Hi Alfonso,
Don't worry too much about having people set their colour preferences
differently. Most people will already have their colour preferences
set to full colour, and if they don't then they will have no reason to
complain about differences in the shell/image colour.
A couple of things that I neglected to mention in my previous message:
I examined the image on a pixel level and found that the text was all
in grayscale (white with anti-aliasing for smoothing) The colours were
all accurate in the text and balanced perfectly for no colour shift --
255,255,255 -- 246,246,246 etc. The pixels in the shell image all were
weighted slightly into the red area. This proves that the problem is
in the file. Monitors might depict colours differently, but the colour
measurements of the file will stay the same regardless of screen
resolution/colour shift problems.
I suspect that the image was processed more than once while in jpg
format, and that one of the last steps was the labelling of the image.
Jpegs can shift toward the red of RGB through processing and multiple
saving. This is a little known effect. To avoid this problem do as
much image processing as possible in either .bmp or .tga format. Only
as a last step should you convert the image to .jpg so that it can be
seen on the web. For emailing images, you can ask if the person would
like the most colour accurate version although the file size will be
larger.
Tga (Targa) is a a television format not often used by graphic
artists. It is very stable and can exist in a compressed form. When
uncompressed it regains its original colour fidelity and is the same
weight as a bitmap file.
For a useful jpg processing tool you can download:
http://www.writer2001.com/lviewpro.zip
Make sure that you download it only from my site as other files on the
web are of later versions, but this one is the very best! Later
version are "dumbed down" and while easier, are less effective and
professional
To process a jpg for colour fidelity: first open the file as a bitmap
or targa (before it ever becomes a jpeg)
Then: File/Properties/JPEG(Normal)
Set DCT Method to floating point
Set Compression quality to 95%
For smaller file size with less accurate colour gradation, but still
with an overall accurate colour, set the compression quality to about
45% to 55% (experiment with the results) This will give you a very
small file size.
You will not see the results until you reopen the file, so always
rename the file so you do not lose the original when working from
jpegs as a source.
Floating point means that the file will average colours to a close
consistent colour. For example if you have adjacent pixels that are,
say, 165,164,165 and 165,165,165, it might change the 164 to 165. This
will reduce the file size while preserving the visual appearance of
the colour. The higher the compression number (in this program) the
more accurate the colour and the less colour averaging. The
compression method limits the number of colours and saves file size
thus.
All jpeg settings in all programs do this differently, and some are
not very good.
Computers handle colours in an arithmetical manner, but colours and
colour perception do not behave this way.
We design colour suites and one project involved creating maps for
emergency evacuations. The colours were designed so that the maps
could be seen at a distance, and the colours created a calming effect
psychologically. It worked on everyone. When colours are handled with
skill they can create the same mental effects through the eyes that
music can through the ears.
If someone wants to buy a shell after seeing its image, it might be
that the particular colour balance resonates with that person's
psychology, or people's psychology generally. This creates an
overwhelming desirability. If they get the shell and find that the
colours are very different, they will be unpleased with their purchase
to some degree. It is all variable according to lighting conditions
etc. Most people will lean toward a higher saturation than is natural,
and find this desirable. You can see how many people have intense
colour on their television sets. So it is a balancing act to maintain
an accurate colour match while making a desirable picture that does
not disappoint when someone sees the real object!
Good luck!
Cheers,
John
> Alfonso Pina wrote:
>
> Dear friends,
>
> First at all, thank you very much for your overwhelming response:
> impossible to acknowledge personally all the people who answer my
> question, so I would like to thanks them all from here.
>
> Also, I have to acknowledge a lot of advices and tips that some
> people has been so kind to tell me and to post in the list, I am
> specially curious about the system that John Hooker propose, and I
> think I will try it, though I have some doubts about his practicity
> because probably most people would prefer to risk a little cast than
> to go changing his display settings continuously, but the method
> looks promising.
>
> And now to the results. Obviously they has a limited validity, but
> can result indicative.
>
> - 80% says pinkish, 20% whitish
> - "Pinkish" answers goes from "very slight pinkish" to "reddish",
> with the mode at "slight pink"
> - Colour vision is independent of geographic zone or origin of the
> equipment (what a pity, I liked my theory...) Percentages remains
> consistent in America, Europe and Australia.
> - It is also independent of the computer used: percentages are the
> same for Mac and PC users.
> - There may be little variations depending of browsers, IExplorer vs
> Netscape, but they don't seem significatives
> - Not enough data about different OS, but there are enough variation
> inside Mac OS and Windows to assume this is not important.
>
> Conclusion: As some of you has already pointed out, the main source
> of difference may be the personal display settings of monitor and
> graphic card (and I should revise mine...)
>
> And another conclusion, next time I think colors from somebody's
> pictures are not accurated, I will revise my settings before to
> complaint.
>
> Though probably not useful, I hope you have found all this a little
> interesting.
> Best regards and thanks again,
>
> Alfonso Pina
> Málaga, Spain
> [log in to unmask]
> www.eumed.net/malakos
--
http://www.writer2001.com/
Hooker & Perron, Total Project Coordination
Database-Web...Graphics...Custom Maps...Colour Suites...Expert Systems
Building the Celtic Coin Index on the Web:
http://www.writer2001.com/cciwriter2001/
|
