IRC logs of Libera.Chat #BZFlag for Sunday, 2022-10-09

*** SpringTank <SpringTank!~quassel@75.28.20.205> has joined #bzflag00:38
*** Cobra_Fast is now away: vacant01:36
*** Cobra_Fast is back01:36
*** Cobra_Fast is now away: vacant02:11
*** Cobra_Fast is back02:11
*** Cobra_Fast is now away: vacant02:47
*** Cobra_Fast is back02:47
BZNotifydocs/bzdb @ bzflag.org: allejo pushed 1 commit (https://github.com/BZFlag-Dev/bzflag.org/compare/dbba8a45fd87...e51e248cb798):03:09
BZNotifydocs/bzdb @ bzflag.org: allejo e51e24: Last batch of BZDB variable descriptions (https://github.com/BZFlag-Dev/bzflag.org/commit/e51e248cb798c807ef188d976c9f7fde24722140)03:09
BZNotifybzflag.org: allejo synchronized pull request #42 "Document BZDB settings" (https://github.com/BZFlag-Dev/bzflag.org/pull/42)03:09
*** I_Died_Once <I_Died_Once!~I_Died_On@c-73-184-170-223.hsd1.ga.comcast.net> has quit IRC (Ping timeout: 252 seconds)03:20
*** FastLizard4 is back04:36
BZNotifydocs/bzdb @ bzflag.org: allejo pushed 2 commits (https://github.com/BZFlag-Dev/bzflag.org/compare/e51e248cb798...44a1888554e0):05:09
BZNotifydocs/bzdb @ bzflag.org: allejo 5bdf34: Add JSON schema for the BZDB Settings file (https://github.com/BZFlag-Dev/bzflag.org/commit/5bdf34320072edb7dc11a28b6d642b6f0779d42f)05:09
BZNotifydocs/bzdb @ bzflag.org: allejo 44a188: Make sure BZDB Settings match our schema for it (https://github.com/BZFlag-Dev/bzflag.org/commit/44a1888554e08e630dc4db27c4e44e1dfc7eb804)05:09
BZNotifybzflag.org: allejo synchronized pull request #42 "Document BZDB settings" (https://github.com/BZFlag-Dev/bzflag.org/pull/42)05:09
*** FastLizard4 is now away: AWAY from keyboard06:14
*** FastLizard4 is now away: IN-GAME - Plate Up!06:14
*** Sgeo <Sgeo!~Sgeo@user/sgeo> has quit IRC (Read error: Connection reset by peer)08:49
*** JunkoAkemi[m] <JunkoAkemi[m]!~junko-ake@2001:470:69fc:105::2:7c57> has quit IRC (Quit: You have been kicked for being idle)09:00
*** FastLizard4 is back10:05
*** FastLizard4 is now away: AWAY from keyboard10:16
*** the_map <the_map!~the_map@user/the-map/x-5158391> has quit IRC (Read error: Connection reset by peer)11:04
*** the-map <the-map!~the_map@user/the-map/x-5158391> has joined #bzflag11:04
*** the-map is now known as the_map11:04
*** I_Died_Once <I_Died_Once!~I_Died_On@c-73-184-170-223.hsd1.ga.comcast.net> has joined #bzflag11:29
*** tupone <tupone!~tupone@gentoo/developer/tupone> has quit IRC (Remote host closed the connection)13:23
*** tupone <tupone!~tupone@gentoo/developer/tupone> has joined #bzflag13:28
*** tupone <tupone!~tupone@gentoo/developer/tupone> has quit IRC (Client Quit)13:28
*** tupone <tupone!~tupone@gentoo/developer/tupone> has joined #bzflag13:28
*** JunkoAkemi[m] <JunkoAkemi[m]!~junko-ake@2001:470:69fc:105::2:7c57> has joined #bzflag14:33
JunkoAkemi[m]What's a good rule of thumb for breaking up large meshes in size/face/vertrex count? (100 world units, 200-300 faces, 500 vertex's?)14:46
blast007not sure about the rendering side, but keep in mind collisions if it's something tanks will be interacting with.  I recall years back how someone had made a metal grate platform out of geometry, and the FPS tanked when driving on (or near) it.  It's possible to have a simplified collision model by having two mesh objects - the complex one with the passable option, and the simplified one just ...15:00
blast007... being transparent.15:01
JunkoAkemi[m]I'm thinking of building some walls/exteriors of a 'building', it's a repeating pattern. Amount of vertex's in a simple section of wall is 32, but all together the amount comes down to easily 1152 vertex points for the "walls" alone of the 'building' alone. (Each section of wall is a single mesh) Could it have similar issues to the metal grate?15:25
JunkoAkemi[m]It's using it solely for geometry in this case though. (Also my bad, each section is three meshes, 16, 8 and 8 in vertex count.)15:27
*** FastLizard4 is back17:25
*** Sgeo <Sgeo!~Sgeo@user/sgeo> has joined #bzflag17:28
*** FastLizard4 is now away: AWAY from keyboard17:40
*** FastLizard4 is now away: GONE - Screen Detached and Disconnected from IRC (I'm probably asleep, at work, or doing something in real life)17:59
AgathaIDK how bz handles it internally, but rendering-wise more meshes is bad. You want to set each separate material (texture) exactly once, then draw all the geometry that uses it, preferably with as few draw calls as possible. Read: group all the triangles that use the same texture into a single mesh. Fewer meshes, more triangles within each.18:38
AgathaAlso, rendering-wise, 1,000-ish vertices / polygons is *nothing*. Even for the ancient hardware bz still supports, anything less than 100,000 in your scene is a rounding error. Any reasonably modern GPU will do 10,000,000 of triangles without breaking stride from 60 Hz.18:38
JunkoAkemi[m]It can render, but playability is my main concern. In Box vs Mesh, Optic wrote that he did some tests and that a model of the starship enterprise was used with varying face counts. (Up to 100k faces.)18:56
JunkoAkemi[m]I think you've addressed my concern though. (I'll try running some tests with larger meshes.)18:59
blast007if you get to the point of needing someone to test it, I do have a variety of hardware I can throw it at19:12
SpringTankthe only thing I can think of that would slow it down with too many meshes is collisoin maths in the CPU20:04
SpringTankso as stated before, have the complex meshes and whatnot as draw only using something like drawinfo, then have invisible meshes/objects that are much simpler for collisions20:05
blast007don't need to use drawinfo for that.  Just a passable mesh, which wouldn't have any collisions.23:57

Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!