July 17, 2019 
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | >
July 17, 2019 [00:09:49] *** LunarJetman <LunarJetman!LunarJetma@94.14.216.120> has quit IRC (Quit: LunarJetman)[00:13:49] *** goiko <goiko!~goiko@unaffiliated/goiko> has joined ##vulkan[01:04:18] *** ColaEuphoria <ColaEuphoria!~ColaEupho@unaffiliated/colaeuphoria> has joined ##vulkan[01:08:39] *** ratchetfreak <ratchetfreak!~ratchetfr@ptr-82s3g7nyo2zjr8pdfah.18120a2.ip6.access.telenet.be> has quit IRC (Ping timeout: 252 seconds)[01:12:09] <ColaEuphoria> Are geometry shaders an efficient way to do flat shading? I heart geometry shaders are slow[01:18:30] <mirv_> do you mean something like cel shading, or vertex output attributes?[01:23:23] <ColaEuphoria> mirv_, just simple flat shading. smooth shading is easy because you can store a single normal per vertex, but with flat shading you would either need to split every single face or generate a normal using the cross product in a geometry shader[01:24:57] <mirv_> you could store a single normal per triangle[01:25:26] <mirv_> or use the flat qualifier when passing normal values to the fragment stage[01:25:43] <ColaEuphoria> mirv_, but everything is done per-vertex[01:25:49] <ColaEuphoria> in the vertex shader[01:26:13] <mirv_> yes[01:26:47] <mirv_> I mean, you could generate normals on the fly I suppose[01:27:51] <mirv_> geometry shader could do that[01:28:11] <mirv_> but honestly, just precalc it from the vertex data and supply in a buffer[01:28:36] <mirv_> modify with whatever modelview transform, and done[01:29:02] <ColaEuphoria> mirv_, you mean like I could send a per-vertex index into a uniform buffer to get the normal?[01:30:12] <mirv_> whatever mesh you have, the vertices will form a triangle - so on the cpu side, just calculate it all there[01:30:22] <mirv_> and make it part of the mesh format[01:30:56] <mirv_> if it's per-vertex, you can either make all the normals the same for a particlar face, or use the flat qualifier[01:31:35] <mirv_> no need to index anything in the vertex shader, it's just per-vertex data[01:31:54] <mirv_> or per-vertex attribute data I should say[01:34:15] <mirv_> but by specifying "flat" in the output variable, the rasteriser should generate the same normal for every fragment of a triangle[01:34:31] <mirv_> giving you flat shading[01:39:11] <ColaEuphoria> mirv_, oh, so the 'flat' qualifier causes the rasterizer to average the variable for the face?[01:42:37] <mirv_> well, takes it from the provoking vertex I believe[01:43:11] <mirv_> if you had, say a cube, you would need to duplicate a few vertices to change their normals[01:43:27] <mirv_> but a sphere, it'd look about the same without doing that[01:45:50] <mirv_> and with apologies, but I gotta get some sleep now[01:46:11] <mirv_> google around for "vulkan flat shading", or look it up in the vulkan spec, it has more info there about it[01:50:19] *** mpersano <mpersano!~mauro@177.32.193.126> has quit IRC (Quit: leaving)[01:51:20] *** derhass <derhass!~derhass@dslb-188-102-016-172.188.102.pools.vodafone-ip.de> has quit IRC (Ping timeout: 272 seconds)[04:45:01] *** pa <pa!~pa@unaffiliated/pa> has quit IRC (Ping timeout: 258 seconds)[04:46:42] *** wildlander <wildlander!~wildlande@unaffiliated/wildlander> has quit IRC (Ping timeout: 245 seconds)[04:58:34] *** pa <pa!~pa@unaffiliated/pa> has joined ##vulkan[05:05:05] *** slime <slime!~slime73@blk-215-81-93.eastlink.ca> has quit IRC (Quit: This computer has gone to sleep)[06:00:31] *** fling <fling!~fling@fsf/member/fling> has quit IRC (Ping timeout: 246 seconds)[07:21:01] <Ralith> it is unlikely that having split faces will be your performance bottleneck[07:32:07] *** glYoda <glYoda!~kristinam@c-73-25-27-206.hsd1.or.comcast.net> has joined ##vulkan[08:36:04] *** Karyon <Karyon!~Karyon@unaffiliated/karyon> has quit IRC (Ping timeout: 276 seconds)[09:23:23] *** Karyon <Karyon!~Karyon@unaffiliated/karyon> has joined ##vulkan[10:04:23] *** englosh <englosh!5bba255f@pool-95.37.186.91.dynamic.wobline-ip.de> has joined ##vulkan[10:08:57] *** englosh <englosh!5bba255f@pool-95.37.186.91.dynamic.wobline-ip.de> has quit IRC (Remote host closed the connection)[10:36:56] <ColaEuphoria> it is when you split every single face on the meshes by drastically increasing memory usage[11:21:07] *** ratchetfreak <ratchetfreak!c351a8d8@195.81.168.216> has joined ##vulkan[11:45:49] *** RicardasSim <RicardasSim!~RicardasS@78-63-172-44.static.zebra.lt> has quit IRC (Quit: Leaving)[12:05:21] *** Deluxe <Deluxe!~Deluxe@2001:67c:1220:80e:e9:1d2:f14f:e47f> has joined ##vulkan[12:39:05] *** Akari <Akari!~Akari@srsfckn.biz> has quit IRC (Quit: ZNC - https://znc.in)[12:47:04] *** Akari <Akari!~Akari@0x0.st> has joined ##vulkan[13:03:43] *** slime <slime!~slime73@24.215.81.93> has joined ##vulkan[13:04:53] *** ratchetfreak <ratchetfreak!c351a8d8@195.81.168.216> has quit IRC (Remote host closed the connection)[13:07:39] *** ratchetfreak <ratchetfreak!c351a8d8@195.81.168.216> has joined ##vulkan[13:19:27] *** slime <slime!~slime73@24.215.81.93> has quit IRC (Quit: This computer has gone to sleep)[14:01:11] <Illya> what are the advantages of using VK_KHR_imageless_framebuffer?[15:04:08] <[exa]> afaik you can specify the actual images from the renderpass, not in advance[15:16:42] <[exa]> (not sure about the real advantage tho)[16:38:37] *** Deluxe <Deluxe!~Deluxe@2001:67c:1220:80e:e9:1d2:f14f:e47f> has quit IRC (Remote host closed the connection)[16:47:57] *** ImQ009 <ImQ009!~ImQ009@unaffiliated/imq009> has joined ##vulkan[17:46:35] *** ratchetfreak <ratchetfreak!c351a8d8@195.81.168.216> has quit IRC (Remote host closed the connection)[17:51:38] *** ratchetfreak <ratchetfreak!c351a8d8@195.81.168.216> has joined ##vulkan[18:12:58] <Ralith> a little less boilerplate for using the same pass on different framebuffers, I guess[18:16:35] *** ratchetfreak <ratchetfreak!c351a8d8@195.81.168.216> has quit IRC (Ping timeout: 260 seconds)[18:20:22] *** glYoda <glYoda!~kristinam@c-73-25-27-206.hsd1.or.comcast.net> has quit IRC (Quit: glYoda)[18:29:32] *** Deluxe <Deluxe!~Deluxe@212.4.150.151> has joined ##vulkan[18:53:30] *** ratchetfreak <ratchetfreak!~ratchetfr@ptr-82s3g7nyo2zjr8pdfah.18120a2.ip6.access.telenet.be> has joined ##vulkan[19:27:07] *** Salek <Salek!~salek@91-154-70-84.elisa-laajakaista.fi> has quit IRC (Read error: Connection reset by peer)[20:58:14] *** ryp <ryp!ryp@gateway/vpn/privateinternetaccess/ryp> has joined ##vulkan[21:04:39] *** Deluxe <Deluxe!~Deluxe@212.4.150.151> has quit IRC (Remote host closed the connection)[21:06:01] *** Deluxe <Deluxe!~Deluxe@212.4.150.151> has joined ##vulkan[21:41:34] *** plasticle <plasticle!~plasticle@cpe-107-15-166-189.nc.res.rr.com> has quit IRC (Quit: WeeChat 2.3)[21:48:53] *** plasticle <plasticle!~plasticle@cpe-107-15-166-189.nc.res.rr.com> has joined ##vulkan[22:03:17] *** Deluxe <Deluxe!~Deluxe@212.4.150.151> has quit IRC (Remote host closed the connection)[22:05:46] *** ImQ009 <ImQ009!~ImQ009@unaffiliated/imq009> has quit IRC (Quit: Leaving)[22:19:47] *** felipealmeida <felipealmeida!~user@191.191.10.198> has joined ##vulkan[22:21:54] *** pa <pa!~pa@unaffiliated/pa> has quit IRC (Ping timeout: 248 seconds)[22:24:09] *** pa <pa!~pa@unaffiliated/pa> has joined ##vulkan[22:28:23] *** slime <slime!~slime73@blk-215-81-93.eastlink.ca> has joined ##vulkan[22:37:56] *** fling <fling!~fling@fsf/member/fling> has joined ##vulkan[23:04:46] *** Salek <Salek!~salek@91-154-70-84.elisa-laajakaista.fi> has joined ##vulkan[23:19:50] *** LunarJetman <LunarJetman!LunarJetma@94.14.216.120> has joined ##vulkan[23:25:09] <Ralith> a little less boilerplate for using the same pass on different framebuffers, I guess