This is Interesting: Free Magazines for Graphics designers and webmasters
Home > Archive > Computer Graphics with Photoshop > December 2005 > Saved images have different color than when in PS
You are viewing an archived Text-only version of the thread.
To view this thread in it's original format and/or if you want to reply to
this thread please [click here]
| Author |
Saved images have different color than when in PS
|
|
| Jon Danniken 2005-12-22, 3:14 am |
| Hello,
When I was making yet another futile attempt at calibrating my monitor, I
caused an unwanted behavior. Specifically, when I save an image in PS (CS),
the saved copy is lighter in color (higher gamma?) than the image was when
in photoshop itself.
This is recent, and I caused it yesterday when I was messing with the color
settings. I have disabled Adobe Gamma loader from starting at bootup, and
tried different color settings, but this is still occuring.
What is the setting that is causing this behavior?
Thanks,
Jon
| |
| Jon Danniken 2005-12-22, 3:14 am |
| "Jon Danniken"
> Hello,
>
> When I was making yet another futile attempt at calibrating my monitor, I
> caused an unwanted behavior. Specifically, when I save an image in PS
(CS),
> the saved copy is lighter in color (higher gamma?) than the image was when
> in photoshop itself.
>
> This is recent, and I caused it yesterday when I was messing with the
color
> settings. I have disabled Adobe Gamma loader from starting at bootup, and
> tried different color settings, but this is still occuring.
>
> What is the setting that is causing this behavior?
Here is an example of what I am talking about; sorry I didn't include it in
the original:
http://img369.imageshack.us/img369/2682/pscolor8uy.jpg
The image on the left is how it looks in PS, the one the right is how it
looks when I go to save it.
Jon
| |
| Mike Russell 2005-12-22, 3:14 am |
| "Jon Danniken" <jonREMOVETHISdanniken@yahoo.com> wrote in message
news:40un9qF1c9h32U1@individual.net...
> Hello,
>
> When I was making yet another futile attempt at calibrating my monitor, I
> caused an unwanted behavior. Specifically, when I save an image in PS
> (CS),
> the saved copy is lighter in color (higher gamma?) than the image was when
> in photoshop itself.
>
> This is recent, and I caused it yesterday when I was messing with the
> color
> settings. I have disabled Adobe Gamma loader from starting at bootup, and
> tried different color settings, but this is still occuring.
>
> What is the setting that is causing this behavior?
Probably the Desired Gamma setting. Use "Windows Default" to set a gamma of
2.2.
--
Mike Russell
www.curvemeister.com
| |
| Jon Danniken 2005-12-22, 6:14 am |
| "Mike Russell" wrote:
> "Jon Danniken" wrote:
I[color=darkred]
when[color=darkred]
and[color=darkred]
>
> Probably the Desired Gamma setting. Use "Windows Default" to set a gamma
of
> 2.2.
Thanks, Mike, but the problem occurs when I *do* use the 2.2 gamma (or any
other "user specifiec" gamma setting through Adobe Gamma).. Only when I go
into system32/spool/color/ and delete the *.icc I created does this problem
go away.
I like the idea of setting up my monitor to a calibrated space, but why when
I do this is my output different than what I am working on? In other words,
why would I want to work on something in PS only to have the levels flip
all over the place when it is output?
Confused,
Jon
| |
| Mike Russell 2005-12-22, 6:14 am |
| "Jon Danniken" <jonREMOVETHISdanniken@yahoo.com> wrote in message
....
> Thanks, Mike, but the problem occurs when I *do* use the 2.2 gamma (or any
> other "user specifiec" gamma setting through Adobe Gamma)..
Here's another possibility: you may have your RGB working space's gamma set
to something other than 2.2. Type ctrl-shift-K and look at your RGB Working
space. Set it to sRGB for the time being.
> Only when I gointo system32/spool/color/ and delete the *.icc I created
> does this problem go away.
I haven't tried this, but I would expect Photoshop would behave as if it
were a non color-aware app.
> I like the idea of setting up my monitor to a calibrated space, but why
> when
> I do this is my output different than what I am working on? In other
> words,
> why would I want to work on something in PS only to have the levels flip
> all over the place when it is output?
You have the logic backwards. The levels are being changed systematically
so that the appearance is the same, if (and only if) you have an accurate
monitor profile. Before Photoshop knew about working color spaces, the same
image would look lighter on the Mac than on the PC. Now that Photoshop is
color aware, the image can be converted from the working space to the
display space, resulting in the same image appearance on the PC and the Mac.
This works, but programs other than Photoshop have failed to follow suit,
for a variety of good reasons. The result is that many images look
different in Photoshop than they do in non color managed applications. This
is only a problem if your images are destined for the web, or if you are
going to share your RGB images with others and there is a chance they may
use them outside of Photoshop.
--
Mike Russell
www.curvemeister.com
| |
| Bill Hilton 2005-12-22, 6:20 pm |
| > Jon writes ...
>
>when I save an image in PS (CS), the saved copy is lighter
>in color (higher gamma?) than the image was when in photoshop
>itself.
I've helped several people track down problems like this and in *every*
case so far it was caused by a bad (inaccurate) monitor profile.
Regenerate the profile from scratch and see if that fixes it.
> Only when I go into system32/spool/color/ and delete the *.icc I
> created does this problem go away.
Just generate a new profile.
Bill
| |
|
|
"Bill Hilton" <bhilton665@aol.com> wrote in message
news:1135268611.624713.287720@f14g2000cwb.googlegroups.com...
>
> I've helped several people track down problems like this and in *every*
> case so far it was caused by a bad (inaccurate) monitor profile.
>
> Regenerate the profile from scratch and see if that fixes it.
>
>
> Just generate a new profile.
>
> Bill
>
Furthermore, the fact that the problem goes away after you have deleted the
icc file is further indication that the profile is bad.
And, previously, the OP said that he delete Adobe Gamma Loader. The purpose
of Adobe Gamma Loader is to load the custom profile into the graphics
driver. Hence, if you delete the loader, it cannot load the profile thus
you will get a problem where the image looks good until you exit PS.
I use Monaco because I really don't trust my eyesight to get a good profile.
This was especially the case before I had cataract surgery.
Jim
| |
| Jon Danniken 2005-12-22, 10:14 pm |
| "Mike Russell" wrote:
> "Jon Danniken" wrote in message
> ...
any[color=darkred]
>
> Here's another possibility: you may have your RGB working space's gamma
set
> to something other than 2.2. Type ctrl-shift-K and look at your RGB
Working
> space. Set it to sRGB for the time being.
Thanks Mike; it is set to sRGB IEC 61966-2.1, but I still have the issue of
the image looking different when I go to save it.
I realize I am missing out on a crucial (and likely very simple) concept
here, and I apologize for my stupidity.
Regardless of how my monitor is set up, I do not understand *why* Photoshop
changes the appearance of an image
*while still in Photoshop*, when it is viewed in the saving window.
I do not understand why it is that Photoshop wants to change the levels of
an
image when it is saved, and why it doesn't save it looking the same way that
I see it while I am working on it.
Hopefully it will make sense at some point, but right now it just isn't.
Thanks though,
Jon
| |
| Mike Russell 2005-12-23, 3:14 am |
| From: "Jon Danniken" <jonREMOVETHISdanniken@yahoo.com>
[re image being darker outside of Photoshop]
> Thanks Mike; it is set to sRGB IEC 61966-2.1, but I still have the issue
> of
> the image looking different when I go to save it.
>
> I realize I am missing out on a crucial (and likely very simple) concept
> here, and I apologize for my stupidity.
You're obviously not stupid. At some point in the future, not many years
away, color management will be dirt simple. Right now it's not, and there
is a large amount of frustration out there. In fact, it's a real mess, and
there are about 32 different ways to set it up incorrectly - you have simply
found one of them.
> Regardless of how my monitor is set up, I do not understand *why*
> Photoshop
> changes the appearance of an image *while still in Photoshop*, when it is
> viewed
> in the saving window.
Photoshop is designed to display the image in a consisistent way on all
monitors. To accomplish this, you must have a correct monitor profile, and
the colors must be altered sightly to compensate for your monitor.
> I do not understand why it is that Photoshop wants to change the levels of
> an image when it is saved, and why it doesn't save it looking the same
> way that
> I see it while I am working on it.
You can use Photoshop this way if you disable color management. Select
"Color Management Off" in the Color Settings dialog. There is nothing wrong
with this, except that you will be swimming hard against the tide of
convention. Up to, but not including version 5.0 this was how everyone used
Photoshop. Although you can do high quality work this way, you may run into
problems when you share your images with others. Much excellent work was
done before color management - a large fraction of all digital publishing,
and no one complained.
> Hopefully it will make sense at some point, but right now it just isn't.
OK - here's another thing to try. You may have saved on top of your sRGB
profile, when you used Adobe Gamma to save your screen profile. Download a
fresh copy from Adobe,
http://www.adobe.com/support/downlo....jsp?ftpID=3148
Then install it in the color folder, and see if things start behaving
themselves.
On my system the color folder is C:\WINDOWS\system32\spool\drivers\color,
YMMV.
Mike Russell
www.curvemeister.com
| |
| rusure@att.net 2005-12-23, 6:14 pm |
| Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Lines: 31
Date: Fri, 23 Dec 2005 13:57:50 GMT
NNTP-Posting-Host: 12.76.192.231
X-Complaints-To: abuse@worldnet.att.net
X-Trace: bgtnsc05-news.ops.worldnet.att.net 1135346270 12.76.192.231 (Fri, 23 Dec 2005 13:57:50 GMT)
NNTP-Posting-Date: Fri, 23 Dec 2005 13:57:50 GMT
Organization: AT&T Worldnet
Xref: number1.nntp.dca.giganews.com comp.graphics.apps.photoshop:305021
Mike Russell wrote:
>
> From: "Jon Danniken" <jonREMOVETHISdanniken@yahoo.com>
> [re image being darker outside of Photoshop]
>
>
> You're obviously not stupid. At some point in the future, not many years
> away, color management will be dirt simple. Right now it's not, and there
> is a large amount of frustration out there. In fact, it's a real mess, and
> there are about 32 different ways to set it up incorrectly - you have simply
> found one of them.
You are overly optimistic. If (not when) color management should become
"dirt simple" (and dirt cheap), many of us would either have gone broke
(from all $$$$$ we wasted on it), or have been locked up in a mental
institution (from the frustration of trying).
Here's my take. There are technical challenges to get color management
right, simple and cheap. And many digital imagers do not need it to
begin with. But Adobe and many vendors not only DELIBERATELY turn color
management into a nightmare, but also try to convince every user that
they need it. Why? To keep revising their products to make ONLY
incremental changes to bait the users. It is their business model to get
the users to keep "upgrading" the sw and hw, and keep wasting consumable
in making test prints.
| |
|
| <rusure@att.net> wrote in message news:43AC018B.3D88A3DF@att.net...
>
> You are overly optimistic. If (not when) color management should become
> "dirt simple" (and dirt cheap), many of us would either have gone broke
> (from all $$$$$ we wasted on it), or have been locked up in a mental
> institution (from the frustration of trying).
>
> Here's my take. There are technical challenges to get color management
> right, simple and cheap. And many digital imagers do not need it to
> begin with. But Adobe and many vendors not only DELIBERATELY turn color
> management into a nightmare, but also try to convince every user that
> they need it. Why? To keep revising their products to make ONLY
> incremental changes to bait the users. It is their business model to get
> the users to keep "upgrading" the sw and hw, and keep wasting consumable
> in making test prints.
Of course, when you say 'Colour Management', you mean 'ICC Colour
Management'. Everyone uses colour management of a sort, it's just that in
the past, each company had their own 'sealed system' - admittedly an
empirical one but a colour management system nevertheless. They knew exactly
what they would get from their own originated material and therefore could
produce excellent results without ICC. That was all well and good when the
original material is on film or hard copy and you can do a direct colour
match between original and output. However, what happens when the original
material is digital capture by an external customer? This is becoming
increasingly the norm. How does such a customer convey correct colour
meaning to the service provider? Without a universal 'colour language',
there isn't a hope in hell that the customer will get anything like
predictable colour. ICC colour management is one such language, perhaps it
is not the best but at the moment it is the only choice for most people.
--
John
Replace 'nospam' with 'todnet' when replying.
| |
| Bill Hilton 2005-12-23, 6:14 pm |
| > Jon Danniken writes ...
>
>I do not understand *why* Photoshop changes the appearance
>of an image *while still in Photoshop*, when it is viewed in the
>saving window.
The 'saving window' you apparently refer to is 'save for web', which
opens ImageReady (based on the link you provided) ... you were saving
as a jpeg and 'save for web' shows you what the image will look like ON
THE WEB, where there is no color management, that is, the ICC working
space profile is ignored for almost all browsers (if you even assign
one) AND the monitor ICC profile is ignored. So basically it's not
"the appearance of an image *while still in Photoshop*" that you're
looking at, it's how that jpeg will look when viewed outside Photoshop
in a non-color managed application like a web browser.
All this is saying is that you have a very bad monitor profile, since
you shouldn't be seeing THAT big a difference if you first convert the
file to sRGB before making it a jpeg. Make a new monitor profile
that's accurate and understand what ImageReady is doing when you 'save
for the web' and the problem is solved.
> I do not understand why it is that Photoshop wants to change the levels
>of an image when it is saved.
It doesn't 'change the levels' at all, it just shows what the image
will look like without profiles when you are saving a jpeg using 'save
for the web'.
>Hopefully it will make sense at some point, but right now it just isn't.
Hope this helps ...
Bill
| |
| Bill Hilton 2005-12-23, 6:14 pm |
| > Jon Danniken writes ...
>
>I do not understand *why* Photoshop changes the appearance
>of an image *while still in Photoshop*, when it is viewed in the
>saving window.
As I said moments ago, it's because you're saving as a jpeg using 'save
for web', which opens ImageReady ... one way to see what's really going
on is to do this 'save for web' and check the 'icc profile' box to keep
the working space profile, then open this jpeg up in Photoshop and make
sure this profile is still assigned (if it's different than the default
you'll get a dialog box asking if you want to drop it, keep it or
change to the working space).
As you'll see, re-opened in Photoshop with the initial working space
profile it should now look pretty much identical to the original image
you had before you did the save. Basically this just verifies that
what you saw in the ImageReady window when you did 'save for web' was
showing you what the image looks like OUTSIDE Photoshop when viewed in
a non-color managed application. And since it's significantly
different than what you see in Photoshop the culprit is almost always a
bad monitor profile, which Photoshop uses but which the ImageReady
preview ignores.
Bill
| |
| Bill Hilton 2005-12-23, 6:14 pm |
| > rus...@att.net writes ...
>
>Adobe and many vendors not only DELIBERATELY turn color
> management into a nightmare ...
Why would a vendor "DELIBERATELY" create problems? Makes no sense to
me ...
>Adobe and many vendors not only DELIBERATELY turn color
>management into a nightmare, but also try to convince every user that
>they need it. Why? To keep revising their products to make ONLY
>incremental changes to bait the users.
Photoshop color management was revamped radically with Version 6, which
came out 3 revs and maybe 5-6 years ago. Since then there haven't been
ANY Photoshop color management changes that I'm aware of, which kind of
kills your conspiracy theory.
And of course you can always turn color management off if you don't
want to use it.
> Here's my take. There are technical challenges to get color
> management right, simple and cheap.
I've helped several total beginners get up to speed ... here's my take
.... I agree with you that there are "technical challenges" but they
aren't overwheliming ... the newbies I've taught needed to do/know
three things, which weren't all that hard once the basics are
understood ...
1) calibrate the monitor accurately ... if you can't do it well enough
with Adobe Gamma or one of the other free software solutions (I never
could, to my satisfaction) then spend $200 on a good hardware solution
like the Sypder 2 or Gretag Eye-One or Monaco Optix.
2) Learn the difference between a device-specific profile (like the
ones for your monitor or printer) and a 'working space' profile, and
don't mix them up. Almost everyone stumbles over this when they start
down the color management path.
3) Buy a printer with good built-in ICC profiles that are accurate for
the papers you'll use and learn how to print to them from Photoshop
(OK, maybe that's four things :). Epson is the leader in this field
but the better Canon and HP photo models are catching up. That's it
.... a decent calibrated monitor, a printer with accurate ICC profiles
that you know how to apply in Photoshop, and understanding the
difference between working space profiles and device-specific profiles
should be enough for most people.
Bill
| |
| Mike Russell 2005-12-23, 6:14 pm |
| <rusure@att.net> wrote in message news:43AC018B.3D88A3DF@att.net...
[re color management being simple someday]
> You are overly optimistic. If (not when) color management should become
> "dirt simple" (and dirt cheap), many of us would either have gone broke
> (from all $$$$$ we wasted on it), or have been locked up in a mental
> institution (from the frustration of trying).
LOL - good points both. Perhaps outpatient status and heavy medication
wouldn't hurt, now that you mention it.
You do have a point, but I'm somewhat more optiministic that color
management will mature, just as networking did, and become much more trivial
to manage.
Meantime, the OP's problem with externally viewed images being substantially
darker presents an interesting puzzle. I don't think it can be explained
merely by a bad device profile. If it is, then what would be the simplest
way to verify this? Perhaps he could email the profile in question to one
of us who has a profile viewer.
--
Mike Russell
www.curvemeister.com
| |
| Bill Hilton 2005-12-23, 6:14 pm |
| > Mike Russell writes ...
>
>the OP's problem with externally viewed images being substantially
>darker presents an interesting puzzle. I don't think it can be
>explained merely by a bad device profile.
Of course it can ... he gave plenty of clues ... it may indeed by
something else but evidence suggests otherwise ...
Clue 1 - "This is recent, and I caused it yesterday when I was messing
with
the color settings." ... he isn't specific as to what he was "messing"
with but when you build a monitor profile first you 'calibrate' the
monitor by getting it to a known state, then 'characterize' it by
measuring colors ... the profile is only accurate for that known state,
once you start "messing with the color settings" you've invalidated the
profile and need to generate another one.
Clue 2 - "Only when I go into system32/spool/color/ and delete the
*.icc I created does this problem go away." ... what can I say?
Obviously this means the profile that was deleted was inaccurate ...
Clue 3 - "I have disabled Adobe Gamma loader from starting at bootup,
and tried different color settings, but this is still occuring." ...
disabling Gamma loader means the system is picking up some default
monitor profile of unknown origins ... you can check to see what this
profile is with Edit - Color settings - in 'working spaces: RGB' box
scroll to the top and see what's listed beside 'monitor RGB' ... this
is the profile being used.
Clue 4 - he wasn't clear about what he meant by "saved" in Photoshop
but the link he provided showed the difference between an image
displayed in Photoshop and the same one in the 'save for web' window of
ImageReady ... the ImageReady window shows how the image will look on
the web without the monitor and working spaces profile, otherwise
there's no difference. Again, this narrows it down ...
> If it is, then what would be the simplest way to verify this?
Generate a new, accurate monitor profile (and if you "mess with the
color settings" again then generate a new profile) ... make sure Gamma
loader is enabled if you use Adobe Gamma, or disabled if using a
hardware solution since their software will load the profile, and check
that this profile is picked up. That should do it.
If you want to see the effect the monitor profile has in Photoshop then
open an image like the test one and do View - Proof Setup - and choose
'monitor RGB' ... this ignores the monitor profile ... then do cntrl-y
to toggle this on/off, showing you the effects of applying the monitor
profile ... if the image is sRGB then you should see just a slight
shift in the saturated colors if you have a good monitor profile ... if
you're seeing a wild jump in colors then the monitor profile is bad ...
so Jon, can you try this? I'll bet you see the image look very similar
to what you get with ImageReady when you do 'save for web' ...
Bill
| |
| Mike Russell 2005-12-23, 6:14 pm |
| "Bill Hilton" <bhilton665@aol.com> wrote in message
>
> Of course it can ... he gave plenty of clues ... it may indeed by
> something else but evidence suggests otherwise ...
> Clue 1 - "This is recent, and I caused it yesterday when I was messing
> with
> the color settings." ... he isn't specific as to what he was "messing"
> with but when you build a monitor profile first you 'calibrate' the
> monitor by getting it to a known state, then 'characterize' it by
> measuring colors ... the profile is only accurate for that known state,
> once you start "messing with the color settings" you've invalidated the
> profile and need to generate another one.
The specific problem is that Photoshop displays a significantly brighter
image than Save for Web. Changing the controls on the monitor will not
affect this comparison, since all images will be equally affected. A bad
monitor profile would explain this, but Jon has stated that he re-created
the profile using a gamma of 2.2. IMHO this means the problem may not be
due to a bad monitor profile..
> Clue 2 - "Only when I go into system32/spool/color/ and delete the
> *.icc I created does this problem go away." ... what can I say?
> Obviously this means the profile that was deleted was inaccurate ...
The profile could be fine. This clue does show that he has re-created the
profile at least once, and Jon has said that he specified a gamma of 2.2.
Since Photoshop must remap from a lower to a higher gamma value to display a
brighter image, I think it's worth checiking his working space profile,
sRGB. Adobe Gamma, among its other virtues, makes it very easy to clobber
the wrong profile.
> Clue 3 - "I have disabled Adobe Gamma loader from starting at bootup,
> and tried different color settings, but this is still occuring." ...
> disabling Gamma loader means the system is picking up some default
> monitor profile of unknown origins.
The System profile will be in place in any case. Adobe Gamma Loader is
responsible for initializing the video LUT to the values specified in Adobe
Gamma.
> ... you can check to see what this
> profile is with Edit - Color settings - in 'working spaces: RGB' box
> scroll to the top and see what's listed beside 'monitor RGB' ... this
> is the profile being used.
Not exactly. The string listed beside monitor RGB is a copy of the
description of the profile that Adobe Gamma loaded initially. It would be
nice if Adobe appended a "modified by Adobe Gamma" to the end of the
description, but since it doesn't, this string is not a reliable way to
determine the monitor profile.
In addition, this profile is not the same thing as the System's display
profile, but a separate profile used by Adobe products for display
conversion purposes. The system display profile is in the Monitor settings
for Windows 2000 and later. Adobe Gamma normally keeps these in sync, but
it's possible to manually change things out from under the Gamma Loader. I
don't think this is Jon's problem, BTW.
> Clue 4 - he wasn't clear about what he meant by "saved" in Photoshop
> but the link he provided showed the difference between an image
> displayed in Photoshop and the same one in the 'save for web' window of
> ImageReady ... the ImageReady window shows how the image will look on
> the web without the monitor and working spaces profile, otherwise
> there's no difference. Again, this narrows it down ...
It certainly does.
>
> Generate a new, accurate monitor profile (and if you "mess with the
> color settings" again then generate a new profile) ... make sure Gamma
> loader is enabled if you use Adobe Gamma, or disabled if using a
> hardware solution since their software will load the profile, and check
> that this profile is picked up. That should do it.
>
> If you want to see the effect the monitor profile has in Photoshop then
> open an image like the test one and do View - Proof Setup - and choose
> 'monitor RGB' ... this ignores the monitor profile ... then do cntrl-y
> to toggle this on/off, showing you the effects of applying the monitor
> profile ... if the image is sRGB then you should see just a slight
> shift in the saturated colors if you have a good monitor profile ... if
> you're seeing a wild jump in colors then the monitor profile is bad ...
> so Jon, can you try this? I'll bet you see the image look very similar
> to what you get with ImageReady when you do 'save for web' ...
Good idea. I'd also like to get hold of one of Jon's files. If it displays
too light on my system, I'll know it's a problem with the embedded profile.
If it doesn't, then the problem has to be, as you suspect, somewhere in the
display setup.
The moral of the story, I think, is that color setup is too arcane. It's
ironic that one of the oldest computer peripherals is still not truly plug
and play. We are waiting for an Alan Oppenheimer of color to make sweeping
simplifications to the setup process.
--
Mike Russell
www.curvemeister.com
| |
| Bill Hilton 2005-12-23, 6:14 pm |
| > Mike Russell writes ...
>
>The specific problem is that Photoshop displays a significantly
>brighter image than Save for Web.
You have it EXACTLY backwards ... here's the link he gave ...
http://img369.imageshack.us/img369/2682/pscolor8uy.jpg ... the
Photoshop image is on the left, the ImageReady image on the right ...
download it and tag it with sRGB and I read about 213/9/10 for the
Photoshop image, and 250/5/4 for the ImageReady image. I don't see how
you can miss this ...
>Changing the controls on the monitor will not
>affect this comparison, since all images will be equally affected.
No, because the ImageReady version isn't using the monitor profile or
the working space tag ... depending on what monitor controls he changed
he could easily have rendered his profile invalid/inaccurate , which
would explain the differences he's seeing.
>A bad monitor profile would explain this, but Jon has stated that he
>re-created the profile using a gamma of 2.2. IMHO this means the
>problem may not be due to a bad monitor profile..
It doesn't matter what gamma he uses if the profile is bad, and a bad
profile is the most likely explanation for the differences he's seeing.
> I think it's worth checiking his working space profile, sRGB. Adobe Gamma,
>among its other virtues, makes it very easy to clobber the wrong profile.
How could using Adobe Gamma "clobber" another profile, unless maybe he
named the newly created profile "sRGB IEC61966-2.1.icc" or similar?
The only other way I know to mess this up is with Edit - Color Settings
and doing a Custom RGB change on the gamma setting and then saving it
as the name of a preset profile by mistake. I'm pretty sure he's doing
this :)
[color=darkred]
>this profile is not the same thing as the System's display
>profile, but a separate profile used by Adobe products for display
>conversion purposes. The system display profile is in the Monitor
>settings for Windows 2000 and later.
Right, with XP you can check this by right-clicking on the open desktop
and doing Properties - Settings - Advanced - Color Management ... the
monitor profile grayed out in the lower box is the one being used and
I've never seen it not match up with what's listed in the Photoshop
window mentioned.
>The moral of the story, I think, is that color setup is too arcane.
You can't say for sure until Jon tries a couple of the suggested fixes
.... it's probably just a bad monitor profile, in which case it wasn't
arcane at all ... at least that's what it was the other 20-something
times people had this same problem in this newsgroup.
Bill
| |
| Mike Russell 2005-12-23, 6:14 pm |
| >> Mike Russell writes ...
From: "Bill Hilton" <bhilton665@aol.com>
[color=darkred]
> You have it EXACTLY backwards ... here's the link he gave ...
> http://img369.imageshack.us/img369/2682/pscolor8uy.jpg ... the
> Photoshop image is on the left, the ImageReady image on the right ...
> download it and tag it with sRGB and I read about 213/9/10 for the
> Photoshop image, and 250/5/4 for the ImageReady image. I don't see how
> you can miss this ...
Right you are, Bill. Thanks for pointing this out. This does not the
reasoning behind what I said, except that the tagged profile's gamma may be
higher, not lower, than 2.2.
Based on what Jon has said, my money is on the embedded profile at the
moment. Yours is on the monitor profile. Either way, I think we are on the
way to solving the problem.
> How could using Adobe Gamma "clobber" another profile, unless maybe he
> named the newly created profile "sRGB IEC61966-2.1.icc" or similar?
> The only other way I know to mess this up is with Edit - Color Settings
> and doing a Custom RGB change on the gamma setting and then saving it
> as the name of a preset profile by mistake. I'm pretty sure he's doing
> this :)
You have described two different ways to clobber the sRGB profile, and I
have encountered people who have done this. This is why I suggested that
Jon reload the sRGB profile.
I suggest that Jon provide one of his images as a way to determine whether
this is the case, and I think it would be informative as well for him to
post his monitor icc file as well.
---
Mike Russell
www.curvemeister.com
| |
| Jon Danniken 2005-12-23, 10:14 pm |
| "Bill Hilton" wrote:
>
> The 'saving window' you apparently refer to is 'save for web', which
> opens ImageReady (based on the link you provided) ... you were saving
> as a jpeg and 'save for web' shows you what the image will look like ON
> THE WEB, where there is no color management, that is, the ICC working
> space profile is ignored for almost all browsers (if you even assign
> one) AND the monitor ICC profile is ignored. So basically it's not
> "the appearance of an image *while still in Photoshop*" that you're
> looking at, it's how that jpeg will look when viewed outside Photoshop
> in a non-color managed application like a web browser.
>
> All this is saying is that you have a very bad monitor profile, since
> you shouldn't be seeing THAT big a difference if you first convert the
> file to sRGB before making it a jpeg. Make a new monitor profile
> that's accurate and understand what ImageReady is doing when you 'save
> for the web' and the problem is solved.
>
>
> It doesn't 'change the levels' at all, it just shows what the image
> will look like without profiles when you are saving a jpeg using 'save
> for the web'.
>
>
> Hope this helps ...
It does, Bill, and thanks. I realize now that PS isn't changing the output
(as seen in the "Save for Web" dialog) but is instead applying a
compensation to what I am working on instead. .
As is probably obvious to yourself and others, I am having issues
calibrating this monitor, not a terribly good one at that, and I had fiddled
with the monitor settings after attempting to set a profile for it.
The problem is that the monitor is mainly used for web browsing, and it is
only when I work with images (digital source, photo lab destination) that I
really want the monitor to match the output (or what the print will look
like when it comes back from the lab).. What would be handy would be a
toggle switch that would switch between a web setup and photolab setup; I am
experimenting with "soft proof" in PS as a means of achieving this. .
I'll need a better monitor to really be able to make any decent attempt at
realistic color matching, but I have a better understanding of what caused
the problem and at least know how to fix it back the next time I go fiddling
about.
Thanks,
Jon
| |
| Jon Danniken 2005-12-23, 10:14 pm |
| "Jon Danniken" wrote:
> As is probably obvious to yourself and others, I am having issues
> calibrating this monitor, not a terribly good one at that, and I had
fiddled
> with the monitor settings after attempting to set a profile for it.
Furthering that thought, why would it even matter? Wouldn't changing my
monitor post-profile affect both "in PS" as well as "Save for Web" images,
or does PS somehow know that I have fiddled with the brightness/contrast
setting on my monitor?
Jon
| |
| Bill Hilton 2005-12-24, 6:14 pm |
| > Jon Danniken writes ...
>
>I realize now that PS isn't changing the output
>(as seen in the "Save for Web" dialog) but is instead applying
>a compensation to what I am working on instead.
OK, so that was it after all?
>I am having issues calibrating this monitor, not a terribly
>good one at that, and I had fiddled with the monitor settings
>after attempting to set a profile for it.
I had the same problems with Adobe Gamma (as do many people, though
others are able to generate good profiles with it) and I eventually had
to fork over $200 for a Sypder several years ago. The difference was
very obvious. Maybe you'll eventually have to do the same ... it helps
to have a fairly new monitor though, the older ones may have trouble
hitting a 6500 K whitepoint and keeping the desired luminance (doing
both can be tough on old monitors). At any rate, now you know to
re-generate the profile after changing the settings ...
>The problem is that the monitor is mainly used for web browsing ...
>What would be handy would be a toggle switch that would switch
>between a web setup and photolab setup
You can do most of the 'web setup' by working with files tagged sRGB
and ignoring the monitor profile, do View - Proof Setup and pick
'monitor rgb'. I think somewhere there's a master switch to turn off
all color management too, though I've never looked for it.
>I'll need a better monitor to really be able to make any decent attempt at
>realistic color matching, but I have a better understanding of what caused
>the problem and at least know how to fix it back the next time I go fiddling
>about.
Glad to help.
Bill
| |
| Bill Hilton 2005-12-24, 6:14 pm |
| > Mike Russell writes ...
>
>Based on what Jon has said, my money is on the embedded profile at
>the moment. Yours is on the monitor profile.
>From his latest post it seems like it was the monitor profile after
all.
>You have described two different ways to clobber the sRGB profile, and I
>have encountered people who have done this. This is why I suggested
>that Jon reload the sRGB profile.
But both were pretty far fetched ... for one you'd have to give your
monitor profile a very unlikely name and for the other you'd need to be
an expert level user to even get there, then make a very careless
non-expert mistake by leaving the newly edited profile with the old
sRGB name ... if you think the sRGB profile is bad because it was
overwritten then I'd suggest first looking at it in My Computer and
checking the date it was created ... probably about the time the
computer was first installed but if it has a recent date stamp (he said
his problem just started recently) then obviously it was over-written.
Another way is to set sRGB as your working RGB space in Edit - Color
Settings (Working spaces: RGB) and then open the 'working spaces: RGB'
drop-down again and go to the top and select 'custom' and it will show
you the numbers for the currently selected working space, where you can
read off the gamma (since you think that's the problem). Here is what
I get for the sRGB specs when I do this ...
http://members.aol.com/bhilton665/srgb_specs.jpg ... by the way, this
is how you can create your very own custom working profile (like
BruceRGB :), if you wish ... when you're done with this dialog box make
sure you cancel out and don't change anything to avoid mucking up the
sRGB profile for real, because here is where you can really hurt
yourself :)
Bill
| |
| Bill Hilton 2005-12-24, 6:14 pm |
| >Jon Danniken writes ...
>
> I am having issues calibrating this monitor, not a terribly good
> one at that, and I had fiddled with the monitor settings after
> attempting to set a profile for it. ...
>
>Furthering that thought, why would it even matter? Wouldn't changing my
>monitor post-profile affect both "in PS" as well as "Save for Web" images,
>or does PS somehow know that I have fiddled with the brightness/contrast
>setting on my monitor?
Yes, it does affect both color managed programs like Photoshop and
non-color managed apps ... but in different ways, because of the effect
of the no-longer accurate icm monitor profile ...
I can explain it better if I skip Adobe Gamma and describe what happens
when you use a hardware device to calibrate ... first you adjust the
settings to a known white point, gamma and contrast, usually adjusting
the red, green and blue guns separately to get the best custom white
balance and luminance (I'll just assume a CRT) ... once the white point
and luminance are as precise as you can get them then it's "calibrated"
and all programs will benefit from this step ... now you attach the
puck to the screen and the program displays various known colors (maybe
as many as 100 in the better cal programs) and measures the actual
color displayed on the screen and compares these measured values with
the known values and devises a simple translation formula so that the
colors in the file are 'translated' on the fly so that what you see
on-screen will look as color-accurate as possible compared to the
colors described in the image file.
To better understand this, read the first couple pages on this web site
.... http://www.creativepro.com/story/feature/13605.html ... (this is
the guy many of us learned color management from via his "Real World
Photoshop" and "Real World Color Management" books) ... in his example
he uses a tan patch, RGB => 247/160/91 in the file and if you just
display 247/160/91 directly on the screen you'll get (in his example) a
slightly off-color, darker image ... so the color management
calibration software figures out that on *this* screen for the present
settings you'll get the same visual color if the numbers are changed to
250/175/100 for display purposes (this is on page 2 of the article).
So that's what the icm file buys you for this example, a translation of
the equivalent numbers so the displayed color is accurate ... now
assume you adjust the settings like you did, say you increase the
brightness by 10 points or whatever (which it looks like you did, based
on the example link you provided), so now the numbers in the file
(247/160/91) will display brighter on the screen, say 254/185/110 (to
make something up) but since you're still using the old icm monitor
profile you won't accurately translate this back to the correct color
(247/160/91) ...
So images will look brighter in both color managed apps and non-color
managed apps, but because the color managed apps are doing a further
color translation based on the profile created with different settings
it can diverge further from the target color.
Bill
| |
| KatWoman 2005-12-26, 6:19 pm |
|
"Bill Hilton" <bhilton665@aol.com> wrote in message
news:1135268611.624713.287720@f14g2000cwb.googlegroups.com...
>
> I've helped several people track down problems like this and in *every*
> case so far it was caused by a bad (inaccurate) monitor profile.
>
> Regenerate the profile from scratch and see if that fixes it.
>
>
> Just generate a new profile.
>
> Bill
I am using 9300 I thought that is the temp of my monitor?
I use Adobe Gamma
I open a good test file print with spectrum of colors gray scale skin tones
etc
open it in windows picture fax viewer
then open same test image in PS
then open and customize the Adobe Gamma and monitor till both look similar
then there is not such a big difference when you go to a non color managed
application
| |
| Bill Hilton 2005-12-26, 10:15 pm |
| > KatWoman writes ...
>
>I am using 9300 I thought that is the temp of my monitor?
Every monitor I've used had presets for 9300, 6500 and 5000 K (and
allowed you to set the individual r,g,b guns for a custom white point
too) ... most people find 9300 too blue so 6500 is probably the most
popular. I think people doing pre-press use either 5000 (more common
on Macs) or 6500, I've never met anyone who used 9300 for print. But
whatever works for you ...
>I use Adobe Gamma
>I open a good test file print with spectrum of colors gray scale skin
>tones etc
>open it in windows picture fax viewer
>then open same test image in PS
>then open and customize the Adobe Gamma and monitor till both look
>similar then there is not such a big difference when you go to a
>non color managed application
Sounds like a good method if you're using Gamma ... with the hardware
solutions like the Gretag Eye-One I see a slight change in saturated
colors when I toggle the monitor profile on/off on test patches ...
every time (so far) there's been a major color shift between CM apps
and non-CM apps it's because the monitor profile is way off or because
an image was saved in a wide-gamut working space and the saturated
colors get dulled down when viewed in an app that doesn't recognize
profiles.
Bill
| |
| rusure@att.net 2005-12-27, 6:19 pm |
|
Bill Hilton wrote:
>
>
> Why would a vendor "DELIBERATELY" create problems? Makes no sense to
> me ...
>
>
> Photoshop color management was revamped radically with Version 6, which
> came out 3 revs and maybe 5-6 years ago. Since then there haven't been
> ANY Photoshop color management changes that I'm aware of, which kind of
> kills your conspiracy theory.
>
> And of course you can always turn color management off if you don't
> want to use it.
My conspiracy theory of course is tongue in cheek. But it certainly
feels that way for some of us. Color management at this stage is as Mike
stated "a mess", and expensive. And as Mike pointed out in another
thread "Cost effective way to calibrate printer/scanner/screen for use
in PS CS2" in alt.graphics.photoshop, not everyone needs calibration.
"If you are a single person operation doing your own images, there is no
need
to spend even one cent on calibration equipment."
>
> I've helped several total beginners get up to speed ... here's my take
> ... I agree with you that there are "technical challenges" but they
> aren't overwheliming ... the newbies I've taught needed to do/know
> three things, which weren't all that hard once the basics are
> understood ...
>
> 1) calibrate the monitor accurately ... if you can't do it well enough
> with Adobe Gamma or one of the other free software solutions (I never
> could, to my satisfaction) then spend $200 on a good hardware solution
> like the Sypder 2 or Gretag Eye-One or Monaco Optix.
This is the first advice the color management "gurus" will give you. But
they won't point out that not all monitors can be calibrated
(accurately). As Jon stated in this thread, his monitor is old. It is
impossible to set good black and white points on some old monitors.
Without that, the calibration tools are useless. Some calibration tools
also require the ability to adjust the rgb channels separately, but most
monitors do not offer this feature.
But of course, the "gurus" would advice to buy a $$$ monitor before
calibration.
> 2) Learn the difference between a device-specific profile (like the
> ones for your monitor or printer) and a 'working space' profile, and
> don't mix them up. Almost everyone stumbles over this when they start
> down the color management path.
>
> 3) Buy a printer with good built-in ICC profiles that are accurate for
> the papers you'll use and learn how to print to them from Photoshop
> (OK, maybe that's four things :). Epson is the leader in this field
> but the better Canon and HP photo models are catching up. That's it
> ... a decent calibrated monitor, a printer with accurate ICC profiles
> that you know how to apply in Photoshop, and understanding the
> difference between working space profiles and device-specific profiles
> should be enough for most people.
It took a long time for Epson get up to speed on providing ICC profiles
and a workflow on how to use them. EX, 1200, 1270 and 1280 in the US
were never shipped with the ICC profiles. Even with the 2200 profiles,
some images just won't print well with some Epson paper. There have been
plenty of such discussions on this. Many suggested that RIP can solve
the problem by replacing the Epson driver. Epson's printer hw has always
been excellent, but their drivers and profiles simply do not measure up.
This leads me again to the conspiracy theory that they are more
interested in selling the consumable.
| |
| KatWoman 2005-12-27, 6:19 pm |
|
"Bill Hilton" <bhilton665@aol.com> wrote in message
news:1135648918.373323.87070@o13g2000cwo.googlegroups.com...
>
> Every monitor I've used had presets for 9300, 6500 and 5000 K (and
> allowed you to set the individual r,g,b guns for a custom white point
> too) ... most people find 9300 too blue so 6500 is probably the most
> popular. I think people doing pre-press use either 5000 (more common
> on Macs) or 6500, I've never met anyone who used 9300 for print. But
> whatever works for you ...
>
>
> Sounds like a good method if you're using Gamma ... with the hardware
> solutions like the Gretag Eye-One I see a slight change in saturated
> colors when I toggle the monitor profile on/off on test patches ...
> every time (so far) there's been a major color shift between CM apps
> and non-CM apps it's because the monitor profile is way off or because
> an image was saved in a wide-gamut working space and the saturated
> colors get dulled down when viewed in an app that doesn't recognize
> profiles.
>
> Bill
Yikes it looks all yellow at 6500
| |
| Bill Hilton 2005-12-28, 6:14 pm |
| > KatWoman writes ...
>
>Yikes it looks all yellow at 6500 (as opposed to 9300 K)
LOL ... run it at 5000 for three days and then 6500 will look pretty
good :)
Bill
| |
| Bill Hilton 2005-12-28, 6:14 pm |
| >> I wrote ...
[color=darkred]
> rus...@att.net writes ...
>
>This is the first advice the color management "gurus" will give you.
>But they won't point out that not all monitors can be calibrated
>(accurately). As Jon stated in this thread, his monitor is old. It is
>impossible to set good black and white points on some old
>monitors.
Once you strip the formal ICC workflow down to its basics, all it's
doing is trying to help you get matching (accurate) colors as you move
from device to device. Bottom line, that's ALL it's really doing. So
I agree that if your monitor is so crappy that it won't display many
colors accurately that the ICC workflow isn't ideal.
What I would wonder though is why someone would spend $600 on a
professional graphics program like Photoshop and then complain when
they get sorry results with an aging $200 thrift shop monitor ... seems
backwards to me, they would actually get better results if then would
spend $700 on a monitor and $100 on an entry level graphics program
like Elements ...
> But of course, the "gurus" would advice to buy a $$$ monitor
>before calibration.
That seems to be the obvious first step to most of us :)
>It took a long time for Epson get up to speed on providing
>ICC profiles and a workflow on how to use them. EX,
>1200, 1270 and 1280 in the US were never shipped
>with the ICC profiles.
This is not true at all ... I'll speak just to the EX, which I used to
own, and the 1280, which I still have though rarely use ... both
shipped with ICC profiles that were in a bundled file so that when you
selected the ICM option in the printer driver it would pick the right
profile based on the paper type you had selected. This was good enough
up to Photoshop 5.5 because there was no way to use the individual
profiles, ie, no soft proofing.
The 1280 also comes with separate ICC profiles for each of the
supported papers. These came out about the time Photoshop 6 with soft
proofing came out ... I can't remember whether you accessed these
profiles by loading the PIM module or whether they were a separate
download Epson made available when they brought out the ColorLife paper
in response to the infamous orange-shift problem, but for sure they are
available (and no doubt are still up on Epson's support download page)
.... I just checked my directory and I have six 1280 Epson profiles
dated Sept 2000 so they've been around for at least half a decade ...
> Even with the 2200 profiles, some images just won't
>print well with some Epson paper.
Not sure what specifically you are referring to, but the softer papers
like Velvet Fine Art and Enhanced Matte don't print reds as well as the
glossy papers with the Photo black inks ... this is a gamut issue and
is reflected in the ICC profiles Epson provided (the matte black ink
with softer watercolor and matte papers has significantly smaller gamut
than glossy papers ... no surprise there, or at least it shouldn't be).
>This leads me again to the conspiracy theory that they
>are more interested in selling the consumable.
If that were true and one of their competitors like Canon or HP
provided something better Epson would be wiped out of the graphics
market in months. But the reason Epson is the leader in this market is
that they are doing a far better job than their competitors ... not
saying they don't have room to improve, just that it's silly to blame
it on a conspiracy.
Bill
| |
| rusure@att.net 2005-12-30, 6:15 pm |
| Bill Hilton wrote:
[color=darkred]
>
> Not sure what specifically you are referring to, but the softer papers
> like Velvet Fine Art and Enhanced Matte don't print reds as well as the
> glossy papers with the Photo black inks ... this is a gamut issue and
> is reflected in the ICC profiles Epson provided (the matte black ink
> with softer watercolor and matte papers has significantly smaller gamut
> than glossy papers ... no surprise there, or at least it shouldn't be).
The best way to illustrate how some images don't print well on some
Epson papers with Epson profiles can be found in this article.
http://daystarvisions.com/Docs/Rvws/EpsonPaper/pg1.html
Some images that look like Fig.2 on a monitor will soft proof and print
like Fig.3 and Fig.4 on Epson matte papers with Epson profiles (1280 and
2200). Fig.2 will soft proof and print much closer to the monitor image
with Luster/Glossy Epson papers and profiles.
Notes:
I think that the article author created these Figures to illustrate the
problem. Actually soft proofing and printing Fig.2 in PS will NOT
produce Fig.3 and 4. But the point is well made.
Many images do soft proof and print very well on Epson matte papers with
Epson profiles. Only image with large areas of deep shadow details
and/or very saturated colors will have the printing problems.
As an experiment, create an image with the rgb colors 75/30/35,
40/35/50, 115/55/0, 230/200/250, 40/20/20, etc. Soft proof the image
with the 2200 EEM (MK) and Premium Luster (PK) profiles, Relative
Colorimetric intent, and you will see a dramatic difference. These
colors are only close to the monitor on the Premium Luster paper and not
on the EEM paper. Can a custom EEM profile fix this?
| |
| Bill Hilton 2005-12-30, 6:16 pm |
| > rus...@att.net writes ...
>
>The best way to illustrate how some images don't print well on some
>Epson papers with Epson profiles can be found in this article.
>http://daystarvisions.com/Docs/Rvws/EpsonPaper/pg1.html
I was expecting one of those rabid anti-Epson screeds like the
orange-shifties write but this was actually a very good article.
>Many images do soft proof and print very well on Epson matte papers with
>Epson profiles. Only image with large areas of deep shadow details
>and/or very saturated colors will have the printing problems.
This is in line with what I already said, except I'm focussing on the
saturated colors and blaming it on limited gamut and you're bringing in
"deep shadow" areas and blaming it on poor profiles ... at least I
think that is the difference between what we are saying, right?
>As an experiment, create an image with the rgb colors
>75/30/35, 40/35/50, 115/55/0, 230/200/250, 40/20/20, etc.
Not trying to give you a hard time (here at least :) but when you give
RGB numbers like this you should always say which 'working space'
profile (sRGB, AdobeRGB, ProPhoto?) is to be used since these numbers
mean different things in different working spaces (see the recent
thread on 'assign' vs 'convert' for examples). Anyway, I assumed you
meant AdobeRGB and whipped out a sample ...
>Soft proof the image with the 2200 EEM (MK) and Premium Luster (PK)
>profiles, Relative Colorimetric intent, and you will see a dramatic difference.
OK, I did this and then I also turned on the 'gamut warning' for both
profiles ... guess what, three of the color patches are out of gamut
for the Enhanced Matte profile (and I'm really surprised a fourth,
230/200/250, is in gamut ... has to be near the boundary) and none of
the patches are OOG for Luster. To me this explains it since by
definition the oog areas aren't going to print accurately ... here's a
screen dump with gamut warning turned on, I also tossed in a LightJet
5000 profile for Calypso's LJ with glossy paper, again as expected all
is in gamut with the LJ too, which has an even wider gamut than
AdobeRGB when properly profiled (this one was profiled by Bill
Atkinson) ... http://members.aol.com/bhilton665/epson_profiles.jpg
>These colors are only close to the monitor on the Premium Luster
>paper and not on the EEM paper.
I think the out of gamut warnings predict this, at least I wouldn't
expect accurate colors there after an oog warning ... it's a problem
with those types of papers ... if you have a program like Color Think
that plots gamuts you can see that the glossy/luster/semi-gloss papers
have a much wider gamut than the matte-type Epson papers (matte, Velvet
Fine Art, Watercolor - Radiant White etc).
> Can a custom EEM profile fix this?
Dunno, it seems like a paper limitation to me ... in the article you
linked to he says to use the Velvet Fine Art profile on the enhanced
matte paper since it has fixed the matte's profile problems (VFA is my
favorite Epson paper with my 2200 and 4000 printers but it still
doesn't hold saturated colors as well as the glossy ones) ... so I
applied VFA's profile from both the 2200 and 4000 and these same three
colors are still out of gamut ...
You are blaming this on Epson's bad profiles (and, tongue in cheek, a
conspiracy to make us waste paper and ink), I'm blaming it on the
limits of the color gamut of these types of papers with this ink set.
Whatever ... anyway, thanks for giving an example rather than just
name-calling. We can agree to disagree agreeably :)
Bill
|
|
|
| | Copyright 2003 - 2008 forum4designers.com Software forum Computer Hardware reviews |
|