binyan

Tried with Firefox Android, works for me. Maybe the control (or the hint) is a bit confusing. When the hint says that you should move your device up to jump it means the following: 1. Move it quickly (bump it) 2. Move it UP on X or Y axis (according to the orientation), while the axes are as described on the following picture Unfortunately it's not an open source project (at least not yet), but if you're interested in something particular I'll be glad to share the details.

Thanks for the feedback, I certainly would think about splitting the game to levels by increasing difficulty. Hardly believe that the error is due to bad loading sequences, I'm starting the game only after all the resources have been loaded. But I'll take a look anyway, thanks for the report. As for the web audio stack, yeah, I'm using the one of Babylon.js and it's great! First I was using howler.js but it had some issues with the latest versions of Chrome so I looked for something else. Fortunately it happened exactly at the same time Web Audio stack was added to Babylon.js and I took it It has a huge advantage of integration with the framework and easy binding to mesh objects so it's perfect for me. If eventually I'll get to bone animations they certainly will be there

Could you elaborate? Which errors exactly? Didn't test it in Edge but it worked in Chrome for sure. Both mobile and desktop, with sounds. Have you tried to reload? Ctrl+F5?

Hi guys, I'd like to present you my project, which I've been working on during the last year. The development is still in progress, so it doesn't have fancy interface and not very optimized, but it's a nice POC of Babylon.js framework. The game supports desktop (mouse + keyboard) and mobile (gyroscope and accelerometer) controls. Share your thoughts, any inputs and comments are highly appreciated.

Well, thanks for putting an effort I'm not sure if it's the same scaling as in BABYLON, but if I set a hardware scaling (through BABYLON) to 0.5 it slows my scene down and if I set it to 2 it speeds the scene up. It actually does make sense... When you set hardware scaling to 2 you reduce the internal canvas buffer size (and due to this the picture quality becomes worse). So high value of scaling should speed up the rendering. Am I missing something?

No, I'm just saying that setting hardware scaling to 1 solves the problem. But the problem still exists (without scaling).

Not sure I understood you well... You say the lower the scaling value the better performance. I'm quite sure of opposite... The higher scaling value the better performance. Say the scaling value is 5, so the image looks like crap, but the performance is awesome. Checked it by myself.

I have only one to check But I checked on two browsers - native and Chrome. Breaks on both of them. First I thought it may be an engine.resize() function (that fixes it up) which called from within engine.setHardwareScaling, but when I tried to do it instead of scaling it didn't work. So somehow the hardware scaling is related...

How do I know my phone GPU resolution?

Interesting thing... If I use a pure 100% x 100% canvas on my LG G2, I get like 7-8 fps. But if along with 100% x 100% canvas I use engine.setHardwareScaling(1), I get a significant boost up to 34-35 fps. Anyone has an idea why?

I have the following CSS for canvas: #canvas { position: absolute; width: 100%; height: 100%; top: 0; margin-bottom: 70px;}I.e. it should appear in full screen. And it does, until I enable the debug layer. Once I've enabled the debug layer, the things are messed up. On desktop everything works as usual, but on mobile the 100% becomes like 30%. I.e. canvas (and all the related controls, background, etc.) that should take 100% of the screen takes only one corner (left&topmost). Upd: If I set engine.setHardwareScaling(1) everything works ok

You mean instead of scaling with CSS?

Hey guys, To boost a performance on mobile devices with Full HD screen resolution I'm doing the following thing: 1. Set canvas size like half of full screen 2. Scale the canvas with CSS As a result, the GPU has to do less work, and the scene runs smoothly. Here is the example css: #canvas { position: absolute; width: 40%; height: 40%; -webkit-transform: scale3d(2.5, 2.5, 1.0); transform: scale3d(2.5, 2.5, 1.0); -webkit-transform-origin: 0 0 0; transform-origin: 0 0 0; top: 0; margin-bottom: 70px; touch-action: none; -ms-touch-action: none; z-index: 0;}Well, this approach worked until I started dealing with picking collisions. The picking mechanism doesn't take in account the scaling. Therefore the pick result is completely wrong. If you run this code without scaling it would work as expected (if you pick sphere, it would write in the console "sphere1"): var createScene = function () { // This creates a basic Babylon Scene object (non-mesh) var scene = new BABYLON.Scene(engine); // Our built-in 'sphere' shape. Params: name, subdivs, size, scene var sphere = BABYLON.Mesh.CreateSphere("sphere1", 16, 3, scene, false); sphere.rotationQuaternion = BABYLON.Quaternion.RotationYawPitchRoll(Math.PI / 2, -Math.PI/2, 0); var camera = new BABYLON.FreeCamera("camera", new BABYLON.Vector3(0, 50, -10), scene); var light = new BABYLON.DirectionalLight("dir1", new BABYLON.Vector3(0, -1, 0), scene); light.position = new BABYLON.Vector3(0, 30, 0); sphere.position = new BABYLON.Vector3(0,1.5,0); camera.target = sphere.position; camera.setTarget(sphere.position); camera.attachControl(canvas, true); scene.onPointerDown = function (evt, pickResult) { console.log("Picked " + pickResult.pickedMesh.name + " in x:" + pickResult.pickedPoint.x + " y:" + pickResult.pickedPoint.y + " z:" + pickResult.pickedPoint.z); } return scene;};But if you run it with the above css, it wouldn't work, i.e. picking sphere would give you a null result. Anyone has any idea how to deal with this?

So maybe it worth to add a precise computation as an option? Just like in the case of the mesh intersection? Also, it maybe useful to let the user define the minimal boundaries (thickness) of the bounding box. I'm not sure this is a problem (maybe it should be this way) but currently there is no way you can get a plane & point intersection.

Have you tried to see the call stack of the warning? Where it comes from?