Is web-safe colour palette still necessary?
Speed User
colours IIRC). Is this still necessary?
Was this important simply to deal with older PCs and older graphics
cards that were too limited in their colour support? Is it now the case
that there are few of those old, incapable cards and we can assume that
97+% of visitors have cards capable of properly displaying a greater
range of colours?
Tapani Kaijanmäki
Hi,
is
http://hotwired.lycos.com/webmonkey/00/37/index2a.html?tw=design
familiar to you?
There exist no really websafe pallette anymore. (exept
http://hotwired.lycos.com/webmonkey/00/37/stuff2a/complete_websafe_216/reall
ysafe_palette.html)
Why not to demand browser makers expand the value range from
0-3-6-9-C-F to 0-3-6-9-C-F-I-L. That should expand those 22 colors
(twenty-two, thats it all) to significantly higher number. the downside?
well, now white is ffffff but then ffffff would change to 36 % grey. Main
thing would be that every site stays readable during some adaptation time.
(unless if scale were 0-2-4-6-9-B-D-F)
Tapani
Richard Grevers
no...@nospammenow.ca said...
default from their windoze installation (non PNP graphics card). They were most pleasantly
surprised when I managed to crank them up to 16-bit
Ryan
--
Ryan Flynn
Practical web development/design
www.ryanflynn.com
"Speed User" <no...@nospammenow.ca> wrote in message
news:3AAE37EA...@nospammenow.ca...
Alan J. Flavell
> is
> http://hotwired.lycos.com/webmonkey/00/37/index2a.html?tw=design
> familiar to you?
Unfortunately, yes
> There exist no really websafe pallette anymore.
There never were.
> (exept
> http://hotwired.lycos.com/webmonkey/00/37/stuff2a/complete_websafe_216/reall
> ysafe_palette.html)
Nope, not even those, and never were.
> Why not to demand browser makers expand the value range from
> 0-3-6-9-C-F to 0-3-6-9-C-F-I-L.
You can demand all you like. It won't happen.
Better to understand the issues. IMHO.
Speed User
colours that will look reasonable (identical is unnecessary) on various
platforms, without nasty dithering?
Alan J. Flavell
> So what is the "solution", so to speak?
First, define what you understand to be the problem. Users of
non-standard palettes are alreadyu accustomed to dithering on other
people's sites, I don't see any reason to make special arrangements
for them. What you mustn't do is risk what Tom Lane called the
'double whammy' of an author dithering to 'the Netscape palette'
and then the user's browser re-dithering to something else.
> How do we go about selecting
> colours that will look reasonable
By all means design palette-based images to 'the Netscape palette'.
>(identical is unnecessary)
is impossible anyway, in a WWW context
> without nasty dithering?
Most important move is to choose the correct format (JPEG if
appropriate).
See Tom Lane's JPEG FAQ first, and then, if you care what I think,
maybe take a look at
http://ppewww.ph.gla.ac.uk/~flavell/www/palette.html
cheers
Jan Roland Eriksson
comp.infosystems.www.authoring.stylesheets Speed User
<no...@nospammenow.ca> wrote:
You can't, it's not an issue that is "under control" from the publishers
side, as is very little else in an environment based on "arbitrary
documents served to arbitrary clients".
Ever tried to "fight" windmills?
Well, we may have better success rate there :)
--
Roland...
Microsoft is not the answer. Microsoft is the question.
NO! is the answer. - Erik Naggum, Norway.
Matt McIrvin
<no...@nospammenow.ca> wrote:
The available surveys are somewhat unscientific and suffer from sampling
bias, but they can be used to get a ballpark idea.
http://www.upsdell.com/BrowserNews/stat_trends.htm
says that about 5% of users are running 8-bit video and another 5% are
using an AOL browser whose ART compression "feature" similarly restricts
colors (the latter shouldn't affect things other than images, I'd think).
16-bit video is very popular, higher depths less so. (I am in fact using
16 bits right now.) I'd say that we're not *quite* at the 97+% level
yet. In another couple of years we may be past that point.
As others have said, the 216-color palette isn't as inviolate as you might
think; for instance, some Unix boxes tend not to respect it. But it works
OK with a variety of user agents across PCs and Macs running 8-bit video.
The web-safe palette shouldn't, in my opinion, be used *too* rigidly. For
instance, if you've got a photograph or continuous-tone image in JPEG form
that uses more colors, it's best to leave it that way and let the
dithering happen at the user's end than to pre-dither it down to 216
colors (for a variety of reasons). On the other hand, I still use the
216-color palette for text and background colors, and GIFs with large
solid blocks of color that would look bad if dithered.
There are actually other considerations that can be even more important
when selecting colors. For instance, you have to think carefully before
doing subtle low-contrast effects in very dark or light colors, because
colors tend to be darker on PCs and lighter on Macs, and (so I have
learned recently) some LCD screens wash out light pastel colors to solid
white. I've been tempted to use lighter pastels than the 216-color
palette can provide, but some users with high-depth graphics may not be
able to see them anyway.
--
Matt McIrvin http://world.std.com/~mmcirvin/
Speed User
> Well I've found a couple of clients running their PC's at the 16 colours that was the
> default from their windoze installation (non PNP graphics card). They were most pleasantly
> surprised when I managed to crank them up to 16-bit
I know what you mean. I'm always amused when I see discussions that
hinge on respecting the users' defined preferences. While I think that's
a good thing, in practice I find huge numbers of my clients and workshop
attendees are positively astounded when I change their screen resolution
or browser font size to something more comfortable for them.
It seems many users will live with settings they hate, because they
either don't know how to alter them or, and this is also common, they
have no idea even that the settings *can* be changed. So much for users'
"preferences". ;)
Speed User
> As others have said, the 216-color palette isn't as inviolate as you might
> think; for instance, some Unix boxes tend not to respect it. But it works
> OK with a variety of user agents across PCs and Macs running 8-bit video.
That's what I've been thinking. I mean, if you follow that WebMonkey
article you could end up throwing your hands up in the air since nothing
works consistently. I guess for the moment it makes sense to follow the
216-colour palette for backgrounds, text colour definitions and the
like?
> The web-safe palette shouldn't, in my opinion, be used *too* rigidly. For
> instance, if you've got a photograph or continuous-tone image in JPEG form
> that uses more colors, it's best to leave it that way and let the
> dithering happen at the user's end than to pre-dither it down to 216
> colors (for a variety of reasons). On the other hand, I still use the
> 216-color palette for text and background colors, and GIFs with large
> solid blocks of color that would look bad if dithered.
Yes, that's the approach I've been using. If the photo looks good, I
leave it alone. But certainly when setting CSS colour values for
backgrounds, text elements, etc., I select colours from the 216-colour
palette. WebMonkey has me thinking that that approach is actually much
less consistently successful than I had believed, but there doesn't seem
to be anything better at this point.
> colors tend to be darker on PCs and lighter on Macs, and (so I have
> learned recently) some LCD screens wash out light pastel colors to solid
> white.
Yes, I've noticed those. LCD screens in particular tend to provide much
less accurate colour than CRTs, which is why they are not good when
you're preparing a job to go to a print shop and need accurate pantone
colours. My laptop's LCD is very comfortable to work with for long
periods, but its colours are often quite different from the same page
displayed on a CRT.
Speed User
> You can't, it's not an issue that is "under control" from the publishers
> side, as is very little else in an environment based on "arbitrary
> documents served to arbitrary clients".
As I'm finding out, that 216-colour palette is a lot less successful
than I had realized, but I guess it's the best compromise for now?
> Ever tried to "fight" windmills?
> Well, we may have better success rate there :)
Hoo boy, ain't that the truth. Good web design is a myriad of
compromises, including trying to educate our clients on why it's
acceptable that a page looks different on different platforms.
Speed User
> First, define what you understand to be the problem.
Ugly, unacceptable dithering of some colours on some platforms/browsers.
> What you mustn't do is risk what Tom Lane called the
> 'double whammy' of an author dithering to 'the Netscape palette'
> and then the user's browser re-dithering to something else.
I'm not familiar with Tom Lane. Is this an article I should read?
> By all means design palette-based images to 'the Netscape palette'.
Different palette from the 216-colour "web safe" palette?
> >(identical is unnecessary)
>
> is impossible anyway, in a WWW context
Yes, I understand and agree.
> > without nasty dithering?
>
> Most important move is to choose the correct format (JPEG if
> appropriate).
>
> See Tom Lane's JPEG FAQ first,
Sorry, wheres that one?
> http://ppewww.ph.gla.ac.uk/~flavell/www/palette.html
I had a look: interesting, although I need to spend more time with it to
understand it. I am mainly concerned with picking appropriate colours
that will render well for CSS definitions of background, text element
colours, border colours, etc. I'd like to know the "best" palette from
which to choose colours for those uses. The actual image palette
creation I normally leave to a graphic artist, but I am interested to
learn more.
Matt McIrvin
<no...@nospammenow.ca> wrote:
>I know what you mean. I'm always amused when I see discussions that
>hinge on respecting the users' defined preferences. While I think that's
>a good thing, in practice I find huge numbers of my clients and workshop
>attendees are positively astounded when I change their screen resolution
>or browser font size to something more comfortable for them.
This is one of the great annoyances of Web design: trying to deal
simultaneously with the two great truths
1. Some users will change all of the available settings
(and would probably like those changes to be respected).
2. Most users will not change any of the available settings
(and want you to code in a good experience for them anyway).
On comp.sys.www.authoring.stylesheets, there tends to be a vocal
advocacy for the interests of the knowledgeable users who fall into
the first class. It's worth remembering that they are a minority,
albeit one which should not be punished for doing intelligent things
with their browsers.
Alan J. Flavell
(various questions to me)
> "Alan J. Flavell" wrote:
>
> > http://ppewww.ph.gla.ac.uk/~flavell/www/palette.html
>
> I had a look:
but it seems you omitted to notice that it contained answers to all
the questions which you asked me.
good luck
Alan J. Flavell
> The web-safe palette shouldn't, in my opinion, be used *too* rigidly. For
> instance, if you've got a photograph or continuous-tone image in JPEG form
> that uses more colors, it's best to leave it that way and let the
> dithering happen at the user's end than to pre-dither it down to 216
> colors (for a variety of reasons).
This seems to be good practical advice, but based on shaky
fundamentals. JPEG isn't a palette based format, it's designed to
represent continuous-tone images. So it would be entirely perverse
(not merely inadvisable, which is what you seem to be suggesting) to
go trying to force a JPEG to 'the' web-'safe' palette. Basically it
would throw away all the advantages of having used JPEG in the first
place, without gaining anything.
> On the other hand, I still use the
> 216-color palette for text and background colors, and GIFs with large
> solid blocks of color that would look bad if dithered.
I'm not going to say you're wrong making that choice, but I'm
comfortable with using other colours by now. Anyone who is both stuck
with an 8-bit display and a poor browser algorithm should proably by
now have located their browser's 'insist on my colours' menu option.
I'm all for maintaining fall-back-ability, but that doesn't mean I'm
always willing to supply the fallback myself: sometimes I think in the
interests of a better result for better browsers, one is entitled to
expect the reader with an impaired browsing situation to play their
part in the deal. Speaking also as a reader, I know that I'm willing
to do that (e.g when on a b/w X terminal display, which I just
occasionally need to use). I wouldn't expect web authors to author in
black/white just for me.
cheers
Philipp Lenssen
mmcirvin-130...@adsl-151-203-70-202.bostma.adsl.bellatlantic.net..
.
>..
> 1. Some users will change all of the available settings
> (and would probably like those changes to be respected).
Well, sometimes they just force those changes to be respected; with a good
browser you can override text-size, colors, and so on to fit your needs. Of
course it's a lot harder to force really bad pages (say, a big table layout
with un-alted images of text, and Flash content inbetween) into becoming
meaningful in different contexts.
(As for the topic, the only place where anyone can expect to find a
"web-safe color palette" is urbanlegends dot com.)
Tapani Kaijanmäki
> There never were.
> Nope, not even those, and never were.
Did not check exact values on spreadsheet!
> > Why not to demand browser makers expand the value range
> You can demand all you like. It won't happen.
> Better to understand the issues. IMHO.
What issues do you mean? There is only two major browsers. That makes only
two identical emails to send and tell what we wish. Most propably nothing
happens but at least we tried...
Tapani
Jeff Thies
> 216-colour palette for backgrounds, text colour definitions and the
> like?
html colors are not dithered. From what I've seen the browser picks what
it thinks the nearest color will be. The color chosen will vary from
browser to browser.
Only worry about web-safe palette in GIF's.
Jeff
Speed User
>
> > I guess for the moment it makes sense to follow the
> > 216-colour palette for backgrounds, text colour definitions and the
> > like?
>
> html colors are not dithered. From what I've seen the browser picks what
> it thinks the nearest color will be. The color chosen will vary from
> browser to browser.
Thank you! That is very useful, practical information. I didn't know
that. So that means that I should be able to choose any RGB or Hex
colour values for backgrounds, text, borders, etc. without worry. While
the result will not be identical on every browser, it will look okay.
Thanks for that.
> Only worry about web-safe palette in GIF's.
Okay, but then as people are pointing out a web-safe palette for GIFs
really doesn't exist ?
John A. Lewis
what
: it thinks the nearest color will be. The color chosen will vary from
: browser to browser.
HTML background colors are dithered on my computer at 16 bit color.
However, this isn't a problem, & my computer doesn't dither HTML
background colors at 8 bit color (it chooses what it thinks the closest
color is).
Just to clarify.
John
Jeff Thies
I'm confused here. Do you mean 16bit, or 16 color (4 bit)? I hadn't seen
this at 16bit, but admittedly it would take good eyes!
Jeff
(I haven't checked 4bit and didn't mention that because web safe colors
are greater than 4 bit)
Jeff Thies
>
> Okay, but then as people are pointing out a web-safe palette for GIFs
> really doesn't exist ?
Now, you are asking the wrong person!
Cheers,
Jeff
John C. Ring, Jr.
>> html colors are not dithered. From what I've seen the browser picks what
>> it thinks the nearest color will be. The color chosen will vary from
>> browser to browser.
>
>Thank you! That is very useful, practical information. I didn't know
>that. So that means that I should be able to choose any RGB or Hex
>colour values for backgrounds, text, borders, etc. without worry. While
>the result will not be identical on every browser, it will look okay.
Well... I'd imagine that the nearest colors *could* end up as colors that
don't look very good together, even the the original colors mixed quite well.
This is only a guess on my part and it may be somewhat unlikely in practice,
although it certainly seems possible.
Worst case would be if the client's machine ended up picking the same color as
the "closest" for both your background and text color selections. But I'd
tend to think that, providing you chose colors which themselves aren't too bad
for readability, i.e. you didn't choose navy over light blue, this wouldn't
happen.
Isn't color fun :)
Matt McIrvin
"Alan J. Flavell" <fla...@mail.cern.ch> wrote:
>This seems to be good practical advice, but based on shaky
>fundamentals. JPEG isn't a palette based format, it's designed to
>represent continuous-tone images. So it would be entirely perverse
>(not merely inadvisable, which is what you seem to be suggesting) to
>go trying to force a JPEG to 'the' web-'safe' palette. Basically it
>would throw away all the advantages of having used JPEG in the first
>place, without gaining anything.
Quite true. What I was thinking of was more the possibility of
turning a photographic JPEG into a pre-dithered GIF or PNG just to
stay within the 216-color palette. This is still a bad idea, but
it's not immediately obvious that it is.
>Anyone who is both stuck
>with an 8-bit display and a poor browser algorithm should proably by
>now have located their browser's 'insist on my colours' menu option.
Maybe the key piece of information is not how many people still have
a certain hardware configuration, but how likely it is that such
people know what they are doing and will have software set to adapt
gracefully.
John A. Lewis
seen
: this at 16bit, but admittedly it would take good eyes!
16 bit, i.e., "thousands of colors." It doesn't matter unless you need a
color to match exactly. The difference between the dithered colors is
small; the dithering mimics the intended color well. The first time I
noticed this I was cropping a zoomed-in screenshot of a web page, but if
I look very closely (esp. at a low resolution), I can see the dithering
with my own eyes.
John
Jeff Thies
> don't look very good together, even the the original colors mixed quite well.
> This is only a guess on my part and it may be somewhat unlikely in practice,
> although it certainly seems possible.
It's not that bad! It certainly isn't worth using the limited palette of
web safe. The colors will always be close to what you expect.
If want to see this, you can set your machine for 256 colors and try my
color picker (IE5,NS4,NS6, maybe IE4):
http://thelimit.com/play/pop_picker.htm
(I use this to set font colors, background colors, any css style... If
you want to see that, email me privately)
Cheers,
Jeff
>
> Worst case would be if the client's machine ended up picking the same color as
> the "closest" for both your background and text color selections. But I'd
> tend to think that, providing you chose colors which themselves aren't too bad
> for readability, i.e. you didn't choose navy over light blue, this wouldn't
> happen.
Well, you shouldn't be doing that anyways!
Stewart Gordon
> colours IIRC). Is this still necessary?
>
Just throwing in my threepence worth....
Since I redesigned my site just before this year I tried to go with the
websafe colours. But I've found one or two drawbacks, like there doesn't
seem to be a decent brown! I've ended up experimenting with a few colours
that are, like, halfway between two websafe colours.
Whoever invented the so-called websafe palette should've at least tried to
give it some gamut. Like, the difference between 00 and 33 is invisible, at
least on my monitor!
Stewart.
--
My e-mail is valid but not my primary mailbox. Please keep replies on the
'group where everyone may benefit.
Tapani Kaijanmäki
"Claus Färber" <claus-usen...@faerber.muc.de> kirjoitti
Tapani Kaijanmäki <ta...@kanetti.com> schrieb/wrote:
> Why not to demand browser makers expand the value range from 0-
> 3-6-9-C-F to 0-3-6-9-C-F-I-L.
This is the most stupid suggestion I've ever heard. You're
basically suggesting to extend the value range of a byte from
0..255 to 0..483 (i.e. from 8 bits to ~8.92 bits) just to be able
to continue not using the full value range. But this won't make
old graphics cards display more than 256 colours either.
If you think that old graphics cards are not a issue, than you can
just use the full range of hex digits from 0 to F. You'll have
about 16.7 million colurs then.
Claus
--
http://www.faerber.muc.de
I was bit over-enthusiastic, wrote one letter too much. My idea was to give
a midpoint-values, like 0bit-2-4bit-6-8bit or 0-bit-4-8bit-12-16bit or
0bit-6-12bit-18-24bit. But thats only 5 value 125 color solution, so that
enthusiasm in my original respond to Speed User just earned some notes. Now
I will always know scale 0-F means hex code numbers. Thank you.
Tapani
Alan J. Flavell
> >This seems to be good practical advice, but based on shaky
> >fundamentals.
[...]
> Quite true. What I was thinking of was more the possibility of
> turning a photographic JPEG into a pre-dithered GIF or PNG just to
> stay within the 216-color palette. This is still a bad idea, but
> it's not immediately obvious that it is.
Right. I think I see what you were aiming at now; my apologies for
the starkly-worded comment.
> >Anyone who is both stuck
> >with an 8-bit display and a poor browser algorithm should proably by
> >now have located their browser's 'insist on my colours' menu option.
>
> Maybe the key piece of information is not how many people still have
> a certain hardware configuration, but how likely it is that such
> people know what they are doing and will have software set to adapt
> gracefully.
The point I was trying to make there is that www readers spend most of
their time on other people's pages, where readers who are in the
position we're discussing will be getting results at least as bad, or
worse, than what they're getting with your pages. So, either they've
already found out how to address that problem, or at least they'll be
no worse off with your web site than they are with all the others
which they visit. In short, as long as you avoid doing anything to
make matters spectacularly worse for them (specifically: dithering
GIFs* to "the" so-called web-safe palette, leading to Tom Lane's
"double whammy") then I'd say you're playing your part in the "www
bargain", and then it's up to them.
*or indexed PNG.
Alan J. Flavell
> What issues do you mean? There is only two major browsers.
If it amuses you to reach that conclusion, I won't bother to argue
with it, but the fact is, those "two major browsers" (I'm assuming you
mean Win MSIE 5.* and that undead zombie Win Netscape 4.*) are used
across a very wide range of settings. In this case their behaviour in
respect of different colour-depths is what we are looking at. And
remember they may (and in fact often do) behave differently as between
background colours, background images, in-lined images, text colours,
and images that are viewed directly i.e not in-lined in HTML.
There's no magic bullet that definitely works best under all
circumstances, but I reckon that if you understand the properties, and
thus the strengths and weaknesses, of what you're working with, you
can make designs that optimise the likelihood of working well in a
wide range of viewing situations, including browsers/versions that
haven't yet been released, or which you've never met.
> That makes only
> two identical emails to send and tell what we wish. Most propably nothing
> happens but at least we tried...
Even if by some miracle they actually did what you demanded, it
certainly won't make any difference to the browser versions that are
already deployed.
I don't see anything so dreadfully wrong with the practical advice I
was offering on my page. In short, understand the properties of
palette-based (indexed) image formats i.e GIF and indexed PNG, and the
properties of JPEGs, choose the right one for your purpose, and design
to its strengths, taking care about its weaknesses.
(More detail at http://ppewww.ph.gla.ac.uk/~flavell/www/palette.html
as already mentioned).
On the other hand if you use techniques that rely on specific bugs in
your Big Two favourite browser versions, it won't be long before
you're scurrying off to remake your images as a new version becomes
popular.
NB this is the technical side of things. I am not a graphic designer,
and not offering any advice on graphic design - I'm doing nothing more
than call attention to some technical features of image representation
for use on the www - use the advice, or not, as you will.
Matt McIrvin
<sm...@my-deja.com> wrote:
>Whoever invented the so-called websafe palette should've at least tried to
>give it some gamut. Like, the difference between 00 and 33 is invisible, at
>least on my monitor!
The palette was not really invented by anybody. It arose more or less by
accident. It was an observation, made several years ago, about what
colors tended to be common to the default 8-bit color palettes used by
common Web browsers under Windows and MacOS (which I think depends more on
the system defaults than on the browsers-- the browser *could* re-map the
colors, but this would result in strange color shifts when changing
applications). The crucial observation seems to have been made by Lynda
Weinman, back in the early days of visual Web design; she claims to have
coined the phrase "browser-safe palette."
Those 216 colors are the crudest arrangement imaginable for an 8-bit color
palette designed to be of general utility: just a regular Cartesian
division of the red-green-blue cube into a 6x6x6 grid. Both Windows and
MacOS reserve the other 40 colors for more carefully targeted OS-specific
purposes, so they vary between the two operating systems.
Furthermore they are not really the same across platforms (or across
monitors, or video cards, or monitor settings...)
As you've noticed, Windows PCs typically have no built-in gamma correction
and the dark colors end up very dark. This doesn't have anything to do
with Web browsers per se, it's more the choices made when designing the
operating system. Macs have built-in gamma correction, which makes all
the colors look somewhat lighter than they do on PCs, especially at the
dark end of the range.
Alan J. Flavell
> The palette was not really invented by anybody.
You're talking about the specific palette commonly known as "the
Netscape web-safe palette"? Then I'm going to have to disagree with
you, though I don't know who to accuse of having "invented" it.
Someone on the NCSA Mosaic team, perhaps?
> It arose more or less by accident.
Well, I'm sure it's true that it was conceived without due
consideration for all the relevant issues; but conceived, by somebody,
it surely was. The evidence is that it must have been a programmer
rather than a colour expert.
> It was an observation, made several years ago, about what
> colors tended to be common to the default 8-bit color palettes used by
> common Web browsers under Windows and MacOS
A purely mathematical exercise, nothing to do with observations. As
you indeed show later in your posting.
> The crucial observation seems to have been made by Lynda
> Weinman, back in the early days of visual Web design; she claims to have
> coined the phrase "browser-safe palette."
But that was only because the browsers were already doing it. So
don't put the observation before the "design".
> Those 216 colors are the crudest arrangement imaginable for an 8-bit color
> palette designed to be of general utility: just a regular Cartesian
> division of the red-green-blue cube into a 6x6x6 grid.
Quite, and makes no allowance for the properties of visual perception.
As I say: a purely mathematical exercise. As you say: "the crudest
arrangement". It would have been better to distribute them more
uniformly in the perception space, rather than uniformly in raw RGB
values.
> Furthermore they are not really the same across platforms (or across
> monitors, or video cards, or monitor settings...)
That's true, but that's quite a separate issue.
> As you've noticed, Windows PCs typically have no built-in gamma correction
> and the dark colors end up very dark.
Or as I've noticed, Mac browsers typically have over-correction, and
colours end up looking very light. There's no absolute standard of
correctness: it's a difference of conventions over what to use as
reference gamma. Which is why PNG format includes provision for a
nominal gamma to be specified, and the display system is supposed to
correct for that.
Tom Lane's JPEG FAQ has pointers to further resources on this topic,
e.g Poynton's gamma FAQ.
all the best
Speed User
> > Those 216 colors are the crudest arrangement imaginable for an 8-bit color
> > palette designed to be of general utility: just a regular Cartesian
> > division of the red-green-blue cube into a 6x6x6 grid.
>
> Quite, and makes no allowance for the properties of visual perception.
> As I say: a purely mathematical exercise. As you say: "the crudest
> arrangement". It would have been better to distribute them more
> uniformly in the perception space, rather than uniformly in raw RGB
> values.
Well, that is but one possible arrangement. Fortunately there are many
sites that have created alternate arrangements of the same palette,
arrangements that are more useful and appealing by organizing according
to hue or position in the colour spectrum or whatever. Those are the
sites I tended to use when selecting colours from the safe palette.
Amit Patel
> > HTML background colors are dithered on my computer at 16 bit color.
>
> I'm confused here. Do you mean 16bit, or 16 color (4 bit)? I hadn't seen
> this at 16bit, but admittedly it would take good eyes!
I can definitely notice it at 16bit. I have an image with a specific
gray (#f0f0f0) and I use the very same gray as my background color.
They match at work (24 bit color) but I can see the difference at
home (16 bit color). 16 bit color probably varies from the 555 to the
565 variety, too. Fortunately it's for an internal web page with
only three users, so I don't care. :-)
- Amit
--
Amit J Patel, Computer Science Department, Stanford University
http://www-cs-students.stanford.edu/~amitp/
``Parkinson's Other Law: Perfection is achieved only
at the point of collapse.''
Jeff Thies
> >
> > I'm confused here. Do you mean 16bit, or 16 color (4 bit)? I hadn't seen
> > this at 16bit, but admittedly it would take good eyes!
>
> I can definitely notice it at 16bit. I have an image with a specific
> gray (#f0f0f0) and I use the very same gray as my background color.
> They match at work (24 bit color) but I can see the difference at
> home (16 bit color).
Actually, the reference was whether 16 bit colors are dithered...
> 16 bit color probably varies from the 555 to the
> 565 variety, too.
Maybe we need a 16K color safe palette!
Jeff
Amit Patel
> > > > HTML background colors are dithered on my computer at 16 bit color.
> > >
> > > I'm confused here. Do you mean 16bit, or 16 color (4 bit)? I hadn't seen
> > > this at 16bit, but admittedly it would take good eyes!
> >
> > I can definitely notice it at 16bit. I have an image with a specific
> > gray (#f0f0f0) and I use the very same gray as my background color.
> > They match at work (24 bit color) but I can see the difference at
> > home (16 bit color).
>
> Actually, the reference was whether 16 bit colors are dithered...
Oops, sorry, I forgot about that part. The 'difference' I was
referring to was dithered vs. not dithered. Oops oops. Anyway the
image got dithered and the background did not, so I could see the
difference between the two. :-(