FIX: Improve channel name rendering in plot_alignment

[aman-coder03]

Reference issue (if any)

Fixes #13772

What does this implement/fix?

The show_channel_names feature added in #13570 had a few rendering issues,

• scale=0.005 was too large, causing labels to appear huge and overlap each other badly
• text color was hardcoded to white (1.0, 1.0, 1.0) which is hard to read against the pale scalp surface
• labels were placed exactly at the sensor position with no offset, causing them to sit on top of the sensor glyphs

this reduces the scale to 0.0015, switches to black text for better contrast, and adds a small outward offset to push each label away from the head surface in the direction away from the centroid

Additional information

tested locally on Windows with the sample dataset using eeg_on_scalp.py. the labels are now readable and reasonably sized. some overlap is still expected with dense electrode layouts (59 EEG channels) but that would require a full label placement algorithm and is out of scope for this fix