Axis labels are out of control

The one thing about ParaView that drives me a bit crazy: when dealing with expansive datasets, the axis labels are out of control, illegible, and majorly slow down rendering.

This is a problem with every dataset I visualize in ParaView as I deal with expansive geospatial data that requires me to zoom in and out of regions throughout the rendering space. The axis grid simply can’t keep up and tries to render what seems like an absurd amount of ticks/label.

Currently, my solutions are to either:

  1. turn off and keep off the axis grid altogether
  2. use a few scripts that set the axis labels to some increment (e.g. 10 km across data bounds)

Is there a better way to manage this? Would it be possible to implement functionality on the backend that prevents ParaView from trying to render overlapping labels?

Here’s an example of what I frequently experience after zooming in just slightly to my data space:

Please share, if possible, a dataset that reproduce this issue and precise which version of ParaView you are using.

@banesullivan, that is indeed correct. the axes grid has gotten a tad worse over the years. there has been discussion in the background to revamp the label rendering infrastructure to overhaul it and avoid such overlapping labels issues for scenes in general. I’d recommend reporting issues on the issue tracker for now (with info about data bounds in world cooridinates so that the issue is easily reproducible).

BTW you can explicitly specify which labels to label on the axes, if I am not mistaken. That’s a lousy workaround, but there, nonetheless.


@utkarsh.ayachit, you’re absolutely correct that we can set the axis labels explicitly. I suppose I should have mentioned that I have a few macros/pvpython scripts set up to handle general axis label specification. These two functions work to take the name of a data source on the pipeline and sets the axis labels at defined intervals for that datasets bounds. (these are macros I made for PVGeo’s pvmacros python module)

def setAxisLabelsFromBounds(name, num=(10, 10, 5)):
    """Sets the axis labels from a given input data source. Use the num argument
    to control the number of labels along each axis. If num is a scalar, then
    a uniform number of labels is used on each axis.

    Author: Bane Sullivan <>

    Originally implemented for PVGeo <>

        name (str): The string name of the input source on the data pipeline
        num (tuple(int) or int): the number of labels for each axis

        >>> import pvmacros as pvm
        >>> pvm.vis.setAxisLabelsFromBounds('TableToPoints1', num=(5, 10, 2))

    import paraview.simple as pvs
    import paraview.servermanager as sm
    import numpy as np
    # Get the input data
    src = pvs.FindSource(name)
    data = sm.Fetch(src)
    xmin,xmax, ymin,ymax, zmin,zmax = data.GetBounds()
    if not isinstance(num, (tuple, list)):
        num = list(num)
    # Cast as ints if needed
    for i, val in enumerate(num):
        if not isinstance(val, int):
            num[i] = int(val)
    # Calculate ranges for each axis
    xrng = np.linspace(xmin, xmax, num=num[0])
    yrng = np.linspace(ymin, ymax, num=num[1])
    zrng = np.linspace(zmin, zmax, num=num[2])

    # Set the axis labels
    customAxisTicks(xrng, axis=0, uniform=False)
    customAxisTicks(yrng, axis=1, uniform=False)
    customAxisTicks(zrng, axis=2, uniform=False)

def customAxisTicks(rng, axis=0, uniform=False):
    """Use to set custom axis ticks in the render view

        rng (list(float)): A list or tuple of floats for the axis ticks
        axis (int): The axis to set (X=0, Y=1, or Z=2)
        uniform (bool): An optional flag to use the given range on all axii
    from paraview.simple import GetActiveViewOrCreate, RenderAllViews
    # note that third parameter is the step size
    # get the active view
    rv = GetActiveViewOrCreate('RenderView')
    if axis is 0 or uniform:
        rv.AxesGrid.XAxisUseCustomLabels = 1
        rv.AxesGrid.XAxisLabels = rng
    if axis is 1 or uniform:
        rv.AxesGrid.YAxisUseCustomLabels = 1
        rv.AxesGrid.YAxisLabels = rng
    if axis is 2 or uniform:
        rv.AxesGrid.ZAxisUseCustomLabels = 1
        rv.AxesGrid.ZAxisLabels = rng
    return None


@mwestphal, I do not have many datasets that I can share… but I have created a state file of some vtkImageData of Landsat imagery which can be downloaded here:

Note: the vtkImageData has an RGB array of a false color scheme so make sure Map Scalars is off to see the false color image.

and a screenshot of the axis labels:

OK, that bug was written up a month ago here: It is scheduled to be fixed in the 5.7 release, out next spring. Required priority.

1 Like