What is a digital signal processor algorithm format?

Digital signal processor (DSP) there are a variety of algorithms. The vast majority of DSP processors use fixed-point arithmetic, figures expressed as an integer, or-1.0 to + 1.0 decimal form. Some processors use floating-point arithmetic, data representation into mantissa and exponent of the form: mantissa × 2 index. Floating point arithmetic is a more complex algorithm, the use of conventional floating point data for large dynamic range of the data (the dynamic range you can use the maximum and minimum ratio to represent). Floating-point DSP applications, design engineers need not care about dynamic range and precision of a class of problems. Floating-point DSP easier than fixed-point DSP programming, but cost and power consumption. Because of the costs and power consumption, General bulk products use fixed-point DSP. Programming and design staff through the analysis of algorithms or simulation to determine the needs of dynamic range and accuracy. If you require ease of development, but also the dynamic range is very wide, high precision, you can consider using floating point DSP. You can also use fixed-point DSP realization by the software under the condition of floating-point calculations, but such software program consumes significant processor time, which are rarely used. Effective solution is the "block floating point", which will have the same index, and mantissa different set of data as the data block for processing. "Block floating point" processing is commonly used software.