Please read the following few lines which explain the structure of the provided phonetics data (Pouplier et al., 2014):


The provided data can be analysed with the proposed method.
The data file phonetics_data.Rdata contains a data.table object in which each row corresponds to one measurement point. 
It can be used as the curve_info object which has to be specified in the main file FLMM_sparse.R.

The provided phonetics data contain the index curves of the nine speakers repeating the 16 target words each five times. Note that due to recording errors, there are some irregularities:

- speaker 1 is missing all curves of target word 13 and target  word 15 (-> 70 curves in total)

- speaker 2 is missing the fifth replication of target word 5 and the fifth replication of target word 14 (-> 78 curves in total)

- speaker 6 is missing the fifth replication of target word 12 (-> 79 curves in total)


For all other subjects, 80 curves are contained in the data table.
Thus, in total there are 707 curves.


A subset of the phonetics data can be found in the same format in file phonetics_data_subset.Rdata containing only the index curves of the first two speakers repeating the first two words.



Covariates (dummy-coded):

- covariate.1 corresponds to the order of the consonants (0: /s#sh/, 1: /sh#s/)

- covariate.2 corresponds to the stress of the final target syllable (0: Strong, 1: Weak)

- covariate.3 corresponds to the stress of the initial target syllable (0: Strong, 1: Weak)

- covariate.4 corresponds to the vowel context (0: ia, 1: ai)


All other columns of curve_info are explained in file FLMM_sparse.R