Re: Panther, sRGB, web browsers
Re: Panther, sRGB, web browsers
- Subject: Re: Panther, sRGB, web browsers
- From: John Gnaegy <email@hidden>
- Date: Thu, 18 Dec 2003 15:31:24 -0800
Can I point out again we're talking about untagged image data? I'm not
convinced that assuming sRGB for untagged image data is significantly
more reliably accurate than assuming the display profile. For untagged
image data you're always just guessing where it came from no matter
what you assume.
Assuming the display profile as source virtually guarantees that the
color will be displayed incorrectly unless it was created in a
non-color-managed application on that machine.
It guarantees failure to a similar degree of probability as assuming
that it's sRGB unless the image was created in sRGB. I would think
that if it were created in explicitly sRGB space that there would be an
embedded profile. sRGB isn't identical to "any unknown monitor space",
it's not a cure all for the problem of untagged image data. Sure I
agree sRGB is one way this could have gone, but I don't think it's
necessary to completely discard it with the argument that assuming sRGB
would have significantly increased your chance of accurately guessing
an untagged image's creation environment.
a.) It's not consistent with how some 90% (or greater) expects
untagged images to behave in web browsers, and increasingly from
digital cameras.
b.) It's not consistent with how W3C says to handle web images.
Yup, you're right there.
c.) It's not consistent with the ColorSync Preferences dialog which
says the selected default profiles will be used as source for
documents that don't have embedded profiles. This window's phraseology
is at best misleading. What applications use these settings "when a
document does not contain embedded profiles?" It does not also tell
the user "your application must be written to specifically ask
ColorSync what these settings are, or these settings don't do
anything."
Ideally at some point it'd be nice to have a user configurable
equivalent of a system wide working space, I agree from a color geek
standpoint. From a performance standpoint it's a win the way it is now
without being completely unreasonable. Remember, everything is matched
as you'd expect if the image data has an embedded profile.
Under Panther, it seems to have reverted to the behavior associated
with previous versions of Mac OS X.
Frankly I'm not clear to what that was referring unless it's something
in the printing architecture. I tried this also on Jaguar 10.2.7 in
Preview, and found the Generic RGB Profile is not used when opening a
.pdf without an embedded profile. I'll have to investigate that one
and let you all know.
_______________________________________________
colorsync-users mailing list | email@hidden
Help/Unsubscribe/Archives:
http://www.lists.apple.com/mailman/listinfo/colorsync-users
Do not post admin requests to the list. They will be ignored.
| References: | |
| | >Panther, sRGB, web browsers (From: Chris Murphy <email@hidden>) |
| | >Re: Panther, sRGB, web browsers (From: Steve Upton <email@hidden>) |
| | >Re: Panther, sRGB, web browsers (From: Chris Murphy <email@hidden>) |
| | >Re: Panther, sRGB, web browsers (From: John Gnaegy <email@hidden>) |
| | >Re: Panther, sRGB, web browsers (From: John Fieber <email@hidden>) |
| | >Re: Panther, sRGB, web browsers (From: Chris Murphy <email@hidden>) |
| | >Re: Panther, sRGB, web browsers (From: John Fieber <email@hidden>) |
| | >Re: Panther, sRGB, web browsers (From: John Gnaegy <email@hidden>) |
| | >Re: Panther, sRGB, web browsers (From: Grant Symon <email@hidden>) |