add strf functionality

[paulsowman]

@larsoner @yigu8115 @JD-Zhu I was wondering about addressing the TRF in the to-do list. Where would be the best place to insert audio-timeseries? If we included some extra miscellaneous channels into the .fif that contained audio time series, the data would be most amenable to later analyses. Thoughts?

Here is a sketch of what I was thinking to do -

https://gist.github.com/paulsowman/8f7bf5bcda9b3657b5a0a8502f81e560

[larsoner]

Hmmmm... probably worth doing in the BIDS-ification step, we could add it as a MISC channel or whatever. We already concat MEG and EEG for example:

natural_conversations/bidsify.py

Line 340 in fe16b77

raw_meg.add_channels([raw_eeg])

This could be one more channel to concat at the end