Ultima7 alternate font

[Ken Kato]

Folks,

It seems that I'm not the only one to find U7's old-fashioned
font style somewhat difficult to read on screen.
So I created an alternate FONTS.VGA with very simple font style.
Just replace the one in STATIC subdirectory of your U7 install
directory with this one.

http://rapidshare.com/files/121379426/FONTS.ZIP.html

Works both with Exult and DosBox.
I'm not sure if it works for Serpent Isle, though.

Regards,

--
Ken

[The Ancient One]

Hallo Ken Katao,

I, too, would like to have an alternate font!
I just downloaded yours to try it in-game.

Would you like to adapt it for our Italian translation? I’m sure Italian players would love to use it.
You should just add a few characters (à è é ì ò ù).

Let me know if you’re interested!
-diego

[Ken Kato]

OK, it turned out that Serpent Isle's FONTS.VGA has a few extra font sets
so I created one for SI. Perhaps this one works for both BG and SI since
BG would simply ignore extra font sets in the file.

http://rapidshare.com/files/121424646/FONTS-SI.ZIP.html

Note that fonts in this file are used in the actual game, not in the
introduction or ending scenes.

To Diego,
I could add generic Western European characters to the file,
however I'm not sure if simply adding those extra characters would do
the trick since I only have English version of U7.
Maybe if you could share your FONTS.VGA file with me...?

--
Ken

[The Ancient One]

Thank you, Ken Kato, for your readiness to help!

I could add generic Western European characters to the file,
however I'm not sure if simply adding those extra characters would do
the trick since I only have English version of U7.

I think that it'd be much simpler if you work directly on our modified fonts.vga.

Maybe if you could share your FONTS.VGA file with me...?

Of course!
Write me at ancient@NOSPAMiagtg.net and I'll gladly send it to you!

[Bob The Cowboy]

What would be the odds of including this font package with Exult? What other work would need to be done?

Just thinking it might be enough to close wishlist item "953520"

Is there a screenshot of this anywhere? I haven't had a chance to try it myself.

[L@Zar0]

Hi,

it seems a good idea to make useful the FONTS.VGA file for Western languages. So, I'm interested in it for the Spanish language.

But, I don't know if it is possible to add more than 127 chars to get the 255 code charset.

If it is possible, please, tell me and I try to put the Spanish chars or re-make/adapt the FONTS.VGA.

I liked a lot the new FONTS.VGA file.

Regards.

[L@Zar0]

I left this codepage info (for spanish and I think that for portuguese, but need some char more) for your interest if want to know more...

Chars:

________________________________________________________
Codepage: DOS CP-437 & CP-850

0x81 -> ü
0x82 -> é
0x90 -> É
0x9A -> Ü
0xA0 -> á
0xA1 -> í
0xA2 -> ó
0xA3 -> ú
0xA4 -> ñ
0xA5 -> Ñ
0xA8 -> ¿
0xAD -> ¡

(not present in codepage DOS CP-437, but in DOS CP-850 | Western Europe)
0xB5 -> Á
0xD6 -> Í
0xE0 -> Ó
0xE9 -> Ú
________________________________________________________

________________________________________________________
Codepage: Western (Windows-1252) as per Exult Studio

0xA1 -> ¡
0xBF -> ¿
0xC1 -> Á
0xC9 -> É
0xCE -> Í
0xD1 -> Ñ
0xD3 -> Ó
0xDA -> Ú
0xDC -> Ü
0xE1 -> á
0xE9 -> é
0xED -> í
0xF1 -> ñ
0xF3 -> ó
0xFC -> ü
0xFA -> ú
________________________________________________________

[Malignant Manor]

I think at least 256 characters are possible. When using a non-serpentine sign gumps and the read flag is set, certain numbers are converted. I'm not sure what Frame number = character number. I'm think Serpentine fonts swap with the read flag set (font 8= 2 or 4 [not sure which], font 9 = 3, font 10 = 6).

Runic conversions with read flag set:

(no spaces between converted numbers)
40 = 84 72
41 = 69 69
42 = 78 71
43 = 69 65
44 = 83 84
124 = (I think this is converted a blank space)

Characters greater than or equal to 97 subtract 32.

[Malignant Manor]

Now that I have more time, my suggestion would be to use the original fonts for signs (1 , 3, 6, 9, 10) if you believe that the very few words that would be contained on them would be understandable without accent marks (or whatever they are called).

Dónde está el baño would be DONDE ESTA EL BANO when translated either manually or with the read flag. For font number 8, put the Roman alphabetical equivalent serpentine symbol in place of characters such as ó, á, and ñ (would place a copy of frames 79, 65, and 78 respectively).

This should allow you to keep the flavor of the game and have the plus side of being a lot less work to do.

[L@Zar0]

The question that appear to me then is:

We have a book with serpentine language (or Britannian language); I agree (absolutely) to use the original serpentine(britannian) fonts. It has no sense to change this language, it's an ancient, old and own language.

Then, when we read the book in understandable language.... ¿would the chars be converted to what they must?

Malignant, I don't know what is the read flag you're talking about. ¿What is it?

[artaxerxes]

What we used for (unofficial) SI-french (Serpent Isle in French) was pretty much the exact same as they did for the official Black Gate translation in French. Since we could not use the characters from 128-255, we tucked the accentuated characters in the first 0-31 characters. We wrote the text using normal accents, then we used a conversion tool to change those to control characters. Naturally, the fonts.vga file had to reflect precisely which character to show for the given control character.

To give an example, suppose we decided on é being assigned to 0x04. Then, the phrase '@Dupré !@' got converted to '@Dup[0x04] !@'. So that when Exult would render the string, it would use the character at position 0x04 in fonts.vga.

It worked well in French because we were able to tuck all the accents within that short range. There may be languages out there that use more than that, so I'm not sure how you would do for those.

Artaxerxes

[L@Zar0]

Then, it seems that the charset could not go up from the font number 127, isn't i? The only possibility is to use the chars before 0x2F (before the char '1')

[Malignant Manor]

The read flag is what is set when using the translation spell or as a cheat in Exult. It turns serpentine or runic font to normal English.


I tested ucc and was able to use character 255 so unless the other compiler doesn't support that high, then there shouldn't be a need to put characters in the wrong place. (I'm unfamiliar with it.)

[marzo]

But, I don't know if it is possible to add more than 127 chars to get the 255 code charset.

It is; you can select a character set in ES for each game, which ES will then use to convert to/from UTF-8. And if you use the characters > 127 in the U7 data files (for example, in the patch dir) for the translation, Exult will display them correctly.

I think at least 256 characters are possible.

I once did some hacks test unicode fonts; there were not many problems with a 65536-character font -- the problems being the size of the file (over 12 MB) and that ES would crash if you switched to frames display. As is, Exult/ES are limited to 256 characters.

[L@Zar0]

1. How can you select the character set the Characer Set in Exult? I could see it in ES and Apply it, but does Exult knows that?

2. If I use this type for Charset... I assume that the old DOS version of Ultima VII will not work correctly with the usecode, isn't i?

[L@Zar0]

OK. I've seen that ES adds a new tag to the .cfg xml named . I've tried it in the SIFixes Mod, but when using this tag, with CP-437 or with Windows-1252 (which includes all the charsets for Spanish language), SI + SIFixes mod does not plays well (at least the title shape is not recognized).

Maybe am i doing something wrong?

[L@Zar0]

More testing:

Finally, I've created a 255 charset for the 0.shp shapes and imported it to the file FONTS.VGA for the Spanish language. This shapes with 255 frames (0.shp), works perfectly with Exult. All the codes are displayed (áéíóú ÁÉÍÓÚ üÜ ñÑ ¡¿) in a Spanish Windows Vista (I think that this must work with other OSs).

The problem that I have now is with the 1.shp (the Britannian runic language - NOT serpentine).

I think that the game (and Exult) uses an UPPER function to get translated the runes to understable language, but I imagine that it is 'limited' to the first 127 chars. I could use in usecode directly the char 'Á' (upper case) and it is displayed well in spanish. Also, I can put directly to the usecode the char 'á' and it is displayed well as runic alphabet.

But when I use the Translate spell on a signpost, example: 'á Fawn' - this has not sense for spanish, but it is an example - the rune, in this case 'á' does not translate, they are left as the runic alphabet.

I could left this texts with Britannian alphabet without the accent punctuation mark, but I don't know if this will have any problem with the ofidic/serpentine runes, because there are some books that must be 'understandable'.

Any suggestions?

Thanks a lot.

[Malignant Manor]

Changing the codeset causes these extra lines which completely screws up The Keyring Mod (doesn't read the correct data).


BLACK GATE
KEYRING


The Exult Menu refuses to show the other codeset for the title or even save it correctly in ES. The codeset probably (or should) only affect the title in the Exult menu.


I told you in a previous post what happens to Runic and Serpentine fonts. If you want to use accented character.

[L@Zar0]

Thank you for your response.

First of all I apologize for my persistence. Sorry.

With the FONTS.VGA charset that I'm doing, finally there is no need to change the codepage. So for the normal texts, it can be used by Exult in ES.

About non-standard fonts, i like to know something:

Coult it be possible to make a patch (or something like that) for the function that 'tells' how to do the conversion from non-standard texts and standard text? Can you tell me which is the function from the exult sources to look at?

Maybe somehow with a re-implementantion of the function, the conversion could be done above the 127 char. I agree to left the runic Britannian alphabet (although it can be interesting have all the text equally), but with the serpentine there are books that need to be understandable.
So, if we change the things for the serpentine, the britannian could be done easily also.

Maybe what I am saying has no sense or can not be done, I don't know.

[Malignant Manor]

Like I said, serpentine swaps fonts with the normal text font equivalent when translate (the read flag) is in effect so that it is extremely simple to use all the characters you want. If a character is "a, A, á, or Á" in the normal font then the serpentine font would need to have the symbol (frame 65) that translates to a in those frames.


Now for runic fonts.

With some math, you could also convert the runic font to use those extra characters.

The frame listed on the left needs to have the character listed on the right.

125 = á
126 = Á
127 = é
128 = É
129 = í
130 = Í
131 = ñ
132 = Ñ
133 = ó
134 = Ó
135 = ú
136 = Ú
137 = ü
138 = Ü

The frame listed on the left will import the runic font frame listed on the right.

157 = 97
158 = 97
159 = 101
160 = 101
161 = 105
162 = 105
163 = 110
164 = 110
165 = 111
166 = 111
167 = 117
168 = 117
169 = 117
170 = 117


¡ and ¿ can fit in empty frames below 97 like:

31 = ¡
62 = ¿


All other runic fonts frames would be the same as the original unless you add more characters than this like lower case letters.

The frame listed on the left needs to have the character listed on the right

171 = a
196 = z

The frame listed on the left will import the runic font frame listed on the right.

203 = 97
228 = 122

[marzo]

Hm. ES should be writing 'display_string' instead of 'mod_title'; this causes the problem with the menu shape not displaying correctly (and probably others too). It will be fixed in the next snapshot.

With that out of the way: right now, Exult is blissfully unaware of the codepage, and only ES cares about it: GTK (the interface toolkit) uses UTF-8 for text everywhere, hence ES has to convert non-7-bit-clean text to a codepage or Exult will choke badly. Given a codepage, and a font with that codepage, Exult will use the correct characters set by ES.

The reason Exult isn't codepage aware yet is that GLib (which is used by ES for the back-and-forth conversion) isn't used or distributed with Exult, and I don't want to add the dependency. But maybe ES can save a translation table for uppercase/lowercase conversion that Exult can use.

[L@Zar0]

@Malignant:

Let's see if I understood well this thing:

* The 'á' is the code 225, at least in Windows Vista Spanish.

- For serpentine symbols (frames from 8.shp), the game gets the frame corresponding to one of the original symbols (some of the 1, 3, 6, 9, 10, I don't know what is).

So, if we have 'á' in a book (frame 225 for ¿2.shp?, I don't know the shape for the texts of the books), the game gets the same frame (225) of the serpentine symbols (frame 225 for 8.shp).

Is it?


- About runic:

I am a little confused about the explanation.

If I use the char 'á' (frame 225, lower case) in usecode, and the frame is the same as the runic alphabet (the frame 97), the same runic symbol is displayed. This is correct.

But, now,

If I use the char 'Á' (frame 193, upper case) in usecode, and the frame is the same as the runic alphabet (the frame 65 plus accent mark, edited by me), the original symbol is displayed. This is correct too.

So, what does it happen when I cast the translate spell?,

Simple, the 'á' is displayed in both cases, with spell or without spell, as runic symbol, there is not conversion. The game does not gets the frame that is at position -32 (no conversion from 225 to 193). The fact, is that in ES the REAL charset is at position -32 really. So it should be easy. It's like if above the frame 127 the rest didn't exist for the conversion.

As I know, the lower case text in usecode tells the game to use runic symbols, and the upper case text tells the game to use original symbols, which both frames/chars, are in the same shape 1.shp, at least for signposts. I don't remember if there are books written in runic symbols now.

Because of that, I don't understand the explanation about the runic texts, because the runic symbols 'Á' or 'á' have a different meaning for the game, isn't it?
Also, I can not see the charset codes written between the frame number described and the frame number that I am using for ES, I think that this could be OS dependant although.

[Malignant Manor]

I think you are grasping the serpentine, but seem off on the runic.

You cannot really do runic using following a character set and still use accented characters, you need to customize it like I was saying since it follows how the original game did it. (It should probably be changed to allow the option to switch fonts for simpler modification.)

With runic font, everything displays the same frame as the character number without translate.


With runic font using translate:

Any frame below 97 will display the same character as the frame number with a few exceptions noted below. Frames 97 and above will display the frame 32 below it.

Runic conversions with read flag set:

The character to the left will be converted to the frame numbers on the right.
(no spaces between converted numbers)
40 = 84 72
41 = 69 69
42 = 78 71
43 = 69 65
44 = 83 84
124 = (I think this is converted a blank space)

[marzo]

Let's see if I understood well this thing:

Yes, that was correct: the frame used by any character is the same regardless of the font used. For serpentine text in books and scrolls, font 8 gets replaced by font 4, but the characters used are the same; serpentine text in signs is also translated only by a font replacement.

The fact, is that in ES the REAL charset is at position -32 really. So it should be easy.

Don't take this for granted; it happens in a few character sets, but it does not happen for most. Performing the conversion requires knowing (1) that a given character is a lower-case letter and (2) that it can be converted to upper case, (3) what is its uppercase form; ES can do that with the help from GLib, Exult can't.

It's like if above the frame 127 the rest didn't exist for the conversion.

Exult performs 'lazy' uppercase conversion, converting only the letters in the 7-bit ASCII character set. This is because the original games didn't have any other characters that needed to be translated in the 128-255 range.

I don't remember if there are books written in runic symbols now.

There aren't.

[Malignant Manor]

If you set it up like I suggested character, for character á to be displayed when translate is used you would need to have the character } used in its place in the text. ~ would be used for Á. The others vary depending on the charset used.

[marzo]

Frames 97 and above will display the frame 32 below it.

That is partially correct, but there is a catch: the conversion only takes place for characters <= 127. The reason is buried in the fact that an std::string is an array of 'char', not 'unsigned char': hence, characters in the 128 to 255 range are actually seen as being in the -128 to -1 range in the comparison you are looking at. But when getting the frame to be used, Exult casts 'back' to 'unsigned char' to obtain the correct frame.

[L@Zar0]

Well, apart from the runic conversion with read flag set, which I don't see the meaning yet,

As you said:

"Frames 97 and above will display the frame 32 below it."

But this must be ONLY for the frames 0-126. This does not apply above the frame 127, isn't it?

If not the problem would be solved now, because the frame 225 it's above the frame 97 and the game does not display the frame 32 below it, the 193, using the translate spell.

So, there is something that prevents make posible the translation, for chars above the 127. That is what I am trying to explain, although my bad english can be worst than the runic alphabet... xDDDDD

[L@Zar0]

About this:

"Exult performs 'lazy' uppercase conversion, converting only the letters in the 7-bit ASCII character set. This is because the original games didn't have any other characters that needed to be translated in the 128-255 range."

Could be a patch (not a modification of Exult, it is very well as it is) be possible for the conversion of the translate spell ONLY?

Also, after having solved the issue with serpentine fonts, which I think that they are the most important, then, it is possble to make the FONTS.VGA for ES (without the runic alphabet). Although I try to left the frames on the shapes for future use.

Well, thank you very much. At least you had solved a lot of questions regarding to the charset/FONT/frames use of Exult/Ultima engines.

[marzo]

Could be a patch (not a modification of Exult, it is very well as it is) be possible for the conversion of the translate spell ONLY?

Modifying Exult to make it behave as expected would be a good thing, I think. But you could do it with an usecode patch, if you prefer (which would be the alternative).

[Malignant Manor]

Lazy is not the right word since those accented characters are worthless for the English Ultima 7 and would be a huge waste on their part in the floppy age. They are extremely rarely used in English.


It would be better to have the optional font substitution coded in the engine for the different runic fonts since it's a lot easier to use.

[Malignant Manor]

I think I misinterpreted the uppercase comment. The translate spell did not exist in BG so there was no reason at all to use upper and lowercase. The text is sparsely used and easily read in all uppercase. Lots of signs in real life have all uppercase letters anyway.

[L@Zar0]

Quote:

"Lazy is not the right word since those accented characters are worthless for the English Ultima 7 and would be a huge waste on their part in the floppy age. They are extremely rarely used in English."

The worst thing is that BG and SI get out to the market in localized versions, like spanish (only the game, the expansions not). It is very rare that this was not implemented for this localized versions... there was the codepage in DOS available in that time (for previous games like U6... never get out in spanish (If now I am not mistaken). Also U9 only go out in english/german? I am now translating the FanPatch to spanish for it. Some people do the spanish translation for the original game, but I think that the FanPatch is more interesting....

I don't know how to make the usecode patch, although I could try to make it in time, but if you expect to change this portion of code for the translate spell of Exult (maybe a new option inside Exult Setup?), I belive that other localized languages would be interested too.

[Malignant Manor]

How many different games had a runic and a serpentine language. I really doubt they had the time and the money to do it and I doubt it would have been worth the money. They didn't even translate it to Japanese.

[Malignant Manor]

Also U9 only go out in english/german?

Who cares? The game is a buggy piece of crap. It should have never been released or at least not in that state.

[marzo]

Yes, I am doing several changes that will change the way Exult handles gumps in general, and sign gumps in particular; adding the code for correct codepage uppercase/lowersace correction would be a minor thing. When it is done, you would have to only select the codepage again and ES would save the table.

Oh, I had forgotten about this:

The Exult Menu refuses to show the other codeset for the title or even save it correctly in ES

This is by design; and I think that the rules aren't strict enough as is: the Exult menu has its own font which cannot be patched and doesn't have characters above 127. ES currently converts the title text according to CP-437 codepage when it really shouldn't have any non-ASCII characters; hence, you can have a menu that doesn't display correctly. That is something I will be fixing (and also adding extra information warning about it).

[L@Zar0]

Mmm....

While adding the frames to the shapes for the FONTS.VGA, I've seen that the serpentine symbols from shapes 9 & 10 have not more than 91 frames (plaques?).

So, I must assume that only there are uppercase chars in the usecode texts that use this symbols?... so it is only needed to add the uppercase chars with accent marks?

I've done the 4.shp and it works perfectly above frame 127 in conjuntion with 8.shp. Only its a replacement frame, not a conversion...

[marzo]

So, I must assume that only there are uppercase chars in the usecode texts that use this symbols?... so it is only needed to add the uppercase chars with accent marks?

There are indeed only uppercase characters in those fonts in the game usecode, as they are used only for plaques/tombstones; the likely reason is that they decided to save some space and didn't add anything which they wouldn't actually use for those fonts.

[L@Zar0]

Well, here is the FONTS.VGA, for the ES (Spanish/Castilian) language. Based on the Alternate font, not the original font from the game:

http://www.telefonica.net/web2/juarapa/FONTS.VGA_ES.zip

It contains the Conversion table and which chars/frames are implemented on the shapes.

Maybe it needs more testing, but the two or three things I've tested, had been working very well.

Enjoy it or use it as you want.

[L@Zar0]

Another odd question....

Where are the fonts of the Intros for BG & SI? They are not in FONTS.VGA, aren't they?

[marzo]

No, they are not. They are split between mainshp.flx, endgame.dat, intro.dat and fonts.vga, depending on which game you want. Not to mention that not all intro text is actual text, drawn with a font: some are shapes, full "pre-rendered" text that gets drawn on-screen.

[L@Zar0]

mmm...

I am talking about the subtitles that appear on the intros of BG & SI, the first, when the Guardian speaks, and the second, when it beggins with an image of the Castle... and say... "Lord British's Castle".... bla bla bla.

This texts are also in the exultmsg.txt, so, the fonts must be somewhere, but I could no find its source.

Also, the subtitles on the Intro for SI are not viewed, after the "... eighteen months..." thing, there is no text viewable (when the guard enters into the castle with the letter of the Guardian for Batlin and shows it to Lord British). Maybe in this case I am doing something wrong...

[L@Zar0]

Hi again,

I've downloaded today the latest snapshot of Exutl 1.4.x win-32

Finally... have you changed the method to assign the fonts with the translate spell? So... the frame 243 from 6.shp now it is "translated" to frame below 211 from 6.shp... or is it another method?

I've done a little testing and see that this frame, 243 (the spanish 'ó') it is "translated" to 'O' (upper case).... I don't know if this is how it must be.

Regards.

[marzo]

have you changed the method to assign the fonts with the translate spell?

Nope, not yet.

[Change]

I can't unzip the FONTS.ZIP file by winrar,7z and windows default unzipper, after downloading it.

any idea?

[Malignant Manor]

Redownload the file again since it is likely corrupt. The zip works fine.

[Gary]

Does anyone have alternate download links for the original fonts.vga at the beginning of this post? I have been scouring the internet trying to find them, but have not had much luck. thx

[me]

http://smellsweird.com/exult/FONTS.ZIP
And for Serpent Isle:
http://smellsweird.com/exult/FONTS-SI.ZIP

[Dominus]

I wonder, why does a new game sometimes *not* show the replacement font but the original one on start? But when you restart Exult and reload the game it always shows the new font?

[Dominus]

found the problem but no solution:

When you have both BG and FoV installed and set up in exult.cfg, start exult, go to setup, go back, start a FoV game.
Exult will not free certain files from BG and will use those instead of the FoV files. For example the fonts.vga.

Edit : this has been fixed some time ago

[Kalos Dragon]

Perhaps I will extract the font from Ultima Online and adapt it for use in Exult.