| --- Log opened Sat Jul 09 00:00:35 2016 |
| 00:08 | | JustLurk [justbob@ServerAdministrator.Nightstar.Net] has joined #code |
| 00:08 | | Guru [ShadyGuru@Nightstar-qfckjl.tv13.ptd.net] has joined #code |
| 00:08 | | Vornotron [Vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 00:08 | | macdjord|slep [macdjord@Nightstar-r9vt2h.mc.videotron.ca] has joined #code |
| 00:09 | | mode/#code [+o macdjord|slep] by ChanServ |
| 00:09 | | Netsplit Kakrafoon.Nightstar.Net <-> Deepthought.Nightstar.Net quits: @Vornicus, @macdjord|slep, @mac, @PinkFreud, @himi, @pjdelport, gizmore, Vornotron, Shady, @Alek, (+4 more, use /NETSPLIT to show all of them) |
| 00:09 | | Netsplit over, joins: Tamber |
| 00:09 | | Guru is now known as Shady |
| 00:13 | | Alek [Alek@Nightstar-9qtiqv.il.comcast.net] has joined #code |
| 00:38 | | Derakon[AFK] is now known as Derakon |
| 00:41 | <@simon_> | how do I copy a lot of files in Windows? rsync was suggested earlier, but I don't have a transfer protocol ready! |
| 00:41 | <@simon_> | i.e. I don't know of a piece of rsync server software, I don't run an sshd on any of the Windows machines, and I frankly don't need the encryption overhead. |
| 00:47 | <&McMartin> | From one Windows machine to another? Can you just use SMB/Windows sharing? |
| 00:51 | | Shady is now known as Shady|afk |
| 01:03 | <@simon_> | McMartin, 2TB, millions and millions of files. would take several days. |
| 01:04 | <@simon_> | I'm going to install FTP. |
| 01:05 | <&McMartin> | Probably for the best |
| 01:11 | | Kindamoody|autojoin [Kindamoody@Nightstar-0lgkcs.tbcn.telia.com] has joined #code |
| 01:12 | | Kindamoody|autojoin is now known as Kindamoody |
| 01:14 | | Kindamoody is now known as Kindamoody[zZz] |
| 01:15 | < ToxicFrog> | simon_: the encryption overhead is basically invisible compared to the disk and network cost unless these are *really* slow machines. |
| 01:15 | < ToxicFrog> | It does require installing sshd or rsyncd, though. |
| 01:30 | | JustLurk is now known as JustBob |
| 01:30 | | Kindamoody|autojoin [Kindamoody@Nightstar-0lgkcs.tbcn.telia.com] has joined #code |
| 01:30 | | pjdelport [sid25146@Nightstar-7he56f.irccloud.com] has joined #code |
| 01:30 | | gizmore [kvirc@Nightstar-dnf252.dip0.t-ipconnect.de] has joined #code |
| 01:30 | | himi [sjjf@Nightstar-v37cpe.internode.on.net] has joined #code |
| 01:30 | | macdjord|slep [macdjord@Nightstar-r9vt2h.mc.videotron.ca] has joined #code |
| 01:30 | | ricer [ricer@Nightstar-52ic2g.wechall.net] has joined #code |
| 01:30 | | PinkFreud [WhyNot@NetworkAdministrator.Nightstar.Net] has joined #code |
| 01:30 | | Vornotron [Vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 01:30 | | ServerMode/#code [+ooooo Kindamoody|autojoin pjdelport himi macdjord|slep PinkFreud] by *.Nightstar.Net |
| 01:30 | | mode/#code [+o Alek] by ChanServ |
| 01:30 | | mode/#code [+o Tamber] by ChanServ |
| 01:30 | | mode/#code [+o JustBob] by ChanServ |
| 01:54 | | gizmore [kvirc@Nightstar-dnf252.dip0.t-ipconnect.de] has quit [Operation timed out] |
| 02:05 | | Derakon is now known as Derakon[AFK] |
| 02:28 | | catalyst [catalyst@Nightstar-bt5k4h.81.in-addr.arpa] has quit [[NS] Quit: Leaving] |
| 02:49 | | catadroid [catadroid@Nightstar-mnhhrc.dab.02.net] has joined #code |
| 05:11 | | Derakon[AFK] is now known as Derakon |
| 05:18 | | Pinkhair [user1@Nightstar-g7hdo5.dyn.optonline.net] has joined #code |
| 05:29 | | crystalclaw is now known as crystalclaw|AFK |
| 05:39 | < Pinkhair> | Anyone happen to be familiar with blending octaves of noise properly? |
| 05:45 | < Vornotron> | how do you mean properly |
| 05:56 | | Shady|afk [ShadyGuru@Nightstar-qfckjl.tv13.ptd.net] has quit [[NS] Quit: Yay, he's gone] |
| 05:57 | < Pinkhair> | As in avoiding a constant rise of fall of overall value |
| 05:58 | < Pinkhair> | I'm having a bit of an issue with higher octaves lowering the overall value. |
| 05:59 | | celticminstrel [celticminst@Nightstar-nhhr58.dsl.bell.ca] has quit [[NS] Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] |
| 06:11 | < Vornotron> | normalize each to -1..1, then multiply by an energy term to get your appropriate noise color? |
| 06:13 | < Vornotron> | then when you're done with that you renormalize to the range you really want. This keeps your expected noise value exactly where you want it, because you're setting the expectation of each octave to 0, so adding it up doesn't move the expectation |
| 06:14 | < Pinkhair> | Hrm, that might be the problem |
| 06:14 | < Pinkhair> | since I was using frac() that will be returning 0-1 instead of -1-1 |
| 06:17 | < Pinkhair> | aha, that was it! Thank you |
| 06:38 | | Derakon is now known as Derakon[AFK] |
| 07:47 | | Kindamoody|autojoin is now known as Kindamoody |
| 07:47 | < Pinkhair> | So yeah, definite improvement. http://i.imgur.com/t3rMn8r.jpg |
| 07:48 | | Kindamoody[zZz] [Kindamoody@Nightstar-0lgkcs.tbcn.telia.com] has quit [Client exited] |
| 07:48 | < Pinkhair> | Now I need to figure out how to generate some normals. |
| 08:12 | < Vornotron> | Nice. Okay, you've got... altitudes? |
| 08:17 | < Pinkhair> | Yes. |
| 08:18 | < Pinkhair> | The problem is that I can't just use sobel on the heights because they aren't in UV space |
| 08:20 | < Vornotron> | What's your map projection here? peters? equirectangular? somethng different? |
| 08:20 | < Vornotron> | ...cube map? |
| 08:21 | < Pinkhair> | They are in world space. |
| 08:22 | < Pinkhair> | 3d noise |
| 08:22 | < Pinkhair> | The shader samples for each pixel |
| 08:22 | < Vornotron> | Oh, Fun. |
| 08:23 | < Pinkhair> | It is great for avoiding seams. Less great for... this. |
| 08:25 | < Pinkhair> | I Since it will probably only be used on spheres, one possibility is to sample around it by rotating worldspace around the center |
| 08:27 | < Vornotron> | hm. you could sample the other pixel's altitude values -- this actually will be terribly inaccurate for stuff near the edges |
| 08:28 | < Vornotron> | though that's also already true, really |
| 08:32 | < Pinkhair> | what edges? |
| 08:33 | < Vornotron> | THe "horizon" of the planet, as we see it |
| 08:35 | < Pinkhair> | Well, it wouldn't be taking the camera into account |
| 08:35 | < Pinkhair> | though actually, I could do that |
| 08:35 | < Pinkhair> | render the noise to world space |
| 08:35 | < Pinkhair> | er, to screen space |
| 08:35 | < Pinkhair> | and cheat that way |
| 08:35 | < Pinkhair> | I'm also considering just doing a massive cheat |
| 08:35 | < Vornotron> | oh, you're ctually rendering -- I'm very confused now, etc |
| 08:36 | < Vornotron> | you're doing the render for each pixel but you're doing this in world space? |
| 08:36 | < Pinkhair> | Yes |
| 08:36 | < Pinkhair> | Ok, here's how it works. |
| 08:37 | < Pinkhair> | For every 'fragment' it takes the world position of the surface and feeds it into the noise. The noise is mapped into a texture lookup to yield a color, and then that's blended into the lighting. |
| 08:38 | < Pinkhair> | I'm considering maybe just completely cheating and making a second, slightly darker copy with the water masked out and shifting it away from the light. That would only work for a single light, but that is probably fine... |
| 08:38 | < Vornotron> | okay, so you're converting screen fragment to world point to noise-based altitude (but not offsetting the pixel according to that altitude) and then getting the texture coordinate accordingly |
| 08:38 | < Pinkhair> | Correct. |
| 08:39 | < Vornotron> | Okay. |
| 08:40 | < Pinkhair> | Currently, the altitude does nothing b ut color- it can also feed into actual vertex displacement but I don't see much need for that other than making things look cartoony |
| 08:40 | < Pinkhair> | Since I can't read the normals of the displaced geometry |
| 08:40 | < Vornotron> | And it doesn't help much anyway if these are "seen from space" -- the variation on normal planets tend not to be enough to notice. |
| 08:41 | < Vornotron> | on earth the highest point is 1/1000 the average diameter |
| 08:42 | < Vornotron> | so let's see. you could run, like -- resample at three points nearby in world space. It gets hairy near the horizon though because you've got massive gappage. |
| 08:42 | < Pinkhair> | Yes, I know. Like I said, cartoony. |
| 08:43 | < Pinkhair> | But if I could read those normals I could generate a tiny displacement and multiply it for the actual shading |
| 08:43 | < Pinkhair> | but I can't |
| 08:47 | < Vornotron> | man I'm not sure. If you could generate geometry, you could -- oh, idunno, build an icosphere, geodesate to a decent value, and then noise those up and render that thing, but it won't be quite as high-detail. |
| 08:48 | < Vornotron> | you'd be throwing all this shader stuff away though :( |
| 08:49 | < Vornotron> | anyway. getting an altitude, you can... hm. try using adjacent fragments' altitudes, maybe, to generate the normal, except that fragments adjacent to "space" fragmentss are going to have problems |
| 08:49 | | catadroid` [catadroid@Nightstar-293.pbu.132.82.IP] has joined #code |
| 08:50 | < Pinkhair> | I think I'm going to try resampling everything with worldspace rotated slightly |
| 08:50 | < Pinkhair> | And then converting those to tangent space |
| 08:50 | < Pinkhair> | I dunno, i'm way over my head here |
| 08:52 | < Vornotron> | I'm in pretty deep too |
| 08:52 | < Vornotron> | Wait - what are these noise things? might it be possible to generate derivatives directly? |
| 08:53 | | catadroid [catadroid@Nightstar-mnhhrc.dab.02.net] has quit [Ping timeout: 121 seconds] |
| 08:53 | | * catadroid` will write render code again one day |
| 08:55 | < Vornotron> | if you ever have you might be able to help here, though i suggest the scond sight to catch up. http://www.starforge.co.uk/irclogs/code/latest.html |
| 08:59 | < Pinkhair> | you'll cry when you see the noise code |
| 09:00 | < Pinkhair> | I cry when I see it. |
| 09:00 | < Pinkhair> | http://pastebin.com/raw/1LS0BxaL |
| 09:02 | < Pinkhair> | Takes a float3(coords) and an int for iterations, which is what I called octaves because I started with a 2d noise someone else had done |
| 09:03 | < Pinkhair> | (I inherited the compounding iteration error from them) |
| 09:04 | < Vornotron> | ¿que? |
| 09:04 | < Pinkhair> | the center of the planet will always be 0,0,0 since I feed the object position in to zero it out |
| 09:04 | < Pinkhair> | the thing you helped me solve initially |
| 09:05 | < Vornotron> | Yeah, I'm looking at the heegybeegy, this is pretty nasty. |
| 09:05 | < Vornotron> | some weirdass magic number |
| 09:05 | < Pinkhair> | Yeah. I had to extrapolate those |
| 09:05 | < Pinkhair> | and sacrifice a chicken |
| 09:07 | < Pinkhair> | I am very strongly considering simply getting normals in screenspace |
| 09:08 | < Pinkhair> | by the way |
| 09:08 | < Pinkhair> | how do the cloud shadows look? |
| 09:13 | < Vornotron> | oh, there are cloud shadows |
| 09:15 | < Vornotron> | uh... in the nearest one they look too far from the clouds themselves. |
| 09:18 | < Vornotron> | correct shadow location is -- locate world space, locate cloud altitude in sun direction from world space, then do cloud noise for that point |
| 09:22 | < Vornotron> | I'm not sure what's happening here but it looks like you're getting way over to the right for some reason |
| 09:22 | < Pinkhair> | They are an extremely cheap trick |
| 09:22 | < Pinkhair> | basically, the shadows are on the surface and the clouds are the same texture with some parallax |
| 09:24 | < Pinkhair> | Doesn't even take the lights into account, other than to mask out lighting in that area |
| 09:24 | < Pinkhair> | I really need a better cloud texture |
| 09:27 | < Pinkhair> | hum, that's a thought |
| 09:28 | < Vornotron> | that noise texture is nasty nasty |
| 09:28 | < Pinkhair> | No shit |
| 09:31 | < Pinkhair> | you know, another option might be to just use a generic normalmap and scale the intensity by the noise texture |
| 09:31 | < Pinkhair> | and mask out the oceans |
| 09:31 | < Vornotron> | a possibility |
| 09:32 | < Pinkhair> | Also, other than the shadow distance, I just took a look at Stellaris and it seems they do their clouds the same way I did |
| 09:54 | | catadroid` [catadroid@Nightstar-293.pbu.132.82.IP] has quit [[NS] Quit: Bye] |
| 10:25 | < Pinkhair> | what a mess |
| 10:25 | < Pinkhair> | I think this shader code might serve a secondary function as an anti-programmer basilisk |
| 10:40 | < Pinkhair> | http://i.imgur.com/HrlgYfq.jpg |
| 11:02 | < Pinkhair> | pretty happy with that. now I need to figure out cities |
| 11:03 | < Pinkhair> | and then the dreaded part |
| 11:03 | < Pinkhair> | optimization >.>; |
| 11:06 | < Pinkhair> | though it's actually not as heavy as I'd feared |
| 11:08 | < Pinkhair> | eight draw calls |
| 11:14 | < Vornotron> | 8 draw calls... for that. |
| 11:14 | < Vornotron> | well, I'm impressed. |
| 11:21 | < Pinkhair> | Well, with one light. Four more per extra light |
| 11:27 | < Vornotron> | how many fricking lights are you planning to use on this thing |
| 11:28 | < Pinkhair> | Probably jsut the one |
| 11:28 | < Pinkhair> | but more are possible |
| 11:30 | < Pinkhair> | But yeah, I'm pretty happy with it |
| 11:43 | | Vornotron is now known as Vornicus |
| 11:44 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 11:45 | | Kindamoody is now known as Kindamoody|afk |
| 11:54 | < Pinkhair> | http://i.imgur.com/Wvti1os.jpg |
| 11:59 | <&VirusJTG> | http://www.albinoblacksheep.com/flash/llama |
| 12:01 | <~Vornicus> | Pinkhair: I wanna live on that trefoil one |
| 12:03 | <~Vornicus> | also |
| 12:03 | <~Vornicus> | Capsule Planet! He's A Hero! |
| 12:04 | < Pinkhair> | gonna take UV seams down to zero |
| 12:04 | < Pinkhair> | (though it actually still uses UVs for the clouds) |
| 12:05 | < Pinkhair> | I figured clouds didn't really need to be procedurally generated since there's a lot less visible variation. I might dilate the texture and make a thickness slider though |
| 12:06 | | Emmy [M@Nightstar-9p7hb1.direct-adsl.nl] has joined #code |
| 12:06 | | mode/#code [+o Emmy] by ChanServ |
| 12:07 | | Emmy [M@Nightstar-9p7hb1.direct-adsl.nl] has quit [Connection closed] |
| 12:07 | <~Vornicus> | I once made a map projection designed to prevent the UV crash at the poles |
| 12:10 | | catadroid [catadroid@Nightstar-sc2a75.dab.02.net] has joined #code |
| 12:12 | < Pinkhair> | Oh? |
| 12:14 | < Pinkhair> | ...and damn, I was livestreaming my desktop this whole time on twitch without noticing |
| 12:14 | < catadroid> | \o/ |
| 12:15 | < Pinkhair> | fortunately all I was doing was shader programming for the two hours I was live |
| 12:16 | <~Vornicus> | Yeah. Basically, the issue with the standard projections is you get this thing where a line on the texture corresponds to a point on the model. This is terrible. So I put the poles in opposite corners and the equator on the other diagonal. |
| 12:17 | <~Vornicus> | Using equidistant lines of latitude you get some area distortion, but not nearly as bad as equirectangular. |
| 12:18 | <~Vornicus> | It's like -- on equirectangular, using a 512x512 texture, a texel at the equator covers 5000 times the area of one at the poles. On mine, it was 2/3. |
| 12:18 | <~Vornicus> | but more importantly, you don't lose stuff into the space between the triangles |
| 12:19 | < Pinkhair> | Yeah, I get what you're saying. |
| 12:32 | | catadroid [catadroid@Nightstar-sc2a75.dab.02.net] has quit [The TLS connection was non-properly terminated.] |
| 12:34 | <~Vornicus> | (the area distortion problem doesn't exist with peters projection, but you still get terrible things happening at the poles with the failure of your triangles to actually meet on the texture) |
| 12:42 | <~Vornicus> | Also you should show that thing with the ridicuplanets to fleet. |
| 12:44 | | Vornotron [Vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 12:44 | < Vornotron> | Also you should show that thing with the ridicuplanets to fleet. |
| 12:46 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has quit [Operation timed out] |
| 12:55 | < Pinkhair> | man, I hate my voice in recordings. |
| 12:55 | < Pinkhair> | I sound so inept. |
| 13:02 | < Pinkhair> | https://www.youtube.com/watch?v=WlUYMACnLlI |
| 13:42 | | gizmore [kvirc@Nightstar-snivg2.dip0.t-ipconnect.de] has joined #code |
| 13:55 | | * Vornotron finds himself wondering how he'd do the three-square corners in a perlin noise thingummy built around the cubification of the sphere |
| 13:58 | < Pinkhair> | Add a fourth dimension and make it a time cube |
| 13:59 | < Pinkhair> | But yeah, the annoying part would be UV mapping it. |
| 14:00 | <@simon_> | Pinkhair, sound is good. |
| 14:00 | < Pinkhair> | Then I'd probably use a rotated sin function to scale the UVs |
| 14:00 | < Pinkhair> | before applying the noise |
| 14:00 | <@simon_> | Pinkhair, considering how horrible people are at talking in youtube tutorials (especially how far it takes for people to get to their actual content), that's way up on the scale. :) |
| 14:01 | < Pinkhair> | Thank you, it is appreciated! |
| 14:01 | | celticminstrel [celticminst@Nightstar-nhhr58.dsl.bell.ca] has joined #code |
| 14:01 | | mode/#code [+o celticminstrel] by ChanServ |
| 14:01 | < Pinkhair> | And yeah, the super padded videos always annoyed me |
| 14:01 | | Pinkhair is now known as ASCII |
| 14:02 | <@simon_> | e.g. https://www.youtube.com/watch?v=1-tQshSno3c - it's not even that bad, but he has what sounds like a Swedish accent and the first 50 seconds is just fluff about his video making career. |
| 14:04 | < ASCII> | ah yeah. He's selling those, too. |
| 14:06 | <@simon_> | it's pretty amazing what he's made. I like his water level slider. |
| 14:07 | <@simon_> | sounds like he also had rendering performance in mind. |
| 14:08 | | * simon_ wants to do procedural generation in Unity, too. hard to prioritize programming outside of work, currently. >_< |
| 14:20 | < ASCII> | Dunno about amazing, but it's a good job certainly. |
| 14:22 | < ASCII> | I'm reaching the point where to go much much further I'd need to add some scripts |
| 14:45 | | Vornotron [Vorn@ServerAdministrator.Nightstar.Net] has quit [Connection reset by peer] |
| 14:50 | <@simon_> | copying these 2TB of small files is ridiculously slow. FTP and rsync are so slow I've copied a total of 2GB in four hours. |
| 14:51 | <@simon_> | this filesystem is really not compatible with this amount of files. |
| 15:14 | <@simon_> | since this is a RAID 1, isn't there a point in splitting up the rsync workload into two? |
| 15:31 | <@simon_> | I'm unsure exactly where the bottleneck is. I think it's filesystem operations. |
| 16:03 | | newbie [kvirc@Nightstar-tgb.uk2.218.85.IP] has joined #code |
| 16:04 | | newbie [kvirc@Nightstar-tgb.uk2.218.85.IP] has quit [[NS] Quit: KVIrc 4.2.0 Equilibrium http://www.kvirc.net/] |
| 16:06 | | catadroid [catadroid@Nightstar-sc2a75.dab.02.net] has joined #code |
| 16:15 | | gizmore [kvirc@Nightstar-snivg2.dip0.t-ipconnect.de] has quit [Ping timeout: 121 seconds] |
| 16:18 | | catadroid [catadroid@Nightstar-sc2a75.dab.02.net] has quit [Ping timeout: 121 seconds] |
| 16:42 | | catadroid [catadroid@Nightstar-2itqor.dab.02.net] has joined #code |
| 16:47 | | catadroid [catadroid@Nightstar-2itqor.dab.02.net] has quit [The TLS connection was non-properly terminated.] |
| 16:50 | | Shady [ShadyGuru@Nightstar-qfckjl.tv13.ptd.net] has joined #code |
| 17:19 | | Derakon[AFK] is now known as Derakon |
| 18:16 | | catalyst [catalyst@Nightstar-bt5k4h.81.in-addr.arpa] has joined #code |
| 18:32 | | gizmore [kvirc@Nightstar-b11psh.dip0.t-ipconnect.de] has joined #code |
| 18:35 | < catalyst> | It's gratifying that my initial attempts to macro up JavaFX are basically the same way the three public libraries I've seen use |
| 18:42 | <&[R]> | simon_: NTFS is pretty bad for speed AIUI |
| 20:38 | | catalyst [catalyst@Nightstar-bt5k4h.81.in-addr.arpa] has quit [[NS] Quit: Leaving] |
| 21:50 | | catadroid [catadroid@Nightstar-2itqor.dab.02.net] has joined #code |
| 22:17 | | Kindamoody|afk is now known as Kindamoody |
| 22:19 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 22:19 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 22:24 | | Shady is now known as Shady|afk |
| 22:56 | | catadroid` [catadroid@Nightstar-05rlhs.dab.02.net] has joined #code |
| 23:00 | | catadroid [catadroid@Nightstar-2itqor.dab.02.net] has quit [Ping timeout: 121 seconds] |
| --- Log closed Sun Jul 10 00:00:50 2016 |