Op 17-mei-2006, om 16:53 heeft Mark Shuttleworth het volgende
geschreven:
Michiel Sikma wrote:
I don't get why you would want that. All that happens with the
scaling is the image being squished and then later stretched back
to normal. The only effect would be a loss in quality; not a loss
in proportions or composition. I don't see why you would want to
maintain two versions.
There is only one version which counts, which is the 600x400 (or
whatever) FINAL IMAGE that gets baked into the .so file. The value
of having a (for example) 1024x768 version is to see what it will
ACTUALLY LOOK LIKE when the framebuffer expands it to fill the
screen. That expansion is "dumb" (there's no smoothing going on) so
the example large-format version should be scaled up with a "dumb"
scaling algorithm so it gives a realistic view.
Mark
Actually, when making a picture that shows how it would look, it's
not very realistic to upscale an image from 640x400 to 640x480 (or
1024x768). That would cause some pixels to be doubled, which would
make it look really weird.
The upscaling that's caused after there's an actual .so file in place
is, like you say, without smoothing. But, it's also important to note
that it's done without altering the pixels. It's only the pixels that
are made to look larger, not the image that's given double pixels.
The most accurate way to emulate how a 640x400 squished image would
look in real usage would be to upscale it to a multiple of 640x480;
1280x960, for example, so there's more room for accurate emulation of
the screen stretching.
Of course, the best way to check would be to somehow get the image to
be displayed in full-screen at 640x400...
Michiel
--
ubuntu-art mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-art
