I have to create a 2D rpg game for my project. In nearly every tutorial I read, they used canvas for drawing the Scene and it was possible with canvas to render the Scene.
Is there any way to render the Scene without canvas?
JavaFX has AnimationTimer, which is equivalent of drawing/updating a scene using canvas in Swing. This allows you to handle sprite movement and animations smoothly.
See: https://docs.oracle.com/javase/8/javafx/api/javafx/animation/AnimationTimer.html
Also, here is a good JavaFX game tutorial. It isn't an RPG game tutorial, but you can take a lot of information from the code, how AnimationTimer is used, and the overall structure of a game using JavaFX:
http://gamedevelopment.tutsplus.com/tutorials/introduction-to-javafx-for-game-development--cms-23835
Related
I'm a newer in webvr. I want to add canvas to scene by A-frame to draw or write any thing online
Thanks.
Esraa
Live Demo
I want to switch between 2 complex scenes for an intro animation. Both scenes load from .gltf or .babylon files and contain camera motion.
The issue that I'm experiencing is that when I switch from scene1 to 2, there's a freeze/stutter before the next scene animation begins. Once the scene is rendered the switching of scenes continues as normal. It may be hard to spot but there's definitely difference between the first switch and the consequent ones.
I've tried using scene2.render(); right after loading the scene and it helps a little but it also makes the first scene stutter while it's trying to render both scenes. Is there something that I may be missing about this? I couldn't find an example of an intro with scene switching that works well online. Only camera switch but that's not what I want.
this is because the shaders of the second scene need to be compiled (as they could be different from the ones used in the first scene).
Babylon.js v4 will support parallel shader compilation but this will not be available on all browsers.
The best option could be to force that compilation beforehand:
scene1.meshes.forEach(mesh => {
mesh.material.forceCompilation(mesh);
});
I'm making fractal creator software. So I need one scene per fractal and these scenes need to be "layered" because of zooming rubber band.
I've already tried to write it the "widget" way, so I had one custom widget called "canvas". In my canvas class I overrided paintEvent, and inside this event I rendered current fractal. Everytime when somebody clicked to menu with another fractal, I called update() method and new fractal was rendered. To zooming I used overriding of mouse events and update() of canvas. At the first time I repainted the whole canvas, but it was very very slow. After that I repainted only the part under the rubber band, but still slow when I'd like to select some bigger area and other problems with repainting.
So I've looked for another way to do it. Layers. I've found the QStackedWidget, but I didn't find way how to make visible both of my layers and the top one to be transparent. After that, I've found QGraphicsScene and this seems to be the best way to do it. But I don't know the correct procedure to do it. Below are two procedures I'm thinking about:
Create QGraphicsView
Instead of the widget, the canvas will be QGraphicsScene
I'll override some QGraphicsScene event (but I don't know which one - drawItems() is obsolete and override update() seems wrong to me, but maybe...)
When other fractal will be chosen, I'll repaint canvas by calling update() the same way as in my "widget" solution
In the foreground layer will be zooming rubber band
or:
Create QGraphicsView
Instead of the widget, the canvas will be QGraphicsScene
Every fractal will be the child of QGraphicsItem
When other fractal will be chosen, I'll remove the old one fractal
item and replace it by new one and probably call invalidate()
In the foreground layer will be zooming rubber band - I think, that
it's common behaviour of the QGraphicsScene isn't it?
Is one of my reasonings correct? Do you suggest anything else? Fractals are complicated in the calculations and It's very important to repaint only if it is necessary. Could you help me, please?
Thank you :-)
Edit: "zooming rubber band" explanation:
I'm sorry for my expression "zooming rubber band". It means scale (zoom) the area below the selection made by the rubber band - zooming the same way as in Photoshop CS5 (for example). And I'd like to know what part of the scene is repainted while selecting this way. If there is repainted whole scene, or the part of the scene below selected area, or there is nothing repainted and rubber band selection is done in separate layer.
I hope my explanation helped :-).
In Qt, a QGraphicsScene can be thought of as a world of items, with a QGraphicsView as a window into that world. Therefore, you should be adding items to the QGraphicsScene, based on QGraphicsItem (or QGraphicsObject if you want signals and slots).
In your situation, I'd create a Fractal class that inherits from QGraphicsItem and add that to the scene. Ensure to override the necessary pure virtual functions such as boundingRect and paint.
Do not calculate the fractal code in the paint function. I suggest the Fractal class stores a QPixmap (or QImage if you're drawing at the pixel level) and render the fractal to this. Then perodically, in the paint function, the Fractal class would render the contents of the QPixmap with a call to painter->drawImage or painter->drawPixmap; whichever is relevant in this case.
As for zooming, your Fractal class can then response to being scaled, appropriately changing the rendering on the internal representation.
I am building a node graph in a QGraphicsView and I am currently implementing panning.
I used the following question "how to pan images in QGraphicsView" to start but the panning is limited by the scrollbar range.
I also tried the translate method but it gives the same result. The view is limited to a certain rectangle.
I would like to pan without limits, the graph can becomes quite large and it is useful to be able to work in different area of the scene (one graph here, another graph over there, etc).
If you take a look at this video, at the 3 minute mark you'll see the demonstration panning the screen. The application here is one I developed and although it doesn't show it, the real estate of the board appears limitless when panning.
What I did for this was to create a QGraphicsScene of 32000 x 32000 and start the application with the view at the centre of the QGraphicsScene. The test team spent ages trying to pan to the edge of the graphics scene and everyone gave up before getting there - perhaps the scene could have been smaller!
The scroll bar policies are set to off and translation is done by moving the QGraphicsView via its translate function, passing in the delta of either touch, or mouse movement that is applied in the mouseMoveEvent.
Done this way, you need not worry about exceeding the scroll bar range and there was no problem creating a very large QGraphicsScene as it's just a coordinate space.
I came across the same issue. However, setting the scene to something big and leaving it I do not think is the best option. I have developed a dynamic way of changing the scene size so it lets you move freely. You can find it in this other stack overflow answer.
You want to plot graphs. Try out this Qt library - QCustomPlot , it will save you hours of hard work.
I'm kind of newbie in both OSG and Qt, still I'm trying to make Qt HUD upon my OSG window, what I want is Qt interface elements fixed inside OSG scene, not spinning with the model. The thing is, I need Qt elements INSIDE osg scene, not OSG scene inside Qt window (like in OSGviewerQt example).
What I've got yet is OSGQtWidgets example with --useWidgetImage --fullscreen arguments, which shows fixed Qt controls ontop of OSG Model. The thing is, it creates new (FIXED) camera for qt element ontop of OSG model -- because of that, user cannot spin and move OSG model because camera is not transparent.
So the question is: is there a way to make transparent camera with useable Qt elements in it? Or is there some other way to achieve my goals?
Thank you in advance!
I am unable to try this out for myself first, but you can try setting a property on the Orthogonal "HUD" camera to let events pass through to the regular OSG viewer:
camera->setAllowEventFocus(false);
If you're using a recent version of osgQtWidgets.cpp, you'd want to add this around line 414.