#jogamp @ irc.freenode.net - 20130403 05:06:02 (UTC)
20130403 05:06:02 -CatOut- Previous @ http://jogamp.org/log/irc/jogamp_20130402050548.html
20130403 05:06:02 -CatOut- This channel is logged @ http://jogamp.org/log/irc/jogamp_20130403050602.html
20130403 07:22:13 * [Mike] (~Mike]@anon) Quit ()
20130403 09:54:13 * sgothel (~sven@anon) has joined #jogamp
20130403 09:54:13 * ChanServ sets mode +v sgothel
20130403 12:23:55 * limbo (~jonas@anon) Quit (Quit: leaving)
20130403 13:47:00 <rmk0> ugh... jmonkey seems awful
20130403 13:47:13 <rmk0> doesn't matter how many times i try it out, i come away thinking "this is regressive crap"
20130403 13:47:30 <rmk0> turns out they don't even have usable lighting yet
20130403 13:47:58 <rmk0> i'd have to implement deferred shading myself
20130403 13:48:10 <rmk0> and then rewrite their shadow mapping code too
20130403 13:50:56 <rmk0> they're claiming that deferred shading is a feature for 3.1
20130403 13:51:03 <rmk0> yet can't actually seem to bring themselves to release 3.0
20130403 13:51:17 <rmk0> so... they essentially admit the engine won't even be usable for another 1-2 years, at this rate
20130403 13:51:30 <rmk0> bit different from the marketing all over their front page
20130403 14:19:26 <xranby> rmk0: it work out of the box on mobile that is one of its greatest strenght nowdays. sure it would be nice if it have more lighting modes.. new modes are invented every year.. right now many people use raymarching instead of deferred shading on high end CPu's
20130403 14:22:11 <xranby> rmk0: usually you allways have more flexibility if you write your own engine, generic solutions such as jme3 and libgdx are good for a quick start. the question is how much time you want to spend on the engine vs your own code
20130403 14:23:54 <rmk0> the main problem is that a quick start is all it is
20130403 14:24:04 <rmk0> they claim it's a usable engine... i disagree
20130403 14:24:29 <rmk0> at least if you're not writing something that might have looked up-to-date about 5-10 years ago
20130403 14:24:54 <rmk0> reading their discussion on deferred shading was thoroughly depressing
20130403 14:27:11 <xranby> rmk0: i disagree, i think if it looks good mostly depend on if a designer is involved in the finished project or not :D
20130403 14:27:36 <rmk0> hehe
20130403 14:27:54 <xranby> usually designers think about if the correct setting is used for the feeling you want to deliver etc
20130403 14:28:34 <rmk0> i guess what i'm trying to say is demonstrated best by their current lighting system
20130403 14:28:46 <xranby> for horror like environments you may want to use hard cut shadows even if its an more old shadowing technique compared to the most advanced soft shadow technique
20130403 14:30:21 <rmk0> their current lighting systems is a poor implementation of the "old" way of doing things
20130403 14:30:34 <rmk0> they have tickets in their tracker that state that it falls over utterly when adding about eight lights
20130403 14:30:44 <rmk0> those tickets are marked as "difficult" and still ignored to this day
20130403 14:31:11 <rmk0> that's not something i'd expect from an engine that claims to be able to compete with the commercial alternatives
20130403 14:31:36 <rmk0> -s
20130403 14:35:57 <xranby> well 8 is a hardware limit http://www.opengl.org/wiki/FAQ#Why_limit_to_8_lights.3F
20130403 14:36:24 <rmk0> am aware... is the "old" way of doing things
20130403 14:36:30 <rmk0> horribly limited
20130403 14:36:56 <rmk0> they aren't actually using the fixed function pipeline
20130403 14:37:11 <rmk0> they just have a poor implementation of forward rendering
20130403 14:39:40 <rmk0> my actual point was that the engine is claimed to compete with the current commercial engines. but given the way it's implemented, i'd be either using their crappy forward rendering, with the bare minimum number of lights it's capable of using without falling over, and baking static light maps like it's 2001... or implementing my own deferred rendering system (including rewriting their now-incompatible shadow mapping system) and basically having to dump t
20130403 14:40:25 <rmk0> it's probably suitable for little mobile 3D games, but nothing more advanced unless the developer using it is willing to basically reimplement a huge part of it
20130403 14:40:55 <rmk0> so... i'm not bothering, will continue work on my own thing
20130403 14:49:17 <xranby> rmk0: i guess mobile game engines may switch to use deffered lightning when the mobile opengl es 3 hardware is ready for it. right now none of the "professional" mobile game engines use it because the hardware available is too slow using it. so yes if you want to use this techniques you may be better off using your own engine
20130403 14:49:53 <rmk0> yeah, is awful trying to implement it using ES2, as well
20130403 14:58:08 <sgothel> good 'morning' .. guess I missed all of this here oops
20130403 14:58:13 <rmk0> lo
20130403 14:59:09 <sgothel> ofc .. everything 'generic API' has it's restriction, surprised though that jME3, which claims to be shader based, uses FFP lighting
20130403 14:59:27 <sgothel> since lighting is orthogonal to core GL ..
20130403 15:00:13 <sgothel> well, you guys know it, lighting methodology is a science for itself :)
20130403 15:01:08 <xranby> I had to look up raymarching after seeing demoscenepassifists latest 4k it was an mind opener
20130403 15:01:49 <sgothel> so .. pls test the new JPEG decoder .. should be rock solid - thx
20130403 15:02:01 <sgothel> *become* :)
20130403 15:02:06 <rmk0> not familiar with it
20130403 15:02:10 <rmk0> ray marching, that is
20130403 15:04:07 <xranby> http://blog.hvidtfeldts.net/index.php/2011/06/distance-estimated-3d-fractals-part-i/
20130403 15:04:22 <xranby> its a way to do quick raytracing using the fragment shader
20130403 15:05:36 <rmk0> pretty
20130403 15:05:54 <xranby> http://www.youtube.com/watch?v=R-iL0o9aJdM
20130403 15:06:03 <sgothel> AFAIK one of our OpenCL demos has this too .. sure the raytracing thingy in hardware is the next big thing
20130403 15:06:56 <xranby> http://www.youtube.com/watch?v=0w_xEUoK79o
20130403 15:09:14 <sgothel> well, Dominik's demos are great - however not really suitable for shader utilization [today] due to massive inner loops etc - it's for sure somewhat a hack using GPUs for such rendering techniques.
20130403 15:13:36 <xranby> sgothel: i have patched processing now, my fixes got merged. so now it starts correctly on opengl es 2 hardware.
20130403 15:14:03 <xranby> i did not manage to clean up the AWT-EDT it was way to complex
20130403 15:14:07 <xranby> for a quick patch
20130403 15:14:51 <xranby> sgothel: this is what got merged https://github.com/processing/processing/pull/1705
20130403 15:15:38 <sgothel> nice, but I try to avoid looking into the changes in detail, otherwise I would jump into work/discussing it :) - but great stuff!
20130403 15:15:56 <sgothel> just found a hack for OSX CALayer position problem - horrible!
20130403 15:16:08 <sgothel> i.e. setSize +1, setSize -1 -> good :)
20130403 15:17:07 <xranby> is this hack endorsed by apple?
20130403 15:17:36 <xranby> sounds like adding that fix will break it in the future
20130403 15:17:47 <xranby> when apple fixes their code
20130403 15:18:01 <sgothel> I don't want to deal w/ OpenJDK folks in this regard .. but well, you may can publish it - I will add a use case [how to reproduce] in the git commit :)
20130403 15:18:11 <sgothel> not Apple, but OpenJDK!
20130403 15:18:20 <sgothel> read: Oracle :)
20130403 15:18:45 * xranby feels ahppy not owning a ma
20130403 15:18:47 <xranby> mac
20130403 15:19:17 <sgothel> that is a given .. :)
20130403 15:19:35 <sgothel> so guys .. test JPEG .. gogogog :)
20130403 15:20:07 <sgothel> since we replaced AWT ImageIO here .. I already see the complaints that it's vertically flipped :)
20130403 15:20:17 <xranby> :D well
20130403 15:20:53 <xranby> i will go on ski vacation tomorrow, so you say i need to code a jpeg photoalbum now?
20130403 15:21:06 <sgothel> haha :)
20130403 15:21:14 <sgothel> don't break your legs ..
20130403 15:21:32 <sgothel> just meaning .. if you have something using JPG textures .. you may try ..
20130403 15:21:47 <sgothel> sure .. now possible on your beloved headless embedded devices
20130403 15:21:55 <sgothel> Ski ? Where is the snow ?
20130403 15:22:09 <xranby> if you travel up the mountains in sweden there is still now
20130403 15:22:22 <sgothel> ohh .. the new Alps 'eh ? :)
20130403 15:23:31 <xranby> actually i walked in snow yesterday by climbing a small mountain that is only 280m above sea-level
20130403 15:24:00 <sgothel> welcome the new ski tourism :)
20130403 15:24:28 <sgothel> here [south Germany, Swiss, Austria, Italy] I guess they have to use artificial snow :)
20130403 15:27:33 * xranby (~xranby@anon) Quit (Quit: Leaving.)
20130403 18:01:32 * [Mike] (~Mike]@anon) has joined #jogamp
20130404 05:06:15 -CatOut- Continue @ http://jogamp.org/log/irc/jogamp_20130404050615.html