---

pp::Filters Namespace Reference

Filters for processing audio data. More...

Compounds
class	pp::Filters::AllPass
class	pp::Filters::BandPass
class	pp::Filters::ButterworthFilter
class	pp::Filters::CombFilter
class	pp::Filters::ConnectInParallel
class	pp::Filters::ConnectInSeries
class	pp::Filters::ConnectorFilter
class	pp::Filters::Convolution
class	pp::Filters::Delay
class	pp::Filters::DynamicDelay
class	pp::Filters::EnvelopeFilter
class	pp::Filters::FindMax
class	pp::Filters::FindOp
class	pp::Filters::FIRComb
class	pp::Filters::GeneralFilter
class	pp::Filters::GeneratorFilter
class	pp::Filters::GenericFilter
class	pp::Filters::HighPassFIR
class	pp::Filters::HighPassIIR
class	pp::Filters::IIRComb
class	pp::Filters::IIRCombLoop
class	pp::Filters::ImpulseResponseFilter
class	pp::Filters::LowPassFIR
class	pp::Filters::LowPassIIR
class	pp::Filters::MultiLowPassFIR
class	pp::Filters::NullFilter
class	pp::Filters::Oscilloscope
class	pp::Filters::PluginFilter
class	pp::Filters::Scale
class	pp::Filters::Schroeder1
class	pp::Filters::Schroeder2
class	pp::Filters::SimpleFilter
class	pp::Filters::TransformFilter
Functions
float	MillisecondsToSamples (int SampleRate, int Milliseconds)
	A simple utility function that converts a time in milliseconds to the number of samples played back during that interval. More...
double	zr (double df, int k=1)
	Returns: the real part of z^k = e^(-2*k*PI*deltaFrequency) z^k = cos(-2*k*PI*deltafrequency).
double	zi (double df, int k=1)
	Returns: the imaginary part of z^k = e^(-2*k*PI*deltaFrequency) z^k = cos(-2*k*PI*deltafrequency) corresponds to the phase.
std::complex<double>	z (double df, int k=1)
	complex version.
template<typenameT, classX> bool	ConvertFilter (SimpleFilter<T>* f, X*& x)
bool	GetSeries (SimpleFilter<short>* f, ConnectInSeries<short>*& s)
bool	GetParallel (SimpleFilter<short>* f, ConnectInParallel<short>*& p)
bool	GetConnector (SimpleFilter<short>* f, ConnectorFilter<short>*& c)

---

Detailed Description

Filters for processing audio data.

Designed for use with the stl algorithms.

You'll probably need to convert unsigned char to signed char before using these filters, preferably using an adapter (there are some defined in Adapters.h). eg.

      Adapter8 adapter(f);
      transform(begin8(), end8(), begin8(), adapter);
   

All are time domain filters.

signed short - or any other signed type - should be okay.

The all pass and recusive comb filters can be used together to create reveberators.

M.R. Schroeder's original designs: 1. 4 comb filters in parallel followed by two all pass filters in series. 2. 4 all pass filters in series.

He later extended his original designs by placing a multitap delay line in parallel with them, to simulate early reflections.

The filter response (H(f)) for a linear filter is given by

   H(f) = Sum[k = 0..M](Ck * z^k)/(1-Sum[j=0..N](Dj * z^j) 

   where z = e^(-2*PI*i*(deltaFrequency))
   

Eulers relation:

 e^(i*x) = cos(x) + i*sin(x) 

for a FIR that is H(f) = Sum[k=0..M](Ck*z^k) ie an FFT

        *** Notch ***

          f = notch center frequency
          r = 0..1, but not 1

          z1x = cos(2*pi*f/samplerate)
          a0a2 = (1-r)^2/(2*(|z1x|+1)) + r
          a1 = -2*z1x*a0a2
          b1 = 2*z1x*r
          b2 = -r*r

          output(t) = a0a2 * (input(t)+input(t-2)) + 
                      a1 * input(t-1)+ 
                      b1 * ouput(t-1) + b2*output(t-2)
  

Reverb can be implemented simplistically using IIRComb with a large delay or Schroeder1 or Schroeder2 the parameters are left up to trial and error.

Chorus and flanging can be implemented using a variable length delay line

---

Function Documentation

float pp::Filters::MillisecondsToSamples ( int SampleRate, int Milliseconds ) [inline]

A simple utility function that converts a time in milliseconds to the number of samples played back during that interval. It may be useful when using the comb filters or other filters that require an interval given by the number of samples.

---