SoftBank Robotics documentation What's new in NAOqi 2.5?

ALAudioDevice API

NAOqi Audio - Overview | API

Namespace : AL

#include <alproxies/alaudiodeviceproxy.h>

Methods

void ALAudioDeviceProxy::disableEnergyComputation()

Disables the computation of the energy on each input channel.

void ALAudioDeviceProxy::enableEnergyComputation()

Enables the computation of the energy on each input channel (this computation is off by default). The result of this computation can be collected by calling:

void ALAudioDeviceProxy::flushAudioOutputs()

This methods clears all the samples that may remain to be sent to the loudspeaker.

float ALAudioDeviceProxy::getFrontMicEnergy()

Returns the signal energy on the front microphone averaged on a 170ms buffer. The computation of the energy must first be enabled with ALAudioDeviceProxy::enableEnergyComputation.

Returns:energy [0,32768]
float ALAudioDeviceProxy::getLeftMicEnergy()

Returns the signal energy on the left microphone averaged on a 170ms buffer. The computation of the energy must first be enabled with ALAudioDeviceProxy::enableEnergyComputation.

Returns:energy [0,32768]
int ALAudioDeviceProxy::getOutputVolume()

Gets the overall output volume of the system. This volume can be set with ALAudioDeviceProxy::setOutputVolume.

Returns:volume [0,100]
int ALAudioDeviceProxy::getParameter(const std::string& parameter)

Deprecated since version 1.22: parameter ‘inputBufferSize’ has no effect now.

This method returns the internal parameter (‘outputSampleRate’). The value -1 is returned if the specified parameter is not valid.

Parameters:
  • parameter – Name of the parameter
Returns:

value
float ALAudioDeviceProxy::getRearMicEnergy()

Returns the signal energy on the rear microphone averaged on a 170ms buffer. The computation of the energy must first be enabled with ALAudioDeviceProxy::enableEnergyComputation.

Returns:energy [0,32768]
float ALAudioDeviceProxy::getRightMicEnergy()

Returns the signal energy on the right microphone averaged on a 170ms buffer. The computation of the energy must first be enabled with ALAudioDeviceProxy::enableEnergyComputation.

Returns:energy [0,32768]
bool ALAudioDeviceProxy::isAudioOutMuted()

Tells if the audio output device has been muted by a call to ALAudioDeviceProxy::muteAudioOut.

Returns:True if the output device has been muted, False otherwise.
void ALAudioDeviceProxy::muteAudioOut(const bool& mute)

Mutes the output device.

Parameters:
  • mute – True to mute, False to unmute.
void ALAudioDeviceProxy::openAudioInputs()

Deprecated since version 1.22: due to technical improvements, closing the audio is not necessary anymore.

Opens the audio device used for capture by ALAudioDevice. This device is open by default. This method is however useful if you used ALAudioDeviceProxy::closeAudioInputs before.

void ALAudioDeviceProxy::openAudioOutputs()

Deprecated since version 1.22: due to technical improvements, closing the audio is not necessary anymore.

Opens the audio device used for playback by ALAudioDevice. This device is open by default. This method is however useful if you used ALAudioDeviceProxy::closeAudioOutputs before.

void ALAudioDeviceProxy::playSine(const int& frequence, const int& gain, const int& pan, const float& duration)

Play a sine wave which the specified characteristics.

Parameters:
  • frequence – Frequence in Hertz
  • gain – Volume between 0 and 100
  • pan – Stereo Pan set to either {-1,0,+1}
  • duration – Duration of the sine wave in seconds
bool ALAudioDeviceProxy::sendLocalBufferToOutput(const int& nbOfFrames, const int& buffer)

By using this method a local module (a module running on the robot) can send a signal to the robot’s loudspeakers. This signal should be formated as a 16 bits stereo interleaved buffer. The size of this buffer should also not exceed 16384.

Parameters:
  • nbOfFrames – Number of stereo frames contained in the buffer
  • buffer – Memory address of the buffer to send.
Returns:

True if the operation is successful - False otherwise

Warning

This interface can crash the robot if not use correctly! Be very careful with the second parameter.

// Example: Play a 1kHz sine wav on the robot's loudspeakers during 2s

int sineTableSize = 1024;
int outputBufferSize = 16384;
float sine[sineTableSize];
int left_phase = 0;
int right_phase = 0;
int nbOfOutputChannels=2;
int sampleRate = 48000;
int freq =1000;
float duration = 2.0;
boost::shared_ptr<ALProxy> audiodeviceProxy = getParentBroker()->getProxy("ALAudioDevice");

signed short fOutputBuffer[nbOfOutputChannels*outputBufferSize];
for ( int i=0; i<sineTableSize; i++ )
{
  sine[i] = float(32767) * sin((double(i)/double(sineTableSize))*M_PI*2.0);
}

float ratio= 1.0 /( 1.0 / float(freq) * float(sampleRate) / float(sineTableSize));
double inc= std::ceil(ratio);
int nbOfBuffersToSend = std::ceil(duration*sampleRate/outputBufferSize);

for (int d=0;d<nbOfBuffersToSend;d++)
{
  for (int i=0;i<nbOfOutputChannels*outputBufferSize;i+=nbOfOutputChannels)
  {
    // left channel
    fOutputBuffer[i] = (signed short)(float(sine[left_phase]));
    // right channel
    fOutputBuffer[i+1] = (signed short)(float(sine[right_phase]));
    left_phase += inc;
    if ( left_phase >= sineTableSize ) left_phase -= sineTableSize;
    right_phase += inc;
    if ( right_phase >= sineTableSize ) right_phase -= sineTableSize;
  }
  audiodeviceProxy->callVoid("sendLocalBufferToOutput",
                             outputBufferSize,
                             (int) &fOutputBuffer);
}

def onInput_onStart(self):
    self.isRunning = True
    self.ad = ALProxy("ALAudioDevice")
    self.addedPaths = [self.behaviorAbsolutePath()]
    for addedPath in self.addedPaths:
        sys.path.append(addedPath)
    from pydub import AudioSegment
    songPath = os.path.join(self.behaviorAbsolutePath(), "audio_file.wav")
    self.song = AudioSegment.from_wav(songPath)
    try:
        # here is important to note that the second parameter is contigus memory audio data!
        self.ad.sendLocalBufferToOutput(int(self.song.frame_count()), id(self.song._data))
    except Exception as e:
        self.log("error for buffer: " + str(e) + str(id(self.song.data)))
    self.onUnload()
    pass
bool ALAudioDeviceProxy::sendRemoteBufferToOutput(const int& nbOfFrames, const AL::ALValue& buffer)

By using this method a remote module (a module running outside the robot) can send a signal to the robot’s loudspeakers. This signal should be formated as a 16 bits stereo interleaved buffer. The size of this buffer should also not exceed 16384.

Parameters:
  • nbOfFrames – Number of stereo frames contained in the buffer
  • buffer – Buffer to send
Returns:

True if the operation is successful - False otherwise
void ALAudioDeviceProxy::setClientPreferences(const std::string& name, const int& sampleRate, const int& channels, const int& deinterleaved)

By using this method a module can specify the format of the signal that will be sent after subscribing to ALAudioDevice. If no call to this method is made, the default format sent to this module is 4 channels interleaved at 48000Hz. Note that for now, only the following combinations are available:

  • four channels interleaved, 48000Hz, (default setting)
  • four channels deinterleaved, 48000Hz
  • one channels (either front, rear, left or right), 16000Hz

This call must be made before the call to ALAudioDeviceProxy::subscribe or startDetection() to be taken into account.

// This shows how setClientPreferences can be used
boost::shared_ptr<ALProxy> proxyAudioDevice = getParentBroker()->getProxy("ALAudioDevice");
proxyAudioDevice->callVoid("setClientPreferences",
                           getName(),                //Name of this module
                           48000,                    //48000 Hz requested
                           (int)ALLCHANNELS,         //4 Channels requested
                           1                         //Deinterleaving requested
                           );
Parameters:
  • name – Name of the module. This name must be the same that the one used in subscribe(const std::string& module).
  • sampleRate – Sample rate requested can either be 16000Hz or 48000Hz.
  • channels – Channel configuration requested can be either AL::ALLCHANNELS, AL::FRONTCHANNEL, AL::LEFTCHANNEL, AL::RIGHTCHANNEL or AL::REARCHANNEL.
  • deinterleaved – Only relevant if AL::ALLCHANNEL is requested.
void ALAudioDeviceProxy::setFileAsInput(const std::string& fileName)

This method notifies ALAudioDevice that audio inputs should be read in the specified file instead. The specified sound file must be a .wav file containing 48000Hz, 16bits, 4 channels interleaved signals.

Parameters:
  • fileName – Absolute path of the file.
void ALAudioDeviceProxy::setOutputVolume(const int& volume)

Sets the overall output volume of the system. This volume can be collected with ALAudioDeviceProxy::getOutputVolume.

Parameters:
  • volume – Volume [0-100]
void ALAudioDeviceProxy::setParameter(const std::string& parameter, const int& value)

Deprecated since version 1.22: parameter ‘inputBufferSize’ has no effect now.

This method sets the internal parameter (‘outputSampleRate’).

Parameters:
  • parameter – ‘outputSampleRate’ can be set to either 16000Hz, 22050Hz, 44100Hz or 48000Hz.
  • value – The value to which the specified parameter should be set.
void ALAudioDeviceProxy::startMicrophonesRecording(const std::string& fileName)

This method records the signal collected on the microphones directly into the specified file.

If the extension of the specified file is ‘wav’, data will be recored as a 16 bits, 48000Hz, 4 channels wav file.

If the extension of the specified file is ‘ogg’, data will be recored as a 16 bits, 16000Hz, 1 channels ogg file.

Parameters:
  • fileName – Absolute path of the file
void ALAudioDeviceProxy::stopMicrophonesRecording()

This method stops the recording started by the call to ALAudioDeviceProxy::startMicrophonesRecording.

void ALAudioDeviceProxy::subscribe(const std::string& module)

This function allows a module which inherits from the ALSoundExtractor class to subscribe to the ALAudioDevice module. Once the module is subscribed, the function ‘process’ of the module (the module needs to contain one) will be automatically and regularly called with raw data from microphones as inputs. The call to this method can be replaced by a call to startDetection() within a NAOqi module that inherits from ALSoundExtractor.

The callback function must be declared as follows:

process(const int & nbOfChannels, const int & nbrOfSamplesByChannel, const AL_SOUND_FORMAT * buffer, const ALValue & timeStamp)
Parameters:
  • module – This module must inherits from ALSoundExtractor
void ALAudioDeviceProxy::unsubscribe(const std::string& module)

This function allows a module which inherits from the ALSoundExtractor class to unsubscribe from the ALAudioDevice module. This stops the regular calls to the module callback “process”. The call to this method can be replaced by a call to stopDetection() within a NAOqi module that inherits from ALSoundExtractor.

Parameters:
  • module – This module must inherits from ALSoundExtractor
void ALAudioDeviceProxy::closeAudioInputs()

Deprecated since version 1.22: due to technical improvements, closing the audio is not necessary anymore.

Closes the audio device used for capture by ALAudioDevice. This method is useful to allow an external program (e.g. arecord) to access the input device of the robot. Note however that a call to this method will prevent every functionality that uses microphones to work correctly (e.g. ALSpeechRecognition, ALAudioSourceLocalization, etc...).

void ALAudioDeviceProxy::closeAudioOutputs()

Deprecated since version 1.22: due to technical improvements, closing the audio is not necessary anymore.

Closes the audio device used for playback by ALAudioDevice. This method is useful to allow an external program (e.g. aplay) to access the output device of the robot. Note however that a call to this method will prevent every functionality that uses microphones to work correctly (e.g. ALTextToSpeech).

bool ALAudioDeviceProxy::isInputClosed()

Deprecated since version 1.22: due to technical improvements, closing the audio is not necessary anymore.

Tells if the audio input device has been closed by a call to ALAudioDeviceProxy::closeAudioInputs.

Returns:True if the input device has been closed, False otherwise.
bool ALAudioDeviceProxy::isOutputClosed()

Deprecated since version 1.22: due to technical improvements, closing the audio is not necessary anymore.

Tells if the audio output device has been closed by a call to ALAudioDeviceProxy::closeAudioOutputs.

Returns:True if the output device has been closed, False otherwise.

Signals

qi::Signal<bool> ALAudioDeviceProxy::speakersPlaying

Signal emitted every 130 ms while the robot is playing sound.