| --- Log opened Tue Jan 02 00:00:25 2007 |
| 00:19 | | Stephenie|AFK is now known as Stephenie |
| 00:56 | | Stephenie is now known as Stephenie|FoodYAY |
| 02:47 | | Stephenie|FoodYAY is now known as Stephenie |
| 03:20 | | Stephenie is now known as TheRoom |
| 03:21 | | TheRoom is now known as Stephenie |
| 03:21 | | Stephenie is now known as TheRoom |
| 03:22 | | TheRoom is now known as Stephenie |
| 05:10 | < Reiver> | Raif |
| 05:10 | < Reiver> | http://svaf.sourceforge.net/svaf.html |
| 05:10 | < Reiver> | Mon[22:22] <McMartin> I'm a little iffy on how I'm handling textures. |
| 05:10 | < Reiver> | Mon[22:22] <McMartin> And I'm also wondering if I should give the self-contained vs. non-self-contained split an explicit definition in its own section and stuff. |
| 05:11 | < Raif> | I was thinking how I would do textures and he seems to be echoing the same basic idea. |
| 05:11 | < Raif> | What's his second line referring to? |
| 05:13 | < Raif> | Looks like he has embedded PNGs and/or embedded TGAs... I guess the texture coords are the sticking point? |
| 05:13 | < Reiver> | As far as I can tell, yes. |
| 05:14 | < Reiver> | I think the plan is that you can have the PNG or TGA either embedded into the file, or have them in a seperate file (More efficient if you have a dozen models all using the same texture). |
| 05:14 | < Raif> | To self-contained vs. non: Yes, explicit definition == good. |
| 05:14 | < Raif> | I think more detail is necessary. |
| 05:15 | < Raif> | There are all kinds of things you can do with textures, but in the end those are almost always hooked onto the geometric data (multitexturing, alpha mapping, etc) |
| 05:15 | < Reiver> | Which part needs the detail, then? |
| 05:16 | < Raif> | He could have something like a "Texture Combination" or something which will essentially index into textures and specify how to use them... vaguely like 3DSMax. |
| 05:17 | < Raif> | The application specific chunk for textures needs a little more hashing out. |
| 05:18 | < Raif> | For instance, if all I stuck in there was data to prevent multiple loading of textures (IE: This is textue 248, use it throughout the program as such), that information doesn't render the model non-self-contained. |
| 05:18 | < Raif> | It's throwaway data. |
| 05:19 | < Raif> | It's good to have a chunk like that though. |
| 05:20 | < Raif> | That clear anything up, or is it just about as transparent as a cement wall? :) |
| 05:21 | < Reiver> | Well, I'm not the writer, soooo... |
| 05:22 | < Reiver> | I just remembered you'd had an interest previously, so thought I'd poke you. |
| 05:22 | < Raif> | In summary: "Seems fine to me" |
| 05:22 | < Raif> | Particularly that the compressed form is in PNG. :) |
| 05:24 | < Reiver> | It's a good format, innit? |
| 05:31 | | * McMartin returns |
| 05:35 | | Netsplit Troika.TX.US.Nightstar.Net <-> DeepThought.NY.US.Nightstar.Net quits: EvilDarkLord, Reiver, @ToxicFrog, Stephenie, @McMartin |
| 05:35 | | Pi [~sysop@Nightstar-6915.hsd1.or.comcast.net] has quit [Ping Timeout] |
| 05:35 | | Serah [~-@87.72.36.ns-26407] has quit [Ping Timeout] |
| 05:35 | | Mahal [~Mahal@Nightstar-12066.worldnet.co.nz] has quit [Ping Timeout] |
| 05:35 | | jerith [~jerith@IRCop.Nightstar.Net] has quit [Ping Timeout] |
| 05:35 | | Raif [~corvusign@Nightstar-24484.hsd1.wa.comcast.net] has quit [Ping Timeout] |
| --- Log closed Tue Jan 02 05:35:34 2007 |
| --- Log opened Tue Jan 02 05:35:48 2007 |
| 05:35 | | TheWatcher[zZzZ] [~chris@Nightstar-29731.dsl.in-addr.zen.co.uk] has joined #code |
| 05:35 | | Irssi: #code: Total of 10 nicks [2 ops, 0 halfops, 0 voices, 8 normal] |
| 05:35 | < Reiver> | Ow. |
| 05:36 | < Reiver> | Might want to try that again, McM |
| 05:36 | | jerith [~jerith@IRCop.Nightstar.Net] has joined #code |
| 05:36 | | mode/#code [+o jerith] by ChanServ |
| 05:36 | | Irssi: Join to #code was synced in 28 secs |
| 05:36 | < McMartin> | Multitexturing and the like is out of range for the intended format -- the "S" is for "Simple". |
| 05:36 | < McMartin> | I'm a little unclear on your texture-ID schema. |
| 05:36 | < McMartin> | Is this like having the texture identifier chunks have ID numbers with them? |
| 07:31 | | Pi [~sysop@Nightstar-6915.hsd1.or.comcast.net] has joined #code |
| 07:31 | | mode/#code [+o Pi] by ChanServ |
| 07:45 | | AnnoDomini [~farkoff@Nightstar-29667.neoplus.adsl.tpnet.pl] has joined #Code |
| 08:28 | | ChalcyLaptop [~Owner@Nightstar-869.bitstream.orcon.net.nz] has joined #code |
| 09:03 | | ChalcyLaptop [~Owner@Nightstar-869.bitstream.orcon.net.nz] has quit [Quit: nigh night.] |
| 10:00 | | You're now known as TheWatcher |
| 10:31 | | Stephenie is now known as Stephenie|ZzZzZ |
| 11:04 | < Raif> | McM: TexID was an example. Let's say you have 100 textures, reused several times in 1000 models, and you only want to load each texture once. |
| 11:04 | < Raif> | You label each texture with an ID when you load it. Whenever you load a new texture, you see if you've already loaded that ID. If so, just use the already loaded one, otherwise, load it. |
| 11:05 | < Raif> | That allows each model to be independent (at the expense of disk space), while still not being ridiculously inefficient in video mem. |
| 11:05 | < Raif> | So each model is still self-contained, there's just some metadata you can tack on for your specific app. |
| 11:05 | < Raif> | I've had this problem before. This would have been the ideal solution. :) |
| 11:06 | < Reiver> | Hm. It remains application-specific, of course. |
| 11:06 | < Reiver> | ...Hey. |
| 11:06 | < Reiver> | What if it was checksum based? |
| 11:06 | < Reiver> | Thus meaning that two different textures can't be misnamed? |
| 11:06 | < Reiver> | And |
| 11:07 | < Reiver> | If you have a 'stock' set of 200 models, who use 100 textures, and use only 50 of the models, you don't have to worry about getting the wrong number? |
| 11:07 | | * Reiver ponders. |
| 11:07 | < Reiver> | The only catch we have here is that each model still stores all of its own textures. This means each file would still be inefficient on storage space. Or is that the point at which you start throwing the textures into another dir? |
| 11:09 | < Raif> | Reiver: That's one way to do it, yes. |
| 11:10 | < Raif> | But checksum is a hash. |
| 11:10 | < Raif> | You can have collisions. |
| 11:10 | < Raif> | This was just a specific example of how one might use that tertiary data chunk attached to each texture, and why that data chunk might not mean that the model isn't self-contained. |
| 11:13 | < Raif> | Also, I'm not sure if this is implied, but the format doesn't seem to specify how many children (subchunks) any given chunk has.... that should be in the format. |
| 11:13 | < Reiver> | Point. |
| 11:13 | < Reiver> | Hm. |
| 11:14 | < Reiver> | The problem is that a handwritten file can have collisions too. |
| 11:14 | < Raif> | For one thing, it makes dealing with corruption much easier, and more importantly eliminates the need to know implicitely how many children your chunk has. |
| 11:14 | < Reiver> | And given humanities knack for patternmatching, we tend to be terrible at inadvertently causing it. But. |
| 11:15 | < Raif> | (I only notice this because I've been working on a few dozen bugs which stem from this very lack, and cause all manner of nastiness in corrupt files) |
| 12:24 | | Reiver is now known as ReivZzz |
| 14:00 | | You're now known as theWatcher[hereish] |
| 15:48 | | You're now known as TheWatcher |
| 16:18 | | Stephenie|ZzZzZ is now known as Stephenie |
| 16:40 | | * McMartin reads backscroll |
| 16:40 | < McMartin> | Hmm. That's actually a totally separate issue. |
| 16:41 | < McMartin> | I'm solving the "arbitrary number of children" problem by using a two-phase file parser, implementing it in a handy library, and releasing it under BSD so other people can steal it. ;) |
| 16:42 | < McMartin> | IFF in general is pretty good about throwing lists together since length information is part of each chunk, so you can hop around quickly. |
| 17:09 | | Stephenie is now known as Stephenie|Groceries |
| 17:14 | | You're now known as TheWatcher[afk] |
| 17:43 | | AnnoDomini [~farkoff@Nightstar-29667.neoplus.adsl.tpnet.pl] has quit [Ping Timeout] |
| 17:49 | | AnnoDomini [~farkoff@Nightstar-29143.neoplus.adsl.tpnet.pl] has joined #Code |
| 18:46 | | You're now known as TheWatcher |
| 19:00 | | Chalcedon [~Chalceon@Nightstar-869.bitstream.orcon.net.nz] has joined #code |
| 19:00 | | mode/#code [+o Chalcedon] by ChanServ |
| 19:54 | | Janus [~Cerulean@Nightstar-10302.columbus.res.rr.com] has joined #Code |
| 20:02 | | Chalcy [~Chalceon@Nightstar-869.bitstream.orcon.net.nz] has joined #code |
| 20:02 | | mode/#code [+o Chalcy] by ChanServ |
| 20:03 | | Chalcedon [~Chalceon@Nightstar-869.bitstream.orcon.net.nz] has quit [Ping Timeout] |
| 20:03 | | Chalcy is now known as Chalcedon |
| 20:26 | | Janus is now known as Jan[caroot] |
| --- Log closed Tue Jan 02 21:46:02 2007 |
| --- Log opened Tue Jan 02 21:46:39 2007 |
| 21:46 | | TheWatcher [~chris@Nightstar-29731.dsl.in-addr.zen.co.uk] has joined #code |
| 21:46 | | Irssi: #code: Total of 9 nicks [2 ops, 0 halfops, 0 voices, 7 normal] |
| 21:46 | | mode/#code [+o TheWatcher] by ChanServ |
| 21:47 | | Irssi: Join to #code was synced in 29 secs |
| 22:01 | | Jan[caroot] [~Cerulean@Nightstar-10302.columbus.res.rr.com] has quit [Quit: Jouets de Dieu, jouets de jouets, les jouets de me, naître Clair enfant voire.] |
| 22:01 | | Chalain [~chalain@Admin.Nightstar.Net] has joined #Code |
| 22:01 | | mode/#code [+o Chalain] by ChanServ |
| 22:01 | | Raif [~corvusign@Nightstar-22484.hsd1.ca.comcast.net] has joined #Code |
| 22:01 | | jerith [~jerith@IRCop.Nightstar.Net] has joined #Code |
| 22:01 | | Serah [~-@87.72.36.ns-26407] has joined #Code |
| 22:02 | | AnnoDomini [~farkoff@Nightstar-29143.neoplus.adsl.tpnet.pl] has joined #Code |
| 22:02 | | jerith is now known as NSGuest-1008 |
| 22:32 | | ReivZzz is now known as Reiver |
| 22:42 | | AnnoDomini [~farkoff@Nightstar-29143.neoplus.adsl.tpnet.pl] has quit [Quit: Juffo-Wup is a *candle*. It is filled with many *candy bars*.] |
| 23:04 | | You're now known as TheWatcher[T-2] |
| 23:09 | | You're now known as TheWatcher[zZzZ] |
| 23:40 | | Pi [~sysop@Nightstar-6915.hsd1.or.comcast.net] has joined #code |
| 23:40 | | mode/#code [+o Pi] by ChanServ |
| 23:53 | | Chalcedon is now known as ChalcyUni |
| 23:54 | | EvilDarkLord [althalas@Nightstar-15301.a88-115-211-62.elisa-laajakaista.fi] has quit [Ping Timeout] |
| 23:57 | | Chalcy [~Chalceon@Nightstar-869.bitstream.orcon.net.nz] has joined #code |
| 23:57 | | mode/#code [+o Chalcy] by ChanServ |
| 23:58 | | ChalcyUni [~Chalceon@Nightstar-869.bitstream.orcon.net.nz] has quit [Ping Timeout] |
| --- Log closed Wed Jan 03 00:00:25 2007 |