void
PacketTest :: testCopyPacket()
{
packet = IPacket::make();
VS_TUT_ENSURE("was able to allocate packet", packet);
// everything else should be garbage.
int64_t position = packet->getPosition();
VS_TUT_ENSURE("position was not set to -1", position == -1);
position = 4;
packet->setPosition(position);
int64_t dts = 28349762;
packet->setDts(dts);
int64_t pts = 82729373;
packet->setPts(pts);
RefPointer<IRational> timeBase = IRational::make(3,28972);
packet->setTimeBase(timeBase.value());
int32_t streamIndex = 8;
packet->setStreamIndex(streamIndex);
int64_t duration = 28387728;
packet->setDuration(duration);
int64_t convergenceDuration = 283;
packet->setConvergenceDuration(convergenceDuration);
// Now, make a copy
RefPointer<IPacket> newPacket = IPacket::make(packet.value(), false);
VS_TUT_ENSURE("should not be empty", newPacket);
VS_TUT_ENSURE_EQUALS("should equal", position, newPacket->getPosition());
VS_TUT_ENSURE_EQUALS("should equal", pts, newPacket->getPts());
VS_TUT_ENSURE_EQUALS("should equal", dts, newPacket->getDts());
VS_TUT_ENSURE_EQUALS("should equal", streamIndex, newPacket->getStreamIndex());
VS_TUT_ENSURE_EQUALS("should equal", duration, newPacket->getDuration());
VS_TUT_ENSURE_EQUALS("should equal", convergenceDuration,
newPacket->getConvergenceDuration());
RefPointer<IRational> newBase = newPacket->getTimeBase();
VS_TUT_ENSURE("should be equal", newBase->compareTo(timeBase.value()) == 0);
}
void
StreamCoderFaacTest :: testDecodingAndEncodingFaacAudio()
{
RefPointer<IAudioSamples> samples = 0;
RefPointer<IVideoPicture> frame = 0;
RefPointer<IRational> num(0);
RefPointer<IStreamCoder> ic(0);
RefPointer<IStreamCoder> oc(0);
RefPointer<ICodec> newcodec;
newcodec = ICodec::findEncodingCodecByName("libfaac");
if (!newcodec)
{
VS_LOG_ERROR("libfaac not supported by this build; ignoring test");
return;
}
int numSamples = 0;
int numPackets = 0;
int numFrames = 0;
int numKeyFrames = 0;
int retval = -1;
LoggerStack stack;
stack.setGlobalLevel(Logger::LEVEL_WARN, false);
// use the 20 second file to make memcheck faster
h->setupReading("testfile_videoonly_20sec.flv");
RefPointer<IPacket> packet = IPacket::make();
hw->setupWriting("StreamCoderFaacTest_testDecodingAndEncodingFaacAudio.mov", "mpeg4", "libfaac", "mov");
RefPointer<IPacket> opacket = IPacket::make();
VS_TUT_ENSURE("! opacket", opacket);
{
// Let's set up audio first.
ic = h->coders[h->first_input_audio_stream];
oc = hw->coders[hw->first_output_audio_stream];
// Set the output coder correctly.
oc->setSampleRate(ic->getSampleRate());
oc->setChannels(ic->getChannels());
oc->setBitRate(ic->getBitRate());
samples = IAudioSamples::make(1024, ic->getChannels());
VS_TUT_ENSURE("got no samples", samples);
retval = ic->open();
VS_TUT_ENSURE("Could not open input coder", retval >= 0);
retval = oc->open();
VS_TUT_ENSURE("Could not open output coder", retval >= 0);
}
{
// now, let's set up video.
ic = h->coders[h->first_input_video_stream];
oc = hw->coders[hw->first_output_video_stream];
// We're going to set up high quality video
oc->setBitRate(720000);
oc->setGlobalQuality(0);
oc->setFlags(oc->getFlags() | IStreamCoder::FLAG_QSCALE);
oc->setPixelType(ic->getPixelType());
oc->setHeight(ic->getHeight());
oc->setWidth(ic->getWidth());
num = IRational::make(1,15);
oc->setFrameRate(num.value());
num = ic->getTimeBase();
oc->setTimeBase(num.value());
oc->setNumPicturesInGroupOfPictures(
ic->getNumPicturesInGroupOfPictures()
);
// Ask the StreamCoder to guess width and height for us
frame = IVideoPicture::make(
ic->getPixelType(),
-1, -1);
VS_TUT_ENSURE("got no frame", frame);
VS_TUT_ENSURE("should not have width", frame->getWidth() <= 0);
VS_TUT_ENSURE("should not have height", frame->getHeight() <= 0);
retval = ic->open();
VS_TUT_ENSURE("Could not open input coder", retval >= 0);
retval = oc->open();
VS_TUT_ENSURE("Could not open output coder", retval >= 0);
}
// write header
retval = hw->container->writeHeader();
VS_TUT_ENSURE("could not write header", retval >= 0);
while (h->container->readNextPacket(packet.value()) == 0)
{
ic = h->coders[packet->getStreamIndex()];
oc = hw->coders[packet->getStreamIndex()];
if (packet->getStreamIndex() == h->first_input_audio_stream)
{
int offset = 0;
numPackets++;
while (offset < packet->getSize())
{
retval = ic->decodeAudio(
samples.value(),
packet.value(),
offset);
VS_TUT_ENSURE("could not decode any audio",
retval > 0);
offset += retval;
VS_TUT_ENSURE("could not write any samples",
samples->getNumSamples() > 0);
numSamples += samples->getNumSamples();
// now, write out the packets.
unsigned int numSamplesConsumed = 0;
do {
retval = oc->encodeAudio(opacket.value(), samples.value(),
numSamplesConsumed);
VS_TUT_ENSURE("Could not encode any audio", retval > 0);
numSamplesConsumed += retval;
if (opacket->isComplete())
{
VS_TUT_ENSURE("could not encode audio", opacket->getSize() > 0);
RefPointer<IBuffer> encodedBuffer = opacket->getData();
VS_TUT_ENSURE("no encoded data", encodedBuffer);
VS_TUT_ENSURE("less data than there should be",
encodedBuffer->getBufferSize() >=
opacket->getSize());
retval = hw->container->writePacket(opacket.value());
VS_TUT_ENSURE("could not write packet", retval >= 0);
}
} while (numSamplesConsumed < samples->getNumSamples());
}
} else if (packet->getStreamIndex() == h->first_input_video_stream)
{
int offset = 0;
numPackets++;
VS_LOG_TRACE("packet: pts: %ld; dts: %ld", packet->getPts(),
packet->getDts());
while (offset < packet->getSize())
{
retval = ic->decodeVideo(
frame.value(),
packet.value(),
offset);
VS_TUT_ENSURE("could not decode any video", retval>0);
num = ic->getTimeBase();
VS_TUT_ENSURE_DISTANCE("time base changed", num->getDouble(), h->expected_time_base, 0.00001);
offset += retval;
if (frame->isComplete())
{
VS_TUT_ENSURE("should now have a frame width", frame->getWidth() > 0);
VS_TUT_ENSURE("should now have a frame height", frame->getHeight() > 0);
numFrames++;
if (frame->isKeyFrame())
numKeyFrames++;
frame->setQuality(0);
retval = oc->encodeVideo(
opacket.value(),
frame.value(),
-1);
VS_TUT_ENSURE("could not encode video", retval >= 0);
VS_TUT_ENSURE("could not encode video", opacket->isComplete());
VS_TUT_ENSURE("could not encode video", opacket->getSize() > 0);
RefPointer<IBuffer> encodedBuffer = opacket->getData();
VS_TUT_ENSURE("no encoded data", encodedBuffer);
VS_TUT_ENSURE("less data than there should be",
encodedBuffer->getBufferSize() >=
opacket->getSize());
// now, write the packet to disk.
retval = hw->container->writePacket(opacket.value());
VS_TUT_ENSURE("could not write packet", retval >= 0);
}
}
}
}
// sigh; it turns out that to flush the encoding buffers you need to
// ask the encoder to encode a NULL set of samples. So, let's do that.
retval = hw->coders[hw->first_output_audio_stream]->encodeAudio(opacket.value(), 0, 0);
VS_TUT_ENSURE("Could not encode any audio", retval >= 0);
if (opacket->isComplete())
{
retval = hw->container->writePacket(opacket.value());
VS_TUT_ENSURE("could not write packet", retval >= 0);
}
retval = hw->coders[hw->first_output_video_stream]->encodeVideo(opacket.value(), 0, 0);
VS_TUT_ENSURE("Could not encode any video", retval >= 0);
if (opacket->isComplete())
{
retval = hw->container->writePacket(opacket.value());
VS_TUT_ENSURE("could not write packet", retval >= 0);
}
retval = hw->container->writeTrailer();
VS_TUT_ENSURE("! writeTrailer", retval >= 0);
retval = h->coders[h->first_input_audio_stream]->close();
VS_TUT_ENSURE("! close", retval >= 0);
retval = h->coders[h->first_input_video_stream]->close();
VS_TUT_ENSURE("! close", retval >= 0);
retval = hw->coders[hw->first_output_audio_stream]->close();
VS_TUT_ENSURE("! close", retval >= 0);
retval = hw->coders[hw->first_output_video_stream]->close();
VS_TUT_ENSURE("! close", retval >= 0);
VS_TUT_ENSURE("could not get any audio packets", numPackets > 0);
VS_TUT_ENSURE("could not decode any audio", numSamples > 0);
VS_TUT_ENSURE("could not decode any video", numFrames >0);
VS_TUT_ENSURE("could not find any key frames", numKeyFrames >0);
if (h->expected_packets)
VS_TUT_ENSURE_EQUALS("unexpected number of packets",
numPackets, 1062);
if (h->expected_audio_samples)
VS_TUT_ENSURE_EQUALS("unexpected number of audio samples",
numSamples, 438912);
if (h->expected_video_frames)
VS_TUT_ENSURE_EQUALS("unexpected number of video frames",
numFrames, 300);
if (h->expected_video_key_frames)
VS_TUT_ENSURE_EQUALS("unexpected number of video key frames",
numKeyFrames, 25);
}
void
MediaPacketTest::testCopyPacket() {
const int32_t size = 512;
packet = MediaPacket::make(size);
TSM_ASSERT("was able to allocate packet", packet);
// everything else should be garbage.
int64_t position = packet->getPosition();
TSM_ASSERT("position was not set to -1", position == -1);
position = 4;
packet->setPosition(position);
int64_t dts = 28349762;
packet->setDts(dts);
int64_t pts = 82729373;
packet->setPts(pts);
RefPointer<Rational> timeBase = Rational::make(3, 28972);
packet->setTimeBase(timeBase.value());
int32_t streamIndex = 8;
packet->setStreamIndex(streamIndex);
int64_t duration = 28387728;
packet->setDuration(duration);
int64_t convergenceDuration = 283;
packet->setConvergenceDuration(convergenceDuration);
// let's get access to the data
RefPointer<Buffer> data = packet->getData();
TS_ASSERT_EQUALS(size+16, data->getBufferSize());
TS_ASSERT_EQUALS(size, packet->getSize());
uint8_t* raw = (uint8_t*) data->getBytes(0, size);
for (int i = 0; i < size; i++)
raw[i] = i % 16;
// Now, make a copy
bool tests[] =
{ true, false };
for (size_t i = 0; i < (sizeof(tests) / sizeof(tests[0])); i++) {
RefPointer<MediaPacket> newPacket = MediaPacket::make(packet.value(), tests[i]);
TSM_ASSERT("should not be empty", newPacket);
// let's make sure that when not copying, the data is the same.
TSM_ASSERT_EQUALS("should equal", position, newPacket->getPosition());
TSM_ASSERT_EQUALS("should equal", pts, newPacket->getPts());
TSM_ASSERT_EQUALS("should equal", dts, newPacket->getDts());
TSM_ASSERT_EQUALS("should equal", streamIndex, newPacket->getStreamIndex());
TSM_ASSERT_EQUALS("should equal", duration, newPacket->getDuration());
TSM_ASSERT_EQUALS("should equal", convergenceDuration,
newPacket->getConvergenceDuration());
RefPointer<Rational> newBase = newPacket->getTimeBase();
TSM_ASSERT("should be equal", newBase->compareTo(timeBase.value()) == 0);
RefPointer<Buffer> buf = newPacket->getData();
TS_ASSERT_EQUALS(size, newPacket->getSize());
TS_ASSERT_EQUALS(size+16, buf->getBufferSize());
uint8_t* d = (uint8_t*) buf->getBytes(0, size);
if (!tests[i]) {
TS_ASSERT_EQUALS(d, raw);
} else {
TS_ASSERT_DIFFERS(d, raw);
}
TS_ASSERT(d);
for (int j = 0; j < size; j++) {
TS_ASSERT_EQUALS(d[j], j % 16);
}
}
}
void
StreamCoderX264Test :: testDecodingAndEncodingH264Video()
{
// This test doesn't pass memcheck but we're disabling for now
const char *memcheck = getenv("VS_TEST_MEMCHECK");
if (memcheck)
return;
return;
RefPointer<IAudioSamples> samples = 0;
RefPointer<IVideoPicture> frame = 0;
RefPointer<IRational> num(0);
RefPointer<IStreamCoder> ic(0);
RefPointer<IStreamCoder> oc(0);
RefPointer<ICodec> newcodec;
newcodec = ICodec::findEncodingCodecByName("libx264");
if (!newcodec)
// we're probably in a LGPL build, and so we shouldn't run this test
return;
int numSamples = 0;
int numPackets = 0;
int numFrames = 0;
int numKeyFrames = 0;
int retval = -1;
LoggerStack stack;
stack.setGlobalLevel(Logger::LEVEL_WARN, false);
// use the 20 second file to make memcheck faster
h->setupReading("testfile_videoonly_20sec.flv");
RefPointer<IPacket> packet = IPacket::make();
hw->setupWriting("StreamCoderX264Test_testDecodingAndEncodingH264Video.mp4", "libx264", "libmp3lame", "mp4");
RefPointer<IPacket> opacket = IPacket::make();
VS_TUT_ENSURE("! opacket", opacket);
{
// Let's set up audio first.
ic = h->coders[h->first_input_audio_stream];
oc = hw->coders[hw->first_output_audio_stream];
// Set the output coder correctly.
oc->setSampleRate(ic->getSampleRate());
oc->setChannels(ic->getChannels());
oc->setBitRate(ic->getBitRate());
samples = IAudioSamples::make(1024, ic->getChannels());
VS_TUT_ENSURE("got no samples", samples);
retval = ic->open();
VS_TUT_ENSURE("Could not open input coder", retval >= 0);
retval = oc->open();
VS_TUT_ENSURE("Could not open output coder", retval >= 0);
}
{
// now, let's set up video.
ic = h->coders[h->first_input_video_stream];
oc = hw->coders[hw->first_output_video_stream];
// We're going to set up high quality video
oc->setBitRate(720000);
oc->setGlobalQuality(0);
oc->setFlags(oc->getFlags() | IStreamCoder::FLAG_QSCALE);
oc->setPixelType(ic->getPixelType());
oc->setHeight(ic->getHeight());
oc->setWidth(ic->getWidth());
num = IRational::make(1,15);
oc->setFrameRate(num.value());
num = ic->getTimeBase();
oc->setTimeBase(num.value());
oc->setNumPicturesInGroupOfPictures(
ic->getNumPicturesInGroupOfPictures()
);
// Ask the StreamCoder to guess width and height for us
frame = IVideoPicture::make(
ic->getPixelType(),
-1, -1);
VS_TUT_ENSURE("got no frame", frame);
VS_TUT_ENSURE("should not have width", frame->getWidth() <= 0);
VS_TUT_ENSURE("should not have height", frame->getHeight() <= 0);
oc->setProperty("coder", "0");
oc->setProperty("flags", "+loop");
oc->setProperty("cmp", "+chroma");
oc->setProperty("partitions", "-parti8x8+parti4x4+partp8x8-partp4x4-partb8x8");
oc->setProperty("me_method", "hex");
oc->setProperty("subq", "3");
oc->setProperty("me_range", "16");
oc->setProperty("g", "250");
oc->setProperty("keyint_min", "25");
oc->setProperty("sc_threshold", "40");
oc->setProperty("i_qfactor", "0.71");
oc->setProperty("b_strategy", "1");
oc->setProperty("qcomp", "0.6");
oc->setProperty("qmin", "10");
oc->setProperty("qmax", "51");
oc->setProperty("qdiff", "4");
oc->setProperty("directpred", "1");
oc->setProperty("flags2", "+fastpskip");
oc->setProperty("cqp", "0");
retval = ic->open();
VS_TUT_ENSURE("Could not open input coder", retval >= 0);
retval = oc->open();
if (retval < 0)
// this fails on Mac builds
return;
VS_TUT_ENSURE("Could not open output coder", retval >= 0);
}
// write header
retval = hw->container->writeHeader();
VS_TUT_ENSURE("could not write header", retval >= 0);
while (h->container->readNextPacket(packet.value()) == 0)
{
ic = h->coders[packet->getStreamIndex()];
oc = hw->coders[packet->getStreamIndex()];
if (packet->getStreamIndex() == h->first_input_audio_stream)
{
int offset = 0;
numPackets++;
while (offset < packet->getSize())
{
retval = ic->decodeAudio(
samples.value(),
packet.value(),
offset);
VS_TUT_ENSURE("could not decode any audio",
retval > 0);
offset += retval;
VS_TUT_ENSURE("could not write any samples",
samples->getNumSamples() > 0);
numSamples += samples->getNumSamples();
// now, write out the packets.
unsigned int numSamplesConsumed = 0;
do {
retval = oc->encodeAudio(opacket.value(), samples.value(),
numSamplesConsumed);
VS_TUT_ENSURE("Could not encode any audio", retval > 0);
numSamplesConsumed += retval;
if (opacket->isComplete())
{
VS_TUT_ENSURE("could not encode audio", opacket->getSize() > 0);
RefPointer<IBuffer> encodedBuffer = opacket->getData();
VS_TUT_ENSURE("no encoded data", encodedBuffer);
VS_TUT_ENSURE("less data than there should be",
encodedBuffer->getBufferSize() >=
opacket->getSize());
retval = hw->container->writePacket(opacket.value());
VS_TUT_ENSURE("could not write packet", retval >= 0);
}
} while (numSamplesConsumed < samples->getNumSamples());
}
} else if (packet->getStreamIndex() == h->first_input_video_stream)
{
int offset = 0;
numPackets++;
VS_LOG_TRACE("packet: pts: %ld; dts: %ld", packet->getPts(),
packet->getDts());
while (offset < packet->getSize())
{
retval = ic->decodeVideo(
frame.value(),
packet.value(),
offset);
VS_TUT_ENSURE("could not decode any video", retval>0);
num = ic->getTimeBase();
VS_TUT_ENSURE_DISTANCE("time base changed", num->getDouble(), h->expected_time_base, 0.00001);
offset += retval;
if (frame->isComplete())
{
VS_TUT_ENSURE("should now have a frame width", frame->getWidth() > 0);
VS_TUT_ENSURE("should now have a frame height", frame->getHeight() > 0);
numFrames++;
if (frame->isKeyFrame())
numKeyFrames++;
frame->setQuality(0);
retval = oc->encodeVideo(
opacket.value(),
frame.value(),
-1);
VS_TUT_ENSURE("could not encode video", retval >= 0);
if (opacket->isComplete())
{
VS_TUT_ENSURE("could not encode video", opacket->isComplete());
VS_TUT_ENSURE("could not encode video", opacket->getSize() > 0);
RefPointer<IBuffer> encodedBuffer = opacket->getData();
VS_TUT_ENSURE("no encoded data", encodedBuffer);
VS_TUT_ENSURE("less data than there should be",
encodedBuffer->getBufferSize() >=
opacket->getSize());
// now, write the packet to disk.
retval = hw->container->writePacket(opacket.value());
VS_TUT_ENSURE("could not write packet", retval >= 0);
}
}
}
}
}
// sigh; it turns out that to flush the encoding buffers you need to
// ask the encoder to encode a NULL set of samples. So, let's do that.
retval = hw->coders[hw->first_output_audio_stream]->encodeAudio(opacket.value(), 0, 0);
VS_TUT_ENSURE("Could not encode any audio", retval >= 0);
if (opacket->isComplete())
{
retval = hw->container->writePacket(opacket.value());
VS_TUT_ENSURE("could not write packet", retval >= 0);
}
retval = hw->coders[hw->first_output_video_stream]->encodeVideo(opacket.value(), 0, 0);
VS_TUT_ENSURE("Could not encode any video", retval >= 0);
if (opacket->isComplete())
{
retval = hw->container->writePacket(opacket.value());
VS_TUT_ENSURE("could not write packet", retval >= 0);
}
retval = hw->container->writeTrailer();
VS_TUT_ENSURE("! writeTrailer", retval >= 0);
retval = h->coders[h->first_input_audio_stream]->close();
VS_TUT_ENSURE("! close", retval >= 0);
retval = h->coders[h->first_input_video_stream]->close();
VS_TUT_ENSURE("! close", retval >= 0);
retval = hw->coders[hw->first_output_audio_stream]->close();
VS_TUT_ENSURE("! close", retval >= 0);
retval = hw->coders[hw->first_output_video_stream]->close();
VS_TUT_ENSURE("! close", retval >= 0);
VS_TUT_ENSURE("could not get any audio packets", numPackets > 0);
VS_TUT_ENSURE("could not decode any audio", numSamples > 0);
VS_TUT_ENSURE("could not decode any video", numFrames >0);
VS_TUT_ENSURE("could not find any key frames", numKeyFrames >0);
if (h->expected_packets)
VS_TUT_ENSURE_EQUALS("unexpected number of packets",
numPackets, 1062);
if (h->expected_audio_samples)
VS_TUT_ENSURE_EQUALS("unexpected number of audio samples",
numSamples, 438912);
if (h->expected_video_frames)
VS_TUT_ENSURE_EQUALS("unexpected number of video frames",
numFrames, 300);
if (h->expected_video_key_frames)
VS_TUT_ENSURE_EQUALS("unexpected number of video key frames",
numKeyFrames, 25);
}
void
StreamCoderTest :: testDecodingAndEncodingFullyInterleavedFile()
{
RefPointer<IAudioSamples> samples = 0;
RefPointer<IVideoPicture> frame = 0;
RefPointer<IRational> num(0);
RefPointer<IStreamCoder> ic(0);
RefPointer<IStreamCoder> oc(0);
int numSamples = 0;
int numPackets = 0;
int numFrames = 0;
int numKeyFrames = 0;
int retval = -1;
h->setupReading(h->SAMPLE_FILE);
RefPointer<IPacket> packet = IPacket::make();
hw->setupWriting("StreamCoderTest_6_output.mov", "mpeg4", "libmp3lame", "mov");
RefPointer<IPacket> opacket = IPacket::make();
VS_TUT_ENSURE("! opacket", opacket);
RefPointer<IMetaData> inputProps = IMetaData::make();
RefPointer<IMetaData> outputProps = IMetaData::make();
{
// Let's set up audio first.
ic = h->coders[h->first_input_audio_stream];
oc = hw->coders[hw->first_output_audio_stream];
// Set the output coder correctly.
oc->setSampleRate(ic->getSampleRate());
oc->setChannels(ic->getChannels());
// Using the MetaData structure
char tmpBuf[1024];
snprintf(tmpBuf, sizeof(tmpBuf), "%d", ic->getBitRate());
inputProps->setValue("b", tmpBuf);
const char* invalidKey="kd82jclkjdi2oc001,ss2-948";
const char* invalidValue="c9xu1nxL28fJ9";
inputProps->setValue(invalidKey, invalidValue);
samples = IAudioSamples::make(1024, ic->getChannels());
VS_TUT_ENSURE("got no samples", samples);
retval = ic->open(0,0);
VS_TUT_ENSURE("Could not open input coder", retval >= 0);
retval = oc->open(inputProps.value(), outputProps.value());
VS_TUT_ENSURE("Could not open output coder", retval >= 0);
// now let's ensure our fake property was not set.
VS_TUT_ENSURE("Should only have one unset setting", outputProps->getNumKeys() == 1)
VS_TUT_ENSURE("Should have expected value", strcmp(outputProps->getValue(invalidKey, IMetaData::METADATA_NONE), invalidValue)==0);
}
{
// now, let's set up video.
ic = h->coders[h->first_input_video_stream];
oc = hw->coders[hw->first_output_video_stream];
// We're going to set up high quality video
oc->setBitRate(720000);
oc->setGlobalQuality(0);
oc->setFlags(oc->getFlags() | IStreamCoder::FLAG_QSCALE);
oc->setPixelType(ic->getPixelType());
oc->setHeight(ic->getHeight());
oc->setWidth(ic->getWidth());
num = IRational::make(1,15);
oc->setFrameRate(num.value());
num = ic->getTimeBase();
oc->setTimeBase(num.value());
oc->setNumPicturesInGroupOfPictures(
ic->getNumPicturesInGroupOfPictures()
);
// Ask the StreamCoder to guess width and height for us
frame = IVideoPicture::make(
ic->getPixelType(),
-1, -1);
VS_TUT_ENSURE("got no frame", frame);
VS_TUT_ENSURE("should not have width", frame->getWidth() <= 0);
VS_TUT_ENSURE("should not have height", frame->getHeight() <= 0);
retval = ic->open();
VS_TUT_ENSURE("Could not open input coder", retval >= 0);
retval = oc->open();
VS_TUT_ENSURE("Could not open output coder", retval >= 0);
}
// write header
retval = hw->container->writeHeader();
VS_TUT_ENSURE("could not write header", retval >= 0);
while (h->container->readNextPacket(packet.value()) == 0)
{
ic = h->coders[packet->getStreamIndex()];
oc = hw->coders[packet->getStreamIndex()];
if (packet->getStreamIndex() == h->first_input_audio_stream)
{
int offset = 0;
numPackets++;
while (offset < packet->getSize())
{
retval = ic->decodeAudio(
samples.value(),
packet.value(),
offset);
VS_TUT_ENSURE("could not decode any audio",
retval > 0);
offset += retval;
VS_TUT_ENSURE("could not write any samples",
samples->getNumSamples() > 0);
numSamples += samples->getNumSamples();
// now, write out the packets.
unsigned int numSamplesConsumed = 0;
do {
retval = oc->encodeAudio(opacket.value(), samples.value(),
numSamplesConsumed);
VS_TUT_ENSURE("Could not encode any audio", retval > 0);
numSamplesConsumed += retval;
if (opacket->isComplete())
{
VS_TUT_ENSURE("audio duration not set", opacket->getDuration() > 0);
VS_TUT_ENSURE("could not encode audio", opacket->getSize() > 0);
RefPointer<IBuffer> encodedBuffer = opacket->getData();
VS_TUT_ENSURE("no encoded data", encodedBuffer);
VS_TUT_ENSURE("less data than there should be",
encodedBuffer->getBufferSize() >=
opacket->getSize());
retval = hw->container->writePacket(opacket.value());
VS_TUT_ENSURE("could not write packet", retval >= 0);
}
} while (numSamplesConsumed < samples->getNumSamples());
}
} else if (packet->getStreamIndex() == h->first_input_video_stream)
{
int offset = 0;
numPackets++;
VS_LOG_TRACE("packet: pts: %ld; dts: %ld", packet->getPts(),
packet->getDts());
while (offset < packet->getSize())
{
retval = ic->decodeVideo(
frame.value(),
packet.value(),
offset);
VS_TUT_ENSURE("could not decode any video", retval>0);
num = ic->getTimeBase();
VS_TUT_ENSURE_DISTANCE("time base changed", num->getDouble(), h->expected_time_base, 0.0001);
offset += retval;
if (frame->isComplete())
{
VS_TUT_ENSURE("should now have a frame width", frame->getWidth() > 0);
VS_TUT_ENSURE("should now have a frame height", frame->getHeight() > 0);
numFrames++;
if (frame->isKeyFrame())
numKeyFrames++;
frame->setQuality(0);
retval = oc->encodeVideo(
opacket.value(),
frame.value(),
-1);
VS_TUT_ENSURE("could not encode video", retval >= 0);
VS_TUT_ENSURE("could not encode video", opacket->isComplete());
VS_TUT_ENSURE("could not encode video", opacket->getSize() > 0);
RefPointer<IBuffer> encodedBuffer = opacket->getData();
VS_TUT_ENSURE("no encoded data", encodedBuffer);
VS_TUT_ENSURE("less data than there should be",
encodedBuffer->getBufferSize() >=
opacket->getSize());
// now, write the packet to disk.
retval = hw->container->writePacket(opacket.value());
VS_TUT_ENSURE("could not write packet", retval >= 0);
}
}
}
}
// sigh; it turns out that to flush the encoding buffers you need to
// ask the encoder to encode a NULL set of samples. So, let's do that.
retval = hw->coders[hw->first_output_audio_stream]->encodeAudio(opacket.value(), 0, 0);
VS_TUT_ENSURE("Could not encode any audio", retval >= 0);
if (opacket->isComplete())
{
retval = hw->container->writePacket(opacket.value());
VS_TUT_ENSURE("could not write packet", retval >= 0);
}
retval = hw->coders[hw->first_output_video_stream]->encodeVideo(opacket.value(), 0, 0);
VS_TUT_ENSURE("Could not encode any video", retval >= 0);
if (opacket->isComplete())
{
retval = hw->container->writePacket(opacket.value());
VS_TUT_ENSURE("could not write packet", retval >= 0);
}
retval = hw->container->writeTrailer();
VS_TUT_ENSURE("! writeTrailer", retval >= 0);
retval = h->coders[h->first_input_audio_stream]->close();
VS_TUT_ENSURE("! close", retval >= 0);
retval = h->coders[h->first_input_video_stream]->close();
VS_TUT_ENSURE("! close", retval >= 0);
retval = hw->coders[hw->first_output_audio_stream]->close();
VS_TUT_ENSURE("! close", retval >= 0);
retval = hw->coders[hw->first_output_video_stream]->close();
VS_TUT_ENSURE("! close", retval >= 0);
VS_TUT_ENSURE("could not get any audio packets", numPackets > 0);
VS_TUT_ENSURE("could not decode any audio", numSamples > 0);
VS_TUT_ENSURE("could not decode any video", numFrames >0);
VS_TUT_ENSURE("could not find any key frames", numKeyFrames >0);
if (h->expected_packets)
VS_TUT_ENSURE_EQUALS("unexpected number of packets",
numPackets, h->expected_packets);
if (h->expected_audio_samples)
VS_TUT_ENSURE_EQUALS("unexpected number of audio samples",
numSamples, h->expected_audio_samples);
if (h->expected_video_frames)
VS_TUT_ENSURE_EQUALS("unexpected number of video frames",
numFrames, h->expected_video_frames);
if (h->expected_video_key_frames)
VS_TUT_ENSURE_EQUALS("unexpected number of video key frames",
numKeyFrames, h->expected_video_key_frames);
}
void
MediaPictureResamplerTest::testRescale() {
TestData::Fixture* fixture=mFixtures.getFixture("testfile_h264_mp4a_tmcd.mov");
TS_ASSERT(fixture);
char filepath[2048];
mFixtures.fillPath(fixture, filepath, sizeof(filepath));
RefPointer<Demuxer> source = Demuxer::make();
source->open(filepath, 0, false, true, 0, 0);
int32_t numStreams = source->getNumStreams();
TS_ASSERT_EQUALS(fixture->num_streams, numStreams);
int32_t streamToDecode = 0;
RefPointer<DemuxerStream> stream = source->getStream(streamToDecode);
TS_ASSERT(stream);
RefPointer<Decoder> decoder = stream->getDecoder();
TS_ASSERT(decoder);
RefPointer<Codec> codec = decoder->getCodec();
TS_ASSERT(codec);
TS_ASSERT_EQUALS(Codec::CODEC_ID_H264, codec->getID());
decoder->open(0, 0);
TS_ASSERT_EQUALS(PixelFormat::PIX_FMT_YUV420P, decoder->getPixelFormat());
// now, let's start a decoding loop.
RefPointer<MediaPacket> packet = MediaPacket::make();
RefPointer<MediaPicture> picture = MediaPicture::make(
decoder->getWidth(),
decoder->getHeight(),
decoder->getPixelFormat());
int32_t rescaleW = decoder->getWidth()/4;
int32_t rescaleH = (int32_t)(decoder->getHeight()*5.22);
PixelFormat::Type rescaleFmt = PixelFormat::PIX_FMT_RGBA;
RefPointer<MediaPicture> rescaled = MediaPicture::make(
rescaleW, rescaleH, rescaleFmt);
RefPointer<MediaPictureResampler> resampler =
MediaPictureResampler::make(rescaled->getWidth(),
rescaled->getHeight(),
rescaled->getFormat(),
picture->getWidth(),
picture->getHeight(),
picture->getFormat(),
0);
resampler->open();
int32_t frameNo = 0;
while(source->read(packet.value()) >= 0) {
// got a packet; now we try to decode it.
if (packet->getStreamIndex() == streamToDecode &&
packet->isComplete()) {
int32_t bytesRead = 0;
int32_t byteOffset=0;
do {
bytesRead = decoder->decodeVideo(picture.value(), packet.value(), byteOffset);
if (picture->isComplete()) {
writePicture("MediaPictureResamplerTest_testRescaleVideo",
&frameNo, picture.value(), resampler.value(), rescaled.value());
}
byteOffset += bytesRead;
} while(byteOffset < packet->getSize());
// now, handle the case where bytesRead is 0; we need to flush any
// cached packets
do {
decoder->decodeVideo(picture.value(), 0, 0);
if (picture->isComplete()) {
writePicture("MediaPictureResamplerTest_testRescaleVideo", &frameNo, picture.value(),
resampler.value(), rescaled.value());
}
} while (picture->isComplete());
}
if ((int32_t)(frameNo/30) > 10)
// 20 pictures should be enough to see if it's working.
break;
}
source->close();
}
void
AudioResamplerTest :: testResamplingAudio()
{
RefPointer<IAudioResampler> resampler = IAudioResampler::make(2, 1,
44100, 22050);
RefPointer<IAudioSamples> samples = 0;
RefPointer<IAudioSamples> resamples = 0;
int numSamples = 0;
int numPackets = 0;
int retval = -1;
samples = IAudioSamples::make(1024, 1);
VS_TUT_ENSURE("got no samples", samples);
resamples = IAudioSamples::make(1024, 2);
VS_TUT_ENSURE("got no samples", samples);
h->setupReading(h->SAMPLE_FILE);
RefPointer<IPacket> packet = IPacket::make();
hw->setupWriting("AudioResamplerTest_3_output.flv", 0, "libmp3lame", "flv");
int outStream = hw->first_output_audio_stream;
VS_TUT_ENSURE("Could not find an audio stream in the output", outStream >= 0);
int inStream = h->first_input_audio_stream;
VS_TUT_ENSURE("Could not find an audio stream in the input", inStream >= 0);
RefPointer<IStreamCoder> ic = h->coders[inStream];
RefPointer<IStreamCoder> oc = hw->coders[outStream];
RefPointer<IPacket> opacket = IPacket::make();
VS_TUT_ENSURE("! opacket", opacket);
// Set the output coder correctly.
int outChannels = 2;
int outRate = 44100;
resampler = IAudioResampler::make(outChannels, ic->getChannels(),
outRate, ic->getSampleRate());
oc->setSampleRate(outRate);
oc->setChannels(outChannels);
oc->setBitRate(ic->getBitRate());
int maxSamples = 10 * ic->getSampleRate(); // 10 seconds
retval = ic->open();
VS_TUT_ENSURE("Could not open input coder", retval >= 0);
retval = oc->open();
VS_TUT_ENSURE("Could not open output coder", retval >= 0);
// write header
retval = hw->container->writeHeader();
VS_TUT_ENSURE("could not write header", retval >= 0);
while (h->container->readNextPacket(packet.value()) == 0
&& numSamples < maxSamples)
{
if (packet->getStreamIndex() == inStream)
{
int offset = 0;
numPackets++;
while (offset < packet->getSize())
{
retval = ic->decodeAudio(
samples.value(),
packet.value(),
offset);
VS_TUT_ENSURE("could not decode any audio",
retval > 0);
offset += retval;
VS_TUT_ENSURE("could not write any samples",
samples->getNumSamples() > 0);
numSamples += samples->getNumSamples();
resamples = IAudioSamples::make((samples->getNumSamples()*2),2);
// now, resample the audio
retval = resampler->resample(resamples.value(), samples.value(), 0);
VS_TUT_ENSURE("could not resample", retval > 0);
VS_TUT_ENSURE("no resamples", resamples->getNumSamples() > 0);
VS_TUT_ENSURE_EQUALS("wrong sample rate", resamples->getSampleRate(),
outRate);
VS_TUT_ENSURE_EQUALS("wrong channels", resamples->getChannels(),
outChannels);
// now, write out the packets.
unsigned int samplesConsumed = 0;
do {
retval = oc->encodeAudio(opacket.value(), resamples.value(),
samplesConsumed);
VS_TUT_ENSURE("Could not encode any audio", retval >= 0);
samplesConsumed += (unsigned int)retval;
VS_LOG_TRACE("packet: %d; is: %d; os: %d",
numPackets, numSamples, samplesConsumed);
if (opacket->isComplete())
{
VS_TUT_ENSURE("could not encode audio", opacket->getSize() > 0);
RefPointer<IBuffer> encodedBuffer = opacket->getData();
VS_TUT_ENSURE("no encoded data", encodedBuffer);
VS_TUT_ENSURE("less data than there should be",
encodedBuffer->getBufferSize() >=
opacket->getSize());
retval = hw->container->writePacket(opacket.value());
VS_TUT_ENSURE("could not write packet", retval >= 0);
}
// keep going until we've encoded all samples in this buffer.
} while (samplesConsumed < resamples->getNumSamples());
}
}
}
// sigh; it turns out that to flush the encoding buffers you need to
// ask the encoder to encode a NULL set of samples. So, let's do that.
retval = oc->encodeAudio(opacket.value(), 0, 0);
VS_TUT_ENSURE("Could not encode any audio", retval >= 0);
if (retval > 0)
{
retval = hw->container->writePacket(opacket.value());
VS_TUT_ENSURE("could not write packet", retval >= 0);
}
retval = hw->container->writeTrailer();
VS_TUT_ENSURE("! writeTrailer", retval >= 0);
retval = ic->close();
VS_TUT_ENSURE("! close", retval >= 0);
retval = oc->close();
VS_TUT_ENSURE("! close", retval >= 0);
VS_TUT_ENSURE("could not get any audio packets", numPackets > 0);
VS_TUT_ENSURE("could not decode any audio", numSamples > 0);
}
void
StreamCoderSpeexTest :: testEncodingSpeexAudio()
{
// disable because broken on Windows
return;
RefPointer<IAudioSamples> samples = 0;
RefPointer<IAudioSamples> outSamples = 0;
RefPointer<IVideoPicture> frame = 0;
RefPointer<IRational> num(0);
RefPointer<IStreamCoder> ic(0);
RefPointer<IStreamCoder> oc(0);
RefPointer<IAudioResampler> resampler;
int numSamples = 0;
int numPackets = 0;
int numFrames = 0;
int numKeyFrames = 0;
int retval = -1;
h->setupReading(h->SAMPLE_FILE);
RefPointer<IPacket> packet = IPacket::make();
hw->setupWriting("StreamCoderSpeexTest_output.ogg",
"libtheora", "libspeex", "ogg");
RefPointer<IPacket> opacket = IPacket::make();
VS_TUT_ENSURE("! opacket", opacket);
{
// Let's set up audio first.
ic = h->coders[h->first_input_audio_stream];
oc = hw->coders[hw->first_output_audio_stream];
// Set the output coder correctly.
oc->setSampleRate(16000);
oc->setChannels(1);
oc->setBitRate(ic->getBitRate());
samples = IAudioSamples::make(1024, ic->getChannels());
VS_TUT_ENSURE("got no samples", samples);
outSamples = IAudioSamples::make(1024, 1);
VS_TUT_ENSURE("got no samples", outSamples);
resampler = IAudioResampler::make(1, ic->getChannels(),
16000, ic->getSampleRate());
retval = ic->open();
VS_TUT_ENSURE("Could not open input coder", retval >= 0);
retval = oc->open();
VS_TUT_ENSURE("Could not open output coder", retval >= 0);
}
{
// now, let's set up video.
ic = h->coders[h->first_input_video_stream];
oc = hw->coders[hw->first_output_video_stream];
// We're going to set up high quality video
oc->setBitRate(720000);
oc->setGlobalQuality(0);
oc->setFlags(oc->getFlags() | IStreamCoder::FLAG_QSCALE);
oc->setPixelType(ic->getPixelType());
oc->setHeight(ic->getHeight());
oc->setWidth(ic->getWidth());
num = IRational::make(1,15);
oc->setFrameRate(num.value());
num = ic->getTimeBase();
oc->setTimeBase(num.value());
oc->setNumPicturesInGroupOfPictures(
ic->getNumPicturesInGroupOfPictures()
);
// Ask the StreamCoder to guess width and height for us
frame = IVideoPicture::make(
ic->getPixelType(),
-1, -1);
VS_TUT_ENSURE("got no frame", frame);
VS_TUT_ENSURE("should not have width", frame->getWidth() <= 0);
VS_TUT_ENSURE("should not have height", frame->getHeight() <= 0);
retval = ic->open();
VS_TUT_ENSURE("Could not open input coder", retval >= 0);
retval = oc->open();
VS_TUT_ENSURE("Could not open output coder", retval >= 0);
}
// write header
retval = hw->container->writeHeader();
VS_TUT_ENSURE("could not write header", retval >= 0);
while (h->container->readNextPacket(packet.value()) == 0)
{
ic = h->coders[packet->getStreamIndex()];
oc = hw->coders[packet->getStreamIndex()];
if (packet->getStreamIndex() == h->first_input_audio_stream)
{
int offset = 0;
numPackets++;
while (offset < packet->getSize())
{
retval = ic->decodeAudio(
samples.value(),
packet.value(),
offset);
VS_TUT_ENSURE("could not decode any audio",
retval > 0);
offset += retval;
VS_TUT_ENSURE("could not write any samples",
samples->getNumSamples() > 0);
numSamples += samples->getNumSamples();
retval = resampler->resample(outSamples.value(),
samples.value(),
samples->getNumSamples());
VS_TUT_ENSURE("Could not resample audio", retval > 0);
// now, write out the packets.
unsigned int numSamplesConsumed = 0;
do {
retval = oc->encodeAudio(opacket.value(),
outSamples.value(),
numSamplesConsumed);
VS_TUT_ENSURE("Could not encode any audio", retval > 0);
numSamplesConsumed += retval;
if (opacket->isComplete())
{
VS_TUT_ENSURE("could not encode audio", opacket->getSize() > 0);
RefPointer<IBuffer> encodedBuffer = opacket->getData();
VS_TUT_ENSURE("no encoded data", encodedBuffer);
VS_TUT_ENSURE("less data than there should be",
encodedBuffer->getBufferSize() >=
opacket->getSize());
retval = hw->container->writePacket(opacket.value());
VS_TUT_ENSURE("could not write packet", retval >= 0);
}
} while (numSamplesConsumed < outSamples->getNumSamples());
}
} else if (packet->getStreamIndex() == h->first_input_video_stream)
{
int offset = 0;
numPackets++;
VS_LOG_TRACE("packet: pts: %ld; dts: %ld", packet->getPts(),
packet->getDts());
while (offset < packet->getSize())
{
retval = ic->decodeVideo(
frame.value(),
packet.value(),
offset);
VS_TUT_ENSURE("could not decode any video", retval>0);
num = ic->getTimeBase();
VS_TUT_ENSURE_EQUALS("time base changed", num->getDouble(), h->expected_time_base);
offset += retval;
if (frame->isComplete())
{
VS_TUT_ENSURE("should now have a frame width", frame->getWidth() > 0);
VS_TUT_ENSURE("should now have a frame height", frame->getHeight() > 0);
numFrames++;
if (frame->isKeyFrame())
numKeyFrames++;
frame->setQuality(0);
retval = oc->encodeVideo(
opacket.value(),
frame.value(),
-1);
VS_TUT_ENSURE("could not encode video", retval >= 0);
VS_TUT_ENSURE("could not encode video", opacket->isComplete());
VS_TUT_ENSURE("could not encode video", opacket->getSize() > 0);
RefPointer<IBuffer> encodedBuffer = opacket->getData();
VS_TUT_ENSURE("no encoded data", encodedBuffer);
VS_TUT_ENSURE("less data than there should be",
encodedBuffer->getBufferSize() >=
opacket->getSize());
// now, write the packet to disk.
retval = hw->container->writePacket(opacket.value());
VS_TUT_ENSURE("could not write packet", retval >= 0);
}
}
}
}
// sigh; it turns out that to flush the encoding buffers you need to
// ask the encoder to encode a NULL set of samples. So, let's do that.
retval = hw->coders[hw->first_output_audio_stream]->encodeAudio(opacket.value(), 0, 0);
VS_TUT_ENSURE("Could not encode any audio", retval >= 0);
if (opacket->isComplete())
{
retval = hw->container->writePacket(opacket.value());
VS_TUT_ENSURE("could not write packet", retval >= 0);
}
retval = hw->coders[hw->first_output_video_stream]->encodeVideo(opacket.value(), 0, 0);
VS_TUT_ENSURE("Could not encode any video", retval >= 0);
if (opacket->isComplete())
{
retval = hw->container->writePacket(opacket.value());
VS_TUT_ENSURE("could not write packet", retval >= 0);
}
retval = hw->container->writeTrailer();
VS_TUT_ENSURE("! writeTrailer", retval >= 0);
retval = h->coders[h->first_input_audio_stream]->close();
VS_TUT_ENSURE("! close", retval >= 0);
retval = h->coders[h->first_input_video_stream]->close();
VS_TUT_ENSURE("! close", retval >= 0);
retval = hw->coders[hw->first_output_audio_stream]->close();
VS_TUT_ENSURE("! close", retval >= 0);
retval = hw->coders[hw->first_output_video_stream]->close();
VS_TUT_ENSURE("! close", retval >= 0);
VS_TUT_ENSURE("could not get any audio packets", numPackets > 0);
VS_TUT_ENSURE("could not decode any audio", numSamples > 0);
VS_TUT_ENSURE("could not decode any video", numFrames >0);
VS_TUT_ENSURE("could not find any key frames", numKeyFrames >0);
if (h->expected_packets)
VS_TUT_ENSURE_EQUALS("unexpected number of packets",
numPackets, h->expected_packets);
if (h->expected_audio_samples)
VS_TUT_ENSURE_EQUALS("unexpected number of audio samples",
numSamples, h->expected_audio_samples);
if (h->expected_video_frames)
VS_TUT_ENSURE_EQUALS("unexpected number of video frames",
numFrames, h->expected_video_frames);
if (h->expected_video_key_frames)
VS_TUT_ENSURE_EQUALS("unexpected number of video key frames",
numKeyFrames, h->expected_video_key_frames);
}
/**
* This test will read ironman (FLV, H263 video and mp3 audio) and transcode
* to MP4 (H264 Video and aac audio).
*/
void
EncoderTest::testTranscode()
{
// enable trace logging
Logger::setGlobalIsLogging(Logger::LEVEL_TRACE, false);
const bool isMemCheck = getenv("VS_TEST_MEMCHECK") ? true : false;
LoggerStack stack;
stack.setGlobalLevel(Logger::LEVEL_INFO, false);
TestData::Fixture* fixture;
fixture=mFixtures.getFixture("testfile.flv");
// fixture=mFixtures.getFixture("testfile_h264_mp4a_tmcd.mov");
// fixture=mFixtures.getFixture("bigbuckbunny_h264_aac_5.1.mp4");
TS_ASSERT(fixture);
char filepath[2048];
mFixtures.fillPath(fixture, filepath, sizeof(filepath));
RefPointer<Demuxer> source = Demuxer::make();
source->open(filepath, 0, false, true, 0, 0);
int32_t numStreams = source->getNumStreams();
TS_ASSERT_EQUALS(fixture->num_streams, numStreams);
// Let's create a helper object to help us with decoding
typedef struct {
MediaDescriptor::Type type;
RefPointer<DemuxerStream> stream;
RefPointer<Decoder> decoder;
RefPointer<MediaSampled> media;
} DemuxerStreamHelper;
// I know there are only 2 in the input file.
DemuxerStreamHelper inputHelpers[10];
for(int32_t i = 0; i < numStreams; i++) {
DemuxerStreamHelper* input = &inputHelpers[i];
input->stream = source->getStream(i);
input->decoder = input->stream->getDecoder();
if (!input->decoder)
// skip
break;
input->decoder->open(0, 0);
input->type = input->decoder->getCodecType();
if (input->type == MediaDescriptor::MEDIA_AUDIO)
input->media = MediaAudio::make(
input->decoder->getFrameSize(),
input->decoder->getSampleRate(),
input->decoder->getChannels(),
input->decoder->getChannelLayout(),
input->decoder->getSampleFormat());
else if (input->type == MediaDescriptor::MEDIA_VIDEO)
input->media = MediaPicture::make(
input->decoder->getWidth(),
input->decoder->getHeight(),
input->decoder->getPixelFormat()
);
}
// now, let's set up our output file.
RefPointer<Muxer> muxer = Muxer::make("EncoderTest_testTranscode.mp4", 0, 0);
RefPointer<MuxerFormat> format = muxer->getFormat();
// Let's create a helper object to help us with decoding
typedef struct {
MediaDescriptor::Type type;
RefPointer<MuxerStream> stream;
RefPointer<MediaResampler> resampler;
RefPointer<MediaSampled> media;
RefPointer<Encoder> encoder;
} MuxerStreamHelper;
MuxerStreamHelper outputHelpers[10];
for(int32_t i = 0; i < numStreams; i++) {
DemuxerStreamHelper *input = &inputHelpers[i];
MuxerStreamHelper *output = &outputHelpers[i];
if (!input->decoder)
// skip
break;
output->type = input->type;
RefPointer<Encoder> encoder;
if (output->type == MediaDescriptor::MEDIA_VIDEO) {
RefPointer<Codec> codec = Codec::findEncodingCodec(Codec::CODEC_ID_H264);
encoder = Encoder::make(codec.value());
// set the encoder properties we need
encoder->setWidth(input->decoder->getWidth());
encoder->setHeight(input->decoder->getHeight());
encoder->setPixelFormat(input->decoder->getPixelFormat());
encoder->setProperty("b", (int64_t)400000); // bitrate
encoder->setProperty("g", (int64_t) 10); // gop
encoder->setProperty("bf", (int64_t)0); // max b frames
RefPointer<Rational> tb = Rational::make(1,2997);
encoder->setTimeBase(tb.value());
} else if (output->type == MediaDescriptor::MEDIA_AUDIO) {
RefPointer<Codec> codec = Codec::findEncodingCodec(Codec::CODEC_ID_AAC);
encoder = Encoder::make(codec.value());
// set the encoder properties we need
encoder->setSampleRate(input->decoder->getSampleRate());
encoder->setSampleFormat(input->decoder->getSampleFormat());
encoder->setSampleFormat(AudioFormat::SAMPLE_FMT_S16);
encoder->setChannelLayout(input->decoder->getChannelLayout());
encoder->setChannels(input->decoder->getChannels());
encoder->setProperty("b", (int64_t)64000); // bitrate
RefPointer<Rational> tb = Rational::make(1,encoder->getSampleRate());
encoder->setTimeBase(tb.value());
// //input->decoder->getTimeBase();
// output->encoder->setTimeBase(tb.value());
}
output->encoder.reset(encoder.value(), true);
if (output->encoder) {
if (format->getFlag(MuxerFormat::GLOBAL_HEADER))
output->encoder->setFlag(Encoder::FLAG_GLOBAL_HEADER, true);
output->encoder->open(0,0);
// sometimes encoders need to change parameters to fit; let's see
// if that happened.
output->stream = muxer->addNewStream(output->encoder.value());
}
output->media = input->media;
output->resampler = 0;
if (output->type == MediaDescriptor::MEDIA_AUDIO) {
// sometimes encoders only accept certain media types and discard
// our suggestions. Let's check.
if (
output->encoder->getSampleRate() != input->decoder->getSampleRate() ||
output->encoder->getSampleFormat() != input->decoder->getSampleFormat() ||
output->encoder->getChannelLayout() != input->decoder->getChannelLayout() ||
output->encoder->getChannels() != input->decoder->getChannels()
)
{
// we need a resampler.
VS_LOG_DEBUG("Resampling: [%"PRId32", %"PRId32", %"PRId32"] [%"PRId32", %"PRId32", %"PRId32"]",
(int32_t)output->encoder->getChannelLayout(),
(int32_t)output->encoder->getSampleRate(),
(int32_t)output->encoder->getSampleFormat(),
(int32_t)input->decoder->getChannelLayout(),
(int32_t)input->decoder->getSampleRate(),
(int32_t)input->decoder->getSampleFormat()
);
RefPointer<MediaAudioResampler> resampler = MediaAudioResampler::make(
output->encoder->getChannelLayout(),
output->encoder->getSampleRate(),
output->encoder->getSampleFormat(),
input->decoder->getChannelLayout(),
input->decoder->getSampleRate(),
input->decoder->getSampleFormat()
);
resampler->open();
output->resampler.reset(resampler.value(), true);
output->media = MediaAudio::make(
output->encoder->getFrameSize(),
output->encoder->getSampleRate(),
output->encoder->getChannels(),
output->encoder->getChannelLayout(),
output->encoder->getSampleFormat());
}
}
}
// now we should be ready to open the muxer
muxer->open(0, 0);
// now, let's start a decoding loop.
RefPointer<MediaPacket> packet = MediaPacket::make();
int numPackets = 0;
while(source->read(packet.value()) >= 0) {
// got a packet; now we try to decode it.
if (packet->isComplete()) {
int32_t streamNo = packet->getStreamIndex();
DemuxerStreamHelper *input = &inputHelpers[streamNo];
MuxerStreamHelper* output = &outputHelpers[streamNo];
if (input->decoder) decodeAndEncode(
packet.value(),
input->decoder.value(),
input->media.value(),
output->resampler.value(),
output->media.value(),
muxer.value(),
output->encoder.value());
++numPackets;
if (isMemCheck && numPackets > 100) {
VS_LOG_WARN("Exiting early under valgrind");
break;
}
}
}
// now, flush any cached packets
for(int i = 0; i < numStreams; i++) {
DemuxerStreamHelper *input = &inputHelpers[i];
MuxerStreamHelper* output = &outputHelpers[i];
if (input->decoder) decodeAndEncode(
0,
input->decoder.value(),
input->media.value(),
output->resampler.value(),
output->media.value(),
muxer.value(),
output->encoder.value());
}
source->close();
muxer->close();
}
