Japanese comics are called manga, and the artists who make them, mangaka. They have a contract with a publisher, and the story and characters they create fall under copyright protection. So far, so good. But how do starting mangaka secure a contract with a publisher? By putting their artistic skills on display. And how? By drawing a comic, with the whole printing/publishing paid out of their own pocket, and selling it at a price intended to cover printing/publishing costs. And one way of ensuring the comic will sell is using known, that is, copyrighted, characters. Such unofficial comics are called doujinshi (singular: doujin), and their makers, doujinshika. Since they're meant to draw attention to the artists' skill and, if using copyrighted characters, somewhat illegal, doujinshi are printed and sold in small numbers. The whole doujinshi industry is tolerated because mangaka whose characters are being used often started as doujinshika themselves. Not all doujinshi use existing characters, though, and not all doujinshika are starting artists; some mangaka also draw doujinshi in which they parody their own manga. The gag/parody type of doujin is called "aniparo" and may contain manga, anime, and even game characters. Like fan fiction, doujinshi often explore the comic/dramatic potential of characters that their own series doesn't have room for, and like fan works in general, doujinshi can be real gems. But because of their unofficial nature, they're rare and hard to find.
The description above is how the situation was before I heard of doujinshi, and is already a little outdated. The financial Powers That Be have noticed that consumers will pay big for doujinshi, and the scene is losing its innocence. Doujinshika may now be contracted by publishers, which basically means that what they produce qualifies as manga, and so the industry is getting a bit less tolerant to having copyrighted characters used by others. I read a letter on the Web from an artist who had been arrested to set an example. It's sad. Hopefully, the aniparo genre will survive.
As indicated above, good doujinshi can be rare and hard to lay one's hands on. No, I don't scan them and burn the scans on CDs and sell them. Not only does that qualify as piracy, but it's just too much work. The scans I have made were for someone who was bidding on the same doujin as I was, and lost. This person offered me scans from her doujin collection if I would scan this doujin for her, and I agreed. Normally one can say "buy your own copy", but in this case that's just not feasible, because there are no other copies to buy. (To date, I've never seen that doujin title up for auction again.) To me, availability is more important than copyright; I wouldn't burn a "free" copy of a currently best-selling game for someone unless I knew that someone to be too poor to buy it, but I would burn a copy of a rare and no longer sold text adventure just to stop it from disappearing. Incidentally, even piracy proper doesn't always hurt the pirated party; I bought volumes 1-11 of the "Gravitation" series and all the "Gravitation Remix" doujinshi because I'd seen the latter on a CD of scanned (ie. pirated) doujinshi I bought, and liked them. No genuine doujin lover will settle for scans if they can get the original. Though better than nothing, scans are a poor substitute for the paper version, as anyone who's made them can tell.
Primo: get a scanner. A good one. This is harder than it sounds: having bought the HP 4670 scanjet to make scans without destroying the book's spine (the scanner is a "window" that can be laid over the book) I found it to be flimsy (it's already been replaced once) and its scan quality absolutely horrendous, so I ended up using a Lexmark printer/scanner combination that came with a new computer. The old parallel-port scanner I had was no longer fit for use, dirt having somehow gotten under the glass plate and producing scans with a smear down the middle, but it must be said that the old scanner drivers at least did their work when scanning colour; both new scanners I have produce worse colour scans than the old one, with the Lexmark greening pink shades and mottling colours, while the HP makes scans that are not only mottled but (both for colour and greyscale scans) banded. It is very fortunate that doujinshi are mostly B/W!
A familiar problem with the HP Scanjet 4670.
Secundo: open the doujin at the page to be scanned and insert under the scanner lid, keeping it as straight as possible but not against the scanner plate's edges, since what's right up against the edge tends not to get scanned. For B/W pages, I've had the following tip to stop the pictures on the other side of the page shining through: put a piece of black paper behind the page being scanned. This also works for coloured pages that have a lot of white in them, but otherwise the black may just muddy the colours. The scanner software will usually have settings to autocorrect the raw scan; I never use autosharpening, but I do use pattern removal to reduce the mottled effect in colour scans, and set the resolution to 300 or 400 DPI, since I'll be reducing the size later anyway. (That is, for blackwhite and two-tone scans; colour scans may lose colour if reduced.) An A4-sized scan will never be as sharp as an A4-sized print on paper; to get the same sharpness, the scan has to be blown up to an impractical size. For B/W pages, I use "Greyscale" instead of "Colour" to keep down scan file size. While scanning, hold the doujin down firmly with both hands, trying to find a happy compromise between getting scans that are smudged where the pages meet and breaking the book's spine. (I advise against scanning thick books, because in scanning a story from a "Lunatic Party" volume, I did break the spine!) After some sixty pages, you should be feeling your back, and the scanner should be hot and creaking ominously. You're not and it isn't? Lucky thing. Now, depending on the scanner software, every scan has to be saved separately or you can make a series of scans and save them afterwards. Beware: the latter may eat up RAM and cause the application to hang or quit, losing all unsaved scans. Either way, save them as PNG or TIFF, or another format with lossless compression; JPG shouldn't be used until after editing, if at all.
I now have several MBs of raw scans sitting in a directory somewhere. Most of them will be at an angle, since it's hard to keep a book exactly straight on the scanner, and the pages may not be exact rectangles to start with. There will be stuff around the edges to trim off. Using the trick of putting a sheet of black paper behind the scanned page, the whole page will be slightly too dark. The pattern removal software will have made the image slightly fuzzy. The image will be TOO BIG. What I want is a scan small enough to (almost) fit on the screen, yet large enough for all the lettering to be legible and to insert translated text for scanlations. In practice, this means making the image twice as large as is practical for viewing, because otherwise small scribbled text becomes illegible. I want all pages of one doujin to be the same size, so first I straighten, crop and resize the cover scans, which are slightly bigger than inside pages because the covers, unlike the pages, can be laid entirely flat. The covers are straightened in the same way as normal pages (see below). When I've resized the back and front cover to the same size, I note the figures somewhere (for instance, 1172x1648) and use those when resizing and cropping the other pages.
(Yes, depending on how the doujin is arranged, the covers can count as pages. This means that the first actual page (the sheet with art on it) is usually numbered "03". To keep numbering consecutive, I therefore numbered the front cover "[name]02.png". But this still suggests missing pages. For any scans I make now, I number the front and back "[name]00.png" and "[name]zz.png".)
The covers, ready to straighten, crop and scale.
Now the very first thing to do is to get that crooked page as straight as possible. I've used Paint Shop Pro 7 to edit the scans, but most graphics applications should have the same functions: notably, layers. I open the layer window and double-click on the "background" layer to make it a real layer, that I can shift around. Next, I rotate the image by a small degree (usually 0.1 to 0.7), undoing the rotation if the image is still tilted and trying a different amount, until it's as straight as I can get it.
Two scanned pages, one almost perfectly straight, the other clearly
skewed. The straight page's top edge shows that the page itself is not exactly
rectangular.
Testing the straightness of edges can be done by selecting the whole image and then adjusting the size of the selection until it neatly encases the page;
The difficult way of testing straight edges.
or, much easier, by shifting the layer against the image window's edge, making sure to scale the page to its full size first to show the thin black margins at the page edges that won't be visible when the still very big image is scaled to fit the screen. I usually move the bottom corner of the scan against the bottom corner of the window, select the window (Ctrl-A) and crop (Shift-R); then I shift the top edge of the scan against the upper edge of the window, select and crop again.
The easy way of testing straight edges.
For either method, I leave a generous margin on the side where the pages meet in the centre. It's almost impossible to get the edges completely straight; I usually crop when they still slope, but the sloping bit can be cut away without losing much of the image. Yes, that means losing about one millimetre of page on all three sides, and this while the art in Japanese comics may extend all the way to the edge of the paper. Just another reason why originals are better than scans. If the page straightness was tested by drawing a selection around the page, I can simply "Crop to selection". If I used the second method, shifting the layer around, there will now be a transparent area around the scan. I select the transparent area with the Magic Wand, invert the selection (Ctrl-Shift-I) and crop.
As I've used the tip of inserting black cardboard behind the scanned page, and as the scanner I use produces such bright light that even the tiny pits in the paper's surface cast shadows on the scan, the page is grey and mottled, as the image below shows. What I want is to change all the light grey stuff that ought to be white, to white. This trick works best for greyscale images, as in colour images the colours are changed: open the Highlight/Midtones/Shadow adjustment dialog (Shift-M) and set Shadow to, depending on the level of murk, 95, 90 or even 85. This clears up the image wonderfully. When I've forgotten to use black cardboard, this also wipes away those very light shadow images from the printing on the other side of the page.
From grey to white.
Now the picture is as clean as I can get it, I can resize (Shift-S). Due to the excess width, I can't just resize to the same size as the covers: I have to fill in the same height as the covers, and let the width calculate itself. In order to do this, "Maintain aspect ratio" should be checked!
The page resized, and hopefully wider than its final width.
The page is now the right height, but resizing it has made the already blurry (due to the pattern removal setting) and washed-out (due to the highlight adjustment) lines even fuzzier. The following menu option will bring crispness back to the image: Effects, Sharpen, Unsharp mask. The settings shown in the picture below are usually sufficient; a strength of 30 already greatly improves the image.
Miraculous improvement after applying the unsharp mask.
The page is now, hopefully, a little too wide. A scanned page will never be quite as wide as the cover (unless I'm prepared to break the spine - which I'm not) and so to keep all scans the same size, a page scan will always include a bit of edge from the opposite page. This is why, in cropping and resizing, I always leave a bit of the other page on. Now is the moment to trim off the extra using "Canvas size":
In this case, the surplus has to be trimmed from the left side.
Sometimes, the page may be too narrow. In that case, I also use canvas size, adding some space on the side of the opposite page, and fill up the empty edge with white or a shade of grey that's not too noticeable.
Finally, to reduce image size a little more, I check if there are any edges to clean up. There are often light shadows around the paper's edge that adjusting the highlights doesn't quite clean up. In fact, there's often patches of murk that adjusting the highlights can't clean up without losing image detail, but which cover lettering and lines, so they can't be removed. A third reason why scans are always inferior to the original! If the murk shows on areas of white, it can simply be whited out by drawing a filled white rectangle over it, or using a big hard-edged white brush, or by selecting and cutting the area and then filling it with white.
Murk around the edges that can be whited out in a variety of ways.
For scanning coloured images, see the next how-I.
