Scribd Logo
Upload

Welcome to Scribd!

  • 80 views

    Finite Word Length Effects

    Uploaded by

    anirobby
    This document discusses various types of errors that can occur due to finite word length effects in digital systems, including quantization error, coefficient quantization error, and zero input limit cycle oscillations. It provides tasks to analyze these errors in systems like analog to digital converters (ADCs) and infinite impulse response (IIR) filters. For example, it asks the reader to calculate quantization error for samples from an ADC, derive the variance of quantization noise, and analyze how coefficient quantization affects pole locations in IIR filters.

    Copyright:

    © All Rights Reserved
    Download as DOCX, PDF, TXT or read online from Scribd

    Finite Word Length Effects

    Uploaded by

    anirobby
    80 views3 pages
    This document discusses various types of errors that can occur due to finite word length effects in digital systems, including quantization error, coefficient quantization error, and zero input limit cycle oscillations. It provides tasks to analyze these errors in systems like analog to digital converters (ADCs) and infinite impulse response (IIR) filters. For example, it asks the reader to calculate quantization error for samples from an ADC, derive the variance of quantization noise, and analyze how coefficient quantization affects pole locations in IIR filters.

    Original Description:

    Finite Word Length Effects

    Copyright

    © © All Rights Reserved

    Available Formats

    DOCX, PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    This document discusses various types of errors that can occur due to finite word length effects in digital systems, including quantization error, coefficient quantization error, and zero input limit cycle oscillations. It provides tasks to analyze these errors in systems like analog to digital converters (ADCs) and infinite impulse response (IIR) filters. For example, it asks the reader to calculate quantization error for samples from an ADC, derive the variance of quantization noise, and analyze how coefficient quantization affects pole locations in IIR filters.

    Copyright:

    © All Rights Reserved
    Download as DOCX, PDF, TXT or read online from Scribd
    Download as docx, pdf, or txt
    80 views3 pages

    Finite Word Length Effects

    Uploaded by

    anirobby
    This document discusses various types of errors that can occur due to finite word length effects in digital systems, including quantization error, coefficient quantization error, and zero input limit cycle oscillations. It provides tasks to analyze these errors in systems like analog to digital converters (ADCs) and infinite impulse response (IIR) filters. For example, it asks the reader to calculate quantization error for samples from an ADC, derive the variance of quantization noise, and analyze how coefficient quantization affects pole locations in IIR filters.

    Copyright:

    © All Rights Reserved
    Download as DOCX, PDF, TXT or read online from Scribd
    Download as docx, pdf, or txt
    You are on page 1of 3

    FINITE WORD LENGTH EFFECTS

    Task 1 Fill up the given b=8 bit data in the following boxes (one bit in each box).
    Data stream is 0.11011011(the digit before the point is reserved for sign, rest of the
    bits represent binary value of a decimal number.

    Identify the operations performed by you in order to accommodate the given data in
    the boxes.
    Task 2- input quantization error (noise) Draw the block of ADC. The input is x
    (t), sample it to obtain x(n), then quantize it to obtain x q(n).
    Assume a sinusoidal signal varying between +1 and -1. Find the dynamic range of
    the signal. The ADC used here uses b+1 bits including sign bit. What is the number
    of levels available for quantizing x(n)? Find the interval q between successive
    levels, where q is called as quantization step size. What will be the q value for the
    given problem?
    Let one of the sample value be 0.7, convert it into binary. Quantize it to 3 bits
    (excluding sign bit).convert the data back to decimal. Now find the error between
    unquantized and quantized values. Denote it as e (n). Write an expression for e (n).
    Prove that the error range falls between q/2 e(n) q/2 (for r) and q e(n) <
    0(for t)
    Prove that the probability of error p(e) is 1/q for both cases mentioned above.(ask
    for a clue!) draw the probability density function.
    Task 3- steady state input noise variance (power)
    Replace the quantizer block by a quantization noise model (use the expression
    obtained for e(n) in task 2). Draw the model.
    Task 4- answer the following:
    What is a stationary random process?
    What is meant by white noise process?
    Task 5- steady state input noise variance (power)
    Derive an expression for variance of e(n) (using probability and random process).
    Prove that for both r and t variance of e(n) is e2 = 2-2b/12.
    Also prove that the signal to noise ratio is 6.02b+10.79+10log x2
    What is the significance of the above expression?

    Steady state input output noise variance (power)


    Obtain the output of a DT LTI system given an input x(n)(draw the block
    diagram).after applying quantization, redraw the block along with the quantization
    noise model for the output side. Obtain an expression for the new output.
    Task 6-Product quantization error (noise)
    What will be the product of two 4 bit numbers? (Say 1100 and 0011). What will be
    the length of the product? Save your answer in the following block.
    What problem do you face? What will you do next?
    Draw the noise model for product quantization.
    Write down the difference equation for a single pole IIR system. Draw the block
    diagram for the same. How many noise sources do you get? Draw the noise model
    for product quantization.
    Write down the difference equation for an II order IIR system. Draw the block
    diagram for the same. How many noise sources do you get? Draw the noise model
    for product quantization. Find the total noise variance due to all noise sources.
    Task 7- coefficient quantization error
    Consider a II order IIR filter with H(z) = 1/(1-0.5z -1)(1-0.45z-1). Find the effect of
    quantization on pole locations of the given system function in direct form and
    cascade form. Assume b=3 bits. Which realization is better? Why?
    Task 8-zero input limit cycle oscillations

    You might also like