Filter Design Toolbox

Filtering Data with Quantized Filters

You can filter data with quantized filters using the filter function.

Example: Filtering Data with a Quantized Filter

• warning on
  randn('state',0);
  x = randn(100,2);
  [b,a] = butter(3,.9,'high');
  Hq = sos(qfilt('ReferenceCoefficients',{0.5* b,0.5*a},...
  'CoefficientFormat',unitquantizer([26 24])));
  y = filter(Hq,x);
  
  Warning: 64 overflows in QFILT/FILTER.
  
                          Max            Min     NOverflows    NUnderflows    NOperations
   Coefficient          1.187             -1              0              0             12
                        1.648             -2              0              0             18
         Input          2.183         -2.202             64              0            200
        Output         0.4345        -0.4477              0              0            200
  Multiplicand              1             -1              0              2            800
                       0.4345        -0.4477              0              0           1000
       Product       0.009246      -0.008869              0              0            800
                       0.7158        -0.7377              0              0           1000
           Sum        0.01274        -0.0122              0              0            600
                 0.4345        -0.4477              0              0           1000
  

Notice that a record of the overflows that occurred in filtering is displayed if you have set warning on.

Use qreport to get this listing when needed as well.

Specifying the Data Formats

Transformation Functions for Quantized Filter Coefficients