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
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
|
/* -*- Mode: C++; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
/* vim:set ts=2 sw=2 sts=2 et cindent: */
/* 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 "mozilla/TaskQueue.h"
#include "FFmpegAudioDecoder.h"
#include "TimeUnits.h"
#define MAX_CHANNELS 16
namespace mozilla
{
FFmpegAudioDecoder<LIBAV_VER>::FFmpegAudioDecoder(FFmpegLibWrapper* aLib,
TaskQueue* aTaskQueue, MediaDataDecoderCallback* aCallback,
const AudioInfo& aConfig)
: FFmpegDataDecoder(aLib, aTaskQueue, aCallback, GetCodecId(aConfig.mMimeType))
{
MOZ_COUNT_CTOR(FFmpegAudioDecoder);
// Use a new MediaByteBuffer as the object will be modified during initialization.
if (aConfig.mCodecSpecificConfig && aConfig.mCodecSpecificConfig->Length()) {
mExtraData = new MediaByteBuffer;
mExtraData->AppendElements(*aConfig.mCodecSpecificConfig);
}
}
RefPtr<MediaDataDecoder::InitPromise>
FFmpegAudioDecoder<LIBAV_VER>::Init()
{
nsresult rv = InitDecoder();
return rv == NS_OK ? InitPromise::CreateAndResolve(TrackInfo::kAudioTrack, __func__)
: InitPromise::CreateAndReject(NS_ERROR_DOM_MEDIA_FATAL_ERR, __func__);
}
void
FFmpegAudioDecoder<LIBAV_VER>::InitCodecContext()
{
MOZ_ASSERT(mCodecContext);
// We do not want to set this value to 0 as FFmpeg by default will
// use the number of cores, which with our mozlibavutil get_cpu_count
// isn't implemented.
mCodecContext->thread_count = 1;
// FFmpeg takes this as a suggestion for what format to use for audio samples.
// LibAV 0.8 produces rubbish float interleaved samples, request 16 bits audio.
mCodecContext->request_sample_fmt =
(mLib->mVersion == 53) ? AV_SAMPLE_FMT_S16 : AV_SAMPLE_FMT_FLT;
}
static AlignedAudioBuffer
CopyAndPackAudio(AVFrame* aFrame, uint32_t aNumChannels, uint32_t aNumAFrames)
{
MOZ_ASSERT(aNumChannels <= MAX_CHANNELS);
AlignedAudioBuffer audio(aNumChannels * aNumAFrames);
if (!audio) {
return audio;
}
if (aFrame->format == AV_SAMPLE_FMT_FLT) {
// Audio data already packed. No need to do anything other than copy it
// into a buffer we own.
memcpy(audio.get(), aFrame->data[0],
aNumChannels * aNumAFrames * sizeof(AudioDataValue));
} else if (aFrame->format == AV_SAMPLE_FMT_FLTP) {
// Planar audio data. Pack it into something we can understand.
AudioDataValue* tmp = audio.get();
AudioDataValue** data = reinterpret_cast<AudioDataValue**>(aFrame->data);
for (uint32_t frame = 0; frame < aNumAFrames; frame++) {
for (uint32_t channel = 0; channel < aNumChannels; channel++) {
*tmp++ = data[channel][frame];
}
}
} else if (aFrame->format == AV_SAMPLE_FMT_S16) {
// Audio data already packed. Need to convert from S16 to 32 bits Float
AudioDataValue* tmp = audio.get();
int16_t* data = reinterpret_cast<int16_t**>(aFrame->data)[0];
for (uint32_t frame = 0; frame < aNumAFrames; frame++) {
for (uint32_t channel = 0; channel < aNumChannels; channel++) {
*tmp++ = AudioSampleToFloat(*data++);
}
}
} else if (aFrame->format == AV_SAMPLE_FMT_S16P) {
// Planar audio data. Convert it from S16 to 32 bits float
// and pack it into something we can understand.
AudioDataValue* tmp = audio.get();
int16_t** data = reinterpret_cast<int16_t**>(aFrame->data);
for (uint32_t frame = 0; frame < aNumAFrames; frame++) {
for (uint32_t channel = 0; channel < aNumChannels; channel++) {
*tmp++ = AudioSampleToFloat(data[channel][frame]);
}
}
} else if (aFrame->format == AV_SAMPLE_FMT_S32) {
// Audio data already packed. Need to convert from S16 to 32 bits Float
AudioDataValue* tmp = audio.get();
int32_t* data = reinterpret_cast<int32_t**>(aFrame->data)[0];
for (uint32_t frame = 0; frame < aNumAFrames; frame++) {
for (uint32_t channel = 0; channel < aNumChannels; channel++) {
*tmp++ = AudioSampleToFloat(*data++);
}
}
} else if (aFrame->format == AV_SAMPLE_FMT_S32P) {
// Planar audio data. Convert it from S32 to 32 bits float
// and pack it into something we can understand.
AudioDataValue* tmp = audio.get();
int32_t** data = reinterpret_cast<int32_t**>(aFrame->data);
for (uint32_t frame = 0; frame < aNumAFrames; frame++) {
for (uint32_t channel = 0; channel < aNumChannels; channel++) {
*tmp++ = AudioSampleToFloat(data[channel][frame]);
}
}
}
return audio;
}
MediaResult
FFmpegAudioDecoder<LIBAV_VER>::DoDecode(MediaRawData* aSample)
{
AVPacket packet;
mLib->av_init_packet(&packet);
packet.data = const_cast<uint8_t*>(aSample->Data());
packet.size = aSample->Size();
if (!PrepareFrame()) {
return MediaResult(
NS_ERROR_OUT_OF_MEMORY,
RESULT_DETAIL("FFmpeg audio decoder failed to allocate frame"));
}
int64_t samplePosition = aSample->mOffset;
media::TimeUnit pts = media::TimeUnit::FromMicroseconds(aSample->mTime);
while (packet.size > 0) {
int decoded;
int bytesConsumed =
mLib->avcodec_decode_audio4(mCodecContext, mFrame, &decoded, &packet);
if (bytesConsumed < 0) {
NS_WARNING("FFmpeg audio decoder error.");
return MediaResult(NS_ERROR_DOM_MEDIA_DECODE_ERR,
RESULT_DETAIL("FFmpeg audio error:%d", bytesConsumed));
}
if (decoded) {
if (mFrame->format != AV_SAMPLE_FMT_FLT &&
mFrame->format != AV_SAMPLE_FMT_FLTP &&
mFrame->format != AV_SAMPLE_FMT_S16 &&
mFrame->format != AV_SAMPLE_FMT_S16P &&
mFrame->format != AV_SAMPLE_FMT_S32 &&
mFrame->format != AV_SAMPLE_FMT_S32P) {
return MediaResult(
NS_ERROR_DOM_MEDIA_DECODE_ERR,
RESULT_DETAIL(
"FFmpeg audio decoder outputs unsupported audio format"));
}
uint32_t numChannels = mCodecContext->channels;
AudioConfig::ChannelLayout layout(numChannels);
if (!layout.IsValid()) {
return MediaResult(
NS_ERROR_DOM_MEDIA_FATAL_ERR,
RESULT_DETAIL("Unsupported channel layout:%u", numChannels));
}
uint32_t samplingRate = mCodecContext->sample_rate;
AlignedAudioBuffer audio =
CopyAndPackAudio(mFrame, numChannels, mFrame->nb_samples);
if (!audio) {
return MediaResult(NS_ERROR_OUT_OF_MEMORY, __func__);
}
media::TimeUnit duration =
FramesToTimeUnit(mFrame->nb_samples, samplingRate);
if (!duration.IsValid()) {
return MediaResult(
NS_ERROR_DOM_MEDIA_OVERFLOW_ERR,
RESULT_DETAIL("Invalid sample duration"));
}
RefPtr<AudioData> data = new AudioData(samplePosition,
pts.ToMicroseconds(),
duration.ToMicroseconds(),
mFrame->nb_samples,
Move(audio),
numChannels,
samplingRate);
mCallback->Output(data);
pts += duration;
if (!pts.IsValid()) {
return MediaResult(
NS_ERROR_DOM_MEDIA_OVERFLOW_ERR,
RESULT_DETAIL("Invalid count of accumulated audio samples"));
}
}
packet.data += bytesConsumed;
packet.size -= bytesConsumed;
samplePosition += bytesConsumed;
}
return NS_OK;
}
void
FFmpegAudioDecoder<LIBAV_VER>::ProcessDrain()
{
ProcessFlush();
mCallback->DrainComplete();
}
AVCodecID
FFmpegAudioDecoder<LIBAV_VER>::GetCodecId(const nsACString& aMimeType)
{
if (aMimeType.EqualsLiteral("audio/mpeg")) {
return AV_CODEC_ID_MP3;
} else if (aMimeType.EqualsLiteral("audio/flac")) {
return AV_CODEC_ID_FLAC;
} else if (aMimeType.EqualsLiteral("audio/mp4a-latm")) {
return AV_CODEC_ID_AAC;
}
return AV_CODEC_ID_NONE;
}
FFmpegAudioDecoder<LIBAV_VER>::~FFmpegAudioDecoder()
{
MOZ_COUNT_DTOR(FFmpegAudioDecoder);
}
} // namespace mozilla
|
