For a long time I was under the impression that the only difference between a game in japanese and in english or any other western language was the text which is written on the screen.
Of course I was completely wrong. I'm now extremely interested at the technical (non linguistical) differences.
With my thread about translating my game to japanese
and the thread about comparing game locatlisations
, I learned a lot of things.
No matter how simple or complex the games are, if some text is going to be written on screen in japanese, different technical choices about how the screen layout is done, etc... is going to be made. This also affects if a game is being translated form a japanese version.
1) If the game uses exclusively kana, it is not too different than english text. But still a lot more can be said in less screen space, and dialogue boxes will be smaller. It is extremely common to leave a blank tile line between 2 lines of text, in order to have the text more readable, but also in order to place dakutens and handakutens. Some games place them as raw characters, but then a space appears in the text, and it can be "weird" to japanese readers.
Kanas have in average more details than our letters, and for this reason they often need the entire 8x8 tile space, as opposed to letters which can easily be squeezed in 7x7 or even less, leaving some blank space for intelligibility. This is not possible with hiragana. However, this is also somewhat possible with katakana, so some games made the choice of using exclusively katakana, which apparently can be "weird" to japanese readers.
Also, even if only a single hiragana or katakana is loaded in the tileset, there is going to be a need for ~50 tiles for text alone, twice as much as what is required for only an alphabet. If both full alphabets will have to be loaded in the tileset, ~100 tiles will be required, almost half of the background tiles available for the NES. Usually this was considered too much huge of a sacrifice, and this path was seldom taken. However all first 3 Final Fantasy game made this choice, in other words, they sacrificed overall graphics quality in order to have more readable japanese text with both full alphabets.
This "sacrifice" is of course not apparent at all in the western version of FF1, which uses barely half the files. Some symbols and "concatenated" text have been made with the unused tiles though.
2) Some games uses several small 8x8 kanjis in their text. Ninja Gaiden for instance. Technically it just makes the game use even more tiles.
3) Finally, gamres which wanted to use fully kanji text had to make even more sacrifices. They can forgot about a 8x8 font, and usually made a 12x16 or 16x16 "VWF". This would turn out in games that have a VWF when being translated into english. Some games also have 8x16 kanas, and some 16x16 kanjus made with two characters in series.
Usually all the characters are the same width in japanese though, but they still made it "VWF" technically for unknown reason. For instance, in FF6, all chars are 12 pixels wide, but it's still apperntly a VWF like in the american version of the game, in the sense that the number "12" is stored a lot of times near the font, and that some rare characters have a smaller size (I think the small ya, yu, yo characters are in this cathegory).
Some games, such as Seiken Densetsu 3 went as far as to use a special high-resolution mode for text alone. This probably increased the readability of kanjis, which are very complex to draw even if they are 12x12 pixels.
This probably affects how the text is encoded and compressed in ROM too. Since there is about 2700 kanjis, we need at least 12 bits per characters, instead of only 8, before compression. Also chances are that different compression algorithms have to be used than with english text in order to be efficient.