mne-tools / mne-python

MNE: Magnetoencephalography (MEG) and Electroencephalography (EEG) in Python

https://mne.tools

BSD 3-Clause "New" or "Revised" License

2.7k stars 1.31k forks source link

mne.viz.set_3d_view with pyvista backend is not consistent! #8812

Closed dasdiptyajit closed 2 years ago

dasdiptyajit commented 3 years ago

I'm trying to do something similar to : https://mne.tools/stable/auto_examples/visualization/plot_mne_helmet.html#sphx-glr-auto-examples-visualization-plot-mne-helmet-py

Describe the bug

However, I am seeing an anti-aliasing effect during interactive mode in ipython and Jupyter-notebook

Expected results: test2

Getting: test1

Additional information

Platform: Linux-4.15.0-132-generic-x86_64-with-debian-buster-sid Python: 3.7.9 (default, Aug 31 2020, 12:42:55) [GCC 7.3.0] Executable: /home/diptyajit/anaconda3/bin/python CPU: x86_64: 4 cores Memory: 15.6 GB

mne: 0.22.0 numpy: 1.19.2 {blas=mkl_rt, lapack=mkl_rt} scipy: 1.5.2 matplotlib: 3.3.2 {backend=Qt5Agg}

sklearn: 0.23.2 numba: 0.51.2 nibabel: 3.0.0 nilearn: 0.6.1 dipy: 1.1.0 cupy: Not found pandas: 1.2.0 mayavi: 4.7.2 pyvista: 0.27.4 {pyvistaqt=0.2.0, OpenGL 3.3 (Core Profile) Mesa 19.2.8 via Mesa DRI Intel(R) HD Graphics 530 (Skylake GT2)} vtk: 9.0.1 PyQt5: 5.14.1

larsoner commented 3 years ago

Can you write a little code snippet using sample data that replicates this? Or is it enough just to run the example above in Python and then in Jupyter to see different outputs?

dasdiptyajit commented 3 years ago

replicate the issue with Jupyter

code

import os.path as op import mne

sample_path = mne.datasets.sample.data_path() subjects_dir = op.join(sample_path, 'subjects') fname_evoked = op.join(sample_path, 'MEG', 'sample', 'sample_audvis-ave.fif') fname_inv = op.join(sample_path, 'MEG', 'sample', 'sample_audvis-meg-oct-6-meg-inv.fif') fname_trans = op.join(sample_path, 'MEG', 'sample', 'sample_audvis_raw-trans.fif') inv = mne.minimum_norm.read_inverse_operator(fname_inv) evoked = mne.read_evokeds(fname_evoked, baseline=(None, 0), proj=True, verbose=False, condition='Left Auditory') maps = mne.make_field_map(evoked, trans=fname_trans, ch_type='meg', subject='sample', subjects_dir=subjects_dir) time = 0.083 fig = mne.viz.create_3d_figure((256, 256)) mne.viz.plot_alignment( evoked.info, subject='sample', subjects_dir=subjects_dir, fig=fig, trans=fname_trans, meg=False, eeg=False, surfaces='white', coord_frame='mri') evoked.plot_field(maps, time=time, fig=fig, time_label=None) mne.viz.set_3d_view(fig, azimuth=0, elevation=0, focalpoint=(0., 0., 0.), distance=1)

Results: r_test1 r_test2

GuillaumeFavelier commented 3 years ago

This bug is described in https://github.com/mne-tools/mne-python/issues/7599 and is related to VTK9. It has been patched for Brain but can probably occur on other viz that do not use the same algorithm. I think it's your case. I happen to work on porting evoked field map feature to use Brain instead. You can track the progress of this work in https://github.com/mne-tools/mne-python/pull/8749