Closed loisfairclough closed 9 months ago
Not directly. I think you should modify the output of the decoder to have now 4 decoders instead of 3, one per formant. To do so, you should then train the new decoder (F4). You would need an annotated dataset to train such decoder.
Hi all, Thanks for this great software. Can the modelling code be adapted to track F4 measurements as well as the first 3 formants?