MP4Reader(const std::string &file_path)
: time_scale(9 * MP4_MSECS_TIME_SCALE)
, file_path(file_path)
, handle(MP4_INVALID_FILE_HANDLE)
, video_track_id(MP4_INVALID_TRACK_ID)
, next_video_sample_idx(1)
, video_sample(nullptr)
, video_timescale(0)
, video_sample_max_size(0)
, video_sample_number(0)
, video_duration(0)
, pSeqHeaders(nullptr)
, pSeqHeaderSize(nullptr)
, pPictHeaders(nullptr)
, pPictHeaderSize(nullptr)
{
handle = MP4Read(this->file_path.c_str());
video_track_id = MP4FindTrackId(handle, 0, MP4_VIDEO_TRACK_TYPE);
if (video_track_id != MP4_INVALID_TRACK_ID) {
video_timescale = MP4GetTrackTimeScale(handle, video_track_id);
video_sample_max_size = MP4GetTrackMaxSampleSize(handle, video_track_id) * 2;
video_duration = MP4GetTrackDuration(handle, video_track_id);
video_sample = new unsigned char[video_sample_max_size];
video_sample_number = MP4GetTrackNumberOfSamples(handle, video_track_id);
if (MP4GetTrackH264SeqPictHeaders(handle,
video_track_id,
&pSeqHeaders,
&pSeqHeaderSize,
&pPictHeaders,
&pPictHeaderSize))
{
printf("Get SPS(%d) and PPS(%d)\n", *pSeqHeaderSize, *pPictHeaderSize);
for(int i = 0; (pSeqHeaders[i] && pSeqHeaderSize[i]); i++) {
printf("SPS(%d): %02x %02x %02x %02x %02x\n", i,
pSeqHeaders[i][0], pSeqHeaders[i][1], pSeqHeaders[i][2],
pSeqHeaders[i][3], pSeqHeaders[i][4]);
}
for(int i = 0; (pPictHeaders[i] && pPictHeaderSize[i]); i++) {
printf("PPS(%d): %02x %02x %02x %02x %02x\n", i,
pPictHeaders[i][0], pPictHeaders[i][1], pPictHeaders[i][2],
pPictHeaders[i][3], pPictHeaders[i][4]);
}
}
}
}
int CMp4File::create_media (CPlayerSession *psptr,
int have_audio_driver,
control_callback_vft_t *cc_vft)
{
uint video_count, video_offset;
uint text_count, text_offset;
uint audio_count, audio_offset;
MP4TrackId trackId;
video_query_t *vq;
audio_query_t *aq;
text_query_t *tq;
uint ix;
codec_plugin_t *plugin;
int ret_value = 0;
uint8_t *foo;
u_int32_t bufsize;
uint32_t verb = MP4GetVerbosity(m_mp4file);
MP4SetVerbosity(m_mp4file, verb & ~(MP4_DETAILS_ERROR));
video_count = MP4GetNumberOfTracks(m_mp4file, MP4_VIDEO_TRACK_TYPE);
audio_count = MP4GetNumberOfTracks(m_mp4file, MP4_AUDIO_TRACK_TYPE);
text_count = MP4GetNumberOfTracks(m_mp4file, MP4_CNTL_TRACK_TYPE);
mp4f_message(LOG_DEBUG, "cntl tracks %u", text_count);
MP4SetVerbosity(m_mp4file, verb);
if (video_count == 0 && audio_count == 0 && text_count == 0) {
psptr->set_message("No audio, video or control tracks in file");
return -1;
}
if (video_count > 0) {
vq = (video_query_t *)malloc(sizeof(video_query_t) * video_count);
memset(vq, 0, sizeof(video_query_t) * video_count);
} else {
vq = NULL;
}
if (have_audio_driver && audio_count > 0) {
aq = (audio_query_t *)malloc(sizeof(audio_query_t) * audio_count);
memset(aq, 0, sizeof(audio_query_t) * audio_count);
} else {
aq = NULL;
}
if (text_count > 0) {
tq = (text_query_t *)malloc(sizeof(text_query_t) * text_count);
memset(tq, 0, sizeof(text_query_t) * text_count);
} else {
tq = NULL;
}
for (ix = 0, video_offset = 0; ix < video_count; ix++) {
trackId = MP4FindTrackId(m_mp4file, ix, MP4_VIDEO_TRACK_TYPE);
const char *media_data_name;
media_data_name = MP4GetTrackMediaDataName(m_mp4file, trackId);
// for now, treat mp4v and encv the same
vq[video_offset].track_id = trackId;
vq[video_offset].stream_type = STREAM_TYPE_MP4_FILE;
vq[video_offset].compressor = media_data_name;
if (strcasecmp(media_data_name, "mp4v") == 0 ||
strcasecmp(media_data_name, "encv") == 0) {
uint8_t video_type = MP4GetTrackEsdsObjectTypeId(m_mp4file, trackId);
uint8_t profileID = MP4GetVideoProfileLevel(m_mp4file, trackId);
mp4f_message(LOG_DEBUG, "MP4 - got track %x profile ID %d",
trackId, profileID);
MP4SetVerbosity(m_mp4file, verb & ~(MP4_DETAILS_ERROR));
MP4GetTrackESConfiguration(m_mp4file, trackId, &foo, &bufsize);
MP4SetVerbosity(m_mp4file, verb);
vq[video_offset].type = video_type;
vq[video_offset].profile = profileID;
vq[video_offset].fptr = NULL;
vq[video_offset].config = foo;
vq[video_offset].config_len = bufsize;
} else if (strcasecmp(media_data_name, "avc1") == 0) {
uint8_t profile, level;
uint8_t **seqheader, **pictheader;
uint32_t *pictheadersize, *seqheadersize;
uint32_t ix;
MP4GetTrackH264ProfileLevel(m_mp4file, trackId, &profile, &level);
MP4GetTrackH264SeqPictHeaders(m_mp4file, trackId,
&seqheader, &seqheadersize,
&pictheader, &pictheadersize);
bufsize = 0;
for (ix = 0; seqheadersize[ix] != 0; ix++) {
bufsize += seqheadersize[ix] + 4;
}
for (ix = 0; pictheadersize[ix] != 0; ix++) {
bufsize += pictheadersize[ix] + 4;
}
foo = (uint8_t *)malloc(bufsize + 4);
memset(foo, 0, bufsize + 4);
uint32_t copied = 0;
// headers do not have the byte stream start code stored in the file
for (ix = 0; seqheadersize[ix] != 0; ix++) {
foo[copied] = 0;
foo[copied + 1] = 0;
foo[copied + 2] = 0;
foo[copied + 3] = 1;
copied += 4; // add header
memcpy(foo + copied,
seqheader[ix],
seqheadersize[ix]);
copied += seqheadersize[ix];
free(seqheader[ix]);
}
free(seqheader);
free(seqheadersize);
for (ix = 0; pictheadersize[ix] != 0; ix++) {
foo[copied] = 0;
foo[copied + 1] = 0;
foo[copied + 2] = 0;
foo[copied + 3] = 1;
copied += 4; // add header
memcpy(foo + copied,
pictheader[ix],
pictheadersize[ix]);
copied += pictheadersize[ix];
free(pictheader[ix]);
}
free(pictheader);
free(pictheadersize);
vq[video_offset].type = level;
vq[video_offset].profile = profile;
vq[video_offset].fptr = NULL;
vq[video_offset].config = foo;
vq[video_offset].config_len = bufsize;
} else {
MP4GetTrackVideoMetadata(m_mp4file, trackId, &foo, &bufsize);
vq[video_offset].config = foo;
vq[video_offset].config_len = bufsize;
}
plugin = check_for_video_codec(vq[video_offset].stream_type,
vq[video_offset].compressor,
NULL,
vq[video_offset].type,
vq[video_offset].profile,
vq[video_offset].config,
vq[video_offset].config_len,
&config);
if (plugin == NULL) {
psptr->set_message("Can't find plugin for video %s type %d, profile %d",
vq[video_offset].compressor,
vq[video_offset].type,
vq[video_offset].profile);
m_illegal_video_codec++;
ret_value = 1;
// possibly memleak for foo here
} else {
vq[video_offset].h = MP4GetTrackVideoHeight(m_mp4file, trackId);
vq[video_offset].w = MP4GetTrackVideoWidth(m_mp4file, trackId);
vq[video_offset].frame_rate = MP4GetTrackVideoFrameRate(m_mp4file, trackId);
vq[video_offset].enabled = 0;
vq[video_offset].reference = NULL;
video_offset++;
}
}
audio_offset = 0;
if (have_audio_driver) {
for (ix = 0; ix < audio_count; ix++) {
trackId = MP4FindTrackId(m_mp4file, ix, MP4_AUDIO_TRACK_TYPE);
const char *media_data_name;
media_data_name = MP4GetTrackMediaDataName(m_mp4file, trackId);
aq[audio_offset].track_id = trackId;
aq[audio_offset].stream_type = STREAM_TYPE_MP4_FILE;
aq[audio_offset].compressor = media_data_name;
if (strcasecmp(media_data_name, "mp4a") == 0 ||
strcasecmp(media_data_name, "enca") == 0) {
uint8_t *userdata = NULL;
u_int32_t userdata_size;
aq[audio_offset].type = MP4GetTrackEsdsObjectTypeId(m_mp4file, trackId);
MP4SetVerbosity(m_mp4file, verb & ~(MP4_DETAILS_ERROR));
aq[audio_offset].profile = MP4GetAudioProfileLevel(m_mp4file);
MP4GetTrackESConfiguration(m_mp4file,
trackId,
&userdata,
&userdata_size);
MP4SetVerbosity(m_mp4file, verb);
aq[audio_offset].config = userdata;
aq[audio_offset].config_len = userdata_size;
}
plugin = check_for_audio_codec(aq[audio_offset].stream_type,
aq[audio_offset].compressor,
NULL,
aq[audio_offset].type,
aq[audio_offset].profile,
aq[audio_offset].config,
aq[audio_offset].config_len,
&config);
if (plugin != NULL) {
aq[audio_offset].fptr = NULL;
aq[audio_offset].sampling_freq =
MP4GetTrackTimeScale(m_mp4file, trackId);
MP4SetVerbosity(m_mp4file, verb & ~(MP4_DETAILS_ERROR));
aq[audio_offset].chans = MP4GetTrackAudioChannels(m_mp4file, trackId);
MP4SetVerbosity(m_mp4file, verb);
aq[audio_offset].enabled = 0;
aq[audio_offset].reference = NULL;
audio_offset++;
m_have_audio = true;
} else {
m_illegal_audio_codec++;
ret_value = 1;
}
}
} else {
if (audio_count)
ret_value = 1;
}
text_offset = 0;
for (ix = 0; ix < text_count; ix++) {
trackId = MP4FindTrackId(m_mp4file, ix, MP4_CNTL_TRACK_TYPE);
const char *media_data_name;
media_data_name = MP4GetTrackMediaDataName(m_mp4file, trackId);
tq[text_offset].track_id = trackId;
tq[text_offset].stream_type = STREAM_TYPE_MP4_FILE;
tq[text_offset].compressor = media_data_name;
plugin = check_for_text_codec(tq[text_offset].stream_type,
tq[text_offset].compressor,
NULL,
NULL,
0,
&config);
if (plugin != NULL) {
tq[text_offset].fptr = NULL;
tq[text_offset].enabled = 0;
tq[text_offset].reference = NULL;
text_offset++;
} else {
m_illegal_text_codec++;
ret_value = 1;
}
}
if (video_offset == 0 && audio_offset == 0 && text_offset == 0) {
psptr->set_message("No playable codecs in mp4 file");
return -1;
}
if (cc_vft && cc_vft->media_list_query != NULL) {
(cc_vft->media_list_query)(psptr, video_offset, vq, audio_offset, aq, text_offset, tq);
} else {
if (video_offset > 0) {
vq[0].enabled = 1;
}
if (audio_offset > 0) {
aq[0].enabled = 1;
}
if (text_offset > 0) {
tq[0].enabled = 1;
}
}
int vidret, audret, textret;
uint start_desc = 1;
vidret = create_video(psptr, vq, video_offset, start_desc);
free(vq);
if (vidret < 0) {
free(aq);
free(tq);
return -1;
}
audret = create_audio(psptr, aq, audio_offset, start_desc);
free(aq);
textret = create_text(psptr, tq, text_offset, start_desc);
free(tq);
if (audret < 0 || textret < 0) ret_value = -1;
char *name;
verb = MP4GetVerbosity(m_mp4file);
MP4SetVerbosity(m_mp4file, verb & ~(MP4_DETAILS_ERROR));
if (MP4GetMetadataName(m_mp4file, &name) &&
name != NULL) {
psptr->set_session_desc(0, name);
free(name);
}
MP4SetVerbosity(m_mp4file, verb);
return (ret_value);
}
AVCDescriptor* MP4Streamer::GetAVCDescriptor()
{
uint8_t **sequenceHeader;
uint8_t **pictureHeader;
uint32_t *pictureHeaderSize;
uint32_t *sequenceHeaderSize;
uint32_t i;
uint8_t profile, level;
uint32_t len;
//Check video
if (!video || video->codec!=VideoCodec::H264)
//Nothing
return NULL;
//Create descriptor
AVCDescriptor* desc = new AVCDescriptor();
//Set them
desc->SetConfigurationVersion(0x01);
desc->SetAVCProfileIndication(profile);
desc->SetProfileCompatibility(0x00);
desc->SetAVCLevelIndication(level);
//Set nalu length
MP4GetTrackH264LengthSize(video->mp4, video->track, &len);
//Set it
desc->SetNALUnitLength(len-1);
// Get SEI informaMP4GetTrackH264SeqPictHeaderstion
MP4GetTrackH264SeqPictHeaders(video->mp4, video->track, &sequenceHeader, &sequenceHeaderSize, &pictureHeader, &pictureHeaderSize);
// Send sequence headers
i=0;
// Check we have sequence header
if (sequenceHeader)
{
// Loop array
while(sequenceHeader[i] && sequenceHeaderSize[i])
{
//Append sequence
desc->AddSequenceParameterSet(sequenceHeader[i],sequenceHeaderSize[i]);
//Update values based on the ones in SQS
desc->SetAVCProfileIndication(sequenceHeader[i][1]);
desc->SetProfileCompatibility(sequenceHeader[i][2]);
desc->SetAVCLevelIndication(sequenceHeader[i][3]);
//inc
i++;
}
}
// Send picture headers
i=0;
// Check we have picture header
if (pictureHeader)
{
// Loop array
while(pictureHeader[i] && pictureHeaderSize[i])
{
//Append sequence
desc->AddPictureParameterSet(pictureHeader[i],pictureHeaderSize[i]);
//inc
i++;
}
}
// Free data
if (pictureHeader)
free(pictureHeader);
if (sequenceHeader)
free(sequenceHeader);
if (sequenceHeaderSize)
free(sequenceHeaderSize);
if (pictureHeaderSize)
free(pictureHeaderSize);
return desc;
}
extern "C" MP4TrackId MP4AV_H264_HintTrackCreate (MP4FileHandle mp4File,
MP4TrackId mediaTrackId)
{
MP4TrackId hintTrackId =
MP4AddHintTrack(mp4File, mediaTrackId);
if (hintTrackId == MP4_INVALID_TRACK_ID) {
return MP4_INVALID_TRACK_ID;
}
u_int8_t payloadNumber = MP4_SET_DYNAMIC_PAYLOAD;
// don't include mpeg4-esid
MP4SetHintTrackRtpPayload(mp4File, hintTrackId,
"H264", &payloadNumber, 0,
NULL, true, false);
/* get the mpeg4 video configuration */
u_int8_t **pSeq, **pPict ;
u_int32_t *pSeqSize, *pPictSize;
char *base64;
uint32_t profile_level;
char *sprop = NULL;
uint32_t ix = 0;
MP4GetTrackH264SeqPictHeaders(mp4File,
mediaTrackId,
&pSeq,
&pSeqSize,
&pPict,
&pPictSize);
if (pSeqSize && pSeqSize[0] != 0) {
// we have valid sequence and picture headers
uint8_t *p = pSeq[0];
if (*p == 0 && p[1] == 0 &&
(p[2] == 1 || (p[2] == 0 && p[3] == 0))) {
if (p[2] == 0) p += 4;
else p += 3;
}
profile_level = p[0] << 16 |
p[1] << 8 |
p[2];
while (pSeqSize[ix] != 0) {
base64 = MP4BinaryToBase64(pSeq[ix], pSeqSize[ix]);
if (sprop == NULL) {
sprop = strdup(base64);
} else {
sprop = (char *)realloc(sprop, strlen(sprop) + strlen(base64) + 1 + 1);
strcat(sprop, ",");
strcat(sprop, base64);
}
free(base64);
free(pSeq[ix]);
ix++;
}
free(pSeq);
free(pSeqSize);
ix = 0;
while (pPictSize[ix] != 0) {
base64 = MP4BinaryToBase64(pPict[ix], pPictSize[ix]);
sprop = (char *)realloc(sprop, strlen(sprop) + strlen(base64) + 1 + 1);
strcat(sprop, ",");
strcat(sprop, base64);
free(base64);
free(pPict[ix]);
ix++;
}
free(pPict);
free(pPictSize);
/* create the appropriate SDP attribute */
char* sdpBuf = (char*)malloc(strlen(sprop) + 128);
sprintf(sdpBuf,
"a=fmtp:%u profile-level-id=%06x; sprop-parameter-sets=%s; packetization-mode=1\015\012",
payloadNumber,
profile_level,
sprop);
/* add this to the track's sdp */
MP4AppendHintTrackSdp(mp4File, hintTrackId, sdpBuf);
free(sprop);
free(sdpBuf);
}
return hintTrackId;
}
int MP4RtpTrack::SendH263SEI(Listener *listener)
{
uint8_t **sequenceHeader;
uint8_t **pictureHeader;
uint32_t *pictureHeaderSize;
uint32_t *sequenceHeaderSize;
uint32_t i;
uint8_t* data;
uint32_t dataLen;
//Not mark
rtp.SetMark(false);
// Get SEI information
MP4GetTrackH264SeqPictHeaders(mp4, track, &sequenceHeader, &sequenceHeaderSize, &pictureHeader, &pictureHeaderSize);
// Get data pointer
data = rtp.GetMediaData();
// Reset length
dataLen = 0;
// Send sequence headers
i=0;
// Check we have sequence header
if (sequenceHeader)
// Loop array
while(sequenceHeader[i] && sequenceHeaderSize[i])
{
// Check if it can be handled in a single packeti
if (sequenceHeaderSize[i]<1400)
{
// If there is not enought length
if (dataLen+sequenceHeaderSize[i]>1400)
{
// Set data length
rtp.SetMediaLength(dataLen);
//Check listener
if (listener)
// Write frame
listener->onRTPPacket(rtp);
// Reset data
dataLen = 0;
}
// Copy data
memcpy(data+dataLen,sequenceHeader[i],sequenceHeaderSize[i]);
// Increase pointer
dataLen+=sequenceHeaderSize[i];
}
// Free memory
free(sequenceHeader[i]);
// Next header
i++;
}
// If there is still data
if (dataLen>0)
{
// Set data length
rtp.SetMediaLength(dataLen);
//Check listener
if (listener)
// Write frame
listener->onRTPPacket(rtp);
// Reset data
dataLen = 0;
}
// Send picture headers
i=0;
// Check we have picture header
if (pictureHeader)
// Loop array
while(pictureHeader[i] && pictureHeaderSize[i])
{
// Check if it can be handled in a single packeti
if (pictureHeaderSize[i]<1400)
{
// If there is not enought length
if (dataLen+pictureHeaderSize[i]>1400)
{
// Set data length
rtp.SetMediaLength(dataLen);
//Check listener
if (listener)
// Write frame
listener->onRTPPacket(rtp);
// Reset data
dataLen = 0;
}
// Copy data
memcpy(data+dataLen,pictureHeader[i],pictureHeaderSize[i]);
// Increase pointer
dataLen+=pictureHeaderSize[i];
}
// Free memory
free(pictureHeader[i]);
// Next header
i++;
}
// If there is still data
if (dataLen>0)
{
// Set data length
rtp.SetMediaLength(dataLen);
//Check listener
if (listener)
// Write frame
listener->onRTPPacket(rtp);
// Reset data
dataLen = 0;
}
// Free data
if (pictureHeader)
free(pictureHeader);
if (sequenceHeader)
free(sequenceHeader);
if (sequenceHeaderSize)
free(sequenceHeaderSize);
if (pictureHeaderSize)
free(pictureHeaderSize);
}
