What I did today

I happened via serendipity (longer story!) to run into Robert Scoble (https://www.facebook.com/RobertScoble) today. He quickly whipped out his iPhone and…

“This Android app makes me jealous. Video effects. Feels like a new age in video.” – Robert Scoble…

I need all the help I can get to tip the crowd funding campaign… Even with Scoble posting to his FB that just got 5 backers 30 to 35… Not going to make it very likely, but anyone on JGO who does a crowd funding campaign has my $5; RPC being my latest funding… Tomb of Tyrants before that and of course Spine… So yeah… I can use all the help I can get at this point w/ 10 days left. Still haven’t got into the press per se…

gJ0kYC4sFo8

;D

Well, I’ve chucked a few bucks your way :slight_smile: Does look kinda fun, although I can’t quite honestly see the application beyond monkeying about!

Cas :slight_smile:

Thank you! Unfortunately, I’m using Java Graphics2d in order to draw so I don’t know how I would replicate crt without shaders

Thanks Cas, you know I got your back on your next one! ;D Yeah… When doing an impromptu demo with Scoble (he literally pulled out his phone and hit record within 30 seconds of engaging him which even caught me off guard a little) and the last one with Android Central what mainly is shown is scrolling between presets and most of the ones I have loaded up are rather gratuitous in style as to be obvious what is happening. There are plenty of opportunities for subtle image enhancement. One fun one is adding a slight bloom around edges (seen in the sizzle reel video with quick shot of the silhouetted girl & the drapes). What really can’t be shown though easily at least in a quick demo is the preset editor where one can spend minutes or hours experimenting with different combinations to tweak out things to come up with original effects composition. For many folks this will be their first time ever doing something like that. I plan to offer a tutorial series describing basic to advanced effects composition and also discuss the GLSL code that makes it happen. I definitely want to enable sharing of effects presets between users.

For game devs it’s certainly possible to prototype an effects composition chain and understand what it’ll look like before implementing it in ones own game / engine.

I think the biggest professional use case is adding lower thirds in the field and immediately uploading completed content to media sites without further handling from offline editors.

Started learning a bit about ray marching. Sorry, no attenuation. Doing it now actually so it could be on by the time someone clicks on it.
Edit: It was easy. Looking into shadows and bloom now.
Edit: Soft shadows working. Looked at this page: http://iquilezles.org/www/articles/rmshadows/rmshadows.htm
It looks super messy and the code is absolutely disgusting but it works and you can kind of get the idea. Later I’ll clean the scene up, put some other primitives in, get reflections working, add bloom and AO, plus other stuff. I like raymarching!

https://www.shadertoy.com/view/Mlf3Rn

worked on a battle system:

Pretty much sums up my day of slaving over a hot, broken bucket of shaders,

I’ve finally implemented a lighting system so that you can still loop lights in the shader, but still have multi-pass infinite-lighting support. With allot of black magic array lists of finite arrays, it works!

95 balls @ 60fps on my GTX 650ti with the hybrid lighting system.
The older method was about 45fps for 12 balls. :confused:

Composed some audio. http://goharsha.com/sounds/song.wav (7.8 MB). Is fast paced, but I can have a slow version of the music too.

Having trouble staying focused and knowing where to start from this week. But I managed to refactor the code into more manageable sections today, and did some more frivolous background work to help set the scene.

Play date on Venus

Mission to the Moon

[quote=“LiquidNitrogen,post:1049,topic:49634”]
Having trouble staying focused and knowing where to start from this week. But I managed to refactor the code into more manageable sections today, and did some more frivolous background work to help set the scene.

WANT! NOW! Please!

Kev

Can’t wait to try this game out, it looks better each day.

Worked on the turret aiming and basic enemy AI for my Tower Defence game. Please disregard the weird turret rotation bug, I’m working on this. Although I’m not completely sure why it is happening.

Messed around with building a little framework for the Google App Engine Datastore for a Java / WebGL project I’m working on. The main object of the framework is to move all Datastore API calls and imports to my DataObject class - I don’t like seeing shopping lists of imports from multiple frameworks in my classes.

GIMME GIMME GIMME!
Seriously though, that’s so awesome! I can’t wait to try this game!

sigh I still have to wait a year :frowning:

Anyway I updated the code from my software that needed desperate performance upgrades.
I rendered this to a single VBO using a custom FloatBuffer handling class , only one area is modified when the tile is put down and the vbo is updated immediately ,the world object has 5 VBOS , one central view one right , left up down etc each of them update the data offscreen when they know that they are coming next , then when the screen fully moves say onto the right VBO , that becomes the centre vbo and the prior centre one becomes the right. This allows for offscreen data loading lowering the chance of an artefact occurring or stuttering from the generation of the landscape.

Hnnnnng.

  • Jev

@home, uninstalled every JRE on my system. No more JRE/JDK confusion for Eclipse and Google App Engine. Ha!

@work, fell in love with C#.

I’ve been telling you guys!
Edit: Look what I found

http://puu.sh/diHy8/84980424ad.png

Experimented with procedural terrain generation again, now with fractals:

https://screenshooter.blob.core.windows.net/engine4files/copiihqlktngnuvtbkenxhkgycqgxmvbtxgmzrstpdwayywnnzsbnxvvejjfqkoxymnpnjujrlyukkiagdtyutnvclhqhmcfpzblkgaefikbzakuajznrpcezsnbbebe

I wrote this generator in 2 hours, but even after such a short time I am able to say that fractal-based generation is superior to noise-based (especially simplex noise). Main advantages of this random generation method:

  • VERY fast generation
  • possibility to generate less detailed, very similar looking maps on demand
  • only small amount of repeating patterns (probably possible to fix with better random function)
  • easily configurable
  • lazy generation

Generation times:
2048X2028 (seen above), full detail - ~2,3s
8192X8192, full detail - 33s
8192X8192, very high detail (almost the same as full detail, with just one generation layer less) - 23s
8192X8192, high detail (first noticeable quality loss) - 14s
8192X8192, low detail (very noticeable quality loss, but still should look acceptable in the far distance in case of 3d games) - 8s

I will try to improve this generator further, this looks very promising.

If you release the code (the pastebin would work, Gist would be better), we could probably speed that up quite a bit. 2s sounds slow, although I’m forgetting what all is involved with fractal gen.
It’s diamond-square, right?

Looks good though!