Subject: Re: Merging or combining fonts query - msg#00007

Previous Message by Date: click to view message preview

Re: Merging or combining fonts query

To merge TrueType "hints" from two fonts, you'll either need to do it manually or you'll need to write a program that will in essence run the TrueType program for each glyph and understand what it going on in each instruction well enough to recognize what is a Control Value and intelligently make substitutions from a merged Control Value Table. It will need to recognize when the two fonts have set up the default graphics state differently and intelligently make allowances for that. It will have to make allowances for the two sets of glyph programs using the Storage Area in different ways and keep them from overwriting each other's values. And for entirely different and incompatible collections of functions. And if you can overcome these difficulties, the results still won't be great, because different programmers will have been making different decisions about stroke widths at various resolutions and other matters. Such problems arise from the fact that TrueType "hints" are not really "hints" at all, but rather programs in a "language" that resembles assembly language. To write a program to merge the "hints" from two TrueType fonts is a similar project to, say, writing one that will analyze the binaries for AbiWord and KWord and merge the two so you can have a word processor with the features of both. The scale of the project is different, but the basic idea is the same. Good luck, Peter Max Rabkin wrote: > On Mon, Mar 3, 2008 at 6:38 PM, Roger Sperberg <rsperberg@xxxxxxxxx> wrote: > >> In both of these trials, I lost all anchor points when copying DejaVu Sans >> glyphs into the KhmerOS font. (Kind of a peripheral question is, Did I lose >> more than that? Like: did I lose hinting information?) >> > > Yes, you lose hinting information. George and I discussed this a while > back, and decided it *might* be possible to preserve hints on merge. > George wasn't interested in doing it. I think it's an interesting > problem, but didn't look into it at the time as nobody seemed to be > desperately in need. > > If you would find it useful, I can take a look. I'm rather busy, so it > might take a while, but if you bug me about it over a few weeks, I'll > either get it done or decide it's impossible. > > >> Roger Sperberg >> >> > > --Max > > ------------------------------------------------------------------------- > This SF.net email is sponsored by: Microsoft > Defy all challenges. Microsoft(R) Visual Studio 2008. > http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ > _______________________________________________ > Fontforge-users mailing list > Fontforge-users@xxxxxxxxxxxxxxxxxxxxx > https://lists.sourceforge.net/lists/listinfo/fontforge-users > ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/

Next Message by Date: click to view message preview

cidmaps

I'd like to get FF to use Adobe cmaps newer than those supplied, ie CNS and GB 1.5 rather than 1.4. Is this something possible to do fairly painlessly? When it says in the docs "There are also many other character sets floating around feel free to install them yourself", is that referring to cmap or cidmap files, or are they the same (they don't seem to be)? And how do you install them? If with a CID font open, I do Change Supplement, then do search, and select an Adobe cmap file, FF crashes. (In any case, would it update glyph information that way? Wouldn't it be better if it was using the newer cmap when it opened the font?)  thanks------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/_______________________________________________ Fontforge-users mailing list Fontforge-users@xxxxxxxxxxxxxxxxxxxxx https://lists.sourceforge.net/lists/listinfo/fontforge-users

Previous Message by Thread: click to view message preview

Re: Merging or combining fonts query

To merge TrueType "hints" from two fonts, you'll either need to do it manually or you'll need to write a program that will in essence run the TrueType program for each glyph and understand what it going on in each instruction well enough to recognize what is a Control Value and intelligently make substitutions from a merged Control Value Table. It will need to recognize when the two fonts have set up the default graphics state differently and intelligently make allowances for that. It will have to make allowances for the two sets of glyph programs using the Storage Area in different ways and keep them from overwriting each other's values. And for entirely different and incompatible collections of functions. And if you can overcome these difficulties, the results still won't be great, because different programmers will have been making different decisions about stroke widths at various resolutions and other matters. Such problems arise from the fact that TrueType "hints" are not really "hints" at all, but rather programs in a "language" that resembles assembly language. To write a program to merge the "hints" from two TrueType fonts is a similar project to, say, writing one that will analyze the binaries for AbiWord and KWord and merge the two so you can have a word processor with the features of both. The scale of the project is different, but the basic idea is the same. Good luck, Peter Max Rabkin wrote: > On Mon, Mar 3, 2008 at 6:38 PM, Roger Sperberg <rsperberg@xxxxxxxxx> wrote: > >> In both of these trials, I lost all anchor points when copying DejaVu Sans >> glyphs into the KhmerOS font. (Kind of a peripheral question is, Did I lose >> more than that? Like: did I lose hinting information?) >> > > Yes, you lose hinting information. George and I discussed this a while > back, and decided it *might* be possible to preserve hints on merge. > George wasn't interested in doing it. I think it's an interesting > problem, but didn't look into it at the time as nobody seemed to be > desperately in need. > > If you would find it useful, I can take a look. I'm rather busy, so it > might take a while, but if you bug me about it over a few weeks, I'll > either get it done or decide it's impossible. > > >> Roger Sperberg >> >> > > --Max > > ------------------------------------------------------------------------- > This SF.net email is sponsored by: Microsoft > Defy all challenges. Microsoft(R) Visual Studio 2008. > http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ > _______________________________________________ > Fontforge-users mailing list > Fontforge-users@xxxxxxxxxxxxxxxxxxxxx > https://lists.sourceforge.net/lists/listinfo/fontforge-users > ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/

Next Message by Thread: click to view message preview

Re: Merging or combining fonts query

Quoting Roger Sperberg <rsperberg@xxxxxxxxx>: > This seems very valuable to me, especially given the difficulty newcomers > have in learning how to hint fonts. I mean, if someone has done that work > already, why discard it? Peter Baker has already given a detailed and accurate account of why hinting work in truetype should be discarded. Please read it. Basically, it boils do to: When truetype hints are moved from one font to another they WILL DO THE WRONG THING -- they will make the glyph look worse, not better. There are some bad things that Peter didn't mention: Something on the stack might be a cvt index, or a function index, or a storage index in one branch of an if statement (ie it needs to be modified), while it's used as a value in another branch of an if statement (ie, it can't be modified). PostScript hints can be merged, and are. So if your goal is to merge fonts, try PostScript ones. > So, yes, I think it's great if you can manage this. I believe it is equivalent to the halting problem, which means that the general case is unsolvable. That's why I don't work on it. ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/