extern "C" bool MP4AV_RfcIsmaHinter(
MP4FileHandle mp4File,
MP4TrackId mediaTrackId,
bool interleave,
u_int16_t maxPayloadSize)
{
// gather information, and check for validity
u_int32_t numSamples =
MP4GetTrackNumberOfSamples(mp4File, mediaTrackId);
if (numSamples == 0) {
return false;
}
u_int32_t timeScale =
MP4GetTrackTimeScale(mp4File, mediaTrackId);
if (timeScale == 0) {
return false;
}
u_int8_t audioType =
MP4GetTrackEsdsObjectTypeId(mp4File, mediaTrackId);
if (audioType != MP4_MPEG4_AUDIO_TYPE
&& !MP4_IS_AAC_AUDIO_TYPE(audioType)) {
return false;
}
u_int8_t mpeg4AudioType =
MP4GetTrackAudioMpeg4Type(mp4File, mediaTrackId);
if (audioType == MP4_MPEG4_AUDIO_TYPE) {
// check that track contains either MPEG-4 AAC or CELP
if (!MP4_IS_MPEG4_AAC_AUDIO_TYPE(mpeg4AudioType)
&& mpeg4AudioType != MP4_MPEG4_CELP_AUDIO_TYPE) {
return false;
}
}
MP4Duration sampleDuration =
MP4AV_GetAudioSampleDuration(mp4File, mediaTrackId);
if (sampleDuration == MP4_INVALID_DURATION) {
return false;
}
/* get the ES configuration */
u_int8_t* pConfig = NULL;
u_int32_t configSize;
uint8_t channels;
if (MP4GetTrackESConfiguration(mp4File, mediaTrackId,
&pConfig, &configSize) == false)
return false;
if (!pConfig) {
return false;
}
channels = MP4AV_AacConfigGetChannels(pConfig);
/* convert ES Config into ASCII form */
char* sConfig =
MP4BinaryToBase16(pConfig, configSize);
free(pConfig);
if (!sConfig) {
return false;
}
/* create the appropriate SDP attribute */
uint sdpBufLen = strlen(sConfig) + 256;
char* sdpBuf =
(char*)malloc(sdpBufLen);
if (!sdpBuf) {
free(sConfig);
return false;
}
// now add the hint track
MP4TrackId hintTrackId =
MP4AddHintTrack(mp4File, mediaTrackId);
if (hintTrackId == MP4_INVALID_TRACK_ID) {
free(sConfig);
free(sdpBuf);
return false;
}
u_int8_t payloadNumber = MP4_SET_DYNAMIC_PAYLOAD;
char buffer[10];
if (channels != 1) {
snprintf(buffer, sizeof(buffer), "%u", channels);
}
if (MP4SetHintTrackRtpPayload(mp4File, hintTrackId,
"mpeg4-generic", &payloadNumber, 0,
channels != 1 ? buffer : NULL) == false) {
MP4DeleteTrack(mp4File, hintTrackId);
free(sConfig);
free(sdpBuf);
return false;
}
MP4Duration maxLatency;
bool OneByteHeader = false;
if (mpeg4AudioType == MP4_MPEG4_CELP_AUDIO_TYPE) {
snprintf(sdpBuf, sdpBufLen,
"a=fmtp:%u "
"streamtype=5; profile-level-id=15; mode=CELP-vbr; config=%s; "
"SizeLength=6; IndexLength=2; IndexDeltaLength=2; Profile=0;"
"\015\012",
payloadNumber,
sConfig);
// 200 ms max latency for ISMA profile 1
maxLatency = timeScale / 5;
OneByteHeader = true;
} else { // AAC
snprintf(sdpBuf, sdpBufLen,
"a=fmtp:%u "
"streamtype=5; profile-level-id=15; mode=AAC-hbr; config=%s; "
"SizeLength=13; IndexLength=3; IndexDeltaLength=3;"
"\015\012",
payloadNumber,
sConfig);
// 500 ms max latency for ISMA profile 1
maxLatency = timeScale / 2;
}
/* add this to the track's sdp */
bool val = MP4AppendHintTrackSdp(mp4File, hintTrackId, sdpBuf);
free(sConfig);
free(sdpBuf);
if (val == false) {
MP4DeleteTrack(mp4File, hintTrackId);
return false;
}
u_int32_t samplesPerPacket = 0;
if (interleave) {
u_int32_t maxSampleSize =
MP4GetTrackMaxSampleSize(mp4File, mediaTrackId);
// compute how many maximum size samples would fit in a packet
samplesPerPacket =
(maxPayloadSize - 2) / (maxSampleSize + 2);
// can't interleave if this number is 0 or 1
if (samplesPerPacket < 2) {
interleave = false;
}
}
bool rc;
if (interleave) {
u_int32_t samplesPerGroup = maxLatency / sampleDuration;
u_int32_t stride;
stride = samplesPerGroup / samplesPerPacket;
if (OneByteHeader && stride > 3) stride = 3;
if (!OneByteHeader && stride > 7) stride = 7;
#if 0
printf("max latency %llu sampleDuration %llu spg %u spp %u strid %u\n",
maxLatency, sampleDuration, samplesPerGroup,
samplesPerPacket, stride);
#endif
rc = MP4AV_AudioInterleaveHinter(
mp4File,
mediaTrackId,
hintTrackId,
sampleDuration,
stride, // stride
samplesPerPacket, // bundle
maxPayloadSize,
MP4AV_RfcIsmaConcatenator);
} else {
rc = MP4AV_AudioConsecutiveHinter(
mp4File,
mediaTrackId,
hintTrackId,
sampleDuration,
2, // perPacketHeaderSize
2, // perSampleHeaderSize
maxLatency / sampleDuration, // maxSamplesPerPacket
maxPayloadSize,
MP4GetSampleSize,
MP4AV_RfcIsmaConcatenator,
MP4AV_RfcIsmaFragmenter);
}
if (!rc) {
MP4DeleteTrack(mp4File, hintTrackId);
return false;
}
return true;
}
bool MP4AV_RfcIsmaConcatenator(
MP4FileHandle mp4File,
MP4TrackId mediaTrackId,
MP4TrackId hintTrackId,
u_int8_t samplesThisHint,
MP4SampleId* pSampleIds,
MP4Duration hintDuration,
u_int16_t maxPayloadSize)
{
// handle degenerate case
if (samplesThisHint == 0) {
return true;
}
u_int8_t auPayloadHdrSize;
// LATER would be more efficient if this were a parameter
u_int8_t mpeg4AudioType =
MP4GetTrackAudioMpeg4Type(mp4File, mediaTrackId);
if (mpeg4AudioType == MP4_MPEG4_CELP_AUDIO_TYPE) {
auPayloadHdrSize = 1;
} else {
auPayloadHdrSize = 2;
}
// construct the new hint
if (MP4AddRtpHint(mp4File, hintTrackId) == false ||
MP4AddRtpPacket(mp4File, hintTrackId, true) == false) return false;
u_int8_t payloadHeader[2];
u_int16_t numHdrBits = samplesThisHint * auPayloadHdrSize * 8;
payloadHeader[0] = numHdrBits >> 8;
payloadHeader[1] = numHdrBits & 0xFF;
if (MP4AddRtpImmediateData(mp4File, hintTrackId,
(u_int8_t*)&payloadHeader,
sizeof(payloadHeader)) == false) return false;
u_int8_t i;
// first the headers
for (i = 0; i < samplesThisHint; i++) {
MP4SampleId sampleId = pSampleIds[i];
u_int32_t sampleSize =
MP4GetSampleSize(mp4File, mediaTrackId, sampleId);
if (auPayloadHdrSize == 1) {
// AU payload header is 6 bits of size
// follow by 2 bits of the difference between sampleId's - 1
payloadHeader[0] = sampleSize << 2;
} else { // auPayloadHdrSize == 2
// AU payload header is 13 bits of size
// follow by 3 bits of the difference between sampleId's - 1
payloadHeader[0] = sampleSize >> 5;
payloadHeader[1] = (sampleSize & 0x1F) << 3;
}
if (i > 0) {
payloadHeader[auPayloadHdrSize - 1]
|= ((sampleId - pSampleIds[i-1]) - 1);
}
#if 0
printf("sample %u size %u %02x %02x prev sample %d\n",
sampleId, sampleSize, payloadHeader[0],
payloadHeader[1], pSampleIds[i-1]);
#endif
if (MP4AddRtpImmediateData(mp4File, hintTrackId,
(u_int8_t*)&payloadHeader,
auPayloadHdrSize) == false)
return false;
}
// then the samples
for (i = 0; i < samplesThisHint; i++) {
MP4SampleId sampleId = pSampleIds[i];
u_int32_t sampleSize =
MP4GetSampleSize(mp4File, mediaTrackId, sampleId);
if (MP4AddRtpSampleData(mp4File, hintTrackId,
sampleId, 0, sampleSize) == false) return false;
}
// write the hint
return MP4WriteRtpHint(mp4File, hintTrackId, hintDuration);
}
static char* PrintAudioInfo(
MP4FileHandle mp4File,
MP4TrackId trackId)
{
static const char* mpeg4AudioNames[] = {
"MPEG-4 AAC main",
"MPEG-4 AAC LC",
"MPEG-4 AAC SSR",
"MPEG-4 AAC LTP",
"MPEG-4 AAC HE",
"MPEG-4 AAC Scalable",
"MPEG-4 TwinVQ",
"MPEG-4 CELP",
"MPEG-4 HVXC",
NULL, NULL,
"MPEG-4 TTSI",
"MPEG-4 Main Synthetic",
"MPEG-4 Wavetable Syn",
"MPEG-4 General MIDI",
"MPEG-4 Algo Syn and Audio FX",
"MPEG-4 ER AAC LC",
NULL,
"MPEG-4 ER AAC LTP",
"MPEG-4 ER AAC Scalable",
"MPEG-4 ER TwinVQ",
"MPEG-4 ER BSAC",
"MPEG-4 ER ACC LD",
"MPEG-4 ER CELP",
"MPEG-4 ER HVXC",
"MPEG-4 ER HILN",
"MPEG-4 ER Parametric",
"MPEG-4 SSC",
"MPEG-4 PS",
"MPEG-4 MPEG Surround",
NULL,
"MPEG-4 Layer-1",
"MPEG-4 Layer-2",
"MPEG-4 Layer-3",
"MPEG-4 DST",
"MPEG-4 Audio Lossless",
"MPEG-4 SLS",
"MPEG-4 SLS non-core",
};
static const u_int8_t mpegAudioTypes[] = {
MP4_MPEG2_AAC_MAIN_AUDIO_TYPE, // 0x66
MP4_MPEG2_AAC_LC_AUDIO_TYPE, // 0x67
MP4_MPEG2_AAC_SSR_AUDIO_TYPE, // 0x68
MP4_MPEG2_AUDIO_TYPE, // 0x69
MP4_MPEG1_AUDIO_TYPE, // 0x6B
// private types
MP4_PCM16_LITTLE_ENDIAN_AUDIO_TYPE,
MP4_VORBIS_AUDIO_TYPE,
MP4_ALAW_AUDIO_TYPE,
MP4_ULAW_AUDIO_TYPE,
MP4_G723_AUDIO_TYPE,
MP4_PCM16_BIG_ENDIAN_AUDIO_TYPE,
};
static const char* mpegAudioNames[] = {
"MPEG-2 AAC Main",
"MPEG-2 AAC LC",
"MPEG-2 AAC SSR",
"MPEG-2 Audio (13818-3)",
"MPEG-1 Audio (11172-3)",
// private types
"PCM16 (little endian)",
"Vorbis",
"G.711 aLaw",
"G.711 uLaw",
"G.723.1",
"PCM16 (big endian)",
};
u_int8_t numMpegAudioTypes =
sizeof(mpegAudioTypes) / sizeof(u_int8_t);
const char* typeName = "Unknown";
bool foundType = false;
u_int8_t type = 0;
const char *media_data_name;
media_data_name = MP4GetTrackMediaDataName(mp4File, trackId);
if (media_data_name == NULL) {
typeName = "Unknown - no media data name";
} else if (strcasecmp(media_data_name, "samr") == 0) {
typeName = "AMR";
foundType = true;
} else if (strcasecmp(media_data_name, "sawb") == 0) {
typeName = "AMR-WB";
foundType = true;
} else if (strcasecmp(media_data_name, "mp4a") == 0) {
type = MP4GetTrackEsdsObjectTypeId(mp4File, trackId);
switch (type) {
case MP4_INVALID_AUDIO_TYPE:
typeName = "AAC from .mov";
foundType = true;
break;
case MP4_MPEG4_AUDIO_TYPE: {
type = MP4GetTrackAudioMpeg4Type(mp4File, trackId);
if (type == MP4_MPEG4_INVALID_AUDIO_TYPE ||
type > NUM_ELEMENTS_IN_ARRAY(mpeg4AudioNames) ||
mpeg4AudioNames[type - 1] == NULL) {
typeName = "MPEG-4 Unknown Profile";
} else {
typeName = mpeg4AudioNames[type - 1];
foundType = true;
}
break;
}
// fall through
default:
for (u_int8_t i = 0; i < numMpegAudioTypes; i++) {
if (type == mpegAudioTypes[i]) {
typeName = mpegAudioNames[i];
foundType = true;
break;
}
}
}
} else {
typeName = media_data_name;
foundType = true;
}
u_int32_t timeScale =
MP4GetTrackTimeScale(mp4File, trackId);
MP4Duration trackDuration =
MP4GetTrackDuration(mp4File, trackId);
double msDuration =
UINT64_TO_DOUBLE(MP4ConvertFromTrackDuration(mp4File, trackId,
trackDuration, MP4_MSECS_TIME_SCALE));
u_int32_t avgBitRate =
MP4GetTrackBitRate(mp4File, trackId);
char *sInfo = (char*)MP4Malloc(256);
// type duration avgBitrate samplingFrequency
if (foundType)
snprintf(sInfo, 256,
"%u\taudio\t%s%s, %.3f secs, %u kbps, %u Hz\n",
trackId,
MP4IsIsmaCrypMediaTrack(mp4File, trackId) ? "enca - " : "",
typeName,
msDuration / 1000.0,
(avgBitRate + 500) / 1000,
timeScale);
else
snprintf(sInfo, 256,
"%u\taudio\t%s%s(%u), %.3f secs, %u kbps, %u Hz\n",
trackId,
MP4IsIsmaCrypMediaTrack(mp4File, trackId) ? "enca - " : "",
typeName,
type,
msDuration / 1000.0,
(avgBitRate + 500) / 1000,
timeScale);
return sInfo;
}
