I have a set of recordings of speakers of two different languages pronouncing certain sounds. Specifically, I have recordings of 6 speakers of language A, and 6 speakers of language B, 3 male and 3 female speakers for each. My goal is to obtain perceptual difference scores for the second element of all possible LangA-LangB sound pairs. I have used a Praat script to create stimuli pairs - files whose filenames are marked for language, speaker and sounds (e.g. EnglishMale1_A_FrenchMale2_U.wav). Pairs are segregated by gender, and there is a number of Lang2-Lang2 'catch' stimuli. Now I want to create pseudorandomized sets.
I want each participant (listener) to receive a pseudorandomized set which will contain each possible Lang1-Lang2 pair and each Lang2-Lang2 pair twice (once for each speaker gender), with speaker pairs chosen at random, but with each speaker being represented at least once in every set. That is, a set would look something like this:
Vowel 1 of Language 1 x Vowel 1 of Language 2 uttered by random male speakers
Vowel 1 of Language 1 x Vowel 1 of Language 2 uttered by random female speakers
... (repeat for all possible pairs of L1xL2 vowels)
Vowel 1 of Language 2 x Vowel 1 of Language 2 uttered by random male speakers
Vowel 1 of Language 2 x Vowel 1 of Language 2 uttered by random female speakers
...(repeat for all possible pairs of L2xL2 vowels)
The order of stimuli in the sets themselves would also be pseudorandomized, of course. The hurdle is creating them in the first place.
I can think of a number of ways to do this, but all of them are clunky, convoluted and take a lot of time (I need at least ten such sets). Is there any software solution, preferably a free one, that will allow me to automatically set the criteria for such a randomization? I only know of Mix but I don't think it can do quite what I want.