Adobe Color Space: Not Lovin' It... (Pics)

Started Jun 29, 2004 | Discussions
Iliah Borg Forum Pro • Posts: 25,750
Re: an attempt to answer :) thank you

I use the "print with preview" and the print profile I usually use
is SRGB. There is also another "print with preview" option: color
management or output.

Color management, and set destination profile (printer profile) to printer/paper/quality combo matching profile.

If I'm using Adobe RGB, at this point and want to make a jpeg for
the web, what method do you recommend? I've been using the save
for the web feature in PSCS. I take my photos as NEFs, bringing
them in to Nikon View 6.22 for initial processing and then in PSCS
where they become 16 bit tiffs. This is the point I then "save for
the web."

Before "save for web" it is a good idea to convert to sRGB or
generic monitor profile. Better to do the conversion in 16-bit
mode, though the difference is not that great.

Where should this adjustment be done?

Image - Mode - Convert to profile

In PS, Shift-Ctrl-K, please, and look at your settings to know what
is going when you open a file. See what printer profiles do you
have installed in the system; how is your driver instructed to deal
with colour; and try "print with preview" with correct profile set
as your printer profile.

I've reviewed this many times to ensure the correct settings are
used, but I think the problem is in software color space .

List your settings please

-- hide signature --

no text

1stAde Regular Member • Posts: 482
Re: I thought I had a rough idea, but that 2nd sentence...

Iliah

If you have the patience then yes I would appreciate it! and lets hope I have the capacity to understand.

It did make me laugh at the time though

Adrian.

Iliah Borg wrote:

1stAde wrote:
But after reading
" While convertion, intent makes the difference. For some profiles it

is perceptual intent that works best; but with many modern profiles
it is relative colorimetric."

I am thinking of going to process B&W for all my photos! LOL!

Oh, yes

On a serious note, should I explain?

Iliah Borg Forum Pro • Posts: 25,750
Re: I thought I had a rough idea, but that 2nd sentence...

If you have the patience

if I have English

First, why we need rendering intent. While converting from one color space to another, we deal wit a problem of "nesting" - source colour space can be larger in some (or all) directions then the destination color space. That is, source colour space may have wider gamut then the gamut of destination.

Some colours from the source colour space can get out of the destination space - what to do with those colours? Rendering intent gives the method of dealing with this problem.

4 major rendering intents are called
absolute colorimetric
relative colorimetric
perceptual
saturation

Colorimetric intents leave the colours that fall into destination gamut fairly intact - the colors are only adjusted for accurate reproduction in destination space. Out of gamut colors are compressed to bring them into destination space.

The difference between abs. and rel colorimetric is in how the white is addressed. For relative colorimetric, white from source is translated into white in destination; so, white is preserved.

Absolute colorimetric is designed for proofing, so it makes an opposite - white from destination is considered white in source. Thus I can proof printing on yellow stock paper with my Epson 2100 on white Luster.

With perceptual intent, all the source colours are compressed (even those that actually fall into destination gamut "as is"). This is done to give more space for out of gamut colours. As a result, all colours are "wrong", but the total impression is "right" - as the relations between colours are kept.

Saturation intent maps all the colours individually, without respect to neighboring colours. As a result, different out of gamut colours can be mapped to the same destination colour. This rendering intent is pretty special. I use it with dull images, in order to gain more saturation.

Now, exept for proofing and special cases, I mostly use rel.colorimetric and perceptual intent. In many cases image colours do not occupy the whole gamut of source space. Using perceptual intent in this case compresses the colours unnessessary. Rel. colorimetric intent in this case give more "life" to the image.

To round this up, here is what I do: if the source gamut is much wider then the destination; and number of colours in the image is large enough so that they substantionally fall out of gamut of destination space - perceptual rendering intent is used. Otherwize - relative colorimetric. Preview in convert dialog helps

-- hide signature --

no text

Dan Clark Senior Member • Posts: 2,703
EXCELLENT Explanation!!! I learned a lot. Many thanks. (nt)

Iliah Borg wrote:

If you have the patience

if I have English

First, why we need rendering intent. While converting from one
color space to another, we deal wit a problem of "nesting" - source
colour space can be larger in some (or all) directions then the
destination color space. That is, source colour space may have
wider gamut then the gamut of destination.

Some colours from the source colour space can get out of the
destination space - what to do with those colours? Rendering intent
gives the method of dealing with this problem.

4 major rendering intents are called
absolute colorimetric
relative colorimetric
perceptual
saturation

Colorimetric intents leave the colours that fall into destination
gamut fairly intact - the colors are only adjusted for accurate
reproduction in destination space. Out of gamut colors are
compressed to bring them into destination space.

The difference between abs. and rel colorimetric is in how the
white is addressed. For relative colorimetric, white from source is
translated into white in destination; so, white is preserved.

Absolute colorimetric is designed for proofing, so it makes an
opposite - white from destination is considered white in source.
Thus I can proof printing on yellow stock paper with my Epson 2100
on white Luster.

With perceptual intent, all the source colours are compressed
(even those that actually fall into destination gamut "as is").
This is done to give more space for out of gamut colours. As a
result, all colours are "wrong", but the total impression is
"right" - as the relations between colours are kept.

Saturation intent maps all the colours individually, without
respect to neighboring colours. As a result, different out of gamut
colours can be mapped to the same destination colour. This
rendering intent is pretty special. I use it with dull images, in
order to gain more saturation.

Now, exept for proofing and special cases, I mostly use
rel.colorimetric and perceptual intent. In many cases image colours
do not occupy the whole gamut of source space. Using perceptual
intent in this case compresses the colours unnessessary. Rel.
colorimetric intent in this case give more "life" to the image.

To round this up, here is what I do: if the source gamut is much
wider then the destination; and number of colours in the image is
large enough so that they substantionally fall out of gamut of
destination space - perceptual rendering intent is used. Otherwize

  • relative colorimetric. Preview in convert dialog helps

Robert Peters Veteran Member • Posts: 5,396
That is the most concise description I have read. Thank

you for taking the time to help.

Bob Peters

Iliah Borg wrote:

If you have the patience

if I have English

First, why we need rendering intent. While converting from one
color space to another, we deal wit a problem of "nesting" - source
colour space can be larger in some (or all) directions then the
destination color space. That is, source colour space may have
wider gamut then the gamut of destination.

Some colours from the source colour space can get out of the
destination space - what to do with those colours? Rendering intent
gives the method of dealing with this problem.

4 major rendering intents are called
absolute colorimetric
relative colorimetric
perceptual
saturation

Colorimetric intents leave the colours that fall into destination
gamut fairly intact - the colors are only adjusted for accurate
reproduction in destination space. Out of gamut colors are
compressed to bring them into destination space.

The difference between abs. and rel colorimetric is in how the
white is addressed. For relative colorimetric, white from source is
translated into white in destination; so, white is preserved.

Absolute colorimetric is designed for proofing, so it makes an
opposite - white from destination is considered white in source.
Thus I can proof printing on yellow stock paper with my Epson 2100
on white Luster.

With perceptual intent, all the source colours are compressed
(even those that actually fall into destination gamut "as is").
This is done to give more space for out of gamut colours. As a
result, all colours are "wrong", but the total impression is
"right" - as the relations between colours are kept.

Saturation intent maps all the colours individually, without
respect to neighboring colours. As a result, different out of gamut
colours can be mapped to the same destination colour. This
rendering intent is pretty special. I use it with dull images, in
order to gain more saturation.

Now, exept for proofing and special cases, I mostly use
rel.colorimetric and perceptual intent. In many cases image colours
do not occupy the whole gamut of source space. Using perceptual
intent in this case compresses the colours unnessessary. Rel.
colorimetric intent in this case give more "life" to the image.

To round this up, here is what I do: if the source gamut is much
wider then the destination; and number of colours in the image is
large enough so that they substantionally fall out of gamut of
destination space - perceptual rendering intent is used. Otherwize

  • relative colorimetric. Preview in convert dialog helps

 Robert Peters's gear list:Robert Peters's gear list
Nikon D70 Nikon 1 V2 Nikon 1 V3 Nikon D750 Nikon AF-S Micro-Nikkor 60mm F2.8G ED +13 more
Rliotta Regular Member • Posts: 378
Re: Time out, folks. Let me try to answer.

tagged

Robert Peters wrote:
All of the following is based on my personal experiences over the
last 4 years. I do not claim to be an expert.

(0) The file name associated with your picture is of the form
DSC_0654, for instance. If the image were recorded in Adobe RGB
the file name would begin with DSC. Have you renamed the files or
was that particular image recorded in sRGB? The file does not
contain any color space tag.

(1) If you are using Photoshop and are recording the images in
Adobe RGB (I'm assuming you are using jpeg format) then the
Photoshop Working Space should be set Adobe RGB.

(2) Open the Potoshop Color Settings and enable ALL the warning
dialogs (missing profile, profile mismatch, etc) so that you dont
get any "color surprises."

(3) If you have a printer such as an Epson Stylus Photo 1280 or any
printer which is supplied with profiles for the various papers you
use then you DO NOT need to convert to sRGB before printing.
Period. However, you MUST let Photoshop do all the color
management and TURN OFF the color management of the printer driver.
If you can't perform the final step then you will probably have to
convert to sRGB (or whatever is recommended for your printer).

(4) If the colors of the Adobe RGB images look drab in your browser
then your browser does not perform proper color management. In
particular, IE on Windows assumes the whole universe is sRGB. If
you use a Macintosh then Safari does color management as does IE.
Drab colors (relative to Photochop) in other applications also are
the result of improper color management.

(5) When you post to the web you must first "Convert to Profile"
with sRGB as the target.

I hope this helps.

Bob Peters

ThomasBricker wrote:

Ok so I shot in Adobe color space.
Color looks pretty grim.
Is it just that it will only look good in Photoshop set to Adobe 1998?
I tried it and it looked a little better.
I personally think that ColorMatch RGB is the best RGB space of all
of them.
What do the rest of you think?

BTW, how do people "load" this "White Wedding Curve I keep hearing
about?
What is it exactly?

These are the Adobe RGB shots:

  • T

fjp Veteran Member • Posts: 5,145
Re: Adobe Color Space: Not Lovin' It... (Pics)

Andrew wrote:

If you are printing photographic, I think most inkjets, or for web,
sRGB is the colour space to use. Adobe rgb has more gamut, but it
wont print properly anyway, it is more for offset type printing
(magazines etc)

Adobe RGB prints just fine if you do it from Photoshop.

-- hide signature --

FJP

1stAde Regular Member • Posts: 482
Brilliant - Thanks

Iliah

Thanks for your patience and efforts. Do you switch between the two solutions through trial and effort or is there an easy way to find out from the source data?

thanks
Adrian

Iliah Borg wrote:

If you have the patience

if I have English

First, why we need rendering intent. While converting from one
color space to another, we deal wit a problem of "nesting" - source
colour space can be larger in some (or all) directions then the
destination color space. That is, source colour space may have
wider gamut then the gamut of destination.

Some colours from the source colour space can get out of the
destination space - what to do with those colours? Rendering intent
gives the method of dealing with this problem.

4 major rendering intents are called
absolute colorimetric
relative colorimetric
perceptual
saturation

Colorimetric intents leave the colours that fall into destination
gamut fairly intact - the colors are only adjusted for accurate
reproduction in destination space. Out of gamut colors are
compressed to bring them into destination space.

The difference between abs. and rel colorimetric is in how the
white is addressed. For relative colorimetric, white from source is
translated into white in destination; so, white is preserved.

Absolute colorimetric is designed for proofing, so it makes an
opposite - white from destination is considered white in source.
Thus I can proof printing on yellow stock paper with my Epson 2100
on white Luster.

With perceptual intent, all the source colours are compressed
(even those that actually fall into destination gamut "as is").
This is done to give more space for out of gamut colours. As a
result, all colours are "wrong", but the total impression is
"right" - as the relations between colours are kept.

Saturation intent maps all the colours individually, without
respect to neighboring colours. As a result, different out of gamut
colours can be mapped to the same destination colour. This
rendering intent is pretty special. I use it with dull images, in
order to gain more saturation.

Now, exept for proofing and special cases, I mostly use
rel.colorimetric and perceptual intent. In many cases image colours
do not occupy the whole gamut of source space. Using perceptual
intent in this case compresses the colours unnessessary. Rel.
colorimetric intent in this case give more "life" to the image.

To round this up, here is what I do: if the source gamut is much
wider then the destination; and number of colours in the image is
large enough so that they substantionally fall out of gamut of
destination space - perceptual rendering intent is used. Otherwize

  • relative colorimetric. Preview in convert dialog helps

-- hide signature --
Iliah Borg Forum Pro • Posts: 25,750
My pleasure

Do you switch between the two
solutions through trial and effort or is there an easy way to find
out from the source data?

At "Color Settings" (Shift-Ctrl-K) I have "Relative Colorimetric" as default in "Conversion Options".

If the file is not 15-bit, I convert it to "16 bits/channel" before converting to another colour space.

Conversion is done by "Image - Moce - Convert to profile", using "Black point compensation". This allows for preview of the result and switching the intents to choose the best suitable.

With some practice the right intent is more or less predictable, taking into account the size and shapes of source and destination gamuts and how the colours from the image before convertion seem to fill the source gamut; but preview in conversion dialog really helps - given the monitor is more or less calibrated and profiled.

-- hide signature --

no text

Misha Hoichman Forum Member • Posts: 96
I do like the colors

I think the colors are very good: not too saturated, without color shifts. Look at the skin tones of the boy. I think they look real. The lips are not too red, as it frequently happens with sRGB.

When I switched from F80 + film scanner to D70, I was a bit upset about the colors: they were oversaturated, too plasticy, too strong red, skin tones leaned towards magenta. Then I switched from sRGB to Adobe RGB, and the right colors were back.

If you want the image be more visually impressive, you should probably increase saturation. If you don't like the yellow cast (I do like), then change the white balance.

Misha

beaucamera Senior Member • Posts: 2,232
Printing Progress

Iliah....I've finally made some progress on the printing front. I used Monaco Optrix/EZ color to create a printer profile for my Tektronix 740 plus. Not an easy task, however, because the printer is networked and the software wasn't designed for network applications and my scanner is attached to a different computer! Eventually I was able to get the printer profile on the network printer on my Windows 2000 Server. Then I set up the printer properties on the Server for color management with the profile I created and ICM management by the host automatically. The hardware and Server both have the same print feature selected for color correction: sRGB (not sure this makes any difference, though).

When I bring up an TIFF image (which I've already postprocessed and stored) in PSCS, I select mode and find that I'm in RGB mode and 16 bit color. Then I select assign profile and change from Adobe RGB (1998) to the Tektronix profile I created. When I do this it lightens the image an seems to add more white.

Now when I print with preview, the Tektronix profile appears. I use same as source as the print space. In the printer dialog box under advanced settting, I have selected ICM handed by host system and printer color as sRGB or automatic.

Now I get printed images that look like what I see on the screen! I don't save the image in the Tektronix profile, though, I just use it temporarily for printer and leave the Tiff in the Adobe RGB (1998).

I used the same method to create a temporary sRGB profile, saving the image as sRGB IEC 1966-2.1 in a separate directory. I am posting that image here so you can see the result in your browser.

There is still much to learn, but I'm coming along slowly but surely!
Thank you for your help.

Iliah Borg wrote:

I use the "print with preview" and the print profile I usually use
is SRGB. There is also another "print with preview" option: color
management or output.

Color management, and set destination profile (printer profile) to
printer/paper/quality combo matching profile.

If I'm using Adobe RGB, at this point and want to make a jpeg for
the web, what method do you recommend? I've been using the save
for the web feature in PSCS. I take my photos as NEFs, bringing
them in to Nikon View 6.22 for initial processing and then in PSCS
where they become 16 bit tiffs. This is the point I then "save for
the web."

Before "save for web" it is a good idea to convert to sRGB or
generic monitor profile. Better to do the conversion in 16-bit
mode, though the difference is not that great.

Where should this adjustment be done?

Image - Mode - Convert to profile

In PS, Shift-Ctrl-K, please, and look at your settings to know what
is going when you open a file. See what printer profiles do you
have installed in the system; how is your driver instructed to deal
with colour; and try "print with preview" with correct profile set
as your printer profile.

I've reviewed this many times to ensure the correct settings are
used, but I think the problem is in software color space .

List your settings please

Iliah Borg Forum Pro • Posts: 25,750
Re: Printing Progress

Then I select assign profile and change from Adobe RGB
(1998) to the Tektronix profile I created.

Normally it should be convert to profile, not assign

-- hide signature --

no text

beaucamera Senior Member • Posts: 2,232
Re: Printing Progress

When I use convert to profile, not assign, there is no color change on screen as I had observed with assign. Guess that solves that mystery.
Would I use the same process then for the web...use convert, rather than assign?

Iliah Borg wrote:

Then I select assign profile and change from Adobe RGB
(1998) to the Tektronix profile I created.

Normally it should be convert to profile, not assign

Iliah Borg Forum Pro • Posts: 25,750
Re: Printing Progress

beaucamera wrote:

When I use convert to profile, not assign, there is no color change
on screen as I had observed with assign. Guess that solves that
mystery.

Convert changes the ink numbers and preserves the look (also preserves the look on the display in PS); assign does the opposite. Normally one converts to printer profile and switcjes off printer colour management.

Would I use the same process then for the web...use convert, rather
than assign?

Exactly right.

Now, a bit of theory. There is a mechanism called CMS - colour management system. When you have an input file with profile attached, and CMS knows what is the profile of the monitor, it converts before outputting to the monitor to monitor profile. That is why (given profiles are good, and monitor is capable) it makes no difference on screen when you convert the input file to different profiles. CMS takes care for you to preserve the look.

-- hide signature --

no text

beaucamera Senior Member • Posts: 2,232
Theory

Okay...so I need to set up my monitor profiles next, right? Then, when I bring a camera image into the CM system (e.g. Nikon View, Nikon Capture or PSCS), the monitor will show the image with the correct monitor profile...and the embedded (camera) profile will remain. Is this correct?

Iliah Borg wrote:

beaucamera wrote:

When I use convert to profile, not assign, there is no color change
on screen as I had observed with assign. Guess that solves that
mystery.

Convert changes the ink numbers and preserves the look (also
preserves the look on the display in PS); assign does the opposite.
Normally one converts to printer profile and switcjes off printer
colour management.

Would I use the same process then for the web...use convert, rather
than assign?

Exactly right.

Now, a bit of theory. There is a mechanism called CMS - colour
management system. When you have an input file with profile
attached, and CMS knows what is the profile of the monitor, it
converts before outputting to the monitor to monitor profile. That
is why (given profiles are good, and monitor is capable) it makes
no difference on screen when you convert the input file to
different profiles. CMS takes care for you to preserve the look.

Iliah Borg Forum Pro • Posts: 25,750
Re: Theory

beaucamera wrote:

Okay...so I need to set up my monitor profiles next, right?

Yes. You can use special profiling software/hardware combinations to obtain more or less reliable monitor profile; you would need to stabilize ambient light at your working place and use a neutral grey deep monitor hood.

Then,
when I bring a camera image into the CM system (e.g. Nikon View,
Nikon Capture or PSCS), the monitor will show the image with the
correct monitor profile...and the embedded (camera) profile will
remain. Is this correct?

Yes, it is correct. Maybe it is a little better to say not "with monitor profile", but "through monitor profile"

-- hide signature --

no text

beaucamera Senior Member • Posts: 2,232
My Homework Assignment

Okay, Illiah, this will take some time. I have several monitors & LCDs. The ones I use the most for imaging work are Sony G500 CRT's. I tried profiling one of these before and it took hours because the adjustment setting are manual. I tried once again and the Monaco stuff would not work, so I got a little discouraged. I don't have a monitor hood....so I'll do the best I can. I'll need to get back to you at another time after I've had a chance to play around with this stuff again. THANKS!

Iliah Borg wrote:

beaucamera wrote:

Okay...so I need to set up my monitor profiles next, right?

Yes. You can use special profiling software/hardware combinations
to obtain more or less reliable monitor profile; you would need to
stabilize ambient light at your working place and use a neutral
grey deep monitor hood.

Then,
when I bring a camera image into the CM system (e.g. Nikon View,
Nikon Capture or PSCS), the monitor will show the image with the
correct monitor profile...and the embedded (camera) profile will
remain. Is this correct?

Yes, it is correct. Maybe it is a little better to say not "with
monitor profile", but "through monitor profile"

Iliah Borg Forum Pro • Posts: 25,750
Re: My Homework Assignment

beaucamera wrote:

Okay, Illiah, this will take some time. I have several monitors &
LCDs. The ones I use the most for imaging work are Sony G500
CRT's. I tried profiling one of these before and it took hours
because the adjustment setting are manual.

You can try restoring default settings, adjusting brightnrss, contrast and colour temperature with Adobe Gamma, then taking Adobe Gamma out of Startup, and installing canned profiles for your monitor model.

I tried once again and
the Monaco stuff would not work

Which stuff exectly was it? BTW, new Spyder seems working...

I
don't have a monitor hood....so I'll do the best I can.

You can make it yourself, IMHO

I'll need
to get back to you at another time after I've had a chance to play
around with this stuff again.

any time
--
no text

Keyboard shortcuts:
FForum MMy threads