Scanlations

[nharding]

I'm writing a replacement CDisplay program (for viewing comics / manga) and I wanted to add support for scanlations. I found a program that makes .ovl files which allows you to draw a replacement balloon and insert the text in the new balloon and then export the image. I was thinking of doing that type of thing but in real time, so you would have the original Japanese image, but the text would be replaced with English (similar to .srt files in videos) which means you could turn it off and see the original text (or edit the English text with a better translation). It will have the option to output the manga as a cbz file for support on non Windows machines.

What software do scanlators currently use, and what would you want in this program. (I don't read Japanese and wanted to contribute something)

Neil Harding

[Sosseres]

Don't scanlators use photoshop, gimp or a similar program to do that? There are so many things that has to be fixed on most scans that the typesetting is just a part of it. You also have to clean the scans, often redraw content that gets lost, especially over 2-page images or when the text you want to replace is over art.

What you want to do is basically allow people to typeset over the image file so people download the raw and the "srt"?

[nharding]

Yes, I was thinking it would make it easier than editing in photoshop and would preserve the original as well (since editing a jpg reintroduces more errors). It would also allow for partial translations to be released so you could have Look Out <>, then someone else could edit the text further (since they could look at the underlying image). It would only work with opaque speech balloons, but most balloons are opaque.

Neil Harding

[Sosseres]

If you want to put out a result quickly I can certainly see the use of what you are saying. I havn't scanlated a lot and havn't done it for several years at all, so I don't know how things are right now.

One point you aren't considering is that most releases use png for greyscale. It doesn't introduce new errors and is often smaller in size than jpeg for a well cleaned page where white is white. Most manga has huge white areas, if those are dotted with very bright grey dots (as is normal after most scans) it increases the file size a lot.

You really have to consider how you would handle pages such as:

http://www.mangafox.com/manga/peony_pavilion/v01/c001/4.html
http://www.mangafox.com/manga/peony_pavilion/v01/c001/6.html
http://www.mangafox.com/manga/peony_pavilion/v01/c001/8.html

It doesn't have to be a perfect solution such as can be made with some time in PS, since that isn't the goal as I understand it.

(All three of those were re-compressed by mangafox, proving that jpegs does make new errors.)

[nharding]

1) The contents page would probably use a rectangular section to be blanked out and the text would then be added as sections.
2) The text page would lose the background image (although I might try and add something to remove text only and leave colored regions)
3) The text on the white areas would be treated as unbounded text, and it would PROBABLY lose the cloud behind the 2nd text.

I wanted to make it easier for translator to work with, and someone else can then do the final layout, or if the results are acceptable then it could be used on it's own. I appreciate the hard work doing translations and wanted it to be easier for them. Plus I wanted a new viewer anyway that has some features that CDisplay is missing.

Neil Harding

[Sosseres]

Well I'll look forward to what comes out of it, hopefully somebody more into things like this can comment as well.

[nstgc]

I saw something before that did what I think you are trying to do. It was a hentai image gallery where users could overlay translations on top of the images. I can't remember where it was though.

That would be nice. When you scan something information is lost and each time they are lossly compressed information is lost. If you can avoid re-compressing the image, the end result will be nicer.

[Xiong Chiamiov]

I saw something before that did what I think you are trying to do. It was a hentai image gallery where users could overlay translations on top of the images. I can't remember where it was though.

Sankaku Complex's imageboard does this; I would provide a link, but, y'know, boobs and loli tits and vaginas everywhere.


I was about to suggest that you hack on Comix, rather than starting from scratch (NIH, anyone?), but it appears to be for UNIX-like OSs only, which I assume is not your target platform.  Python, PIL and PyGTK are all cross-platform, though, so it (theoretically) shouldn't be too hard to port.  I'd be interested in helping if you chose this route.

[nharding]

I'm a games programmer with over 20 years experience, and wanted to learn C#, so I am using XNA. I've got a basic viewer working already, it handles directory of images, zip or rar files and uses 3D card and shaders to apply real time contrast / brightness / gamma correction to the image as well. It uses the same keys as CDisplay, although those will be possible to change. It also correctly handles zip files with MacOS files in them (CDisplay normally crashes on these).

Neil Harding

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

One other thing that can be added is the ability to flip the image so it can be read left to right as well, so that could be an option for some people to read it in modified format or original format.

Neil Harding

EDIT - newy: Please edit your post if no one has posted after you, thanks.

[Sosseres]

Oh, something to consider outside of the scanlation portion. History of files read. It is nice to be able to read 3-4 series and just jump right back at the correct point using the reader itself. Though that might be a bit much, just a history over last read files that you can increase to x amount would be a feature I at least used. If you don't want history, then set it to 0 or disable.

[nharding]

Yes, the history is one of the things I want to change about CDisplay (CDisplay allows you to resume where you left off, but I like to check a scan when I download it and that means it loses the existing resume).

Neil Harding

[Xiong Chiamiov]

I'm a games programmer with over 20 years experience,

Ah, well, that explains your NIH syndrome.

and wanted to learn C#, so I am using XNA.

I have no experience with XNA, but doesn't it seem a bit odd to use a game development library for a simple image viewer?

I've got a basic viewer working already, it handles directory of images, zip or rar files and uses 3D card and shaders to apply real time contrast / brightness / gamma correction to the image as well.

Because that's totally necessary, amirite?  Granted, I haven't tried this out (so I may be completely in the wrong), but it seems like you could do all of that stuff in a rather naive way in a scripting language and still be fine.  You're not calculating softshadows here; write the simplest thing that can possibly work, and optimize it when you need to.

Sorry if this is a bit abrasive; I'm really not trying to discourage you from this project, or even force you into creating something that I could use (not being on Windows).  I just think that doing this all from scratch seems silly.

[nharding]

I am using 3D card, so that I can use the pixel shader (as well as resize the image to screen size). I've got it so that it can adjust the color at the spine of the image, enhance grays colors, deyellow the page, and do the contrast, gamma, and saturation. It's also going to have a database included, so it's more than a simple viewer.

Neil Harding

[xShadow]

I'm a games programmer with over 20 years experience,

Ah, well, that explains your NIH syndrome.

and wanted to learn C#, so I am using XNA.

I have no experience with XNA, but doesn't it seem a bit odd to use a game development library for a simple image viewer?

I've got a basic viewer working already, it handles directory of images, zip or rar files and uses 3D card and shaders to apply real time contrast / brightness / gamma correction to the image as well.

Because that's totally necessary, amirite?  Granted, I haven't tried this out (so I may be completely in the wrong), but it seems like you could do all of that stuff in a rather naive way in a scripting language and still be fine.  You're not calculating softshadows here; write the simplest thing that can possibly work, and optimize it when you need to.

Sorry if this is a bit abrasive; I'm really not trying to discourage you from this project, or even force you into creating something that I could use (not being on Windows).  I just think that doing this all from scratch seems silly.

Man, unless you're willing to do this yourself in a better way, why don't you just wait to complain until he finishes it and we get to look at it? I don't really understand what your problem is here.

You don't seem to have a clear understanding of exactly what all he wants this program to do, and how it already functions, and why it functions that way, yet you're already Brooklyn Ragin'!

Sorry, it's just a little annoying.