2011-06-04

The making of, Part 5: Putting it together.

Several people have asked me about the tools I used to create The Mouse and The Bean. Here is a brief rundown.

To start with, my computer is running Fedora Linux. I avoid using Windows (except for gaming) because I value the freedom which Linux provides. When I run Linux I know that the people who made the software usually have my best interests at heart. There are no anti-features, or features which serve to benefit some foreign business at my expense. There is also a wide variety of professional tools for just about any purpose. Some of these tools are the best in their categories. For graphics design and publishing there are some decent tools available.

  1. I typed the text using a text editor called jEdit. Just about any text editor would do, but I like jEdit and I'm used to its features.
  2. I scanned the drawings using Xsane. This is not the easiest tool to use, but it works and is fairly powerful. Depending on your scanner, you might have much better Windows software available. I found that the Epson software which came with my scanner was really good. But while in Linux, Xsane does the trick.
  3. I edited the photos using Gimp. Gimp is pretty good. It has some flaws. But for my needs it was more than enough. Sometimes its user-interface design gets in the way, but feature-wise it's hard to beat.
  4. I put all this together in a program called Scribus, which is a desktop publishing program with lots of features. I probably used 1% of the features. Scribus is awesome. But it's not super-easy to use. However, it does let you do layout and publish to PDF, which is what I needed. Scribus let me import the text from the text file and put it into linked text frames, so I could adjust the size and position of the text frames on the pages and the rest of the document would re-flow.

Creating the layout of The Mouse and The Bean was, for me, the most time-consuming part. Writing the story was actually much easier because it's a more straight-forward process. You put words on paper (or, in my case, into the text-editor) and then you change the words until you like the result. Doing the layout for an actual book requires a skill which I have not practised much. It took lots of tweaking to make it happen the way I wanted.

The layout of the original prototype book was very basic. Essentially, each page had a 4.5" top margin, and given the font size that made a certain number of pages, which I printed in booklet form. Then I put in an illustration for each page, according to whatever was on that page. It was easy, but not that professional. Ideally, you pace the story and the illustrations according to what is happening, not how many inches of paper you have.

For the new version, I had a few significant changes to make, layout-wise. First, the book had to fit into at least 32 pages. I added some pages, such as a title page, copyright page, dedication, about the author, etc, but I still had a certain minimum number of pages to fill.

Second, I had new illustrations, some of which were based on the original illustrations, and some which were brand-new. Even though the original illustrations didn't fit the new pacing, when LeBinh sent me her sketches I knew I couldn't discard any, so I re-worked the layout to fit all the illustrations in. This meant that I had several typical layouts throughout the book: pages with just text and the illustration on the facing page, pages with a single illustration spanning both pages, and pages with one or more illustrations on a single page along with the text.

Finally, the new story was about two-thirds longer than the prototype. This meant that I did, indeed, have to ask LeBinh to discard some drawings as they didn't fit with the new direction of the story. Fortunately, she was pleased with the new ending and didn't mind making new illustrations.

Once we'd decided on the layout, and fit all the illustrations in place, I had to edit all the images to adjust their final appearance on the page. The most common problem was that the images didn't have a pure-white background. Because they were scanned, the background which appeared white was slightly off-white. The text, however, was on a transparent background and printed on pure-white paper, so the non-white of the images really stood out against the pure white background. Every single image which didn't completely cover the page had to be meticulously cleaned up. Any details which couldn't be cleaned up had to be faded, blurred, or cut. I tried several things before settling on some techniques which I found attractive.

My first approach would be to simply cut out any not-white and make it white. If this was possible it gave the best results, as the foreground of the image would stand out against the page. This was difficult for me to do reliably, but I managed to do it properly for the spiderweb on the last page and several other pages. If this failed, I moved on to the next approach: blurring.

When an image couldn't have the not-white removed, I would try instead to blur or blend it into the white background. This would give a better transition between the pure white and the image itself. I even blurred edges which weren't supposed to be white, to make the images site better within the page. But sometimes, this wasn't enough to make it look good.

When I couldn't blend, I had to cut. At least one image didn't fit properly on the page because the background wasn't drawn all the way to one side. This background would have let me cut out the exterior and leave the interior, but as the line didn't reach the edge of the page it wouldn't work. So instead I cut out that whole line and blended the interior of the image to the page. The end result was probably not what LeBinh envisioned when she sent me the image, but it made for a nice effect and she said she liked it.

After all the cutting and blending and cropping and tweaking was done, I was ready to save the PDF and upload it to Lulu. At the last minute I realized that I had misread one of the export options and had been over-compressing the resulting PDF, thus harming the quality. I quickly fixed my error and was shocked with a 500MB PDF file. Luckily, Lulu has an FTP service, where you can upload large files fairly easily. I uploaded the book and waited eagerly for my copy to arrive.

When the first copy arrived, I was generally pleased with how it looked. I made a few minor edits to the text, I adjusted the image on the last page of the story, and I adjusted all the images (in a batch) to make them brighter (more gamma). The reason is that the Lulu process gives fairly dark prints, much darker than what appears on my computer monitor. I knew this would happen, but I was surprised to see just how dark the images were. I tried printing the images locally to see how they'd look, but my printouts were so different from the book that it was meaningless to compare. Instead, I adjusted the images by 20% gamma and hoped for the best. The next copy was much better, colour-wise, and thus I made the book available for sale.