Winhex has a function to show the number of occurences of each byte value... dunno if that'd help you though.

Give this a try.  Look for a word in the game that's just hiragana or katakana (I'm going to assume it's hiragana).  An ideal word would have no dakuten or handakuten marks. 

Now, open up the font in TLP and adjust it so that it makes sense (I think you've already done this).  That's going to be the order they'll show up in the table.  So, if you see あいうえお... in the font, think of あ as 1, い as 2, etc. 

Now, open up the rom in your hex editor (ideally windhex) and open up the relative search function.  As an example, if your word is あおい, then you'll want to relative search for 1 5 2.  If that doesn't find anything, try putting wildcards between them: 1 * 5 * 2 (sometimes each character is two bytes instead of one, so you'll need the wildcards). 

If that doesn't net you any results, then I'd guess the script is compressed (in which case a relative search won't do you much good  )

The table you've got in TLP , the one in 1BPP, is not the table used to put the text in the dialogs. If you look good there are characters in the dialog that don't exist in the table, just check the brackets in pink, or for example in the beggining talk to the guy with gray hair and look for the first character in the braquets, the one that resembles a "different of" sign.

I doubt... If you chack the characters I've mentioned, the one that looks like a different, it isn't even in that table. Most of the game graphics are in 4BPP planar (in tlp choose SMS and you'll see them) And for that table just see the size in pixels of each character in tlp, then compare to the ones in the text, the first are 8x8, the ones in the dialogs are bigger, something like 12x12

EDIT: One hint: just keep playing the game until some word(more than 3 letters) using the characters A-z appear. Then scan relative that word in translhextion, then if he find something ask him to build a table based in that

EDIT2: I've found the Capital Letter's table. A starts at 2A and Z is at 43. I've got it going up the stairs into the computer room. There a guy says NOW PLAYING. The tricky part was to discover that each letter is separated using a space. Then I relative searched the word playing and came out a result. That phrase in the rom is at position 0x541ed4.

do this:
. open the rom in translhextion
. search -> scan relative
. enter(with the spaces between the letters): P L A Y I N G
. it should have got one offset,  choose generate table and save it
. script -> open thingy table-> open the table you saved
. then select thingy view active

and there it is your text

Now based in those found letters try to find the other that are used in the same sentences.

sry for doublepost...

thank you again

I got into it a little bit in the last hour...it's gonna be a long night ^^

but without you I would have never got so far, thank you very much (again)

when I'm done with it you'll be the first to play it (if you want of course)