Book Module, blurb and color management

[digitaldog]

So you can't be bothered to click on the web site icon then ?

Confirms my suspicions that you like to argue with people who have vastly more experience and understanding of a topic (color management) than you.
Hole 2 feet deeper. Let's move on; your concepts are based on color science fiction.

[Rhossydd]

I really don't like to argue with anyone, especially someone who is reasonably knowledgable about a subject.

However you're telling me I'm wrong with the evidence I'm seeing here.
When I give you measurements to back up my observations you say they're wrong, but won't say why or how one could make measurements that you'd accept.

I've not disagreed with your statement that the Blurb profile is in any way a definitive tool for soft proofing.
I simply don't understand why you can't see that there is a compromise view of it being useful without it being a 'correct' solution at the same time.
For a colour expert you seem to taking a very black and white attitude to this. It's really very unhelpful to just say there's no way at all of knowing what you might get from Blurb, when that doesn't seem to be the case in my, and others, experience.

[digitaldog]

I've not disagreed with your statement that the Blurb profile is in any way a definitive tool for soft proofing.

It's a totally inadequate one! It's color science fiction. I prefer Stephen King!
I just measured ONE patch from the two books from Blurb printed over time. The patch was from an image of a MacBeth Color Checker (purple). It's is nearly a dE of 8!
NO profile, custom and certainly not canned, that doesn't define the print conditions can be used for soft proofing when the output varies this much.

I simply don't understand why you can't see that there is a compromise view of it being useful without it being a 'correct' solution at the same time.

You've been told repeatedly why and more data just above but again, it will likely fall one one person's deaf ears. I'm OK with that.

For a colour expert you seem to taking a very black and white attitude to this.

Unlike you, I'm taking a colorimetric attitude and all the measured data thus far, show that Blurb doesn't supply a profile for their process and their process control kind of sucks. So yeah, pick ANY ICC output profile to soft proof and throw that dart blindfolded and enjoy the wasted time doing so. I'm here to help others; you made up your mind about how useful sRGB or a profile that doesn't define device behavior is for you. As for others, let em waste their time or not. The data is as clear as the nose on your face; Blurb's profile and Blurb's process control doesn't provide a sound color managed path. So long.

[Chris Kern]

Delta-E and color accuracy

In this 7 minute video I'll cover: What is Delta-E and how we use it to evaluate color differences. Color Accuracy: what it really means, how we measure it using ColorThink Pro and BableColor CT&A. This is an edited subset of a video covering RGB working spaces from raw data (sRGB urban legend Part 1).

Low Rez: https://www.youtube.com/watch?v=Jy0BD5aRV9s&feature=youtu.be
High Rez: http://digitaldog.net/files/Delta-E%20and%20Color%20Accuracy%20Video.mp4

Thanks for posting this: commendably clear.

[herminy]

Wowza! Er...thanks for responses...?  This is all helpful. As I am new to this forum and to using this type of forum in general, I thought I would get a notification via e-mail if anyone responded so I didn't realize all this was going on, and in the interim I did my own feeble experiment and sent in a "test" book to blurb for printing. Of note: My ultimate goal is to make 100 50 page books as part of a larger project my husband and I are working on. We're not rich, so I am researching the POD option (winds up being about $19 per book with Blurb)...which is leading to much hair pulling! Here's what I did: I tried 3 different options for several images, side by side in the book. Option 1, corrected in Photoshop using the ICC provided by blurb with gamut warning turned on. Option 2, same as previous but then pulled those images into LR and corrected further (if necessary) using sRGB profile as recommended by Blurb. Option 3, I just corrected for print in Photoshop as though I were going to print my image on my Pro-1000 at home, doing all corrections that I would normally do prior to adding an ICC profile for whatever paper I would use (so NO ICC profile other than what I set up for my monitor). I bet some of you can guess what I got back...They all basically looked the same, except that the ones with the sRGB correction looked kind of dead and overly desaturated in some areas. The good news? I get a proof copy before I go ahead and get 100 made. The bad news, I don't even feel very confident that the bulk order will match the proof copy! Mind you, some of the pics looked pretty great, as I chose the "Proline Pearl" paper option and it seems nice for the price. I don't know yet whether I will wind up going the Blurb route. Do any of you recommend any other POD options over Blurb? I am looking into AsukaBook, but I don't know yet if there is an option for such a small run of books that can meet what I'm hoping for print quality-wise. Thanks for any and all help here.

[Eric Myrvaagnes]

Hannah,

If you decide to go with Blurb, after you have a reasonably satisfactory sample copy, be patient! Wait for Blurb to offer a discount of 35 to 40% and then order your big batch.

Once you have bought a book, they will email you frequently urging you to order more, and 40% discounts seem to be offered several times each year, as are lesser discounts. I haven't seen any in excess of 40%.

Good luck with your project!

-Eric M.

[digitaldog]

I've got a Blurb book on it's way which was once printed over a year ago. Identical book sent through Lightroom except, a year or so apart. In the books are more than one page of a Macbeth which I"ll measure and report back. First report will be to compare the 24 patches printed this week versus a year ago. This will tell us how consistent the output is. Then I'll use the actual supplied Blurb profile to convert the Macbeth to get Lab values and compare what the profile predicts versus what the output is! This will tell us the accuracy of the supplied profile (on a mere 24 patches but that's a start). Stay tuned.

[herminy]

If you decide to go with Blurb, after you have a reasonably satisfactory sample copy, be patient! Wait for Blurb to offer a discount of 35 to 40% and then order your big batch.

Once you have bought a book, they will email you frequently urging you to order more, and 40% discounts seem to be offered several times each year, as are lesser discounts. I haven't seen any in excess of 40%.

Good luck with your project!

-Eric M.

Hi, Eric! Thanks! I actually contacted their customer service department for a quote on the larger order when I started the process. The book I hope to order, if I do get it from them, would "normally" cost about $55 for one copy, they are quoting me about $19 per for the bigger order, so it's a pretty good deal...if it turns out looking nice!

Andrew, thanks for your info and I will be curious to see what you come up with. I am getting a total crash course in color management since buying a Canon Pro-1000 printer a few months ago. It's not as painful as I expected...but it's still a little painful.

[digitaldog]

New book from Blurb with images printed through LR in the past. Here's just the color differences between the same image (MacBeth 24 patch target) on the outside cover and the inside:

--------------------------------------------------


dE Report


Number of Samples: 24


Delta-E Formula dE2000


Overall - (24 colors)
--------------------------------------------------
Average dE:   5.84
    Max dE:   9.13
    Min dE:   1.40
 StdDev dE:   2.12


Best 90% - (21 colors)
--------------------------------------------------
Average dE:   5.39
    Max dE:   8.52
    Min dE:   1.40
 StdDev dE:   1.87


Worst 10% - (3 colors)
--------------------------------------------------
Average dE:   8.96
    Max dE:   9.13
    Min dE:   8.74
 StdDev dE:   0.20


--------------------------------------------------
--------------------------------------------------


SAME Blurb ICC profile for download is said to describe the two; really?

[digitaldog]

New book versus old book, just the MacBeth from inside each:



--------------------------------------------------


dE Report


Number of Samples: 24


Delta-E Formula dE2000


Overall - (24 colors)
--------------------------------------------------
Average dE:   6.30
    Max dE:  19.24
    Min dE:   1.19
 StdDev dE:   4.34


Best 90% - (21 colors)
--------------------------------------------------
Average dE:   5.05
    Max dE:  12.14
    Min dE:   1.19
 StdDev dE:   2.70


Worst 10% - (3 colors)
--------------------------------------------------
Average dE:  15.01
    Max dE:  19.24
    Min dE:  12.82
 StdDev dE:   3.67


--------------------------------------------------
--------------------------------------------------
Max dE is pretty awful!

[Tim Lookingbill]

Now show a before and after of an actual photo of the two prints that comprise a wide range of colors in the form of actual photographed scenes (not just Xrite CCchart flat colors) so we all know what the visual differences these Delta numbers reflect.

Lab numbers can shift by as much as 5 points without showing a visual difference and Lab is the reference for Delta number mismatches.

I'm sure I'll get a straight forward and respectfully toned answer from Andrew as well as the A/B print proof I requested to back up what he says.

[digitaldog]

Now show a before and after of an actual photo of the two prints that comprise a wide range of colors in the form of actual photographed scenes (not just Xrite CCchart flat colors) so we all know what the visual differences these Delta numbers reflect.

[Tim Lookingbill]

And where's an actual photo taken of a real scene containing a wide range of colors to show how off the Blurb print is from the actual digital file.

If Blurb is using a CMYK offset commercial printer and no custom profile it's guaranteed it's not going to match "perfectyl" due to the very small gamut of commercial presses compared to even sRGB except with cyans. It's how off and objectionable it makes the photo look is what's important. Consistency between prints is second in importance. Split hair precision to the original sRGB file is wishful thinking with offset printing unless the paper is printed on glossy paper and even that will show anomalies down into the darker colors.

Inkjet printing would give a closer match but still not as precise as a custom profile across ALL possible photographed scenes.

And thanks, Andrew for the straightforward answer.

[digitaldog]

And where's an actual photo taken of a real scene containing a wide range of colors to show how off the Blurb print is from the actual digital file.

Some yes. It shows a visual difference. A very big one!
And as of the moment I'm typing this, you haven't even viewed the numbers:

[Tim Lookingbill]

I viewed the numbers but I don't and can't see visual differences in color patches that will translate into how screwed up it makes the printed photos compared to the original digital file.

Examining color patches in relation to Delta numbers for me is a waste of time when it comes to how it changes the aesthetics of the printed photo. Perfect match is the goal but how much time is spent getting there I have to way against whether it will increase sales or desire for the printed photo.

People viewing the printed photo will either say it looks like crap or it looks beautiful without having the original digital file to compare against. If they see the original first and there is a slight mismatch, if the photo is desirable they may just chock it up to the limitations of the technology.The decision on whether they want to buy the print whether it is in a book or hanging on a wall will be influenced by other factors other than just Delta E number mismatches to the original digital file.

Your posted Xrite CCchart prints shows it is far worse than I thought, but I don't know how much a custom ICC profile will fix it. I certainly will not be using Blurb to print books if it's that bad.

[digitaldog]

What the worst (dE 19) patch looks like:

[digitaldog]

People viewing the printed photo will either say it looks like crap or it looks beautiful without having the original digital file to compare against

They don't match. That's what you're missing. They should; they are suppose to both define, by the same profile, something useful for soft proofing OR even considering a match? They don't match. By a lot.

[Tim Lookingbill]

They don't match. That's what you're missing. They should; they are suppose to both define, by the same profile, something useful for soft proofing OR even considering a match? They don't match. By a lot.

And you're missing the point on whether those that would use Blurb to print photos they care about or want to promote as a way to increase sales should expect a perfect match to the original file or just be happy the print reflects the original intended aesthetic.

I've seen some color photos of our local parks printed on my local papers newsprint and have been surprised they're as good as they look considering they've been printed on what amounts to toilet paper. Oh, I doubt they match the original but I don't know how much is on account the newspaper uses a custom ICC profile.

[Tim Lookingbill]

I think what would help quite a bit for those deciding to use a printing service like Blurb is to present an original digital photo that looks fabulous and pleasing that contains a wide range of colors and show how the aesthetics (the like factor) of the photo are altered or reduced by showing twp versions of the print, one without a custom profile next to the one printed using an ICC profile.

That would really show how useful this technology is.

[digitaldog]

I think what would help quite a bit for those deciding to use a printing service like Blurb is to present an original digital photo that looks fabulous and pleasing that contains a wide range of colors and show how the aesthetics (the like factor) of the photo are altered or reduced by showing twp versions of the print, one without a custom profile next to the one printed using an ICC profile.

That would really show how useful this technology is.

By all means, please do so. On your dime. I've printed two books now on my own dime and provided my data. Please start your own print output project and provide data.
Without data, you're just a person with an opinion.