Sunday, January 24, 2010

Are you ronesome tonight~♪

Not much to say about the KS development as usual (I am currently working on directing Lilly's Act 2, and if you think that this is any kind of progress indicator, you're wrong). But, there are a couple of other things worth mentioning that might be of interest to people not intimately familiar with what's going on.

Personas

Firefox 3.6 is out now (and if you're using Firefox, you should have gotten the update prompt by now), and with that, Personas have become standard. These are essentially wallpapers for your browser window chrome. Since they're easy to make, you can now now have your KS character of choice staring at you all the time. Cool, right? In fact it's not cool at all, but anyway, here they are:

Misha
Emi
Hanako
Lilly
Rin
Shizune
SD Girls

Rin is missing because her Persona contained the mural, hence boobies, hence it was not approved. (if you are curious or want to stick it to The Man by making your own, it looked like this). IT'S ART YOU PHILISTINES! I made a new one but it's stuck in the approval pipeline. Once it's through you'll find it along with the rest at my account page. There are also a couple of tweaks to the existing ones based on feedback submitted, but they'll hopefully have the same ID.

Dakimakura

In case you haven't been paying attention, there's been a trend in the KS fanartist scene to draw hug pillows. So far, pretty much every major character but Hanako has been (un)covered, a fact that amuses me to no end. Even though these are technically not unsafe for work, you'll probably not want to be caught looking at scantily clad disabled girls (or men) at work in any case, so proceed with caution. The exception to this is of course Lilly, which is not safe for work in any definition of the term.

by pimmy:
Misha
Emi
Rin
Shizune

by climatic:
Lilly (double NSFW)

by konflikti (DO NOT VIEW WHILE DRIVING OR OPERATING HEAVY MACHINERY):
Nomiya
Nurse

by VCR:
Hideaki

Twitter

We have a Twitter now, for things that are too unimportant, irrelevant or vapid to warrant a blog post. In case you don't like unimportant, irrelevant or vapid things... well, then you are probably not using Twitter anyway. If nothing else, it's a decent way to keep up with Mishimmie uploads. Also, in keeping with the theme of this post, it will be like you really have friends.
FOLLOW US ON TWITTER!

–delta

Friday, January 1, 2010

Happy New Year!

1990 was 20 years ago.

Saturday, December 26, 2009

Am I Going to Be Popular Now?

Hello and hope you all had a great Christmas. Here at 4LS, we love to redo old work and throw dozens of manhours worth of production material away, if it's for the sake of art and ultimately, a better game. A lot of that stuff doesn't really strike the eye because you never see the old versions that have been discarded. However, one change is definitely going to make some waves, so I thought I'd announce the fruits of a couple of weeks of hard toiling by moekki and delta:

We upgraded your loli.


The new Emi has a bit more mature air than her earlier design, perhaps more appropriate of a teenaged girl  rather than a REALLY 18 YEAR OLD pseudo-loli. She grew almost 3 inches, but is still the shortest of the characters (so we can make bad jokes about it) and still of the same general build and shape. The reasons for redesigning a main character had mostly to do with stylistic consistency, we felt keeping the old design drove her too far away from rest of the art that features her (as you can see from the 100% completion CG in Act 1). Anyway, apart from maybe promo pictures and such you won't see more of her until full releases, any further Act 1 versions, released mostly for various translation projects, will retain the old sprites.

 

-Aura

Thursday, December 24, 2009

Merry Christmas

Reimu wears red and white, that makes this a Christmas image.

Sunday, December 13, 2009

Act 1 V2 Release




Four Leaf Studios would like to announce the official release of Act 1 V2.

Torrents:
Windows Act 1 V2
Linux Act 1 V2
Mac Act 1 V2
Help in seeding these would be greatly appreciated, in lieu of the previous Act 1 torrent.

Direct Download:
Links will come later if/when we find volunteers.
UPDATE: Thanks to Eksopl and magz, DDLs are now available from the download page.

This updated version of the previous Act 1 game preview includes:

Additions:
* Traditional Chinese localization
* Simplified Chinese localization
* Italian manual
* New Windows installer, KS now by default starts up in the language the installer was run in

Fixes:
* No longer raises DEP errors under Vista
* Should now at least start on low-resolution displays
* Various typo and grammar fixes
* Flow fixes, all scenes should now unlock properly

This release is not recommended for those who already played the initial Act 1 release - the changes to the English part are minimal. However, for archival or new players, this effectively replaces the old release.

We hope that everyone, especially Chinese speakers, enjoy this updated release and will be looking forward to the full game.

Sunday, December 6, 2009

Status and Backgrounds

Hello, everyone. Blog post time.

As far as general status goes, there isn't a lot to report; paths are still being written, scenes are being directed, art is being drawn, and music is being composed. And vitriol is still being flung, but nobody really cares about that.

So I guess to give this blog post some point other than "we're still alive, and still working", I'll prattle on about backgrounds.


Artist's rendition of Delta working on a game background. Image by Moekki.

Here is an example of one of our backgrounds, for context:



This is the result of a somewhat longwinded process that came about due to many experiments in different styles, varied and vitriolic arguments, and tears. But in the end, it all finally coalesced into the result above which we were all happy with.

It all started with... it not starting. As people began to write and the first of our artists began to produce sprites, one large omission was made. Expecting that, in the fullness of time, they would be done.

That said, a few months saw some movement. To be precise, a very basic tech demo.



The less is said about this, the better.

After a while, with the game proper in development, the first batch of actual game backgrounds were produced. Created largely as placeholders, they were inserted into the early prototype of what would become Act 1.



Pretty simple - an oil effect and colour wash on a monochrome image. These backgrounds were grabbed from around the internet and therefore generally lacked any consistency, and were also taken from a variety of heights and angles. These held us over while we waited for a background artist.

Which, as you can see from our current backgrounds, never materialised.

Well, they did, but not for a long time. Backgrounds were for months entirely passed off as a future consideration as writing progressed on every path, which seemed to be fine. Until we took stock of what would be needed after we were done writing. What had happened is a complete explosion in the sheer amount of art assets required, especially in terms of backgrounds. A certain amount of rewriting has helped, but the fact remained that we suddenly found ourselves requiring a ton of backgrounds we simply didn't have.

It was then that we made the hard call to give up on the prospect of drawn backgrounds, and have a hard look at how we could accomplish our needs via modified photos. This lead to a couple of blog posts you may remember.

This decision wasn't made lightly. Everybody on the developer team wanted drawn backgrounds. The problems though were many; we needed over one hundred backgrounds, which were going to be an absolute hell to get even with photography, it would have meant growing the developer team even more when communication was already an issue, going back and changing things is simply not really possible for drawn backgrounds, we wanted an artist (or artists) that could make relatively good looking backgrounds, rather than accepting drawings of any quality just so we could say "we have drawn backgrounds!", and there was the simple fact that we didn't have any offers until far too late.

Even then, after deciding on photographic backgrounds, we were in despair about the sheer number of backgrounds needed, some of which would be quite hard to get. That is, until the heavens parted and a dedicated photographer joined our development team: Yujovi.

While a number of public domain photos from the internet and photos from developer's trips to Japan helped supplement our needs, Yujovi took a large majority of the photos we used for backgrounds. For the first time in a long time, the background situation seemed solvable.

But one question remained: How to modify the photos to suit Katawa Shoujo's art style the best?

This lead to a few weeks of experimentation, arguments, and general fuss. In the end, the decision came down to two candidates:

Crosshatch Hell:



We decided that these were too dark, the enhanced object outlines didn't really bring out the sprites very well, and that the crosshatch effect just didn't look very good in general. Vibrating trees: not so good.

12-Step-Recipe:



Proposed by Moekki, this was the general process we decided on using (albeit with some tweaks). It gave the backgrounds a flatter, more simplified look whilst keeping the eye directed towards the sprites. Note that wasn't a game background, just a general example image.

Several other ideas were proposed, and thrown out in quick measure; A simple cutout filter messed up the background outlines too badly, and made expanses of colour (such as skies) look awful, while paint daubs and other such single-step filters looked 'cheap'.

But filtering is not the end of it, as there were a number of other obstacles to overcome.

The next step in the saga of backgrounds was the fact that, despite having a bunch of photos and a modification process, backgrounds still needed to be selected. For this, some standards were put forth.
* All images must be at least 1000x600, to allow horizontal panning and background manipulation.
* All images had to be taken at a close to horizontal angle, from normal eye height. This is pretty much common sense, as the backgrounds and sprites, at least within KS, simulate seeing the world through the protagonist's eyes.
* All images had to be reasonably clean, without too much blurring or distortion. Some of this can be covered over during the BG-ification process, but too much and the background will simply look bad.

And so, selection began. After a few weeks, most of the required backgrounds that had popped up during writing were selected from the available pool, and a naming scheme was devised - location_place_time.jpg (for example, school_dormhisao_ni.jpg).

With that sorted, the modification of each photo began to take place. With Delta and Moekki's skillful work, the images were not only filtered, but also prepared beforehand to look the best they could (via adjusting saturation, brightness, removing people, fixing image tilts, changing and removing signs, fixing of vertical lines, etc), and had variants made for different times of day. This often took much more time than making the photos into backgrounds, as many photos required quite heavy photo manipulation.

The conveyor belt of background requesting -> photo taking -> selection -> image preparation -> background-ification -> game insertion cranked to life, and soon we were in business.

And that's the story of how Katawa Shoujo's backgrounds are made, and the history of how they came to be this way. Thanks for listening, chiiiiiildren.

Oh, and Dragon Age: Origins is a great game.

- Suriko

Wednesday, November 11, 2009

# need another event cg here —δ

One of the big realizations during the work on this game was that while consuming a lot of a particular media does help, it does not actually give you the experience in creating it. While this seems painfully obvious, it's also disheartening, and even if you try not to think about it it will hit you eventually. And it's not like you can go around and ask professionals or read an authoritative book on the topic of VN creation. So all of us are just trying things and hoping that they will work out for the best. I'm not even talking about the big things — it's obvious that a good story needs to adhere to a few well-known ground rules like pacing, and what kind of art is pleasant to look at, because those are pretty intuitive. The devil is in the details.

So today I'm writing a bit about event CG. But you can't start talking about event CG without considering what the alternative to event CG is, and that is almost always sprites in front of backgrounds. These are the foundation of virtually all VNs, because they "only" require initial effort, and then can be used for the rest of the game. Sprites and backgrounds will, in a long, not ridiculously-budgeted game, always be your basic bread and butter.

Of course, when everything is the same sprites and backgrounds over and over again, it's hardly exciting. That's where event CG comes into play. Basically, an event CG is something that is specifically created for an event in the game, something that is not universally usable or modular. But as we found out, determining when to use and when not to use an event CG is not as easy as it sounds.

We had two lines of thinking in the dev team: One saying that event CG should be less specific and more universally usable, and one saying that it should be very specific to get maximum impact from them when they happen. Much discussion was spent on this, and we ended up trying both. As always, it became apparent that both had their downsides as well - the real "bulky" scenes in the game are dialog scenes, and spending event CG on those seems less than ideal, because you could have done them with sprites. However, the other way, making an event CG for every "non-generic" happening, is just not possible — Ideally, you would do the entire game with nothing but event CG, and it would always fit without ever getting boring (and some games indeed do just that). But event CGs are "expensive", if not in money, at least in effort. So you will never have as many as you would like.

Basically, there are two types of scenes that are even considered for event CGs: Important ones that deserve the additional visual impact, and ones that just cannot be visualized otherwise. If those were always the same, it would be easier. However, the smaller the amount of CG you want to spend, the more they drift apart, to the point where you have to cut corners for the sake of feasibility. So some kind of abstraction, instances where something in the text is just impossible to visualize, will definitely happen. And sadly, eventually you have the choice between doing something that "deserves" an event CG but could be done without it and something that will have problems with the visual representation but isn't really important enough - and in this case, the latter will probably be the one that gets chosen.

Now, what did we end up with? We did not even have the (not that deep, to be honest) realizations above early on, so what we decided on was to let the writers decide where they want event CG. Which sometimes worked, sometimes didn't. But what can you do? Basically you only really realize you need an event CG when you try directing a scene with sprites and it just doesn't work. Then again, directing only happens when the scene is written, and the artists should already be working on the event CG. And sometimes you just have to bite the bullet and even drop text, as good as it may have been, just because it's not feasible in a visual novel. Of course, an experienced VN director could see those things coming (and an experienced writer would not write scenes that lead to the dilemma above), but I think all of us are more than ready to admit that we're just barely entering a level of experience that makes this possible. This is, in fact, a big part of the reason why the development process seems unrealistically slow sometimes.

But it's the way it is. We're still learning and trying new things every day. Not thinking in absolutes and theories, but in the compromises and realities of actually creating something. It's not always easy or fun, but it's what makes this project so interesting in the first place.

Today's blog art by our man to go to for emergency event CGs, climatic.


— delta