I recently used print-on-demand technology to publish a book. This is the first time I used this technology, so my experience with it is not extensive. Nonetheless, my findings may be useful to other authors. These findings are based exclusively on my experience with this book, and I make no claim that they are representative of the typical quality of books produced by these companies. The books shown below are proofs of an earlier manuscript version, not the final book version.
Not having a direct experience with this type of publishing, I initially selected two companies, CreateSpace and Lulu. Both appear to be among the largest in this type of industry. CreateSpace operates in tight connection with Amazon (according to other web sites, CreateSpace is owned by Amazon, but I did not seek a direct confirmation of this). Lulu as a company is independent of Amazon, although it does offer distribution through the latter channel.
In order to directly compare the book quality and appearance offered by these companise, I created book projects on both web sites, and went through the process of developing the same book with either company, up to the stage of ordering printed proofs. The projects were prepared so that they are as similar to each other as possible.
The same PDF file was used for the interior text. The covers were prepared in slightly different formats, as required by the different submission rules and different book thickness. As a result of the different paper types, the CreateSpace product (above, left) was significantly thicker than the Lulu one (42 mm versus 34 mm for a total of 708 numbered pages). CreateSpace uses a white, quite thick paper that feels like a printout coming straight from an office laser printer. Lulu uses a thinner light cream paper which looks and feels more similar to a conventional book paper. Aside for the different color of the glue used for binding, the binding type and cover thickness feel practically identical.
The nominal book format is 6" by 9". In both cases, the height is very close to 9", while the width is closer to 5" 7/8. I further noticed the following differences:
- The CreateSpace color on the cover is lighter and more similar to the original file. The Lulu cover is darker (and remained this dark also in subsequent book copies and hardcover bindings). Cover image quality is very good in both cases. The printing raster pattern can be seen only if one looks closely at the gray background of the cover image. It is invisible in patterned areas of the image. There is a slightly uneven mottling of this gray background, which probably represents the limit imposed by the paper quality and printing process. If you are concerned about this, try to avoid broad areas of uniform gray or neutral color in the cover image. Uniform areas of strong pure color (like the red background in my case) look solid, with no visible pattern. The edges between white characters and red background are very sharp. The rear cover carries white 11 pt "blurb" text on red background, also very sharp.
- Equations and foreign (Japanese) script were included in the text as 600 DPI bitmaps. They look fully as sharp as the text. There is no difference between the CreateSpace and Lulu products in this respect. Ordinary text is also identical.
- Drawings were included as 600 DPI 8-bit grayscale bitmaps. Their black lines and black lettering look very sharp. Gray areas are rastered in the same way as photographs (see below). As a whole, both companies produced almost identical results in this respect.
- Photographic illustrations were included as 300 DPI bitmaps. The raster in gray areas is quite visible (above details). The dynamic range is also relatively poor in both examples. I used this knowledge to prepare a high-resolution version of selected book illustrations, which can be downloaded for free by book buyers. The printed illustrations alone would not be sufficient to judge fine picture detail or overall picture quality. Perhaps color illustrations and/or photo books are printed differently, but I have no experience with the latter. Given the large page count of my book, I was forced to use ordinary paper with either company. As a whole, Lulu's raster is more uniform, finer and slightly less "offensive" to the eye (above). Black lettering often is printed with a visible raster along its edges, which is slightly disturbing but does not make text illegible (except for very small fonts). In the above figures, the text at the bottom of the figures is printed at 8 pt, to give an idea of the image magnification (note that this is not rastered text, and not part of the illustration).
- The differences between photographic illustrations from CreateSpace and Lulu that I observed may or may not be typical. For instance, I found substantial differences among different book copies printed by Lulu on different occasions (the above are from copies of the same book edition, albeit they may have been printed at different geographic locations). In the above example, the leftmost illustration is about right in tonality and contrast, while the rightmost one is definitely darker and "muddier". The background areas in the rightmost example are also less uniform. These examples were scanned from two different paper copies at the same time and with the same scanner settings. I don't have multiple copies of CreateSpace illustrations to compare with the above. As a whole, this suggests that the image quality observed in a set of proofs may not be indicative of all copies of the same book.
In the end, I decided to publish and distribute my book through Lulu instead of CreateSpace. The two companies provide significantly different services in several respects not mentioned above. For instance, CreateSpace offers a higher percentage of profits from book sales to authors, especially for what concerns Amazon sales. On the other hand, Lulu offers the option of hardcover binding (for certain book formats only), which is not available from CreateSpace. Hardcover binding and a more reasonable paper quality and thickness (especially important for a book of this size) were the primary factor that made me choose Lulu, in spite of the lower profit margins they offer to authors. Different projects may very well have different requirements, so my choice is by no means the "best" one in all respects.
Lulu uses a rather complicated algorithm to determine the text that will be printed on the spine of hardcover books (on the cover itself, not on the dust jacket, which is supplied by the author). One can specify both book title and author name when configuring the cover (as I did in the above example), but this is not what was printed. It is quite difficult to predict what the spine will look like, and I am aware of even worse examples by other authors, in which some book copies have a vertical text and other copies of the same book a horizontal text. It seems from the scarce Lulu information that it is in fact impossible to have both book title and author name simultaneously on the spine, which seems extremely odd to me. On the other hand, most customers likely will never remove the dust cover, and will not notice the problem. If you wish to entirely avoid this problem, choose a hardcase binding, which has a full-color, non-removable cover instead of the removable dust jacket.
page hits since June 13, 2011
Copyright (c) by Enrico Savazzi, 1990-2011. This page was last updated June 14, 2011
This site is ad-free. If you see any ads here, they are added by your ISP, or by spyware on your computer, or you are visiting this site through frames of another site.