What is a 2D experiment?

As it was originally designed, Audapter allowed you to perturb the F1 and F2 formants. However, F2 was the only variable which controlled how F1 and F2 were perturbed. This was a limitation; experiments could not control how F1 and F2 were perturbed based on F1.

So, our lab modified Audapter to support 2D perturbations; that is, you can independently perturb F1 and F2, and you can use Audapter's input signal's F1 and F2 to control that perturbation amount and direction. As of this writing, our 2D code changes have not been merged with the main Audapter code base, and the Audapter manual has no information for 2D configurations.

How do I configure my experiment for 2D perturbation?

This document assumes you already know what variables to set for a typical Audapter experiment. I will only be describing the parameters to set as compared to a 1D experiment.

• First, in your Audapter parameters variable (which is often called p), set the flag bShift2D = 1 . This tells Audapter to look for the values in the 2D variables which are coming up.
• Then, instead of setting pertAmp and pertPhi, you should set pertAmp2D and pertPhi2D . Note that these variables will need to be two-dimensional arrays of size 257x257 instead of just a vector of size 1x257.
• In addition to the normal pertF2, there will need to be a pertF1 of the same format which fulfills the same purpose but for F1 instead of F2.

And that's it! If you're converting an existing 1D experiment to a 2D experiment, make sure any references to 1D variables like pertAmp get replaced with the 2D variant.

If you need more help, run_varModOut2_expt and run_varMod_audapter are examples of 2D experiments. (Note that run_varMod_audapter supports 1D and 2D.)