0 users browsing Hacking. | 1 guest  
Main » Hacking » Why is ROM translation so (technologically) hard?
Pages: First Previous 1 2
Posted on 19-06-20, 22:46 (revision 1)

Post: #136 of 210
Since: 10-29-18

Last post: 390 days
Last view: 362 days
Posted by sureanem
You can put more than two characters in a kanji.

You're clearly trying to compare radicals to kanji compounds, which is just hilarious scrambling considering how radicals work. This isn't Hangul, buddy.
$gf tells me you have too much time on your hands.
Posted on 19-06-20, 23:12
Stirrer of Shit
Post: #424 of 717
Since: 01-26-19

Last post: 277 days
Last view: 275 days
I'm not talking about any radicals or compounds, man. That sounds like something you'd want to advertise your sports drink has got lots of (or none at all, I wouldn't know - free radicals cause cancer, right?).

You can fit more than two (2) alphabetic characters in the area of the sprite that would ordinarily be used to render one (1) discrete kanji logogram.

Provided a kanji logogram is encoded with two bytes and an alphabetic character ordinarily would be encoded with one, this reduces the amount of space needed to encode a given sequence of alphabetic characters.

There was a certain photograph about which you had a hallucination. You believed that you had actually held it in your hands. It was a photograph something like this.
Posted on 19-06-21, 00:23
Stirrer of Shit
Post: #425 of 717
Since: 01-26-19

Last post: 277 days
Last view: 275 days
Posted by Kawa
Unfortunately, the amount of 90s/00s games on consoles and handhelds that use Unicode in any form since Unicode's inception (1991) can be counted on one hand, and UCS-2 (as "classic UTF-16 aka "Unicode"" is properly called) is considered wasteful.

Fun fact about dedicating tile space to digraphs: at least one Final Fantasy fan translation that I've seen did this, with about ten at most character values being digraphs like 'll', 'il', or 'th'. I'll bet biscuits to an asskicking that this was done primarily for the visual aspect, and that the actual text was still repointed to fit.

Well, yeah, then, uh, case closed. Personally, I think UCS-2 should be called either Unicode or wchar_t, but "classic UTF-16" seems like a reasonable compromise to minimize confusion.

So why is repointing so important, then? It's a slight improvement, but you sure could do without it if the ratios are as you say. How come it can preclude games from getting translated?


There was a certain photograph about which you had a hallucination. You believed that you had actually held it in your hands. It was a photograph something like this.
Posted on 19-06-21, 01:13
Custom title here

Post: #526 of 944
Since: 10-30-18

Last post: 20 hours
Last view: 7 hours
Because translation isn't an exact science. Some statements will be shorter, others longer. Some will have to be reworked to fit the game's output, which can change the length.
Some games just have insane lovecraftian nightmares where one would expect the text engine code to be.


I don't know what specific games you're thinking of, so I can only speak in vague generalities.

--- In UTF-16, where available. ---
Posted on 19-06-21, 06:24
Lotus Eater

Post: #279 of 521
Since: 10-29-18

Last post: 15 days
Last view: 8 hours
Chrono Trigger's text encoding included a large swathe of dictionary lookup bytes, mapping one byte value to two or more characters, along with the general "insert name here" bytes. This would let entire parts of words like "pedia" be saved as one byte in the original text string, but decode into the full version for display in the dialogue box. The dictionary is not based on the top 30 of a given language, but tailored to the needs of the game.

Unfortunately, I don't know what the Japanese version's text encoding is like, only that the names and dictionary lookups are there too, so I don't know how it handles kanji.

Does CT have kanji?
Posted on 19-06-21, 10:46

Post: #137 of 210
Since: 10-29-18

Last post: 390 days
Last view: 362 days
Posted by sureanem
I'm not talking about any radicals or compounds, man. That sounds like something you'd want to advertise your sports drink has got lots of (or none at all, I wouldn't know - free radicals cause cancer, right?).

You can fit more than two (2) alphabetic characters in the area of the sprite that would ordinarily be used to render one (1) discrete kanji logogram.

Provided a kanji logogram is encoded with two bytes and an alphabetic character ordinarily would be encoded with one, this reduces the amount of space needed to encode a given sequence of alphabetic characters.

So you basically don't know what you're talking about. Gotcha.
Posted on 19-06-21, 11:33
A punch comes free with most major felinies

Post: #281 of 521
Since: 10-29-18

Last post: 15 days
Last view: 8 hours
What's worse is, with all this talk of specific encoding schemes...

repointing to make room for trivially-encoded strings is the easiest way, even on systems with "nasty" pointers.



Seriously though, there's basically only one reason for Chrono Trigger to have dictionary lookup bytes and that's "ROM is expensive".
Posted on 19-06-21, 18:45

Post: #69 of 100
Since: 10-30-18

Last post: 295 days
Last view: 59 days
love 2 unironically write the string of text "discrete kanji logogram"
Posted on 19-06-21, 22:07

Post: #138 of 210
Since: 10-29-18

Last post: 390 days
Last view: 362 days
Yeah, why can't we just say singular adopted logographic Chinese character?
Posted on 19-06-22, 08:29

Post: #70 of 100
Since: 10-30-18

Last post: 295 days
Last view: 59 days
formally numberless conceptual-morphemic orthographical unit of the middle empire
Posted on 19-06-22, 12:02
Secretly, I'm Matvey Soloviev

Post: #282 of 521
Since: 10-29-18

Last post: 15 days
Last view: 8 hours
Yup, this is a byuuboard alright.
Posted on 19-07-21, 02:31

Post: #45 of 49
Since: 10-29-18

Last post: 415 days
Last view: 300 days
Can confirm firsthand what Kawa wrote about Chrono Trigger text stored in a tailored dictionary. One of the things I wrote in my nascent (and recently untouched) CT ROM parser is the ability to read the text data stored, and read early on in Geiger's notes about the compression/decompression method and it checked out.
Posted on 19-07-21, 09:43
I said, put the bunny back in the box!

Post: #304 of 521
Since: 10-29-18

Last post: 15 days
Last view: 8 hours
I'd damn well hope so considering the amount of research time I put in that post.
Pages: First Previous 1 2
Main » Hacking » Why is ROM translation so (technologically) hard?
Yes, it's an ad.