Scribd Logo
Upload

Welcome to Scribd!

  • 63 views

    DSP Assignment

    Uploaded by

    Saruchi Attri
    This document discusses different methods of fixed point number representation, including sign magnitude form, one's complement form, and two's complement form. It then discusses truncation and rounding of fixed point numbers, and analyzes the errors introduced by truncation and rounding. The key points are: 1. Truncation discards less significant bits and approximates the number to the highest quantized level below it, introducing a negative or zero error. 2. Rounding chooses the number closest to the original, introducing an error between -2-b/2 and 2-b/2. 3. The maximum relative error from truncation is between -2-b and 0, while from rounding it is between -

    Copyright:

    Attribution Non-Commercial (BY-NC)
    Download as DOCX, PDF, TXT or read online from Scribd

    DSP Assignment

    Uploaded by

    Saruchi Attri
    63 views8 pages
    This document discusses different methods of fixed point number representation, including sign magnitude form, one's complement form, and two's complement form. It then discusses truncation and rounding of fixed point numbers, and analyzes the errors introduced by truncation and rounding. The key points are: 1. Truncation discards less significant bits and approximates the number to the highest quantized level below it, introducing a negative or zero error. 2. Rounding chooses the number closest to the original, introducing an error between -2-b/2 and 2-b/2. 3. The maximum relative error from truncation is between -2-b and 0, while from rounding it is between -

    Original Title

    Dsp Assignment

    Copyright

    © Attribution Non-Commercial (BY-NC)

    Available Formats

    DOCX, PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    This document discusses different methods of fixed point number representation, including sign magnitude form, one's complement form, and two's complement form. It then discusses truncation and rounding of fixed point numbers, and analyzes the errors introduced by truncation and rounding. The key points are: 1. Truncation discards less significant bits and approximates the number to the highest quantized level below it, introducing a negative or zero error. 2. Rounding chooses the number closest to the original, introducing an error between -2-b/2 and 2-b/2. 3. The maximum relative error from truncation is between -2-b and 0, while from rounding it is between -

    Copyright:

    Attribution Non-Commercial (BY-NC)
    Download as DOCX, PDF, TXT or read online from Scribd
    Download as docx, pdf, or txt
    63 views8 pages

    DSP Assignment

    Uploaded by

    Saruchi Attri
    This document discusses different methods of fixed point number representation, including sign magnitude form, one's complement form, and two's complement form. It then discusses truncation and rounding of fixed point numbers, and analyzes the errors introduced by truncation and rounding. The key points are: 1. Truncation discards less significant bits and approximates the number to the highest quantized level below it, introducing a negative or zero error. 2. Rounding chooses the number closest to the original, introducing an error between -2-b/2 and 2-b/2. 3. The maximum relative error from truncation is between -2-b and 0, while from rounding it is between -

    Copyright:

    Attribution Non-Commercial (BY-NC)
    Download as DOCX, PDF, TXT or read online from Scribd
    Download as docx, pdf, or txt
    of 8

    DIGITAL SIGNAL PROSESSING

    ASSIGNMENT 01

    SUBMITTED TO :Mrs. Richa Sharma

    SUBMITTED BY:Saruchi Attri 10902121 E-4

    THE FIXED POINT NUMBER REPRESENTATION


    In fixed point arithmetic, the position of the binary point is fixed. The bit to the right represent the fractional part of the number and those to the left represent the integer part. For eg. binary no 01.1100 has the value 1.75 in decimal. THREE TYPES OF FIXED POINT ARITHMATIC: Sign magnitude form Ones-compliment form Twos-compliment form

    SIGN MAGNIYUDE FORM:In sign magnitude representation, the most significant bit is set to 1 to represent the negative sign. As an eg. the decimal number -1.75 is representation as 11.110000 and 1.75 is represent as 01.110000. In sign magnitude form, the no. zero has two representations, i.e. 00.000000 or 10.000000 with b bits only (2b-1) numbers can be represented. ONES-COMPLIMENT FORM In ones-compliment form, the positive no. is represented as in the sign magnitude notations. However the negative no. can be obtained by complimenting all the bits of the positive numbers. Example: -0.875 can be represented as under: (0.875) 10 = (0.111000)2 (-0.875)10 = (1.000111)2 Note that this is the same as subtracting the magnitude from 2-2b, where b is the number of bits (without sign bits) 2-2b = 10.000000-0.000001 = 1.111111 Now subtract 0.875= (0.111000)2

    1.111111 0.111000 1.000111 = (-0.875)10 in ones-compliment form In ones- compliment form, the magnitude of the negative no. is given as 1-bi=1 ci2-i 2-b

    TWOS COMPLIMENT FORM In twos compliment representation, positive numbers are represented as in sign magnitude form and ones compliment. However, the negative no. can be obtained by complimenting all the bits of the positive no. and adding one to the least significant bit. As an eg. we have (0.875)10 = (0.111000)2

    1.000111 + 0.000001 (-0.875)10 = (1.001000)2 This is same as subtracting the magnitude from 2 2.0 = 10.000000 -0.111000 1.001000 = (-0.875)10 in twos compliment form The magnitude of the negative number is given as follows: 1-bi=1ci2-i If the MSB of binary is one then no. is negative. However, if the MSB is zero, no. is positive fraction. So MSB is called sign bit.

    TRUNCATION
    Truncation is a process of discarding all bits less significant than least significant bits that is retained. Suppose, if we truncated the following binary no. from 8 bit to 4 bits, we obtain 0.00110011 to 0.0011 8 bits 4 bits

    When we truncate the number, the signal value is approximated by the highest quantized level which is not greater than signal.

    ROUNDING
    Rounding of a number of b bits is accomplished by choosing the rounded result as the b bits number closest to the original number unrounded. As an example, 0.11010 rounded to three bits is either 0.110 or 0.111.

    ERROR INTRODUCED DUE TO TRUNCATION


    If the quantization method is that of truncation, then the number is approximated by the nearest level that does not exceed it. In this case, the error xT-x is negative or zero where xT is truncated value of x and assumed |x| <= 0. The error made by truncating a number to b bits following the binary, point satisfies the following inequality: 0 xT-x>-2-b ....... (01)

    Equation (01) holds for both sign magnitude, ones compliment and twos compliment if x> 0. If x<0 we have to find whether the equation (01) holds good for all types of representations. Now let us consider first the twos compliment representation. From equation (01) the magnitude of the negative no. is:

    x = 1-

    ci2-i

    If we truncate the number to N bits, then we have xT = 1- ci2-i

    the change in magnitude will be xT-x = or ci2-i - ci2-i = ci2-i

    xT x >=0

    ........... (02)

    From the (02), it may be observed that the truncation increase the magnitude, which implies that the error is negative and satisfy the following inequality: 0 xT x -2-b ........... (03)

    For ones compliment representation, the magnitude of negative number with b bits is given as: x = 1 - bi=1 ci2-i 2-b when number is truncated to N bits, then we have xT = 1-Ni=1 ci2-i 2-N the change in magnitude due to truncation will be xT x = bi=N ci2-i (2-N 2-b) xT x < 0 Therefore, the magnitude decreases with truncation which implies that error is positive and satisfy the following inequality 0 xT-x < 2-b If x=2c.M then we have XT = 2c.MT Error e = xT-x = 2c (MT-M)

    Using equation (03) with twos compliment representation of mantissa, we have 0MT-M>-2-b

    0e>-2-b2c We define relative error as = (xT-x)/x = e/x

    ................. (04)

    now equation can be written as follows: 0 x > -2-b2c 0 2cM > -2-b2c 0 M > -2-b If M= the relative error is maximum. Therefore we have 0 > -2.2-b If M= - the relative error range will be 0 < 2.2-b In ones compliment representation: 0 MT-M > -2-b 0 e > -2-b2c e = 2cM For negative mantissa values error will be 0 MT-M < -2-b 0 e < -2-b2c. If M= - the maximum relative error range will be 0 > 2.2-b Which is the same as positive M.

    ERROR INTRODUCED DUE TO ROUNDING


    In fixed point the error due to rounding a no. to b bits produces an error e=xT-x which satisfies the following inequality: -2-b/2 xT x 2-b/ 2 This is because of the fact that with rounding if the value lies the half way between two levels, it can be approximated to either nearest high level or by nearest lower level. x = M. 2c xT = MT. 2c then e= xT-x = (MT-M). 2c but for rounding -2-b/2 MT M 2-b/ 2 ............. (06) ............. (05)

    Using (05) eq. (06) can be written as: (-2c)-2-b/2 xT x 2-b/ 2(2c) (-2c)-2-b/2 x 2-b/ 2(2c) We have x = 2cM Then (-2c)-2-b/2 2cM 2-b/ 2(2c)

    Which gives -2-b/2 M 2-b/ 2 The mantissa satisfies M< 1 If M= the maximum range of relative error is : -2-b < 2-b.

    PDF for truncation.

    PDF for rounding.

    You might also like

    pFad - Phonifier reborn

    Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

    Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


    Alternative Proxies:

    Alternative Proxy

    pFad Proxy

    pFad v3 Proxy

    pFad v4 Proxy