Ad
Ad
Ad
Pages: [1]   Bottom of Page
Print
Author Topic: Making custom profiles for ImagePrint RIP  (Read 4917 times)
PetterStahre
Newbie
*
Offline Offline

Posts: 40



WWW
« on: December 02, 2006, 06:11:03 AM »
ReplyReply

When I create custom profiles for ImagePrint the profiling software (Gretag Eye-One Match) always reports in the end that my profile might not work properly (after calculating the values). But when I try to use it it seems to work ok.

However, this makes me feel a little uncertain about the quality about the profiles I make. Specially since the IP manual is rather limitied on this subject.

Can anyone with experience from producing custom made profiles for ImagePrint see any errors in this process:

1)
In the Color Management window, tab "Bitmap": Missing profile/Untagged image for RGB is selected as "None"

2)
I import the first of the two reference charts (TC9.18) and position it.

3)
In the Color Management window, I choose a paper that is as close to the one I'm using. In my case I'd like to profile a Hahnemuhle PhotoRag Bright White, so I choose ImagesPrints own profile for that paper, in this case "ep2400mk_HahnPhotoRagBrightWht_RDAY"*

4)
I print the chart, and then the second one. I read them with the spectrophotometer in the Eye-One Match software and when it has calculated the values it reports concerns about the measurements (I don't remember the exact words).

5)
Upon inspection of the profile (in ColorSync) it seems ok but a little bit smaller than the profile I get when doing the same process for Photoshop-printing (less shadow separation it seems and generally smaller). When I try the profile in IP it seems ok, but I will do more testing on this.

This has happened approx. 4 out of 5 times when I try to do profiles for the IP. Is this "normal"?

* In the IP manual it states that the only thing used from this profile, when printing untagged images, will be the ink set. This seems a bit odd, but I tried other profiles like the "_TUNG" and it seems right since no differences can be seen on the output. Strange method, however (and why so little information about this in the manual?).

// Petter
Logged
digitaldog
Sr. Member
****
Offline Offline

Posts: 8574



WWW
« Reply #1 on: December 02, 2006, 07:22:25 AM »
ReplyReply

Step 1 (bitmap to none) is key here. This will push the file through the RIP without applying color management.

As to the errors, it happens and you might want to actually test the profile. But it shouldn't happen that often and there's error checking to see that the measured data and reference are not a mile off. Also run the EyeOne Diagnostic just to ensure your Spectrophotometer is up to snuff.
Logged

Andrew Rodney
Author “Color Management for Photographers”
http://digitaldog.net/
PetterStahre
Newbie
*
Offline Offline

Posts: 40



WWW
« Reply #2 on: December 02, 2006, 07:35:54 AM »
ReplyReply

Quote
Step 1 (bitmap to none) is key here. This will push the file through the RIP without applying color management.

As to the errors, it happens and you might want to actually test the profile. But it shouldn't happen that often and there's error checking to see that the measured data and reference are not a mile off. Also run the EyeOne Diagnostic just to ensure your Spectrophotometer is up to snuff.
[a href=\"index.php?act=findpost&pid=88227\"][{POST_SNAPBACK}][/a]

Thank you. I will do the diagnostic test, but since I make profiles for Photoshop with no problem I suspect something else. (But you never know in The Binary World   )

I will do another test and get back.

/ Petter
Logged
ericbullock
Jr. Member
**
Offline Offline

Posts: 67


WWW
« Reply #3 on: December 02, 2006, 07:54:47 AM »
ReplyReply

Quote
Thank you. I will do the diagnostic test, but since I make profiles for Photoshop with no problem I suspect something else. (But you never know in The Binary World   )

I will do another test and get back.

/ Petter
[a href=\"index.php?act=findpost&pid=88229\"][{POST_SNAPBACK}][/a]

You may also try setting the Printer/Paper pull-down to "Select Correct Printer Profile" and see if that makes an appreciable difference. Since the paper profile you chose has an "R" in it, it means its a recipe'd profile, so additional voodoo is going on in there. Just make sure in the Setup window you're choosing the Matte Black inkset (KMCMYcm).

If you're getting better color in Photoshop, this begs the question. Why use ImagePrint?

-eric-
Logged
PetterStahre
Newbie
*
Offline Offline

Posts: 40



WWW
« Reply #4 on: December 02, 2006, 08:39:21 AM »
ReplyReply

Quote
You may also try setting the Printer/Paper pull-down to "Select Correct Printer Profile" and see if that makes an appreciable difference. Since the paper profile you chose has an "R" in it, it means its a recipe'd profile, so additional voodoo is going on in there. Just make sure in the Setup window you're choosing the Matte Black inkset (KMCMYcm).

If you're getting better color in Photoshop, this begs the question. Why use ImagePrint?

-eric-
[a href=\"index.php?act=findpost&pid=88233\"][{POST_SNAPBACK}][/a]

Thanks, I forgot to ask about that one ... I can't find a profile without "R" in it to use for my R2400 or 7800. Maybee I need to download some special profiles from the IR-site which doesn't contain "R" in it. I will also try your recommendation and see what happens.

About the IR vs Photoshop  Yes, I've thought about it since I have such problems with the custom made profiles (and I don't regard the IR-made ones as good as I would like them to be, except from the grey profiles). But when outputting images with very subtle tonal gradations and saturated colors I find IR to do a better job. (Due to the 16 bit vs 8 bit output difference or the different rasterizing methods used -  I don't know, but it looks good.) Also it's a great time saver when I want to output a lot of images on roll media.

/ Petter
Logged
madmanchan
Sr. Member
****
Offline Offline

Posts: 2101


« Reply #5 on: December 02, 2006, 08:46:34 AM »
ReplyReply

Hi, I once had Colorbyte make profiles for Moab Kokopelli Semi-Gloss for an Epson 2200. (They didn't have profiles yet for this paper, so they had me print the targets and send the prints to them.) Here are the instructions they gave me:

2) Go to Image->ColorManagement. Your settings should be as following:

Output tab:

Printer - "select correct printer profile" (there is such option in that drop-down list)

Colorized - off

Proofer - NONE (very important!)

BITMAP tab:

Missing profiles:

RGB - NONE (very important!)

You can leave CMYK, Embedded and Gray fields as they are.

3) Select Quality that you want to use when the profile is built. 1440 is recommended. Select Page size big enough to fit the target.

4) Go to File->Print->Setup. Make sure Ink is set to KCMYcm Pigment (to
print on photo paper with Photo Black ink installed in the printer).
Print.
Logged

PetterStahre
Newbie
*
Offline Offline

Posts: 40



WWW
« Reply #6 on: December 02, 2006, 08:57:04 AM »
ReplyReply

Quote
Hi, I once had Colorbyte make profiles for Moab Kokopelli Semi-Gloss for an Epson 2200. (They didn't have profiles yet for this paper, so they had me print the targets and send the prints to them.) Here are the instructions they gave me:

2) Go to Image->ColorManagement. Your settings should be as following:

Output tab:

Printer - "select correct printer profile" (there is such option in that drop-down list)

Colorized - off

Proofer - NONE (very important!)

BITMAP tab:

Missing profiles:

RGB - NONE (very important!)

You can leave CMYK, Embedded and Gray fields as they are.

3) Select Quality that you want to use when the profile is built. 1440 is recommended. Select Page size big enough to fit the target.

4) Go to File->Print->Setup. Make sure Ink is set to KCMYcm Pigment (to
print on photo paper with Photo Black ink installed in the printer).
Print.
[a href=\"index.php?act=findpost&pid=88240\"][{POST_SNAPBACK}][/a]

This is very interesting, thanks. The main thing which differs from my settings is again the "Select correct printer profile" (and then check the ink setting carefully). I will do new tests during the weekend.

Thank you all so far for fast and helpful comments!

/ Petter
Logged
digitaldog
Sr. Member
****
Offline Offline

Posts: 8574



WWW
« Reply #7 on: December 02, 2006, 09:33:31 AM »
ReplyReply

Quote
About the IR vs Photoshop  Yes, I've thought about it since I have such problems with the custom made profiles (and I don't regard the IR-made ones as good as I would like them to be, except from the grey profiles). But when outputting images with very subtle tonal gradations and saturated colors I find IR to do a better job. [a href=\"index.php?act=findpost&pid=88238\"][{POST_SNAPBACK}][/a]

That's my take exactly! The canned profiles are OK but blues are not so great compared to custom profiles. The dither is better but you really need to look close to see it. But it's visible. The new Print Through Application (PTA) is pretty nice once you actually get it setup.
Logged

Andrew Rodney
Author “Color Management for Photographers”
http://digitaldog.net/
PetterStahre
Newbie
*
Offline Offline

Posts: 40



WWW
« Reply #8 on: December 02, 2006, 10:56:52 PM »
ReplyReply

I have now tried making new profiles using suggestions in this thread. The main difference was the "Select correct printer profile".

However, this doesn't solve the problem. This is what happens:

1) Profiling using "Select correct printer profile"
- This produces a profile that the Gretag software accepts (meaning no error message when calculating the profile). The problem is that it has severe problems with tonal transitions in shadows. It looks awful. I started to profile another paper just to check but since that printed charts seemed to have the same problem I stopped (some patches looks much to light when I compare to charts I profiled before for the same paper).

2) Profiling using "p2400mk_HahnPhotoRagBrightWht_RDAY"
- This is the ImagePrint profile for the paper I try to profile. When I print the charts using this setting the Gretag software report calculation errors. However, when I use this profile I actually get better tonal transitions in the shadows and the rest of the image than when I try the method above, but not as good as the IR-made profile itself. Otherwise I could have accepted the error report - but the error is visible on prints for shure. Now, I know from this discussion and the manual that I shouldn't use profiles with "R", but I can't find ANY profile for either my 2400 or my 7800 which doesn't have an "R" in it. (I downloaded all profiles for the 2400+7800 three months ago so the library is quite fresh.)

All other settings are double checked. I use the matte ink set "KMCMYcm" for this paper.

I feel I'm quite stuck here. To me it seems the problem has to do with the selected profile. The "Select correct printer profile" produces charts with some patches looking to light. And when I use the IM-made profile instead to print the charts I get patches that looks better (as in more similar to what I'm used to from Photoshop). But clearly the Gretag software doesn't agree with me.

When I first made the new profile using "Select correct printer profile" and printed 3 images it looked very good. Accurate flesh tones, neutral greys and over all very pleasing. But when I tried an image with most of the information in the shadows I got very disappointed. Brutal transitions.

Any suggestions?

/ Petter
Logged
ericbullock
Jr. Member
**
Offline Offline

Posts: 67


WWW
« Reply #9 on: December 03, 2006, 08:50:59 AM »
ReplyReply

Quote
Now, I know from this discussion and the manual that I shouldn't use profiles with "R", but I can't find ANY profile for either my 2400 or my 7800 which doesn't have an "R" in it. (I downloaded all profiles for the 2400+7800 three months ago so the library is quite fresh.)
[a href=\"index.php?act=findpost&pid=88350\"][{POST_SNAPBACK}][/a]

I did not mean to infer that you should avoid the Recipe'd profiles altogether. In fact for some printers and papers you have to use them in order to get the right amount of ink on the paper. All the printer profile does in IP is select the right black ink mode, and in the case of an "R" profile applies any necessary ink limiting.

What target are you printing when you make your own profiles through IP? I find that most of the time the TC9.18 is sufficient, although lately I've been using a 1728 patch target.

That is one area where the IP RIP is superior...shadow detail. The Epson driver way over-inks, and IP's RAW behavior is much more well-behaved. Their dithering is excellent too.

-eric-
Logged
digitaldog
Sr. Member
****
Offline Offline

Posts: 8574



WWW
« Reply #10 on: December 03, 2006, 09:48:34 AM »
ReplyReply

Make sure the targets are indeed untagged documents! That, along with the Bitmap tab is key to sending the numbers through the product without applying a profile (doesn't matter what color paper profile is selected in the popup). IF the file is tagged, the bitmap tab setting is invalidated, you apply profile to data.
Logged

Andrew Rodney
Author “Color Management for Photographers”
http://digitaldog.net/
ericbullock
Jr. Member
**
Offline Offline

Posts: 67


WWW
« Reply #11 on: December 03, 2006, 09:56:16 AM »
ReplyReply

Quote
...doesn't matter what color paper profile is selected in the popup...
[a href=\"index.php?act=findpost&pid=88412\"][{POST_SNAPBACK}][/a]

Except if you choose a recipe'd profile. Right? I'm under the impression that the recipe'd profiles do apply some underlying ink limiting, even if the chart is untagged and you set the RGB bitmap to "none".

-eric-
Logged
digitaldog
Sr. Member
****
Offline Offline

Posts: 8574



WWW
« Reply #12 on: December 03, 2006, 10:03:58 AM »
ReplyReply

Quote
Except if you choose a recipe'd profile. Right? I'm under the impression that the recipe'd profiles do apply some underlying ink limiting, even if the chart is untagged and you set the RGB bitmap to "none".

-eric-
[a href=\"index.php?act=findpost&pid=88413\"][{POST_SNAPBACK}][/a]

I'm not sure what you mean by recipe'd profile. All can say is everytime I've built a profile through IP, the doc is untagged and bitmap is set to none. Presumably setting a gray paper profile might be a bad move but with color paper profiles, it works on this end (I guess it's possible ColorByte changed something between versions).
Logged

Andrew Rodney
Author “Color Management for Photographers”
http://digitaldog.net/
PetterStahre
Newbie
*
Offline Offline

Posts: 40



WWW
« Reply #13 on: December 03, 2006, 12:05:39 PM »
ReplyReply

Quote
Make sure the targets are indeed untagged documents! That, along with the Bitmap tab is key to sending the numbers through the product without applying a profile (doesn't matter what color paper profile is selected in the popup). IF the file is tagged, the bitmap tab setting is invalidated, you apply profile to data.
[a href=\"index.php?act=findpost&pid=88412\"][{POST_SNAPBACK}][/a]

Yes, I've checked that. I've now measured the chart again and made an average from three different readings - still the same problem. I've also printed out new charts but they are identical in color to the one I've used for measuring. I've also checked the spectrophotometer using the diagnostic software.

I guess I will have to contact Colorbyte about this (to bad I didn't experience this during the support period).

And about the "R" profiles, here's what the manual says: "Important Note: Some profiles are of a special type, containing “recipes” for different combinations of inks and papers that require additional information in the profile to print correctly.  These profiles will have the letter R at the end of the name, just before the light temperature designation (such as RF3, or RDAY).  Using a profile with a recipe in it WILL affect the calibration print.  If profiles with “R” at the end of the name are available in your profile listing, and you are not sure if you should be using one when
outputting your target, call or email ColorByte Software to confirm. " (The last sentence has to do with if you send your target to CB to have a profile created.)

And just to be clear about this: I've used the "Select Correct Printer Profile" instead.

Anyhow, lot's of thanks to all who have contributed to this thread. I will repeat the profiling process for my 7800 and see if I can see any improvements, otherwise I will contact Colorbyte. Maybe a specific R2400-problem?

Cheers,
Petter
Logged
madmanchan
Sr. Member
****
Offline Offline

Posts: 2101


« Reply #14 on: December 03, 2006, 12:24:12 PM »
ReplyReply

Petter, it sounds like you're choosing all the right settings. I am wondering if you have tried printing a test target (as opposed to a profiling target) such as the ones that Bill Atkinson offers for download here:

http://homepage.mac.com/billatkinson/FileSharing2.html

Go to the Profile Test Images section, download the file, and look for the directory titled "Print Without Profile".  You will see a file called RGB Hues Target.tif. Try printing this using the same settings that you are using to print your profile targets (again, color management disabled in IP).  Assuming the IP driver for the 2400 is behaving nicely for your particular paper, you should see 12 distinct gray patches (i.e., nothing too weird between the darkest patches) that are roughly perceptually evenly spaced. If the darkest patches are all looking the same, then there's probably too much ink being put down, in which you might investigate using the Ink Limiting options in IP.

Also try printing the RGB Blends 1024x768.tif file and you should see smooth color gradations through the image. This is a diagnostic image to check for unexpected driver behavior. You might also find it helpful to download Bill's Targets FAQ (also from the same page) and read some of his info there about looking for profile problems. His context is using the native Epson driver, but the concepts apply to IP as well.

Out of curiosity, have you tried your profiling procedure using the native Epson 2400 driver?

Eric
Logged

PetterStahre
Newbie
*
Offline Offline

Posts: 40



WWW
« Reply #15 on: December 03, 2006, 02:43:39 PM »
ReplyReply

Quote
Petter, it sounds like you're choosing all the right settings. I am wondering if you have tried printing a test target (as opposed to a profiling target) such as the ones that Bill Atkinson offers for download here:

http://homepage.mac.com/billatkinson/FileSharing2.html

Go to the Profile Test Images section, download the file, and look for the directory titled "Print Without Profile".  You will see a file called RGB Hues Target.tif. Try printing this using the same settings that you are using to print your profile targets (again, color management disabled in IP).  Assuming the IP driver for the 2400 is behaving nicely for your particular paper, you should see 12 distinct gray patches (i.e., nothing too weird between the darkest patches) that are roughly perceptually evenly spaced. If the darkest patches are all looking the same, then there's probably too much ink being put down, in which you might investigate using the Ink Limiting options in IP.

Also try printing the RGB Blends 1024x768.tif file and you should see smooth color gradations through the image. This is a diagnostic image to check for unexpected driver behavior. You might also find it helpful to download Bill's Targets FAQ (also from the same page) and read some of his info there about looking for profile problems. His context is using the native Epson driver, but the concepts apply to IP as well.

Out of curiosity, have you tried your profiling procedure using the native Epson 2400 driver?

Eric
[a href=\"index.php?act=findpost&pid=88435\"][{POST_SNAPBACK}][/a]

Thank you. I have now printed both testimages you mentioned, using the same settings as when I printed the chart, and they both look good. Good separation between the patches and the gradation tests looks smooth.

Yes, I'm used to profiling through the native Epson driver (from within Photoshop) and when I compare the profiles made for the same paper, the profile for the Epson/Photoshop combo is much smoother and natural in the shadows. (And not by a small margin!)

But even if the patches/gradations looks good I could maybe benefit from experimenting with less ink since it's the Hahnemuhle PhotoRag Bright White I'm profiling. (But then again - there already is a distinct difference between all grey patches.)

/ Petter
Logged
Pages: [1]   Top of Page
Print
Jump to:  

Ad
Ad
Ad