Data for: Quantitative Analysis of Multimodal Speech Data
Description
Data for "Quantitative Analysis of Multimodal Speech Data," including raw data, analysis scripts, and example optical flow videos. This data can be used to replicate the analyses in the manuscript according to the 'steps to reproduce' below.
Files
Steps to reproduce
To obtain Peak Velocity magnitudes/locations and peaks per second averages by participant for velocity signals, run the script "AllPeaks_JPhonArticle.m" on participant velocity signal data, organized in folders named Participant_VelocityPeakData. This script requires "gfl.m" and "ReadPraatTier.m," available in the Analysis Scripts folder, to run. The script should produce two spreadsheet files, one with Velocity Peak information and another with Peaks per Second information, similar to the ones already present in each participant's folder. To obtain CMA and PrPosCorr information, run the script "PrPosCorr.m" on the signal data in the CorrelationMapData folder. This script requires "gfl.m" and "corrmap_histogram.m," available in the Analysis Scripts folder, and "correlation_map_gui.m," provided in the avspeech/cma-matlab github directory, which is also linked to this dataset. The script should produce a spreadsheet file similar to the one already present in the CorrelationMapData folder. The script which contains descriptive statistics and the other analyses we performed in R is available in Jphon_revised_analysis_v3.R.