Prominent

remember that Y axis is positive down, so if the "up" vector is positive Y, you're defining things flipped. The reason for this is obvious, so that the view-port of construct's editor lines up with Q3D's, and of course maintaining the fact that everything use right-handed cross products / coordinate systems (which is the de-facto standard in most 3D fields)

Viewports only support full transparent / opaque backgrounds because of how they're rendered. It's not something that can easily be changed due to limitations in three.js/webGL and Construct. Doing this was the "best" choice I could make in terms of speed as well.

Q3D Master has a lot of deprecated features, that have niche uses you normally would need. Always look at the actual available objects first as they're better optimized and require less micro-management from you and are mostly guaranteed to work. I have no idea how you were even trying to pick the Q3DModels with "change" parent as they don't have any internal ID's available to Q3D.

gorgonzola3000

That's weird. I haven't tested anything in 214, but that bug in particular means something fundamentally isn't working. Could be how you installed the plugins, or could just be how you set up the scene, I have no way of telling. You always need an instance of Q3DMaster in the layout to do anything at all, if you're missing it that kind of bug can occur, since Q3D isn't loading any of the necessary components (In this case the ENTIRE three.js library isn't loaded for likely this reason). Do you get that error trying to run examples?

that example is unneeded now, it's included as a reference for people who could be using the older plugin versions. Just use Q3DLight instead, it's easier and better. In the word document I linked before they're explained quite thoroughly.

Ethan, which examples aren't working? If they've been deprecated i'll delete them, but there's no real reason for the deprecated examples to work, they've been superseded by better systems already (e.g. Q3Dlight object)

I'd hardly call the updates I put in small also. You may have bought the plugin recently, but it's come a long way for the initial purchasers, with additions that enable full skeletal animation and physics, which we're by no means small additions. Its not as if I've turned around suddenly and dropped support either, It's a fully functional product as is for it's intended purpose (3D web applications).

three.js is a free library and anyone can write their own plugin that interfaces it with Construct 2, but by no means is this a small undertaking. I've supported the bulk of the features you'd need for a game and then some (even pushing changes into three.js to optimize it for games, and adding major upgrades to oimo.js, the library i used for the physics). I don't think the price is unfair for what's already there, documentation or not. Even though anyone *can* make a similar plugin, and distribute it for free, no one has yet. I've put over a year and a half of development into Q3D, and try to maintain it every few months. I can't work on it weekly...

I'm extremely busy these days, and don't have any time to work on updates. Q3D is immensely complex and construct 2 is constantly upgrading so i can't get it to always work with the latest versions that release bi-weekly, but the SDK is too limited for me to take Q3D much further. I stated after the last large updated I wouldn't be around for a while because of other projects I needed to focus on. I have some builds with minor fixes, but it can't be expected that I'd be able to work on the plugin for years. It's already got a lot of functionality and isn't/was never meant to be a competitor to unity etc. just a supplement to construct 2. Q3D has most of the features you'd need for basic 3D games, but you can't expect it to have all the features of an engine like unity which is made by a huge team of people.

Ethan

I was working on a large "how to" manual this summer, but got caught up with other things and never finished it. Even in it's incomplete state it should be of help to some users, and covers some of the basics.

https://dl.dropboxusercontent.com/u/829 ... 2-4_2.docx

I apologize for it abruptly ending and incompleteness, but I just don't have the time to write 100+ pages more at this time. I had hoped plugin users would themselves make examples and share them/help each other out in my absence.

Minify will break everything. Other than that, using too recent a version of construct may cause bugs as Q3D's current version was only tested with versions coinciding with it's release (around April 22nd 2015), I can't really future proof releases cause construct 2 updates can change things unexpectedly.

Rhed

as long as you're feeding in x/y coordinates that are from Q3D, the distance values you get should be right and in the right coordinate system, although you must realize it's only the planar distance between objects, not accounting for Z. To get 3D distance you must use the Pythagorean theorem extended to 3D, distance=sqrt((x2-x1)^2+(y2-y1)^2+(z2-z1)^2)

—

It'd be helpful if you could pm me a simple .capx demonstrating the problem, so i could troubleshoot.

From what i understand though, physics are still working, its just the rotated collider isn't rotated anymore?

—

I can't really figure out what you mean from those images.

—

you can only use primitives or combinations of primitives (to use combinations you must set up the extra colliders with events). This is for performance reasons.

kmsravindra

you can do this by changing the "model fit"/"model center" properties. By default Q3D centers models, but you can adjust the center or leave it unchanged/default, and also use the model offset parameters.

—

Debug bodies shouldn't affect the physics at all, and in my testing they never did. Are you 100% sure this is whats happening?

frozze

crossy road is very possible, it's essentially 2D gameplay so you could get away without even using 3D physics. The rendering of course is also very easy to attain

that fracturing however is not easily attainable. You can kind of fake it but getting the real effect wouldn't be possible cause the physics only supports simple primitives or built up objects made of simple primitives (box,sphere,cyclinder).

kmsravindra

I'm still too busy to work on Q3D atm.

Polygallon

You can fake cell shading somewhat, by having a larger shell mesh with reversed normals/backface rendering. The shading itself wont be cell but you'll get the "toon" outline, and could play with the specularity/ambient settings to obtain some kind of toonish look.

Prominent

yea, you need the custom depth material for it to work, but as i've said the shader loader is incomplete and doesn't allow for these kinds of adjustments yet (you can only affect the main material, so some things can't be changed yet). This is why i don't recommend using it. Even though you've reverse engineered it, it'll probably be changed in a future update so you'll have issues when migrating if you're using the current qfx loader.

Also, systems like custom depth material which give low level access are a bit iffy in three.js itself at the moment if i recall. I need to investigate more.