Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
add in-place
mt_pgram!
, and multitaper spectrogram, multitaper csd,…
… multitaper coherence (#401) * Initial multitaper spectrogram code ported from a private Beacon repo Co-authored-by: Alex Arslan <[email protected]> Co-authored-by: Kalyan Palepu <[email protected]> Co-authored-by: kolia <[email protected]> * fix size of `tmp`, rename fields, return spectrogram object * format * port WIP coherence and cross spectral code Co-authored-by: Alex Arslan <[email protected]> Co-authored-by: Kalyan Palepu <[email protected]> Co-authored-by: kolia <[email protected]> wip wip wip wip add WIP cross-spectra code so I don't lose it Co-authored-by: Alex Arslan <[email protected]> Co-authored-by: Kalyan Palepu <[email protected]> Co-authored-by: kolia <[email protected]> fix bug, change syntax for 1.0 format wip wip wip fix Rename `sample_rate` -> `fs` fixes * transfer tests Co-authored-by: Alex Arslan <[email protected]> Co-authored-by: Kalyan Palepu <[email protected]> Co-authored-by: kolia <[email protected]> * add some MATLAB reference tests Co-authored-by: mich11 <[email protected]> swap order to match Onda wip align better with existing api fix test Fix MATLAB reference test Co-authored-by: mich11 <[email protected]> * fix csd bugs, add tests fix bugs rename add coherence reference test * remove `d`s since these aren't normalized as densities * whoops, rm extraneous file * fix typo * infer `MTSpectrogramConfig` eltype from `MTConfig` eltype * rm `n_freq_bins`, `n_time_points`, add some checks to the MTSpectrogramConfig constructor * try to fix bugs... * fix find-and-replace bug * reformat * test magically works now (?) * handle `n_samples < samples_per_window` * tweak `coherence_from_cs!` * improve test coverage * tweak `MTSpectrogramConfig` API, add `mt_spectrogram` method, improve docstrings * coherence should be 1 on the diagonal * allow and test mixed-precision operations * allocate less, use aligned memory * make `demean=false` the default * don't export `allocate_output` * don't shadow `DSP.freq` * don't average over frequencies in the coherence * Update src/periodograms.jl * add note crediting MNE-python * change names, allow taper weight configuration * update docstrings * adjust docstring, make loadable on 1.0 * fixes * add test and note that two-sided FFTs not supported in CSD * add tests * argh Julia 1.0 * increase codecov * fix very broken tests, make them more robust * add docs, cleanup docstrings * rm old keyword argument from docstring * rm old comment * should be slightly faster to do the `real` inside rather than outside Co-authored-by: Alex Arslan <[email protected]> Co-authored-by: Kalyan Palepu <[email protected]> Co-authored-by: kolia <[email protected]> Co-authored-by: mich11 <[email protected]>
- Loading branch information