asked in 社會與文化語言 · 9 years ago


¡ In order to operate on real-world data, DSPs must be able to handle calculations with real number. The number formats used by DSPs to represent real numbers fall into two major categories:fixed point and floating point. Each of these formats allow a DSP to store and combine real numbers, but each has important advantages most often rely on binary or hexadeci-mal formats to represent numbers ¡ Conversions between decimal (base 10), binary (base 2), and hexadecimal (base 16) numbers are explained in Appendix A.11. As indicated in the ap-pendix, numbers can be marked by a subscript according to their bases, and hexadecimal numbers are frequently marked with the annotations “0x” or “h.”¡ To compare fixed point and floating point and formats requires the concepts of dynamic range and resolution. Dynamic range, reported in dB, refers to the ratio of the maximum possible range of the signal, called the full scale range, to the smallest signal difference that can be resolved, the quantization step size.¡ An alternative form of this definition, which gives a slightly larger result, is:¡ Resolution refers to the size of the step between two adjacent numbers in a number system. The smaller the step size, the better the resolution of the system and the more closely a real number can be approximated. ¡ For a given full scale range, the better the resolution, the greater the dynamic range of the system. Resolution may be increased by increasing the number of fractional bits in a number format. In a binary number, fractional bits are the bits to the right of the binary point (an analog of the decimal point). ¡ Since a limited number of bits is available, quantization error affect all number represented in fixed or floating point formats, no matter how many fractional bits are used. Except for the very largest signals in range, these errors can never be greater than half of the step size between adjacent quantization levels, as explained in Section 2.3.

1 Answer

  • piglet
    Lv 6
    9 years ago
    Favorite Answer


    固定點和浮點運算。這些格式可以讓每一個 DSP存儲,並結合實數,但每個都有最重要的優勢往往依賴於二進制或hexadeci -不善的格式來表示數字

    ¡十進制之間的轉換(基10),二進制(底數 2),和十六進制(基數為 16)號解釋在附錄 A.11。正如在AP - pendix,數字可以標記由標根據他們的基地,和十六進制數經常標有註釋“0x”或“小時”






    2011-01-05 23:07:12 補充:


Still have questions? Get your answers by asking now.