luca82 wrote at 2013-04-09 11:56:

objo wrote at 2013-04-15 12:16:

luca82 wrote at 2013-04-15 14:08:

cedrelo wrote at 2013-04-16 08:35:

luca82 wrote at 2013-04-16 12:53:

BogusException wrote at 2014-07-02 20:23:

objo wrote at 2014-07-11 21:52:

ejleigh wrote at 2012-06-12 03:28:

I have a terrain model with objects on top and I'm trying to find a good way to prevent the user from rotaing the scene such that he can look underneath the terrrain model. It seemed like the best way would be to derive a new class from CameraController and override the OnManipulationDelta method, but unfortunately you cannot set the camera contoller in the HelixViewPort3D. At the moment I'm setting the RotationSensitivity to a low value when the camera Up vector approaches level, but that only kindof works. It's not ideal. Is there a better way?

objo wrote at 2012-06-15 06:38:

The CameraController is created in the control template of the HelixViewport3D, so you have to override the control template. http://wpftutorial.net/templates.html

Or, use the Camera's Position and LookDirection coerce value callbacks, and enforce the Z values to to be positive and negative, respectively.

http://msdn.microsoft.com/en-us/library/ms754209.aspx
http://msdn.microsoft.com/en-us/library/system.windows.coercevaluecallback.aspx

I have not tried this, but I think it should work.

lightxx wrote at 2013-02-26 10:11:

objo wrote at 2013-03-13 07:12:

jst7 wrote at 2012-03-27 12:57:

Hey,

I'm succesfully catching the doubleclick event from the helixview BUT this event is not triggered when double clicking the middlemousebutton.

Any thoughts on what i could be missing?

p.s. same story when overriding the previewmouseevent!

objo wrote at 2012-04-05 01:35:

I guess this behaviour is caused by the ResetCameraGesture property being set to "MiddleDoubleClick". Try to change it to "None" or some other MouseAction!

jst7 wrote at 2012-04-05 10:05:

No results yet, but thanks for the reply

Elemental wrote at 2012-02-19 11:16:

Hello,

the method CameraController.AddRotateForce(double dx, double dy) always works as CameraMode.Inspect.

In my case, I have CameraMode set to WalkAround, which works fine for mouse controls, but keyboard control is wrong in that case.

I changed the method to

        public void AddRotateForce(double dx, double dy)
        {
            if (!this.IsRotationEnabled)
            {
                return;
            }

            this.PushCameraSetting();

            if (this.CameraMode == CameraMode.WalkAround)
            {
                this.rotationPoint3D = this.CameraPosition;
            }
            else
            {
                this.rotationPoint3D = this.CameraTarget;
            }
            this.rotationPosition = new Point(this.ActualWidth / 2, this.ActualHeight / 2);
            this.rotationSpeed.X += dx * 40;
            this.rotationSpeed.Y += dy * 40;
        }

which does kind of work, but rotating left/right via keyboard is inverted compared to mouse control.

What is the reason for the inverted controls?

Best Regards

Bernd

btw: Just downloaded the latest source code and changing the CameraMode in CameraControlDemo does not work anymore.

objo wrote at 2012-02-21 20:59:

thanks for the feedback! I have only used the "Inspect" mode recently, it seems like the "Walkaround" mode is not working properly. I added it as issue 9949!

Elemental wrote at 2012-02-22 08:08:

Thank you.

Elemental wrote at 2011-12-12 08:10:

I am running the CameraControlDemo and have activated the FPS counter. The FPS stays around 55FPS when I do nothing. When I rotate via mouse, it rises to about 110-120FPS. When I rotate via keyboard, it goes down to about 40FPS.

My question is:

Why do FPS rise when rotating via mouse, but go down when rotating via keyboard?

Best Regards

Bernd

objo wrote at 2011-12-12 20:40:

hi Bernd, the fps counter is counting frames by the CompositionTarget.Rendering event, which should fire at around 60Hz when the model is not too big. I think the mouse events have higher frequency and will force more frequent updates of the model, but I have not investigated how this works. I don't know why it goes down when using the keyboard, could it have something to do with the keyboard repetition rate?

http://msdn.microsoft.com/en-us/library/system.windows.media.compositiontarget.rendering.aspx 

Elemental wrote at 2011-12-13 08:33:

The background of my question is that FPS calculation seems to be wrong for larger objects.

When I have objects with like 200 000 triangles, the FPS when rotating via keyboard is like 5 FPS.

The FPS when rotating via mouse is like 80 FPS. The 80 FPS cannot be true, because it does not look smooth anymore.

I was just wondering why there is this difference in FPS calculation, but it is not really a problem...

Best Regards

Bernd

objo wrote at 2011-12-13 18:42:

Right, using the CompositionTarget.Rendering event is probably not an accurate method of calculate the frame rate.

Let us know if you find a better way to achieve this! (is there some other event in the WPF3D render loop that could be used?)

lioneloddo wrote at 2014-03-30 14:07:

objo wrote at 2014-04-29 10:37:

sandualbu wrote at 2011-12-15 22:36:

This is a really cool project, i like it. I'm thinking about integration with kinect SDK.  Wouldn't it be nice to zoom in/out and rotate objects using only hands movement?

objo wrote at 2011-12-19 09:16:

Yes, a demo using the Kinect SDK would be very interesting! Are you only thinking manipulating objects or also manipulating the camera? Which gestures could be used? Do you have links to some good kinect demos of 3D interaction?

sandualbu wrote at 2011-12-19 14:22:

i'm thinking about both..camera and object rotation... you can take a look at kinesis project. it performs zooming/scrolling and chart rotation (not on 3d objects, of course, but the idea can be developed more). Kinesis works with menus, but something which combine several movements can be built for manipulating 3d objects.

you can perform the rotatation operation as a real time rendering, depending on distance / elevation between hands, and not composed from multiple oprations (up/down/right/left), like kinesis is doing.

The most significant aspect i can think now, is to make kinect interaction as a kind of plugin, and not make helix dependent of it.

objo wrote at 2011-12-19 14:35:

Have a look at the SpaceNavigator implementation. Maybe the Kinect implementation can be done in a similar way (I don't think it should be handled as mouse/touch events).

I agree, a Kinect implementation should be in a separate assembly - suggesting the name HelixToolkit.Kinect.

I'll add the kinect sensor to my xmas wish list, so I can test this :)

sandualbu wrote at 2011-12-19 14:49:

is too hard what the space navigator is doing...I mean he wants to construct something....i think you want just to present 3dobjects, not build them with hands. I'm thinking more like what the guy is doing when presenting the universe with kinect (http://www.youtube.com/watch?v=GMZcz_tlGqw)

objo wrote at 2011-12-21 21:50:

Did you see
http://channel9.msdn.com/coding4fun/kinect/Kinect--3D--Fusion4D
http://www.interlab.pcs.poli.usp.br/fusion4d/

sandualbu wrote at 2011-12-22 07:49:

thta's just amazing. The only disadvantage is the red-cyan glasses. But the gestures used are pretty much like what I was thinking about your project. Of course, making the application render on 120hz display and a pair of active glasses will improve a lot the quality of image.

MrNikito wrote at 2013-09-12 17:33:

objo wrote at 2013-09-29 09:07:

jeeva_sjce wrote at 2013-09-03 06:12:

objo wrote at 2013-09-05 07:56:

jeeva_sjce wrote at 2013-09-06 11:50:

jeeva_sjce wrote at 2013-09-30 10:20: