Jamoma API  0.6.0.a19
TTDCBlock Class Reference

TTDCBlock in an audio processor that filters out DC Offsets from an input. More...

#include <TTDCBlock.h>

+ Inheritance diagram for TTDCBlock:
+ Collaboration diagram for TTDCBlock:

Protected Member Functions

TTErr updateMaxNumChannels (const TTValue &oldMaxNumChannels, TTValue &)
 This method gets called when the inherited mMaxNumChannels attribute is changed. More...
 
TTErr calculateValue (const TTFloat64 &x, TTFloat64 &y, TTPtrSizedInt channel)
 Standard single value calculate method as used by DSP objects. More...
 
TTErr processAudio (TTAudioSignalArrayPtr inputs, TTAudioSignalArrayPtr outputs)
 Standard audio processing method as used by DSP objects. More...
 
TTErr clear ()
 Resets the DC-Blocking filter. More...
 
- Protected Member Functions inherited from TTAudioObjectBase
TTErr setProcess (TTProcessMethod processMethod)
 Set the audio processing routine to point to a method that is defined as an arg to this function. More...
 
TTErr setCalculate (TTCalculateMethod newCalculateMethod)
 Set the sample calculate routine to point to a method that is defined as an arg to this function. More...
 
TTErr setBypass (const TTValue &value)
 Bypass the audio processing routine and copy all input samples to the output unchanged. More...
 
TTErr resetBenchmarking ()
 Reset the benchmarking accumulation used to calculate the results of getProcessingBenchmark(). More...
 
TTErr getProcessingBenchmark (TTValueConstRef, TTValueRef v)
 Return the average time spent by this object processing audio since the last reset. More...
 
 TTAudioObjectBase (const TTValue &arguments)
 Constructor. More...
 
- Protected Member Functions inherited from TTObjectBase
 TTObjectBase (const TTValue arguments)
 Constructor. More...
 

Protected Attributes

TTSampleVector mLastInput
 Feedback values used for the audio filter.
 
TTSampleVector mLastOutput
 Feedback values used for the audio filter.
 
- Protected Attributes inherited from TTAudioObjectBase
TTUInt32 sr
 Current sample rate being used by this object.
 
TTFloat64 srInv
 1.0 over the current sample rate (inverse)
 
TTFloat64 srMill
 1/1000 of the current sample rate (samples per millisecond)
 
TTChannelCount mMaxNumChannels
 This is the maximum number of channels that can be guaranteed to work.
 
TTBoolean unused
 Old var that is not used anymore, but we want to keep the struct size the same.
 
TTBoolean attrBypass
 Are we bypassing the processMethod?
 
TTBoolean attrMute
 Mute the processMethod.
 
TTProcessMethod processMethod
 This function pointer points to the active (non-bypass) processing routine.
 
TTProcessMethod currentProcessMethod
 This function pointer always points to the current processing routine.
 
TTCalculateMethod calculateMethod
 This function pointer points to the active (non-bypass) calculate routine.
 
TTCalculateMethod currentCalculateMethod
 This function pointer always points to the current calculate routine.
 
TTObject inputArray
 If the process method is passed a signal, rather than an array of signals, we wrap the signal in this array.
 
TTObject outputArray
 If the process method is passed a signal, rather than an array of signals, we wrap the signal in this array.
 
TTFloat64 startProcessingTime
 The time at which this object's process method was last invoked (for benchmarking)
 
TTFloat64 accumulatedProcessingTime
 The amount of time spent in this object's process method (for benchmarking)
 
TTFloat64 accumulatedProcessingCalls
 The number of times the process method has been called (for benchmarking)
 
- Protected Attributes inherited from TTObjectBase
TTList observers
 List of all objects watching this object for life-cycle and other changes.
 

Additional Inherited Members

- Public Member Functions inherited from TTAudioObjectBase
TTErr setMute (const TTValue &value)
 Mute the audio processing routine and zero all output. More...
 
TTErr setMaxNumChannels (const TTValue &newValue)
 Setter for the maxNumChannels attribute. More...
 
TTErr adaptMaxNumChannels (const TTChannelCount newMaxNumChannels)
 Increase the maxNumChannels attribute, if neccessary. More...
 
TTErr setSr (const TTValue &newValue)
 Setter for the sample-rate attribute. More...
 
TTErr setSampleRate (const TTUInt32 &newSampleRate)
 Convenience method for updating the sample-rate. More...
 
virtual ~TTAudioObjectBase ()
 Destructor.
 
TTErr calculate (const TTFloat64 &x, TTFloat64 &y)
 Calculate a single sample of output for a single sample of input. More...
 
TTErr calculateMessage (TTValueConstRef input, TTValueRef output)
 
TTErr process (TTAudioSignal &in, TTAudioSignal &out)
 Process the input signal, resulting in an output signal. More...
 
TTErr process (TTAudioSignal &out)
 Process the an output signal only, e.g. More...
 
TTErr process (TTAudioSignal &in1, TTAudioSignal &in2, TTAudioSignal &out1, TTAudioSignal &out2)
 Process two input signals, resulting in two output signals. More...
 
TTErr process (TTAudioSignal &in1, TTAudioSignal &in2, TTAudioSignal &out)
 Process two input signals, resulting in an output signal. More...
 
TTErr process (TTAudioSignalArrayPtr inputs, TTAudioSignalArrayPtr outputs)
 Process an array of audio signals. More...
 
TTErr bypassProcess (TTAudioSignalArrayPtr inputs, TTAudioSignalArrayPtr outputs)
 The default audio processing method, which simply copies a signal through with no modifications. More...
 
TTErr bypassCalculate (const TTFloat64 &x, TTFloat64 &y, TTPtr data)
 Produces a result for calls to calculate when the unit is bypassed. More...
 
TTErr muteProcess (TTAudioSignalArrayPtr inputs, TTAudioSignalArrayPtr outputs)
 A muted audio processing method, which simply copies zeroes to the output. More...
 
TTErr defaultCalculateMethod (const TTFloat64 &x, TTFloat64 &y, TTPtr data)
 A default calculate method, which simply calls the process method with 1 channel signals that are 1 sample in length. More...
 
TTFloat64 hertzToRadians (const TTFloat64 hz)
 Convert Hertz to radians. More...
 
TTFloat64 radiansToHertz (const TTFloat64 radians)
 Convert radians into Hertz. More...
 
TTFloat64 degreesToRadians (const TTFloat64 degrees)
 Degrees-to-radians conversion. More...
 
TTFloat64 radiansToDegrees (const TTFloat64 radians)
 Radians-to-degrees conversion. More...
 
TTFloat64 decayToFeedback (const TTFloat64 decay_time, TTFloat64 delay)
 Decay Time (seconds) to feedback coefficient conversion. More...
 
TTFloat64 feedbackToDecay (const TTFloat64 feedback, const TTFloat64 delay)
 Return the decay time based on the feedback coefficient. More...
 
virtual TTErr test (TTValue &)
 Default (empty) template for unit tests. More...
 
- Public Member Functions inherited from TTObjectBase
virtual ~TTObjectBase ()
 Destructor.
 
TTUInt16 getReferenceCount ()
 Query an object to get its current reference count. More...
 
TTUInt32 getObserverCount () const
 Query an object to get the number of observers. More...
 
TTErr registerAttribute (const TTSymbol name, const TTDataType type, void *address)
 Register an attribute. More...
 
TTErr extendAttribute (const TTSymbol name, const TTObjectBasePtr extendedObject, const TTSymbol extendedName)
 Extend the attribute of an existing TTObjectBase to this TTObjectBase (using another attribute name) More...
 
TTErr removeAttribute (const TTSymbol name)
 Remove an attribute. More...
 
TTErr findAttribute (const TTSymbol name, TTAttribute **attr)
 Find an attribute. More...
 
TTErr setAttributeValue (const TTSymbol name, TTValue &value)
 Set an attribute value for an object. More...
 
template<class T >
TTErr set (const TTSymbol aName, T aValue)
 Set an attribute value for an object This is the same as calling setAttributeValue(). More...
 
TTErr getAttributeValue (const TTSymbol name, TTValue &value)
 Get an attribute value for an object. More...
 
template<class T >
TTErr get (const TTSymbol aName, T &aReturnedValue)
 Get an attribute value for an object This is the same as calling getAttributeValue(). More...
 
TTSymbol getAttributeType (const TTSymbol name)
 Return the type of an attribute as a symbol. More...
 
TTErr getAttributeGetterFlags (const TTSymbol name, TTAttributeFlags &value)
 Get the getterFlags of an attribute. More...
 
TTErr setAttributeGetterFlags (const TTSymbol name, TTAttributeFlags &value)
 Set the getterFlags of an attribute. More...
 
TTErr getAttributeSetterFlags (const TTSymbol name, TTAttributeFlags &value)
 Get the setterFlags of an attribute. More...
 
TTErr setAttributeSetterFlags (const TTSymbol name, TTAttributeFlags &value)
 Set the setterFlags of an attribute. More...
 
TTErr registerAttributeProperty (const TTSymbol attributeName, const TTSymbol propertyName, const TTValue &initialValue, TTGetterMethod getter, TTSetterMethod setter)
 Register an attribute property. More...
 
TTErr registerMessageProperty (const TTSymbol messageName, const TTSymbol propertyName, const TTValue &initialValue, TTGetterMethod getter, TTSetterMethod setter)
 Register an message property. More...
 
TTErr getAttribute (const TTSymbol name, TTAttributePtr *attributeObject)
 Search for and locate an attribute. More...
 
TTErr getMessage (const TTSymbol name, TTMessagePtr *messageObject)
 Search for and locate a message. More...
 
void getAttributeNames (TTValue &attributeNameList)
 Return a list of names of the available attributes. More...
 
void getMessageNames (TTValue &messageNameList)
 Return a list of names of the available messages. More...
 
TTSymbol getName () const
 Return the name of this class. More...
 
TTErr registerMessage (const TTSymbol name, TTMethod method)
 Register a message with this object. More...
 
TTErr findMessage (const TTSymbol name, TTMessage **message)
 Find a message registered with this object. More...
 
TTErr removeMessage (const TTSymbol name)
 Remove a message. More...
 
TTErr sendMessage (const TTSymbol name)
 TODO: Document this function. More...
 
TTErr registerObserverForMessage (const TTObject &observingObject, const TTSymbol messageName)
 Register an observer for a message. More...
 
TTErr registerObserverForAttribute (const TTObject &observingObject, const TTSymbol attributeName)
 Register an observer for an attribute. More...
 
TTErr registerObserverForNotifications (const TTObject &observingObject)
 Register an observer. More...
 
TTErr unregisterObserverForMessage (const TTObject &observingObject, const TTSymbol messageName)
 Unregister an observer for a message. More...
 
TTErr unregisterObserverForAttribute (const TTObject &observingObject, const TTSymbol attributeName)
 Unregister an observer for an attribute. More...
 
TTErr unregisterObserverForNotifications (const TTObject &observingObject)
 Unregister an observer for notifications. More...
 
TTErr sendNotification (const TTSymbol name, const TTValue &arguments)
 Send a notification. More...
 
TTErr logMessage (TTImmutableCString fmtstring,...)
 Log messages scoped to this object instance. More...
 
TTErr logWarning (TTImmutableCString fmtstring,...)
 Log warnings scoped to this object instance. More...
 
TTErr logError (TTImmutableCString fmtstring,...)
 Log errors scoped to this object instance. More...
 
TTErr logDebug (TTImmutableCString fmtstring,...)
 Log messages (scoped to this object instance) to output only if the basic debugging flag is enabled in the environment. More...
 
TTErr lock ()
 Lock the object in order to ensure thread-safe processing. More...
 
TTErr unlock ()
 Unlock the object when thread-safe processing is no longer required. More...
 
TTBoolean isLocked ()
 Query if the object currently is locked for thread-safe processing. More...
 
TTBoolean waitForLock ()
 If the object is locked (e.g. More...
 
- Public Attributes inherited from TTObjectBase
TTBoolean valid
 If the object isn't completely built, or is in the process of freeing, this will be false.
 
TTBoolean track
 enable the print of a log message to ease the track of reference counting mechanism on an instance More...
 

Detailed Description

TTDCBlock in an audio processor that filters out DC Offsets from an input.

This class is a good example of a very simple audio filter that can process any number of parallel audio channels. It has a couple of simple messages, but defines no attributes beyond those that are inherited by all TTAudioObjectBases.

TTDCBlock implements a first-order highpass filter algorithm that is used pretty much everywhere (STK, ChucK, RTCMix, SuperCollider, Max, Pd, etc), with the difference equation:

y(n) = (1 * x(n))  +  (-1 * x(n-1))  -  (-0.9997 * y(n-1))  ,  n = 0, 1, 2, 3, ...

which can be simplified to:

y(n) = x(n) - x(n-1) + (0.9997 * y(n-1))

and thus characterized by the Z-transform:

Y(z) = X(z)  -  X(z) * z^(-1)  +  Y(z) * 0.9997 * z^(-1)

meaning the transfer function is:

H(z) = [1  -  z^(-1)]  /  [1  +  0.9997 * z^(-1)]

and resulting in the frequency response:

H( e^(i*omega*T) ) = [1  -  e^(-i * omega * T)]  /  [1  +  0.9997 * e^(-i * omega * T)]

where $i$ is the sqrt(-1), e is Euler's log base, T is the sampling interval, and omega is 2*pi*frequency.

In Max, it usually shows up simply as [biquad~ 1.0 -1.0 0.0 -0.9997 0.0]. In other places it usually shows up with the feedback coefficient set to -0.995 (e.g in SuperCollider and in [JOS, 2007, pp 273]). The higher coefficient is desirable so as to not attenuate lowish frequencies in the spectrum, but with the caveat that it also won't respond as quickly to varying amounts DC Offset.

The power is attenuated by -3 dB at a normalized frequency of 0.1612 * pi @ 0.9997. At fs=44100 this translates to cf = 22050 * 0.1612 = 3554.46 Hz.

The power is attenuated by -3 dB at a normalized frequency of 0.1604 * pi @ 0.995. At fs=44100 this translates to cf = 22050 * 0.1604 = 3536.82 Hz.

For reference, in this last case, the power is attenuated by -6 db (magnitude attenuated by -12 dB) @ 0.0798 * pi, which at fs=44100 translates to 1759.59 Hz.


So what are benefits of this filter? In the case of biquad~ in Max, we are processing with 2 coefficients zeroed, meaning that we burn the cpu for a second-order filter when we are only getting the benefit of a first-order filter. Why would we want to do this instead of doing something like a using a Butterworth filter for DC Blocking?

We could implement a nice third-order Butterworth with a cf=20 Hz @ fs=48000 that would not disturb the gain of the spectrum, but... the filter does not have linear phase.

Since we are outside of Max, we can evaluate this filter for what it is: a first-order filter. The clear benefit then is its speed.

  • however, if we used a pure FIR filter then we could potentially use SSE instructions, so: how would a second-order FIR compare to this classical first-order FIR/IIR combination?

Definition at line 84 of file TTDCBlock.h.

Member Function Documentation

TTErr TTDCBlock::calculateValue ( const TTFloat64 x,
TTFloat64 y,
TTPtrSizedInt  channel 
)
inlineprotected

Standard single value calculate method as used by DSP objects.

Definition at line 64 of file TTDCBlock.cpp.

References kTTErrNone, mLastInput, and mLastOutput.

Referenced by processAudio().

TTErr TTDCBlock::clear ( )
protected

Resets the DC-Blocking filter.

This algorithm uses an IIR filter, meaning that it relies on feedback. If the filter should not be producing any signal (such as turning audio off and then back on in a host) or if the feedback has become corrupted (such as might happen if a NaN is fed in) then it may be neccesary to clear the filter by calling this method.

Returns
Returns a TTErr error code.

Definition at line 56 of file TTDCBlock.cpp.

References kTTErrNone, mLastInput, mLastOutput, and TTAudioObjectBase::mMaxNumChannels.

Referenced by updateMaxNumChannels().

TTErr TTDCBlock::processAudio ( TTAudioSignalArrayPtr  inputs,
TTAudioSignalArrayPtr  outputs 
)
protected

Standard audio processing method as used by DSP objects.

Definition at line 74 of file TTDCBlock.cpp.

References calculateValue().

+ Here is the call graph for this function:

TTErr TTDCBlock::updateMaxNumChannels ( const TTValue oldMaxNumChannels,
TTValue  
)
protected

This method gets called when the inherited mMaxNumChannels attribute is changed.

Definition at line 47 of file TTDCBlock.cpp.

References clear(), kTTErrNone, mLastInput, mLastOutput, and TTAudioObjectBase::mMaxNumChannels.

+ Here is the call graph for this function:


The documentation for this class was generated from the following files: