summaryrefslogtreecommitdiff
path: root/dom/media/AudioCaptureStream.cpp
blob: f72d025a0e0c5b3a9b4149df4db13abd453f6513 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
/* -*- Mode: C++; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*-*/
/* This Source Code Form is subject to the terms of the Mozilla Public
 * License, v. 2.0. If a copy of the MPL was not distributed with this file,
 * You can obtain one at http://mozilla.org/MPL/2.0/. */

#include "MediaStreamGraphImpl.h"
#include "MediaStreamListener.h"
#include "mozilla/MathAlgorithms.h"
#include "mozilla/Unused.h"

#include "AudioSegment.h"
#include "mozilla/Logging.h"
#include "mozilla/Attributes.h"
#include "AudioCaptureStream.h"
#include "ImageContainer.h"
#include "AudioNodeEngine.h"
#include "AudioNodeStream.h"
#include "AudioNodeExternalInputStream.h"
#include "webaudio/MediaStreamAudioDestinationNode.h"
#include <algorithm>
#include "DOMMediaStream.h"

using namespace mozilla::layers;
using namespace mozilla::dom;
using namespace mozilla::gfx;

namespace mozilla
{

// Mixing to mono.
// NOTE: MONO is a misnomer, it's not a boolean but the number of channels.
static const uint32_t MONO = 1;

AudioCaptureStream::AudioCaptureStream(TrackID aTrackId)
  : ProcessedMediaStream(), mTrackId(aTrackId), mStarted(false), mTrackCreated(false)
{
  MOZ_ASSERT(NS_IsMainThread());
  MOZ_COUNT_CTOR(AudioCaptureStream);
  mMixer.AddCallback(this);
}

AudioCaptureStream::~AudioCaptureStream()
{
  MOZ_COUNT_DTOR(AudioCaptureStream);
  mMixer.RemoveCallback(this);
}

void
AudioCaptureStream::Start()
{
  class Message : public ControlMessage {
  public:
    explicit Message(AudioCaptureStream* aStream)
      : ControlMessage(aStream), mStream(aStream) {}

    virtual void Run()
    {
      mStream->mStarted = true;
    }

  protected:
    AudioCaptureStream* mStream;
  };
  GraphImpl()->AppendMessage(MakeUnique<Message>(this));
}

void
AudioCaptureStream::ProcessInput(GraphTime aFrom, GraphTime aTo,
                                 uint32_t aFlags)
{
  if (!mStarted) {
    return;
  }

  uint32_t inputCount = mInputs.Length();
  StreamTracks::Track* track = EnsureTrack(mTrackId);
  // Notify the DOM everything is in order.
  if (!mTrackCreated) {
    for (uint32_t i = 0; i < mListeners.Length(); i++) {
      MediaStreamListener* l = mListeners[i];
      AudioSegment tmp;
      l->NotifyQueuedTrackChanges(
        Graph(), mTrackId, 0, TrackEventCommand::TRACK_EVENT_CREATED, tmp);
      l->NotifyFinishedTrackCreation(Graph());
    }
    mTrackCreated = true;
  }

  if (IsFinishedOnGraphThread()) {
    return;
  }

  // If the captured stream is connected back to a object on the page (be it an
  // HTMLMediaElement with a stream as source, or an AudioContext), a cycle
  // situation occur. This can work if it's an AudioContext with at least one
  // DelayNode, but the MSG will mute the whole cycle otherwise.
  if (InMutedCycle() || inputCount == 0) {
    track->Get<AudioSegment>()->AppendNullData(aTo - aFrom);
  } else {
    // We mix down all the tracks of all inputs, to a stereo track. Everything
    // is {up,down}-mixed to stereo.
    mMixer.StartMixing();
    AudioSegment output;
    for (uint32_t i = 0; i < inputCount; i++) {
      MediaStream* s = mInputs[i]->GetSource();
      StreamTracks::TrackIter track(s->GetStreamTracks(), MediaSegment::AUDIO);
      if (track.IsEnded()) {
        // No tracks for this input. Still we append data to trigger the mixer.
        AudioSegment toMix;
        toMix.AppendNullData(aTo - aFrom);
        toMix.Mix(mMixer, MONO, Graph()->GraphRate());
      }
      for (; !track.IsEnded(); track.Next()) {
        AudioSegment* inputSegment = track->Get<AudioSegment>();
        StreamTime inputStart = s->GraphTimeToStreamTimeWithBlocking(aFrom);
        StreamTime inputEnd = s->GraphTimeToStreamTimeWithBlocking(aTo);
        AudioSegment toMix;
        if (track->IsEnded() && inputSegment->GetDuration() <= inputStart) {
          toMix.AppendNullData(aTo - aFrom);
        } else {
          toMix.AppendSlice(*inputSegment, inputStart, inputEnd);
          // Care for streams blocked in the [aTo, aFrom] range.
          if (inputEnd - inputStart < aTo - aFrom) {
            toMix.AppendNullData((aTo - aFrom) - (inputEnd - inputStart));
          }
        }
        toMix.Mix(mMixer, MONO, Graph()->GraphRate());
      }
    }
    // This calls MixerCallback below
    mMixer.FinishMixing();
  }

  // Regardless of the status of the input tracks, we go foward.
  mTracks.AdvanceKnownTracksTime(GraphTimeToStreamTimeWithBlocking((aTo)));
}

void
AudioCaptureStream::MixerCallback(AudioDataValue* aMixedBuffer,
                                  AudioSampleFormat aFormat, uint32_t aChannels,
                                  uint32_t aFrames, uint32_t aSampleRate)
{
  AutoTArray<nsTArray<AudioDataValue>, MONO> output;
  AutoTArray<const AudioDataValue*, MONO> bufferPtrs;
  output.SetLength(MONO);
  bufferPtrs.SetLength(MONO);

  uint32_t written = 0;
  // We need to copy here, because the mixer will reuse the storage, we should
  // not hold onto it. Buffers are in planar format.
  for (uint32_t channel = 0; channel < aChannels; channel++) {
    AudioDataValue* out = output[channel].AppendElements(aFrames);
    PodCopy(out, aMixedBuffer + written, aFrames);
    bufferPtrs[channel] = out;
    written += aFrames;
  }
  AudioChunk chunk;
  chunk.mBuffer = new mozilla::SharedChannelArrayBuffer<AudioDataValue>(&output);
  chunk.mDuration = aFrames;
  chunk.mBufferFormat = aFormat;
  chunk.mVolume = 1.0f;
  chunk.mChannelData.SetLength(MONO);
  for (uint32_t channel = 0; channel < aChannels; channel++) {
    chunk.mChannelData[channel] = bufferPtrs[channel];
  }

  // Now we have mixed data, simply append it to out track.
  EnsureTrack(mTrackId)->Get<AudioSegment>()->AppendAndConsumeChunk(&chunk);
}
}