FIX: Fix grand_average to support BaseSpectrum objects

[tsbinns]

Reference issue (if any)

Fixes #13374

What does this implement/fix?

Allows equalize_channels to support BaseSpectrum objects.

Additional information

Also adds a setter for BaseSpectrum.nave which is also required for grand_average().

[tsbinns]

Small note: can't import BaseSpectrum directly from the time_frequency module, and it's not included in the API reference, but BaseTFR is.

[tsbinns]

This follows the existing tests for checking equalize_channels with Raw, Evoked, CSD, etc...

[tsbinns]

mne-python/mne/channels/tests/test_channels.py

Lines 557 to 561 in f65c6ff

	def test_equalize_channels():
	"""Test equalizing channels and their ordering."""
	# This function only tests the generic functionality of equalize_channels.
	# Additional tests for each instance type are included in the accompanying
	# test suite for each type.

Going over this, I noticed this comment isn't technically true anymore, as equalize_channels supports TFR objects, but there isn't an equivalent test. However, there is one instance where equalize_channels is called on TFR data as part of another test:

mne-python/mne/time_frequency/tests/test_tfr.py

Line 373 in f65c6ff

power_pick, power_drop = mne.equalize_channels([power_pick, power_drop])

Is it worth adding a dedicated test as the comment alludes to?