mailing list