<?xml version="1.0" encoding="utf-8"?>
<!DOCTYPE document PUBLIC "-//CNX//DTD CNXML 0.5 plus MathML//EN" "http://cnx.rice.edu/cnxml/0.5/DTD/cnxml_mathml.dtd">
<document xmlns="http://cnx.rice.edu/cnxml" xmlns:m="http://www.w3.org/1998/Math/MathML" xmlns:bib="http://bibtexml.sf.net/" xmlns:md="http://cnx.rice.edu/mdml/0.4" id="id2255528">
  <name>Extension to Non-lowpass Filters</name>
  <metadata>
  <md:version>1.2</md:version>
  <md:created>2008/05/20 13:54:46 GMT-5</md:created>
  <md:revised>2008/07/27 14:24:12.538 GMT-5</md:revised>
  <md:authorlist>
      <md:author id="jrt">
      <md:firstname>John</md:firstname>
      <md:othername>R</md:othername>
      <md:surname>Treichler</md:surname>
      <md:email>jrt@appsig.com</md:email>
    </md:author>
  </md:authorlist>

  <md:maintainerlist>
    <md:maintainer id="jrt">
      <md:firstname>John</md:firstname>
      <md:othername>R</md:othername>
      <md:surname>Treichler</md:surname>
      <md:email>jrt@appsig.com</md:email>
    </md:maintainer>
    <md:maintainer id="dcwill">
      <md:firstname>Daniel</md:firstname>
      <md:othername>Collins</md:othername>
      <md:surname>Williamson</md:surname>
      <md:email>dwilliamson1285@gmail.com</md:email>
    </md:maintainer>
    <md:maintainer id="richb">
      <md:firstname>Richard</md:firstname>
      <md:othername>G.</md:othername>
      <md:surname>Baraniuk</md:surname>
      <md:email>richb@rice.edu</md:email>
    </md:maintainer>
    <md:maintainer id="cburrus">
      <md:firstname>C.</md:firstname>
      <md:othername>Sidney</md:othername>
      <md:surname>Burrus</md:surname>
      <md:email>csb@rice.edu</md:email>
    </md:maintainer>
    <md:maintainer id="rswagner">
      <md:firstname>Raymond</md:firstname>
      
      <md:surname>Wagner</md:surname>
      <md:email>rwagner@rice.edu</md:email>
    </md:maintainer>
  </md:maintainerlist>
  
  

  <md:abstract/>
</metadata>
  <content>
    
      
      <para id="id2255548">All of the discussion to this point has focused on lowpass filters.
Practical applications require other types, of course, including highpass,
bandpass, and bandstop designs. In fact the analysis presented in the
previous sections applies to all of these design criteria and the rules
for filter length estimation can be used almost directly. In general
  <cnxn document="m17234" target="uid2">Equation 1</cnxn>

and <cnxn document="m17234" target="uid3">Equation 2 from the module titled "Filter Sizing"</cnxn>

apply when one of
the equal ripple specifications dominates all others and when one of the
transition band specifications dominates all others. As a practical matter
this means that <m:math overflow="scroll"><m:msub><m:mi>δ</m:mi><m:mi>i</m:mi></m:msub></m:math> dominates if it is less than one-tenth of all
other rippple specifications and that <m:math overflow="scroll"><m:mrow><m:mi>Δ</m:mi><m:msub><m:mi>f</m:mi><m:mi>i</m:mi></m:msub></m:mrow></m:math> dominates if it is
simply less than all others. Suppose we define <m:math overflow="scroll"><m:mi>δ</m:mi></m:math> and <m:math overflow="scroll"><m:mrow><m:mi>Δ</m:mi><m:mi>f</m:mi></m:mrow></m:math> by
the equations:</para>
      
      <para id="element-486"><m:math mode="display" overflow="scroll">
		<m:mtable displaystyle="true">
			<m:mtr>
				<m:mtd columnalign="right">
					<m:mi>δ</m:mi>
				</m:mtd>
				<m:mtd>
					<m:mo>=</m:mo>
				</m:mtd>
				<m:mtd columnalign="left">
					<m:mrow>
						<m:mi>m</m:mi>
						<m:mi>i</m:mi>
						<m:mi>n</m:mi>
						<m:mo>{</m:mo>
						<m:msub>
							<m:mi>δ</m:mi>
							<m:mi>i</m:mi>
						</m:msub>
						<m:mo>}</m:mo>
						<m:mtext>, for all pass and stopbands i, and </m:mtext>
					</m:mrow>
				</m:mtd>
			</m:mtr>
			<m:mtr>
				<m:mtd columnalign="right">
					<m:mrow>
						<m:mi>Δ</m:mi>
						<m:mi>f</m:mi>
					</m:mrow>
				</m:mtd>
				<m:mtd>
					<m:mo>=</m:mo>
				</m:mtd>
				<m:mtd columnalign="left">
					<m:mrow>
						<m:mi>m</m:mi>
						<m:mi>i</m:mi>
						<m:mi>n</m:mi>
						<m:mo>{</m:mo>
						<m:msub>
							<m:mi>Δf</m:mi>
							<m:mi>k</m:mi>
						</m:msub>
						<m:mo>}</m:mo>
						<m:mtext>for all transition bands k</m:mtext>
					</m:mrow>
				</m:mtd>
			</m:mtr>
		</m:mtable>
	</m:math>
</para><para id="id2255818">If so then equation <cnxn document="m17234" target="uid2">Equation 1 from the module titled "Filter Sizing"</cnxn>

can be used directly and the
equation for <m:math overflow="scroll"><m:mi>α</m:mi></m:math> becomes</para>
      <equation id="id2255835">
        <m:math mode="display" overflow="scroll">
          <m:mrow>
            <m:mi>α</m:mi>
            <m:mo>=</m:mo>
            <m:mn>0</m:mn>
            <m:mo>.</m:mo>
            <m:mn>22</m:mn>
            <m:mo>-</m:mo>
            <m:mfrac>
              <m:mrow>
                <m:mi>l</m:mi>
                <m:mi>o</m:mi>
                <m:msub>
                  <m:mi>g</m:mi>
                  <m:mi>e</m:mi>
                </m:msub>
                <m:mi>δ</m:mi>
              </m:mrow>
              <m:mi>π</m:mi>
            </m:mfrac>
            <m:mo>.</m:mo>
          </m:mrow>
        </m:math>
      </equation>
      <para id="id2254919">A final hint - Watch out for the implicit boundary conditions present in
the design of linear phase FIR digital filters in two cases: even order,
symmetric response and odd order, antisymmetrical response. In both of these
cases the underlying equations for the filter's frequency response constrain
it to equal exactly zero at <m:math overflow="scroll"><m:mfrac><m:msub><m:mi>f</m:mi><m:mi>s</m:mi></m:msub><m:mn>2</m:mn></m:mfrac></m:math>. This is obviously not a problem
for lowpass filters, since the desired gain at <m:math overflow="scroll"><m:mfrac><m:msub><m:mi>f</m:mi><m:mi>s</m:mi></m:msub><m:mn>2</m:mn></m:mfrac></m:math> is zero
already. However, in the design of multiband and highpass filters an
inordinate amount of engineering time has been spent trying to design
even-order filters when in fact it is impossible to do so. The Parks-McClellan
algorithm will gamely try, but will fail. As a rule, use odd values of <m:math overflow="scroll"><m:mi>N</m:mi></m:math>
for highpass and multiband filters requiring nonzero response at
<m:math overflow="scroll"><m:mfrac><m:msub><m:mi>f</m:mi><m:mi>s</m:mi></m:msub><m:mn>2</m:mn></m:mfrac></m:math> and use even-order filters for differentiators.</para>
    
  </content>
</document>
