SBAM - SPM Bit Accurate Modelling
The SBAM library also contains a Matlab interface which allows rapid prototyping of algorithms.
- Uses Object Orientated programming to interface to the c++ SBAM library.
- Allows runtime configuration of numeric type parameters.
- Possible to view underlying binary representation of numbers.
- Currently only FloatingPoint type has been implemented.
- One function call needed to convert all variables in the current workspace from Matlab double format to cmat:
Low-pass Filter Example
A 2nd order Butterworth low-pass filter was implemented in MATLAB, the input to the filter was a low frequency sine wave, with high frequency noise. Below is a plot of the output of the filter using the SBAM floating point type in various configurations. The green plot corresponds to the IEEE double precision number format, the cyan is configured with IEEE single precision, finally the magenta is a highly restricted format with 6 exponent and 6 mantissa bits.