MUSTIG is a signal analysis oriented language. It allows you to experiment with algorithms and is an excellent teaching tool. MUSTIG is :

• non procedural : the analysis is described in the form of a dependence graph of outputs designed from inputs, and not through a classical program including series of instructions executed in a sequential form.
• interactive : the graph is displayed as a real diagram on the screen, it may be extended or modified after an evaluation, and only the necessary calculations will be executed again.

The language provides easy interactive multi-dimensional signal analysis. It allows the user to easily plot an evolution graph according to parameters entered as additional dimensions. A complete library of operators is provided as basic building blocks and a partial list appears below.

• multidimensional : the set of elements handled by MUSTIG belongs to a space with any number of dimensions. Sections of the graph can be automatically repeated along some of these dimensions.

As far as signal analysis is concerned, these signals can be images (functions of two space variables) or sets of temporal signals like outputs of a network antenna (function of time and of one or two space variables).

Each "wire" of the description graph analysis propagates a signal S(u1, u2, …, uN) in which N represents the number of dimensions (or of variables) of the signal. The order N of the signal is between 0, for a constant, and some number (limited basically by the available memory). This order can change between a module input and output ; its identity can be changed as well : the Fourier transform changes time into frequency.

- Demultiplexing and displaying of acquired data

- Smoothing of temporal series, regression lines.

- Integration, differentiation of signals.

- Non linear adjustment, least squares.

- Recursive, transverse filtering.

- Lattice filter.

- Filter gauging.

- Simulations with a set of coefficients in a form of series (like in a spread sheet)

- Generating binary sequences of maximal length.

- Spectral analysis of signals : weighting function FFT, ...

- Adaptive filters.

- Whitening of signal.

- Wiener filtering.

- Kalman filtering.

- Programming autoregressive methods (or maximum entropy methods).

- Antenna analysis, super-resolution.

- Image analysis

- …