Embedded software engineers typically prefer to design and prototype control systems in floating-point arithmetic, but final designs are usually implemented in

fixed-point arithmetic.

Fixed-point arithmetic is often used for the computations because of the cost reasons--floating point calculations significantly increase the size of the design and make it more complex.

After testing numerous competing products, we found the Filter Design Toolbox provided the best range of

fixed-point arithmetic parameters, filter architectures and operational statistics," said Richard Hewitt, senior DSP and software consultant at Toracomm.

Any data word smaller than 40-bits or

fixed-point arithmetic substantially degrades the sound quality.

Engineers can use these blocks to quickly create PID controllers, select PID structure, include integrator anti-windup logic, convert from continuous-time to discrete-time domain, and prepare the controller for implementation on a processor using

fixed-point arithmetic.

The company is also presenting a paper titled, "Exploration of Least-Squares Solutions of Linear Systems of Equations with

Fixed-Point Arithmetic," on Monday November 14th, at 2:00 p.

Until now, performing matrix factorization and inversion in hardware has been difficult because the algorithms are complex and highly sensitive to numerical problems, particularly when using

fixed-point arithmetic preferred in high-performance ICs.

The DSP Studio Toolbox adds

fixed-point arithmetic to MATLAB and supports common processor-specific operations, such as multiply with shift and saturate.

Performance improvements have also been made to

fixed-point arithmetic when using 64-bit emulation systems.

A|RT Library provides a set of C++ data classes and operators that encapsulate the characteristics of

fixed-point arithmetic and that accurately account for quantization and overflow effects.

Most hardware implementation, on the other hand, is done using

fixed-point arithmetic because of significant cost-saving and increased performance.

16-bit and 32-bit

fixed-point arithmetic supported (rounding and