Package net.cspeech.DSP

Class NonCausal

java.lang.Object

net.cspeech.DSP.ConsoleObj

net.cspeech.DSP.NonCausal

All Implemented Interfaces:

SISOFilter, SISOThread.FilterIO

public class NonCausal
extends ConsoleObj
implements SISOFilter, SISOThread.FilterIO

Constructor Summary

Constructors

Constructor	Description
NonCausal()

Method Summary

Modifier and Type	Method	Description
void	dispose()
int	extraN()
float[]	outIn(float[] sig, int nzero)
float[]	outInFilter(float[] sig, int nzero)
void	setMs(double sample_ms)
void	setParams(double[] params)

Methods inherited from class net.cspeech.DSP.ConsoleObj

setConsoleOut, setSystemOut

Methods inherited from class java.lang.Object

equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Methods inherited from interface net.cspeech.DSP.SISOFilter

setConsoleOut, setSystemOut

Constructor Detail

NonCausal

public NonCausal()

Method Detail

setParams

public void setParams(double[] params)

Specified by:

setParams in interface SISOFilter

Parameters:

params - is an array of double-precision floating-point values configuring the filter. These values of type double have sufficient precision to pass integer values as required. Implementation of this method typically checks the length of the params array: if params.length is less than the required number of parameters, default values are used.

setMs

public void setMs(double sample_ms)

Specified by:

setMs in interface SISOFilter

Parameters:

sample_ms - is the time interval between input samples to the filter, expressed in milliseconds (ms).

Invoking this method computes filter coefficients from the parameters specified in a prior call to setParams(). This action prepares the filter for processing a sequence of input samples supplied to outIn() below.

Reinvoking setMs() not only recalculates the filter coefficients, which is needed if another call is made to setParams. It also reinitializes any buffers holding samples in the "memory" of the filter.

extraN

public int extraN()

Specified by:

extraN in interface SISOFilter

Returns:

The number of extra number of input samples required to produce a given number of output samples.

To output nBuf samples of from the filter, for example, nBuf + extraN() samples of input are required. For a non-causal filter that needs to look ahead to input x[n+M] to compute output y[n], extraN() == M immediately after setMs() is called to initialize the filter.

outInFilter

public float[] outInFilter(float[] sig,
                           int nzero)

Specified by:

outInFilter in interface SISOThread.FilterIO

dispose

public void dispose()

outIn

public float[] outIn(float[] sig,
                     int nzero)

Specified by:

outIn in interface SISOFilter

Parameters:

sig - is the array of input samples.

nzero - extends the input with zero-valued samples. This "padding" is useful if sig contains a pulse waveform and you want to compute the "tail" of the filter response past the end of that pulse.

Returns:

The array of filtered waveform samples.

You can call outIn(sig, 0) repeatedly and the filter will keep track of the state of the filter between successive calls. A call to setMs() resets the state of the filter to the initial condition, typically that where the input to the filter has been zero until supplying the next values in sig as input.

If you want to apply a non-causal filter with M-sample look ahead to an array sig in one step, and if you want sig.length samples of output to match the sig.length samples of input, invoke

float[] out = mySISOFilter.inOut(sig, mySISOFilter.extraN());

to properly extend input sig with extraN() zeroes to obtain the required number of output samples.