3

I'm trying to write a vispy program that will display an image and superimpose two circles on top; ultimately the information for both of these will come from real-time data steams.

To start with, I'm trying to write something that will just display the circles periodically. I modified the real time signal example from the Vispy gallery and this worked o.k. Since it seems that I want to use scenegraphs in order to combine this with an image, I changed the Canvas to sceneCanvas. Parts of the new class are here; I've tried to include only what I think may be relevant.

class Canvas(scene.SceneCanvas):
    def __init__(self, lsl_inlet):
        scene.SceneCanvas.__init__(self, keys='interactive', size=(800, 600))
        ps = self.pixel_scale
        self.unfreeze()
        self.inlet = lsl_inlet

        # Create vertices
        n = 2
        self.data = np.zeros(n, [('a_position', np.float32, 2),
                            ('a_bg_color', np.float32, 4),
                            ('a_fg_color', np.float32, 4),
                            ('a_size', np.float32, 1)])
        self.data['a_fg_color'] = 255, 0, 255, 1
        self.data['a_bg_color'] = 0, 0, 255, 200
        self.data['a_size'] = np.random.uniform(5*ps, 10*ps, n)
        u_linewidth = 1.0
        u_antialias = 1.0

        self.translate = 5
        self.program = gloo.Program(vert, frag)
        self.view = translate((0, 0, -self.translate))
        
        self.model = np.eye(4, dtype=np.float32)
        self.projection = np.eye(4, dtype=np.float32)

        self.apply_zoom()

        self.program.bind(gloo.VertexBuffer(self.data))
        self.program['u_linewidth'] = u_linewidth
        self.program['u_antialias'] = u_antialias
        self.program['u_model'] = self.model
        self.program['u_view'] = self.view
        self.program['u_size'] = 5 / self.translate

        self.theta = 0
        self.phi = 0

        gloo.set_state('translucent', clear_color='white')

        self.timer = app.Timer(0.5, connect=self.on_timer, start=True)
        self.freeze()
        self.show()

    def read_from_lsl(self):
        sample, time = self.inlet.pull_sample()
        return np.array(sample[:2]), np.array(sample[2:4]), sample[4]

    def on_timer(self, event):
        x, y, diam = self.read_from_lsl()
        self.data['a_position'] = np.array([x,y])
        self.data['a_size'] = diam*np.ones(2)
        self.program.bind(gloo.VertexBuffer(self.data))
        self.program.draw()
        self.update()

What I'm seeing is that the canvas will only update very irregularly and seems to freeze after a little bit (but the image might update if I move the frame). This was working o.k. when I was using an app.Canvas instead of the SceneCanvas. Does anyone see what I might be doing wrong? Is it the view? I noticed a lot of programs used something like central_widget.add_view() but when I tried that setting the 'u_view' for self.program didn't work.

Any ideas?

Community
  • 1
  • 1
jdbrody
  • 533
  • 7
  • 16
  • 1
    Typically when you use the SceneCanvas you shouldn't be creating the low-level `gloo` objects any more or be calling any `.draw()` methods. The SceneCanvas will handle that when you add *Visual* objects to it. Let me know if you need any help on this still (I know it's been a long time). – djhoese Oct 03 '20 at 11:59

0 Answers0