Fix tfCorrection with numPeriodGrouping

[jusack]

Currently it is not possible to use numPeriodGrouping together with tfCorrection (in both getMeasurements and getMeasurementsFD) because the sampled TF data inside the MDF is for a single period.

We have different options:

1. Just interpolate the TF based on the data inside the MDF and apply it to the grouped periods
2. Do TF correction on each period before period grouping, this might cause jumps in the signal
3. MPIMeasurements: Save more samples into the TF if multiple periods are measured -> not according to MDF standard

[nHackel]

I think option 1 sounds good. Don't we already have most of that functionality with the TransferFunction struct, sampleTF function and so on?

We could extend the rxTransferFunction call to include the period grouping. I think at the moment that might clash with the setters of MDF. We could change those to rxTransferFunction!, but maybe the dispatch just works out fine with an integer or keyword argument .

Additionally or alternatively, we could also give the tf as a new keyword argument getMeasurement(f, ...; tfCorrection = !isnothing(tf), tf = rxTransferFunction(f, ...), ...) and then a user could supply a TF for the requested period grouping