I had asked about this as well, but no answer ever came up. Can be quite useful at times.

Both are excellent updates!!! I can't wait to play around with the new version.

If I don't finish the partset by the end of this week I'll just put up what's already there so you guys can play around with the new things.

How did you get it to export? PLEASE PLEASE PLEASE PLEASE PLEASE tell me!

Seems like Burvey already helped you out, but in the next bigger version the quick help should be back to avoid this sort of panic <img src="{SMILIES_PATH}/icon_e_smile.gif" alt=":)" title="Smile">

This is looking better and better every time I see it. You are doing such an amazing job with this Somebody!

Thanks! Still plenty to do, but it's getting somewhere.

Event: Every 0.25 sec simulate m pressing = automatic mirror update (only activ when user is activ).

I get the idea, but it would just strangely flicker if done like this (besides confuse the undo system). I might look into a real-time mirror "modifier", but in the future.

-

So that this isn't just replying to positive feedback here's a little status update.

Colors

Since the HSL effect workaround was a little messy (the pixels came back on extra high or low brightness) I started doing some research and ended up on a thread with a similar problem and a solution provided. The result is a new effect: Adjust HSB - effect-adjust-hsb_t123286

So, hopefully, color mush is a thing of the past.

Selection (Or - A Story of optimization)

So one thing that keeps on changing is the way selected shapes are shown - it started with just the gizmos, then went onto shaking, then somewhat rough outlines, then outlines that were basically achieved by scaling the part so at times the corners would be ugly up to fairly decent ones in the latest published version.

How selection outline basically worked was - a layer above the rest with additive blending - we draw a larger or shifted white copies of the original shape and then draw a black copy of the original shape. In the latest version this was 5 extra shapes for each selected shape - since the background white shape was shifted in a cross formation. In code it looked like this:

If you did a Select All a more complex project the slowdown was tangible - since we basically had x5 of all shapes, each with shader effects runing.

So I went around looking for a solution - R0J0 had a rather nice outline effect, but it was somewhat imited - but with a little work (read: blindly pasting things in from other things and hoping they work - after about 30 versions they did, heh) I had a new version. It takes your shape:

Zeroes the original pixels to black and adds 8 shifted copies around it and makes them white:

Then we just place this on an additive layer and have a really neat selection effect:

I hope this works on all systems, but if it doesn't I can put the previous clumsy version in as a toggleable backup solution. Performance-wise this should be rather good as the GPU does all the heavy lifting.

More to come.

Edit: Here's the effect, if you need it: https://www.dropbox.com/s/vwvgvb2syvpvb ... addon?dl=0

Finally got around to have a look at the SDK and made the easy-to-use c2addon package:

https://www.dropbox.com/s/lc785d06b9hsx ... addon?dl=0

Save and just drag into an open C2 window to install.

Or use the add action and global * 0.1

Hi, R0J0, you said edits are welcome - for my purposes the effect didn't suffice because it didn't have the expanded draw rectangle and had all the extra parameters for colors, size, etc.

So I went ahead a kinda re-purposed it to do a single thing - fill the original image with black and add a 2px outline. This lets use create a neat selection effect if the shapes are placed on an additive layer:

I have but one question - does this seem to be the optimal way to do it?

/////////////////////////////////////////////////////////
// Selection outline effect
//
varying mediump vec2 vTex;
uniform lowp sampler2D samplerFront;

precision lowp float;
uniform lowp float pixelWidth;
uniform lowp float pixelHeight;
uniform lowp float width;

void main(void)
{
	lowp vec4 front = texture2D(samplerFront, vTex);
	
	front = vec4(0.0,0.0,0.0,front.a);

	float dx = pixelWidth*2.0;
	float dy = pixelHeight*2.0;
    float dxdiag = 0.7071*dx;
	float dydiag = 0.7071*dy;
	
    float a0 = texture2D(samplerFront, vTex + vec2(-dxdiag, dydiag)).a;
    float a1 = texture2D(samplerFront, vTex + vec2(dxdiag, -dydiag)).a;
    float a2 = texture2D(samplerFront, vTex + vec2(-dxdiag, -dydiag)).a;
    float a3 = texture2D(samplerFront, vTex + vec2(dxdiag, dydiag)).a;
    float a4 = texture2D(samplerFront, vTex + vec2(-dx, 0.0)).a;
    float a5 = texture2D(samplerFront, vTex + vec2(dx, 0.0)).a;
    float a6 = texture2D(samplerFront, vTex + vec2(0.0, dy)).a;
    float a7 = texture2D(samplerFront, vTex + vec2(0.0, -dy)).a;
	
    float ina=max(max(max(max(max(max(max(a0,a1),a2),a3),a4),a5),a6),a7)-front.a;

    float outa = ina + front.a*(1.0-ina);
    vec3 outrgb = (vec3(1.0,1.0,1.0)*ina + front.rgb*front.a*(1.0-ina));
    gl_FragColor = vec4(outrgb, outa);
}[/code:j34146av]

I did extensive performance testing and despite my perceived "performance optimizations"  (2x less multiplication, etc) it doesn't really change. And expanding the draw rectangle instantly cuts fps in half. Still, it gets teen fps on my modest machine with 1000 objects, that's more than enough for its intended use.

Any feedback is welcome.

R0J0hound - could you try the demo capx in my bug report up there - it's really obvious there.

For what it's worth in your example I get the same behaviour - overeager mipmapping when one value is lower than some threshold.

Uh, why would you need that? Just place your text on the topmost layer?

Ah, well, if the original creator gives the art away it's a-ok.

Tyrian was the bomb back in the day, they should really make a new one for the mobile platforms with all the wackiness of the original.

As for your game - it seems alright, you could probably add some respawning coins or things in risky places that give extra points, but increase the chances of terrible tragedy.

I suspect first you get a guy like this: "Niko Derome – Server Architect. A veteran on designing large-scale game servers, Niko has spent the last 15 years in developing real time and asynchronous systems that serve tens of millions of users at companies like Codeonline, Sulake, and Digital Chocolate." - then that guy tells you what the system that communicates with his servers needs and you go from there. The front end could probably be C2, no problem.

Gotta love these "How do I make game X" threads...

A game like that has a budget with a great deal of zeroes - they spend a good deal of that on ads (funny ones, with quality CGI and voice over work) so it's certain the game will get plenty of exposure and thus plenty of players and thus plenty of revenue from IAPs.

So they can afford to have the servers and the decent programmers and the art folk and everything else needed to keep the game running. And when it's been running for a while this happens (Wikipedia):

"The game has been very successful for Supercell. By April 2013, Supercell had only two games on the App Store, Clash of Clans and Hay Day, which together had grossed $279 million. Daily revenue was at $2.4 million,[12] with 8.5 million daily players, and Forbes projected a revenue of between $800 million and $1 billion by the end of 2013.[13][14] They ultimately earned $892 million (compared to $101 million in 2012).[15]"

Long story short - you probably won't make a game with millions of players that earns BILLIONS on your own, especially if you need to ask these questions. Start sensible, start small.

I was just wondering - can you use assets from Tyrian in something that potentially makes money (with ads or otherwise)?

Beaverlicious - somehow, no matter how much I stare at it, it's not moving