Home

Musical Gestures Toolbox is a Matlab toolbox which contains functions for the analysis and visualization of video, mocap, and audio data. Musical Gestures Toolbox integrates Mocap tool box and Mirtoolbox which aim for the analysis the motion capture data and audio or music input data respectively. So before using Musical Gestures Toolbox, make sure Mocap tool box and MIRtoolbox are correctly installed. Mocap tool box is available on the website(www.jyu.fi/music/coe/materials/mocaptoolbox). The URL of MIRtoolbox website is (www.jyu.fi/music/coe/materials/mirtoolbox). The steps of installation are the same as MGT above. To use all the functions in the MGT, Signal Processing Toolbox and Image Processing Toolbox are needed as well. The Musical Gestures Toolbox comes with no warranty. It is free software.

Basics

Musical Gestures Toolbox(MGT) requires Matlab 2015a environment or more recent. Matlab software is available on the website(www.mathworks.com). To install MGT is quite simple. Just add MGT in the Matlab path:

• In Matlab 2015a or more recent,on the Home tab, in the Environment section, click Set Path.

• In the new Set Path window,click on the second button Add with Subfolders.

• In the file browser,select the folder you downloaded and unzipped and click Open.Then click Save button. The addresses of the folder are added.

Functions

The Musical Gestures Toolbox contains a set of functions for the analysis and visualization of video, audio, and mocap data. There are four categories of functions:

• Data input and edit functions
• Data preprocessing functions
• Visualization functions
• Middle and higher level feature extraction functions

Data Structures

The Musical Gestures Toolbox uses only one struct data structure. This structure contains three fields: video, audio, and mocap, which corresponds to these three types of data, respectively. An MGT data structure is created with the function:

mg=mginitstruct;

This produces a structure mg with three fields. The video field (mg.video) contains data and general parameters of the video file:

• .obj: stores each frame of the video
• .gram: after running the function mgmotion, motiongram data are written to .gram.x and .gram.y.
• .qom: quantity of motion of each frame
• .com: centroid of motion of each frame
• .nframe: the number of frames in the video file. The framerate of the video is stored in the field framerate. This is most often 25 or 30 frames per second. Duration field is for length of the video. The framerate of video is stored in the field framerate. Usually, it is 30 frames per second. Duration field is for length of the video.
• .method: the toolbox uses two general methods to estimate the motion: either based on frame differencing ('Diff') or optical flow ('OpticalFlow').

The data structures for mocap and audio are copied from the MoCap and MIR toolboxes.