@GuillaumeFavelier locally on master with a script similar to:

import os.path as op
import mne
data_path = mne.datasets.sample.data_path()
sample_dir = op.join(data_path, 'MEG', 'sample')
subjects_dir = op.join(data_path, 'subjects')
inv = mne.minimum_norm.read_inverse_operator(op.join(
    sample_dir, 'sample_audvis-meg-vol-7-meg-inv.fif'))
    # sample_dir, 'sample_audvis-meg-oct-6-meg-inv.fif'))
evoked = mne.read_evokeds(op.join(sample_dir, 'sample_audvis-ave.fif'))[0]
evoked.apply_baseline((None, 0)).pick_types(meg=True)
stc, res = mne.minimum_norm.apply_inverse(
    evoked, inv, return_residual=True, pick_ori='vector')

initial_time = 0.1
# stc, kwargs = stc, dict(overlay_alpha=0.5, brain_alpha=0.5, vector_alpha=0.)
stc, kwargs = stc.magnitude(), dict(alpha=0.5, surface='white')
kwargs['src'] = inv['src']
meth = stc.plot_3d if isinstance(stc, mne.VolSourceEstimate) else plot
brain = meth(
    subjects_dir=subjects_dir, initial_time=initial_time,
    clim='auto', hemi='both', views=['sag', 'axi', 'cor'],
    view_layout='horizontal', smoothing_steps='nearest', verbose=True,
    time_viewer=False, size=(800, 300), background='k',
    **kwargs)

I see this "flashing" (sorry the window disappears -- I'm talking about the meshes and outlines appearing and disappearing):

And on this branch and latest PyVista master it doesn't seem to be improved much. Do you see the same thing? Can you see if it can be improved?

Actually with this example script it's a bit worse -- the display doesn't update until I click on the window (ignore the "blue brain" in the background -- the window produced by the script above will appear on top of it, flash for a bit, then not display the final image until I click on the window to bring it to focus):

Thanks for reporting, I can reproduce. I will investigate in this PR (already opened).

For me, 0372cd2 improved the situation greatly. In plot_source_estimates(), show() is called last, when everything else is set basically. Could you try it now @larsoner ?

@GuillaumeFavelier I'm +1 for backporting, any reason not to?

I see no problem here 👍

Thanks @GuillaumeFavelier

Nope, not backporting -- the conflict due to the TimeViewer / Brain refactor was ugly

Ouch, my bad. I thought it was straightforward to cherry-pick this.

Actually I think I figured it out: 35662d0