Forums

what's the feasibility of making the gd3 of a vgm formatted in xml? currently the gd3 is a sequence of 11 strings (using 16-bit chars, each string terminated by an L'\0') and attempting to write a cross-platform solution to handling a vgm's gd3 is pissing me off to no end, mainly due to the difference in wchar implementations btwn OSes. even tho it'll probably be more work, i think it'll make gd3 more extensible and we'd be able to add per-system, per-game, and even per-song custom elements. also, this would be a possible base for a solution to cataloguing instrument presets that appear in songs for the eventual vgm-to-midi conversion i'm adding to vgmtool.

example song gd3-xml:

<?xml version="1.0" encoding="UTF-8"?>
<gd3>
<system>
<en>Sega Mega Drive / Genesis</en>
<jp>セガメガドライブ</jp>
</system>
<game>
<en>Zero Wing</en>
<jp>ゼロウィング</jp>
</game>
<composer>
<en>Tatsuya Uemura</en>
<jp>上村達也</jp>
</composer>
<composer>
<en>Toshiaki Tomizawa</en>
<jp>富沢敏明</jp>
</composer>
<composer>
<en>Masahiro Yuge</en>
<jp>弓削雅稔</jp>
</composer>
<releasedate>1991/05/31</releasedate>
<ripper>-DJSW-</ripper>
<trackname>
<en>Introduction (Part 1)</en>
</trackname>
</gd3>

the above was just an example of a single-song gd3. there's likely improvements that can be made in places, but one of the plusses is the potential to use entities and such in text. if a proper doctype was created we could have rudimentary xml validation as well.

tell me what y'all think abt this.

I was thinking of doing that myself. The problem is a lack of an editor, and of course backwards compatibility. It might be good to instead use a standard format like Vorbis Comment and have standard field names, rather than the structure of XML. Vorbis Comment seems to be a concatenated set of Pascal-style strings in UTF-8.

in_vgm 2 is in active development, so now is a good time to make suggestions, especially if they make these damn tags less awful to parse. I could even perhaps allow editing in Winamp... which would give automagical mass tagging for free.

since the gd3 would be xml, it theoretically could be a file separate from the vgm, therefore editable in any plain text editor (preferably one that supports utf-8 or utf-16).

in all seriousness, tho dude, the gd3 format as it is now is pissing me off cuz i can't come up w/a cross-platform solution to reading it and writing it for the vgmtool update. ATM the only way i can guarantee reading it as a string of 16-bit chars is to loop thru using "unsigned short" type, and even then it won't display in the console/terminal as a unicode string (i WANTED to use wchar.h and its functions, but wchar_t is 32 bits on non-windows :( ).

i'm so tempted to hold off on porting gd3 tagging & editing on the cross-platform vgmtool until AFTER i do everything else because of this :/ i'm seriously, you guys!

16-bit "Unicode" is also known as UCS-2. Pass it through iconv to get your UCS-4.

Author	Message
neologix Joined: 13 Nov 2007 Posts: 93 Location: New York, NY, USA	vgm/gd3 format & xml? Posted: Fri Jul 10, 2009 5:45 am
	what's the feasibility of making the gd3 of a vgm formatted in xml? currently the gd3 is a sequence of 11 strings (using 16-bit chars, each string terminated by an L'\0') and attempting to write a cross-platform solution to handling a vgm's gd3 is pissing me off to no end, mainly due to the difference in wchar implementations btwn OSes. even tho it'll probably be more work, i think it'll make gd3 more extensible and we'd be able to add per-system, per-game, and even per-song custom elements. also, this would be a possible base for a solution to cataloguing instrument presets that appear in songs for the eventual vgm-to-midi conversion i'm adding to vgmtool. example song gd3-xml: <?xml version="1.0" encoding="UTF-8"?> <gd3> <system> <en>Sega Mega Drive / Genesis</en> <jp>セガメガドライブ</jp> </system> <game> <en>Zero Wing</en> <jp>ゼロウィング</jp> </game> <composer> <en>Tatsuya Uemura</en> <jp>上村達也</jp> </composer> <composer> <en>Toshiaki Tomizawa</en> <jp>富沢敏明</jp> </composer> <composer> <en>Masahiro Yuge</en> <jp>弓削雅稔</jp> </composer> <releasedate>1991/05/31</releasedate> <ripper>-DJSW-</ripper> <trackname> <en>Introduction (Part 1)</en> </trackname> </gd3> the above was just an example of a single-song gd3. there's likely improvements that can be made in places, but one of the plusses is the potential to use entities and such in text. if a proper doctype was created we could have rudimentary xml validation as well. tell me what y'all think abt this.

Maxim Site Admin Joined: 19 Oct 1999 Posts: 14761 Location: London	Posted: Fri Jul 10, 2009 10:00 am
	I was thinking of doing that myself. The problem is a lack of an editor, and of course backwards compatibility. It might be good to instead use a standard format like Vorbis Comment and have standard field names, rather than the structure of XML. Vorbis Comment seems to be a concatenated set of Pascal-style strings in UTF-8. in_vgm 2 is in active development, so now is a good time to make suggestions, especially if they make these damn tags less awful to parse. I could even perhaps allow editing in Winamp... which would give automagical mass tagging for free.

neologix Joined: 13 Nov 2007 Posts: 93 Location: New York, NY, USA	Posted: Sat Jul 11, 2009 6:00 am
	since the gd3 would be xml, it theoretically could be a file separate from the vgm, therefore editable in any plain text editor (preferably one that supports utf-8 or utf-16). in all seriousness, tho dude, the gd3 format as it is now is pissing me off cuz i can't come up w/a cross-platform solution to reading it and writing it for the vgmtool update. ATM the only way i can guarantee reading it as a string of 16-bit chars is to loop thru using "unsigned short" type, and even then it won't display in the console/terminal as a unicode string (i WANTED to use wchar.h and its functions, but wchar_t is 32 bits on non-windows :( ). i'm so tempted to hold off on porting gd3 tagging & editing on the cross-platform vgmtool until AFTER i do everything else because of this :/ i'm seriously, you guys!

Maxim Site Admin Joined: 19 Oct 1999 Posts: 14761 Location: London	Posted: Sat Jul 11, 2009 8:57 am
	16-bit "Unicode" is also known as UCS-2. Pass it through iconv to get your UCS-4.

Forums

View topic - vgm/gd3 format & xml?