blob: 9242f88a636ad91e1d7aefb8e91b301e19709a0c [file] [log] [blame]
/*
* Copyright (C) 2010, Google Inc. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY APPLE INC. AND ITS CONTRIBUTORS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
* WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
* DISCLAIMED. IN NO EVENT SHALL APPLE INC. OR ITS CONTRIBUTORS BE LIABLE FOR ANY
* DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
* (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
* LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON
* ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
* SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#include "config.h"
#include "PannerNode.h"
#if ENABLE(WEB_AUDIO)
#include "AudioBufferSourceNode.h"
#include "AudioBus.h"
#include "AudioContext.h"
#include "AudioNodeInput.h"
#include "AudioNodeOutput.h"
#include "HRTFDatabaseLoader.h"
#include "HRTFPanner.h"
#include "ScriptExecutionContext.h"
#include <wtf/IsoMallocInlines.h>
#include <wtf/MathExtras.h>
namespace WebCore {
WTF_MAKE_ISO_ALLOCATED_IMPL(PannerNode);
static void fixNANs(double &x)
{
if (std::isnan(x) || std::isinf(x))
x = 0.0;
}
PannerNode::PannerNode(AudioContext& context, float sampleRate)
: AudioNode(context, sampleRate)
, m_panningModel(PanningModelType::HRTF)
, m_lastGain(-1.0)
, m_connectionCount(0)
{
setNodeType(NodeTypePanner);
// Load the HRTF database asynchronously so we don't block the Javascript thread while creating the HRTF database.
m_hrtfDatabaseLoader = HRTFDatabaseLoader::createAndLoadAsynchronouslyIfNecessary(context.sampleRate());
addInput(makeUnique<AudioNodeInput>(this));
addOutput(makeUnique<AudioNodeOutput>(this, 2));
// Node-specific default mixing rules.
m_channelCount = 2;
m_channelCountMode = ClampedMax;
m_channelInterpretation = AudioBus::Speakers;
m_distanceGain = AudioParam::create(context, "distanceGain", 1.0, 0.0, 1.0);
m_coneGain = AudioParam::create(context, "coneGain", 1.0, 0.0, 1.0);
m_position = FloatPoint3D(0, 0, 0);
m_orientation = FloatPoint3D(1, 0, 0);
m_velocity = FloatPoint3D(0, 0, 0);
initialize();
}
PannerNode::~PannerNode()
{
uninitialize();
}
void PannerNode::pullInputs(size_t framesToProcess)
{
// We override pullInputs(), so we can detect new AudioSourceNodes which have connected to us when new connections are made.
// These AudioSourceNodes need to be made aware of our existence in order to handle doppler shift pitch changes.
if (m_connectionCount != context().connectionCount()) {
m_connectionCount = context().connectionCount();
// Recursively go through all nodes connected to us.
HashSet<AudioNode*> visitedNodes;
notifyAudioSourcesConnectedToNode(this, visitedNodes);
}
AudioNode::pullInputs(framesToProcess);
}
void PannerNode::process(size_t framesToProcess)
{
AudioBus* destination = output(0)->bus();
if (!isInitialized() || !input(0)->isConnected() || !m_panner.get()) {
destination->zero();
return;
}
AudioBus* source = input(0)->bus();
if (!source) {
destination->zero();
return;
}
// HRTFDatabase should be loaded before proceeding for offline audio context when panningModel() is "HRTF".
if (panningModel() == PanningModelType::HRTF && !m_hrtfDatabaseLoader->isLoaded()) {
if (context().isOfflineContext())
m_hrtfDatabaseLoader->waitForLoaderThreadCompletion();
else {
destination->zero();
return;
}
}
// The audio thread can't block on this lock, so we use std::try_to_lock instead.
std::unique_lock<Lock> lock(m_pannerMutex, std::try_to_lock);
if (!lock.owns_lock()) {
// Too bad - The try_lock() failed. We must be in the middle of changing the panner.
destination->zero();
return;
}
// Apply the panning effect.
double azimuth;
double elevation;
getAzimuthElevation(&azimuth, &elevation);
m_panner->pan(azimuth, elevation, source, destination, framesToProcess);
// Get the distance and cone gain.
double totalGain = distanceConeGain();
// Snap to desired gain at the beginning.
if (m_lastGain == -1.0)
m_lastGain = totalGain;
// Apply gain in-place with de-zippering.
destination->copyWithGainFrom(*destination, &m_lastGain, totalGain);
}
void PannerNode::reset()
{
m_lastGain = -1.0; // force to snap to initial gain
if (m_panner.get())
m_panner->reset();
}
void PannerNode::initialize()
{
if (isInitialized())
return;
m_panner = Panner::create(m_panningModel, sampleRate(), m_hrtfDatabaseLoader.get());
AudioNode::initialize();
}
void PannerNode::uninitialize()
{
if (!isInitialized())
return;
m_panner = nullptr;
AudioNode::uninitialize();
}
AudioListener* PannerNode::listener()
{
return context().listener();
}
void PannerNode::setPanningModel(PanningModelType model)
{
if (!m_panner.get() || model != m_panningModel) {
// This synchronizes with process().
std::lock_guard<Lock> lock(m_pannerMutex);
m_panner = Panner::create(model, sampleRate(), m_hrtfDatabaseLoader.get());
m_panningModel = model;
}
}
DistanceModelType PannerNode::distanceModel() const
{
return const_cast<PannerNode*>(this)->m_distanceEffect.model();
}
void PannerNode::setDistanceModel(DistanceModelType model)
{
m_distanceEffect.setModel(model, true);
}
void PannerNode::getAzimuthElevation(double* outAzimuth, double* outElevation)
{
// FIXME: we should cache azimuth and elevation (if possible), so we only re-calculate if a change has been made.
double azimuth = 0.0;
// Calculate the source-listener vector
FloatPoint3D listenerPosition = listener()->position();
FloatPoint3D sourceListener = m_position - listenerPosition;
if (sourceListener.isZero()) {
// degenerate case if source and listener are at the same point
*outAzimuth = 0.0;
*outElevation = 0.0;
return;
}
sourceListener.normalize();
// Align axes
FloatPoint3D listenerFront = listener()->orientation();
FloatPoint3D listenerUp = listener()->upVector();
FloatPoint3D listenerRight = listenerFront.cross(listenerUp);
listenerRight.normalize();
FloatPoint3D listenerFrontNorm = listenerFront;
listenerFrontNorm.normalize();
FloatPoint3D up = listenerRight.cross(listenerFrontNorm);
float upProjection = sourceListener.dot(up);
FloatPoint3D projectedSource = sourceListener - upProjection * up;
projectedSource.normalize();
azimuth = 180.0 * acos(projectedSource.dot(listenerRight)) / piDouble;
fixNANs(azimuth); // avoid illegal values
// Source in front or behind the listener
double frontBack = projectedSource.dot(listenerFrontNorm);
if (frontBack < 0.0)
azimuth = 360.0 - azimuth;
// Make azimuth relative to "front" and not "right" listener vector
if ((azimuth >= 0.0) && (azimuth <= 270.0))
azimuth = 90.0 - azimuth;
else
azimuth = 450.0 - azimuth;
// Elevation
double elevation = 90.0 - 180.0 * acos(sourceListener.dot(up)) / piDouble;
fixNANs(elevation); // avoid illegal values
if (elevation > 90.0)
elevation = 180.0 - elevation;
else if (elevation < -90.0)
elevation = -180.0 - elevation;
if (outAzimuth)
*outAzimuth = azimuth;
if (outElevation)
*outElevation = elevation;
}
float PannerNode::dopplerRate()
{
double dopplerShift = 1.0;
// FIXME: optimize for case when neither source nor listener has changed...
double dopplerFactor = listener()->dopplerFactor();
if (dopplerFactor > 0.0) {
double speedOfSound = listener()->speedOfSound();
const FloatPoint3D &sourceVelocity = m_velocity;
const FloatPoint3D &listenerVelocity = listener()->velocity();
// Don't bother if both source and listener have no velocity
bool sourceHasVelocity = !sourceVelocity.isZero();
bool listenerHasVelocity = !listenerVelocity.isZero();
if (sourceHasVelocity || listenerHasVelocity) {
// Calculate the source to listener vector
FloatPoint3D listenerPosition = listener()->position();
FloatPoint3D sourceToListener = m_position - listenerPosition;
double sourceListenerMagnitude = sourceToListener.length();
double listenerProjection = sourceToListener.dot(listenerVelocity) / sourceListenerMagnitude;
double sourceProjection = sourceToListener.dot(sourceVelocity) / sourceListenerMagnitude;
listenerProjection = -listenerProjection;
sourceProjection = -sourceProjection;
double scaledSpeedOfSound = speedOfSound / dopplerFactor;
listenerProjection = std::min(listenerProjection, scaledSpeedOfSound);
sourceProjection = std::min(sourceProjection, scaledSpeedOfSound);
dopplerShift = ((speedOfSound - dopplerFactor * listenerProjection) / (speedOfSound - dopplerFactor * sourceProjection));
fixNANs(dopplerShift); // avoid illegal values
// Limit the pitch shifting to 4 octaves up and 3 octaves down.
if (dopplerShift > 16.0)
dopplerShift = 16.0;
else if (dopplerShift < 0.125)
dopplerShift = 0.125;
}
}
return static_cast<float>(dopplerShift);
}
float PannerNode::distanceConeGain()
{
FloatPoint3D listenerPosition = listener()->position();
double listenerDistance = m_position.distanceTo(listenerPosition);
double distanceGain = m_distanceEffect.gain(listenerDistance);
m_distanceGain->setValue(static_cast<float>(distanceGain));
// FIXME: could optimize by caching coneGain
double coneGain = m_coneEffect.gain(m_position, m_orientation, listenerPosition);
m_coneGain->setValue(static_cast<float>(coneGain));
return float(distanceGain * coneGain);
}
void PannerNode::notifyAudioSourcesConnectedToNode(AudioNode* node, HashSet<AudioNode*>& visitedNodes)
{
ASSERT(node);
if (!node)
return;
// First check if this node is an AudioBufferSourceNode. If so, let it know about us so that doppler shift pitch can be taken into account.
if (node->nodeType() == NodeTypeAudioBufferSource) {
AudioBufferSourceNode* bufferSourceNode = reinterpret_cast<AudioBufferSourceNode*>(node);
bufferSourceNode->setPannerNode(this);
} else {
// Go through all inputs to this node.
for (unsigned i = 0; i < node->numberOfInputs(); ++i) {
AudioNodeInput* input = node->input(i);
// For each input, go through all of its connections, looking for AudioBufferSourceNodes.
for (unsigned j = 0; j < input->numberOfRenderingConnections(); ++j) {
AudioNodeOutput* connectedOutput = input->renderingOutput(j);
AudioNode* connectedNode = connectedOutput->node();
if (visitedNodes.contains(connectedNode))
continue;
visitedNodes.add(connectedNode);
notifyAudioSourcesConnectedToNode(connectedNode, visitedNodes);
}
}
}
}
} // namespace WebCore
#endif // ENABLE(WEB_AUDIO)