[thelist] "muddy" images

Carol Stein techwatcher at accesswriters.com
Sun Mar 23 03:18:18 CST 2003


Hi, Amy --

> I'm working on Windows XP home designing a website for a customer who
views it on a Mac with AOL.  The images I develop look 
> fine to me, but when viewed on her machine she says they look "muddy"
(basically the background appears blotchy).  
> My first question is if anyone has ever heard of this issue and what they
found to resolve it.  
Two things: First, follow the instructions of those responders who know
about AOL's image compression hack (kludge?). Second, be aware that if
you're using a "Web-safe palette" -- you're using a gif, or indexed color
-- the PC has a DIFFERENT set of system (and thus "Web-safe") colors than
the Mac. This can cause problems, especially if you have "dithered" colors
ON. You're better off using JPG to avoid this sort of problem, if you can.

> My second question is what is the best way to scan (I am working with
images at 72dpi)  I get the images at higher resolutions 
> from her and I size them down.  She is scanning objects off of the
scanner, not images.
As to scanning, have your client scan images at 96dpi, and keep them at
that resolution. The exception is if your client wants her Web viewers to
be able to PRINT downloaded images -- then you should have her scan them at
(at least) 300dpi, and you should reduce them to about 96dpi for the
thumbnail and screen view, but have the much larger file available as a
download for a print-only version.

Please note that if you do reduce the resolution on an image, PhotoShop
(and presumably other programs) offer various ways of "sampling" the
original pixels to come up with the pixel value for the new image. Again,
if you're working with indexed color (GIF), you want to make the sure new
pixel is NOT some calculated part of other pixels. So, choose some option
like "nearest neighbor" instead of some option like "bicubic."

If any of this doesn't make sense, drop me a line.

Cheers --
Carol




More information about the thelist mailing list