developer meetup on irc 3rd service

July 1, 2012 posted by: joreg

dear coders,

this time of the month again… please join your fellow vvvv developers later today on irc to tell us what you’re working on or let us know whats bugging you.

chat log archive

04 06 12: developer-meetup-on-irc-2nd-service 07 05 12: developer-meetup-on-irc


Comments:

Comments are no longer accepted for this post.

joreg
02.07.2012 - 16:57

…and here is todays log:

[](16:09:20) <iorec__> 1) supporting sprite sheets for Texture
[](16:09:20) <iorec__> 2) test framework
[](16:09:20) <iorec__> 3) depth rendering
[](16:09:20) <iorec__> 4) keyboard/mouse state
[](16:09:20) <iorec__> 5) testing kontrolleur on android 4
[](16:09:43) <iorec__> so it seems we are a small group today..
[](16:09:53) <iorec__> please mino start with 1)
[](16:10:16) <mino_> ok
[](16:10:48) <mino_> i found this: http://www.codeandweb.com/texturepacker
[](16:11:46) <mino_> from new way to GPU rendering with actionscript
[](16:11:59) <mino_> called "Starling".
[](16:12:31) <mino_> it's only package of multi-images and xml.
<http://16:12:39) --> sebl_ (~quassel@kvm2.johannwimmer.at> hat ##vvvv-meetup betreten
<http://16:12:44) <-- sebl_ (~quassel@kvm2.johannwimmer.at> hat das Netzwerk verlassen (Client Quit)
[](16:12:59) <iorec__> never heard of such a thing..
<http://16:13:22) --> sebl_ (~quassel@kvm2.johannwimmer.at> hat ##vvvv-meetup betreten
[](16:13:36) <mino_> it will save memory and faster than spreads.
[](16:13:36) <iorec__> so you give it a bunch of images and it puts those on a large texture atlas and returns an .xml which describes the layout?
[](16:14:59) <mino_> TexturePacker did that.
[](16:15:39) <iorec__> ok. so it should be easy to write a plugin that decodes that info and returns texture + texturetransforms?
[](16:16:00) <mino_> right :)
[](16:16:16) <iorec__> sounds good. you're welcome.
[](16:16:45) <iorec__> if noone objects 3...2...1..this goes to mino
[](16:16:52) <mino_> ok!
[](16:17:23) <iorec__> then i continue with 4) keyboard/mouse state
[](16:17:29) <iorec__> (also for mino)
[](16:17:59) <iorec__> remember you asked once for your gui plugin how to enter mouse/keyboard info to the plugin
[](16:18:17) <mino_> yes
[](16:18:35) <iorec__> latest alpha has new plugins/modules: Keyboard (System Global State), Mouse (System Global State)...
[](16:19:13) <iorec__> so you can now use those single connections to put keyboard/mouse states to a plugin
[](16:20:11) <iorec__> and create an ISpread<KeyState> and ISpread<MouseState>
[](16:21:30) <iorec__> (oh seems probably not yet in latest alpha..but will be later today..)
[](16:21:53) <mino_> sounds cool, then we can get set of properties of key/mouse, right?
[](16:22:11) <iorec__> right
[](16:22:32) <iorec__> TypeWriter (String) is using it already (in addonpack with sources)
[](16:23:04) <m4d> nice :)
[](16:23:16) <m4d> also hello all. just dropping by..
[](16:23:30) <iorec__> next: 3) depthrendering
[](16:23:34) <iorec__> good you're here m4d
[](16:23:56) <m4d> just read about INTZ in the  log :)
[](16:23:58) <iorec__> latest dx9ex - alpha has depthrendering of complete scenes available
[](16:24:08) <m4d> mighty cool!
[](16:24:18) <m4d> big thanks :)
[](16:24:19) <iorec__> just set a DX9Texture to Format: INTZ
[](16:24:54) <m4d> *heading over to  alphasection*
[](16:25:47) <iorec__> drawbacks: no antialiasing (as i understand it could work on ATI cards with RESZ, but not testet), didn't manage to move it via texture-sharing, no idea to realize depth/cube rendering
[](16:25:57) <iorec__> anyone with an ATI card to test?
[](16:27:01) <m4d> i have a firegl in my   laptop
[](16:27:07) <mrvux> @iorec, can you rebind it to another renderer?
[](16:27:16) <m4d> but it's really not latest gen
[](16:27:51) <m4d> +1000 for rebinding
[](16:28:13) <m4d> ==traditional deferred  rendering   in vvvv dx9
[](16:28:21) <iorec__> not sure what you mean..you should be able to use it as an ordinary texture
[](16:28:26) <m4d> (sorry, my space key  is broken)
[](16:29:17) <m4d> we're talking  about using that depthbuffer for early z rejection in  a following pass
[](16:29:21) <m4d> (e.g. lighting)
[](16:29:53) <mrvux> or stenciling
[](16:30:16) <iorec__> not familiar with those..does it need more than depth as a texture?
[](16:30:36) <mrvux> no, you need to be able to set that resource back as depth stencil
[](16:30:46) <mrvux> 2-3 main uses
[](16:31:13) <mrvux> 1/ you render your scene, every pixel that passes the depth test does an inc on stencil
[](16:31:41) <mrvux> 2/ when you apply your light pass (deffered), you rebind that stencil as input
[](16:31:51) <mrvux> and you change stencil test
[](16:32:13) <mrvux> so only pixel that got written previously will qualify for light calculation
[](16:32:38) <iorec__> ok, but does it require something that is not there yet in vvvv? or can it all be done in a pixelshader?
[](16:32:52) <mrvux> idea is it doesn't enter pixelshader at all
[](16:33:02) <m4d> yup
[](16:33:58) <m4d> it's mostly about optimization and more or less common practice in gamedevelopment for the last years
[](16:34:12) <mrvux> only thing that miss will be a few new renderstate nodes
[](16:34:47) <iorec__> k. but i am still missing the point where/how you want to "bind" that texture in vvvv terms
[](16:34:53) <mrvux> as renderer input
[](16:35:17) <mrvux> so instead of using a "blank" depth/stencil
[](16:35:24) <mrvux> it uses the already written one
[](16:35:39) <m4d> yeah
[](16:35:40) <iorec__> k, so the renderer would need a stencil-input?
[](16:35:49) <m4d> that would be  quite cool
[](16:35:53) <mrvux> depth and stencil is generally same resource
[](16:36:03) --> iptz|2 (~kvirc@188.134.48.8) hat ##vvvv-meetup betreten
[](16:36:09) <mrvux> just depth will be first 3 comps and stencl last one
[](16:36:21) <iorec__> oui..will see about that.
[](16:36:28) <m4d> nice :)
[](16:36:32) <mrvux> it's also quite cool for post proc
[](16:36:45) <iorec__> and how about antialising? that not so important with depth/stencil?
[](16:36:45) <mrvux> you can generate a stencil mask and save lot of lookup
[](16:36:58) <m4d> saves  looots of gpu cycles
[](16:37:26) <m4d> youmean antialising of the actual depthbuffer?
[](16:37:50) <iorec__> ya. or doesn't that make sense?
[](16:38:03) <m4d> no, i don't think so
[](16:38:16) <mrvux> it does, but not sure that's possible in 9
[](16:39:15) <iorec__> to my understanding this: http://aras-p.info/texts/D3D9GPUHacks.html suggests that it should be possible with RESZ on ati cards
[](16:39:27) <mrvux> and you'd need to be able to sample at subpixel level (for nice edge detection in things like depth of field, or if you do depth-> world reconstruction for lights)
[](16:40:49) <mrvux> if you find easy way to add feel free :)
[](16:41:44) <iptz|2> hey vux, do you know how to transform this INTZ depth for world pos reconstruction?
[](16:41:48) <m4d> i personally would think that it still could be yuite useful without multisampling etc
[](16:42:12) <m4d> yuite=quite jsyk..
[](16:42:16) <iptz|2> i had some problems.. it kinda needs these near/far positions from camera/projection
[](16:43:13) <iptz|2> helloes to everyone btw
[](16:43:13) <iorec> suggestion: move deph-reconstruction to ##vvvv
[](16:43:23) <iptz|2> ok
[](16:43:28) <mrvux> yep, will move that after
[](16:43:28) <m4d> heyunc
[](16:43:48) <iorec> let me continue with 2) test framework
[](16:44:06) <iorec> just wanted to make sure mrvux, woei and all have seen: [/forum-alpha/automated-help-patch-tests](/forum-alpha/automated-help-patch-tests)
[](16:44:32) <iorec> would be great to receive fixing pullrequests for those
[](16:44:50) <iorec> this obviously only did a quick check on the helppatches
[](16:45:19) <iorec> a more general patch testing framework is underway..
<http://16:45:53) --> mrvux_ (~IceChat77@AAnnecy-158-1-29-65.w90-42.abo.wanadoo.fr> hat ##vvvv-meetup betreten
[](16:46:02) <mrvux_> will check those
[](16:46:18) <mrvux_> on latest alphas i noticed very often i have some red nodes first frame
[](16:47:12) <iorec__> ja? if those are not cool and you could pinpoint two revisions that caused that change for you we'd be happy to check.
<http://16:48:18) <-- mrvux (~IceChat77@AAnnecy-158-1-23-44.w90-42.abo.wanadoo.fr> hat das Netzwerk verlassen (Ping timeout: 265 seconds)
[](16:49:11) <iorec__> oui, which brings us to the last point for today: 5) testing kontrolleur on android 4
[](16:49:25) <mrvux_> ok, seems mostly some renaming
[](16:49:33) <mrvux_> just double checked
[](16:49:50) <iorec__> just wanted to check if someone has an android 4 device and would be willing to give my latest fixes a try, also see: [forum](forum)
[](16:50:03) <iorec__> (@vux: yep)
[](16:50:26) <mrvux_> think i'm on 3.2
[](16:50:36) <m4d> 2.3:7
[](16:50:48) <mrvux_> and from what i remember, app from market was just crashing on start
[](16:51:24) <iorec__> right, that crash is supposed to be fixe.d
[](16:51:35) <mrvux_> k, can check now sec
[](16:51:41) <iorec__> but it is only guesswork since i cannot test
[](16:52:01) <iorec__> (but change is only in source yet..so you'd need to build yourself)
[](16:53:16) <mrvux_> ok
[](16:53:43) <woei> @iorec__ i can give you a device for a while
[](16:53:53) <iorec__> ah, right, that could help
[](16:54:40) <iorec__> well, then i think we can close here for today.
[](16:54:59) <iorec__> thanks everyone for coming by and have a great july.
[](16:55:22) <m4d> aye, same to all devvvvs!
[](16:55:47) <woei> cu
[](16:55:52) <mino_> sayonara!

Contact


Imprint

vvvv - Dießl & Gregor GbR
Oranienstrasse 10
10997 Berlin/Germany
VAT: DE275566955

groupӘvvvv.org

Follow us

Sign up for our Newsletter

Your subscription could not be saved. Please try again.
Your subscription has been successful.