How to tackle time series with different length ?

It can be noticed that time series dataset with unequal length is used for test in this paper, i.e. JapaneseVowels in classification task.

So, I wonder whether the operations (e.g. make a batch, vectorization) on time series with different length is same as others? 