After piloting your experiment, here's a few tools you can use to validate your data.

• [onset_bad, offset_bad] = validate_onsetOffset(dataPath, badThresh, bInterpret)
• WHAT IT DOES: This function shows histograms of how close a speaker's voice onset and offset were to the edge of the recording. Use to validate that you were recording the whole utterance.
• WORKS FOR: Audapter experiments.
• Optionally returns two output arguments, which are vectors of trial numbers too close to the onset/offset respectively.

• validate_formantShift(dataPath, plotParams, bInterpret)
• WHAT IT DOES: Plots the average formants from signalIn and compares it against the average formants from signalOut on an [F1 F2] field. Shows different plots for different conditions in expt.conds. Use to visually validate that the proper F1/F2 perturbations were applied in different conditions.
• WORKS FOR: Formant perturbation experiments, especially ones where the same perturbation is applied over the whole utterance.

• [Tables] = validate_exptSetup(dataPath, groupings, bInterpret)
• WHAT IT DOES: Shows the number of trials in pairs of groupings; for example, how many trials were both expt.word{1} and expt.cond{1}, expt.word{2} and expt.cond{1}, etc. Useful for making sure your counterbalancing across different manipulations is happening properly.
• WORKS FOR: Any experiments with counterbalanced groupings. E.g., words + conds (various), words + conds + colors (stroop), words + conds + dots (attentionComp).

• [figHandle] = plot_audapterFormants(data, plotParams)
• WHAT IT DOES: Simply displays the wave form (top), spectrogram (bottom), signalIn formant track (bottom; cyan), and signalOut formant track (bottom; magenta).
• WORKS FOR: Any Audapter experiment.
• The first input parameter takes a data.mat file. However, this function is intended to run on 1-10 trials at a time. You should view only a subset of the data file at a time, such as plot_audapterFormants(data(1:10), []). 

• [] = check_trialDurations(dataPath, bPlot)
  • WHAT IT DOES: Plots the duration of audio recorded for all trials in an experiment. Helps validate that trial duration was consistent across all trials.
  • WORKS FOR: Any experiment that records Audapter audio
  • It's expected that there are 10-20 ms variations in trial length across some trials. 
  • If your experiment is steadily increasing in length, the #1 culprit is that you aren't clearing the MATLAB figures which display stimuli between trials. If you don't regularly clear the figure, the figure variable will hold more and more data and take longer to load. To fix this, make sure you are clearing the experiment text between trials, e.g.,  delete_exptText(h_fig,h_text); clear h_text. Example in context.