So far the results always dissapoint a little when I receive the book. I understand Blurb uses a CMYK printing process and know I can't expect Steidl quality B&W but I would still like to get as close as possible.
The problem isn't CMYK per se, it's partially the Indigo press and it's screening and equally Blurb who's output isn't anywhere as good as it could be. I know exactly what you're talking about, I've made many tests to a number of book publishers including Blurb and getting neutral B&W is very difficult if not impossible. If you print an entire page or better spread of a fixed neutral, you'll see shifting from green to magenta across the page. Nothing you can do about it. If you are on Mac and have access to Aperture, you might try using their printing process, while it can't overcome the Indigo screening issues, I've seen vastly better color and neutral output from whoever Apple is using to print their books. Or, try an overall warm or cool tone instead of shooting for dead nuts neutral which simply isn't possible.
I second DD's remark. This is a lost cause from the main book printing sources if you want dead neutral. They simply do not control their process well enough. There are a few places such as bay photo or adorama that print (or used to, haven't done it in a while) using photographic printers and they do better but they are much more expensive. They basically print on photographic paper (even on real black and white paper if you want!) and then bind the prints for you. If you go with blurb, your best bet is to not fight it and dial in a little toning in develop so that you feel like you have some control over the look. Just make sure you do the same toning for all your images, otherwise it is going to look terrible.
This won't help with the LR Book module issue, but I've gotten good B&W results using InDesign with the Blurb InDesign plugin. All placed images are ProPhoto RGB Exports from LR and the book is then Exported to PDF using the Blurb PDF X-3 Export Preset with CMYK conversion to the Blurb.icc profile (not sRGB).
That's a good idea if only because you might force the output to only have black ink in it by that conversion and therefore end up with as neutral a conversion as possible. The blurb profile is complete fantasy otherwise as there is no guarantee they use the specific printer indicated by it at all nor that they even maintain their printers to the status indicated by the profile. I think digitaldog tested this out with blurb quite a bit and found that the profile is no good but I might be misremembering.
If interested there's a lengthy post discussing issues with the Blurb icc profile and how Blurb processes book PDFs. Essentially there is only one Blurb.icc file and its paper type agnostic. I ordered the Blurb Swatch Kit, which contains both B&W and Color images on each of the five (5) paper types. When looking at the B&W swatch sample images there are obvious contrast and color tone differences (yes!). The Standard paper contrast falls in the middle of the pack (12345) and has the most neutral B&W tone (i.e. no color). Andrew Rodney measured the swatch samples and I created a patch chart showing those measurements.
My guess is that the single Blurb.icc profile is actually the Blurb Standard paper profile. During RIP Blurb applies a paper specific profile regardless of whether the PDF is sRGB or CMYK converted with the Blurb.icc profile. In the latter case that means there are two CMYK conversions applied to the images. In both cases (sRGB, CMYK) color shift and gamut clipping are a distinct possibility.
Essentially there is only one Blurb.icc file and its paper type agnostic.
No such thing really. Profiles must take paper into account. Like you, I have that swatch book and measured each. They are all over the map in terms of their measured values for paper and worse, OABs. One size doesn't fit all but can if you don't mind suboptimal to really poor results!
My guess is that the single Blurb.icc profile is actually the Blurb Standard paper profile.
It isn't even a blurb standard paper profile, it's based on GRACoL which I'm fairly certainly isn't what they are producing on all those papers given the standards. What they are doing is supplying a profile for soft proofing that has no bearing on the output to make those who feel they kind of, sort of understand color management and want to soft proof feel better. If you owned an Epson printer and output to Luster paper, how would supplying a profile for that printer but based on Matt work out? Not well. This is akin to the silly photo lab, RGB workflow where you are given an output profile you can't use and told to send sRGB. Rather pointless but if it makes the user feel that by soft proofing they've done something useful, better than a sharp stick in the eye.
Based on sending the same images to differing vendors who produce books using Indigo (at least for the inside of the book), Blurb is far from best in class in my experience. And for B&W, on top of all the issues I've outlined, even the best shop is going to have issues with the Indigo screening which produces crossover in terms of color over the press sheet. That alone is a problem unless HP has updated their screening as they have promised for some time. Dead nuts neutral is hard to do!
Agreed on all points. The LR Book module outputs an sRGB profiled PDF, so even with a proper ICC paper profile there's bound to be color shift and gamut clipping with certain image and paper types.
Below are the Paper White Lab numbers you measured. The differences in Paper White alone are pretty apparent.
Thanks for all your helpfull and insightfull responses. What i'm thinking after reading them is that the B&W results I get from Blurb are not going to get better or more predictable by using another method then the Lightroom Book Module. Except maybe by using InDesign, maybe... I think I might need to give another printer a try when it comes to B&W and see if the results differ much. To bad the book module only supports Blurb.