commit 3fbdec07724f4806f88dd93ac3927162b2cdaabd
Author: Joel Smith <[email protected]>
Date: Thu Nov 8 14:23:35 2012 -0700
add mp4 chapter reading support to Atomic Parsley bridge
libs/atomic-parsley/AtomicParsleyBridge.cpp | 125 +++++++++++++++++++++++++--
1 files changed, 119 insertions(+), 6 deletions(-)
---
diff --git a/libs/atomic-parsley/AtomicParsleyBridge.cpp
b/libs/atomic-parsley/AtomicParsleyBridge.cpp
index ef1a182..704fd3b 100644
--- a/libs/atomic-parsley/AtomicParsleyBridge.cpp
+++ b/libs/atomic-parsley/AtomicParsleyBridge.cpp
@@ -146,7 +146,10 @@ void AP_read_metadata(const char *filePath, Track *track) {
FILE *mp4File;
Trackage *trackage;
uint8_t track_cur;
+ uint8_t txttrack_cur;
gboolean audio_or_video_found = FALSE;
+ gboolean has_quicktime_chaps = FALSE;
+ uint32_t timescale = 0;
APar_ScanAtoms(filePath, true);
mp4File = openSomeFile(filePath, true);
@@ -156,11 +159,8 @@ void AP_read_metadata(const char *filePath, Track *track) {
for (track_cur = 0; track_cur < trackage->total_tracks; ++track_cur) {
TrackInfo *info = trackage->infos[track_cur];
- // FIXME no chapter information implemented yet
-
- if (info->track_type && audio_or_video_found == FALSE
- && ((info->track_type & AUDIO_TRACK) || (info->track_type &
VIDEO_TRACK)
- || (info->track_type & DRM_PROTECTED_TRACK))) {
+ if ((info->type_of_track & AUDIO_TRACK) || (info->type_of_track &
VIDEO_TRACK)
+ || (info->type_of_track & DRM_PROTECTED_TRACK)) {
/*
* the info->duration is in the track's timescale units so must be
divided by that
@@ -171,10 +171,123 @@ void AP_read_metadata(const char *filePath, Track
*track) {
track->bitrate = APar_calculate_bitrate(info);
track->samplerate = info->media_sample_rate;
+ audio_or_video_found = TRUE;
+ break;
+ }
+ }
+ for (txttrack_cur = 0; audio_or_video_found && txttrack_cur <
trackage->total_tracks; ++txttrack_cur) {
+ TrackInfo *txtinfo = trackage->infos[txttrack_cur];
+ char buf[128];
+ // search for chapter track
+ if (!(txtinfo->type_of_track & TEXT_TRACK))
+ continue;
+ // see if the AV track's chap refers to this text track
+ // chap: 0: atom size 4: 'chap' 8,12,...,8+(N-1)*4: (0: referenced
track ID)
+ snprintf(buf, sizeof(buf), "moov.trak[%u].tref.chap", track_cur + 1);
+ AtomicInfo* chapAtom = APar_FindAtom(buf, false, SIMPLE_ATOM, 0);
+ if (!chapAtom)
+ continue;
+ int entry_count = (chapAtom->AtomicLength - 8) / 4;
+ for (int i = 0; i < entry_count; ++i) {
+ if (APar_read32(buf, mp4File, chapAtom->AtomicStart + 8 + i * 4)
== txtinfo->track_id) {
+ has_quicktime_chaps = TRUE;
+ timescale = txtinfo->media_sample_rate;
+ break;
+ }
}
+ if (has_quicktime_chaps)
+ break;
+ }
+ if (has_quicktime_chaps) {
+ // found a chapter... now get the chapter data from the text track
+ char buf[128];
+
+ // stts: 0: atom size 4: 'stts' 8: version 12: entry count
16,24,...,16+(N-1)*8: (0: frame count 4: duration)
+ snprintf(buf, sizeof(buf), "moov.trak[%u].mdia.minf.stbl.stts",
txttrack_cur + 1);
+ AtomicInfo* sampleAtom = APar_FindAtom(buf, false, VERSIONED_ATOM, 0);
+
+ // stsz: 0: atom size 4: 'stsz' 8: version 12: size of all (or 0)
16: entry count 20,24,...,20+(N-1)*4: (0: sample size)
+ snprintf(buf, sizeof(buf), "moov.trak[%u].mdia.minf.stbl.stsz",
txttrack_cur + 1);
+ AtomicInfo* sampleSizeAtom = APar_FindAtom(buf, false, VERSIONED_ATOM,
0);
+
+ // stco: 0: atom size 4: 'stco' 8: version 12: entry count
16,20,...,16+(N-1)*4: (0: sample byte offset)
+ snprintf(buf, sizeof(buf), "moov.trak[%u].mdia.minf.stbl.stco",
txttrack_cur + 1);
+ AtomicInfo* sampleOffsetAtom = APar_FindAtom(buf, false,
VERSIONED_ATOM, 0);
+
+ // We must have a valid sampleAtom to know chapter times. If
sampleSizeAtom or sampleOffsetAtom is invalid,
+ // we can do without them (and instead create a default chapter name).
+ if (sampleAtom && sampleAtom->AtomicLength >= 16) {
+ Itdb_Chapterdata *chapterdata = itdb_chapterdata_new();
+ uint32_t stts_entry_count = APar_read32(buf, mp4File,
sampleAtom->AtomicStart + 12);
+ uint32_t stsz_entry_count = !sampleSizeAtom ||
sampleSizeAtom->AtomicLength < 20 ? 0 :
+ APar_read32(buf, mp4File, sampleSizeAtom->AtomicStart + 16);
+ uint32_t stco_entry_count = !sampleOffsetAtom ||
sampleOffsetAtom->AtomicLength < 16 ? 0 :
+ APar_read32(buf, mp4File, sampleOffsetAtom->AtomicStart + 12);
+ uint32_t stsz_all_size = !sampleSizeAtom ||
sampleSizeAtom->AtomicLength < 16 ? 0 :
+ APar_read32(buf, mp4File, sampleSizeAtom->AtomicStart + 12);
+
+ uint32_t start_time = 0;
+
+ u_int32_t max_frame_size = stsz_all_size; // if stsz_all_size
specified, use only that size
+ for (int i = 0; !stsz_all_size && i < stsz_entry_count; ++i) {
+ uint32_t chap_name_len = APar_read32(buf, mp4File,
sampleSizeAtom->AtomicStart + 20 + i * 4);
+ if (chap_name_len > max_frame_size)
+ max_frame_size = chap_name_len;
+ }
+ max_frame_size += 1; // for trailing '\0' (unneeded?), and to make
sure that malloc() gets passed at least 1
+ char * namebuf = (char *)malloc(max_frame_size * sizeof(char));
+ for (int i = 0; i < stts_entry_count; ++i) {
+ gchar *title = NULL;
+ uint32_t chap_name_len = stsz_all_size;
+ uint32_t chap_offset = 0;
+ if (stsz_all_size == 0 && i < stsz_entry_count)
+ chap_name_len = APar_read32(buf, mp4File,
sampleSizeAtom->AtomicStart + 20 + i * 4);
+ if (i < stco_entry_count)
+ chap_offset = APar_read32(buf, mp4File,
sampleOffsetAtom->AtomicStart + 16 + i * 4);
+ if (chap_offset != 0)
+ APar_readX(namebuf, mp4File, chap_offset, chap_name_len);
+ else // If the location of the chapter name is unknown,
trigger default chapter naming
+ chap_name_len = 0;
+ if (chap_name_len > 2) {
+ int titlelength = (namebuf[0] << 8) + namebuf[1];
+ // if the stsz atom and the title value disagree, use the
smaller one for safety
+ titlelength = (titlelength > chap_name_len) ?
chap_name_len : titlelength;
+ // If a title begins with 0xFFFE, it's a UTF-16 title
+ if (titlelength >= 2 && namebuf[2] == 0xff && namebuf[3]
== 0xfe)
+ title = g_utf16_to_utf8((const gunichar2 *)
&namebuf[4], titlelength - 2, NULL, NULL, NULL);
+ else
+ title = g_strndup(&namebuf[2], titlelength);
+ }
+ else
+ {
+ // chapter title couldn't be found; create our own titles
+ // (and some ipods don't display them anyway)
+ title = g_strdup_printf("Chapter %3d", i);
+ }
+
+ if (!timescale) // assume 1000, also, don't divide by 0
+ timescale = 1000;
+ double duration_ms = (double)start_time * 1000.0 /
(double)timescale;
+
+ itdb_chapterdata_add_chapter(chapterdata, duration_ms, title);
+ g_free(title);
- audio_or_video_found = TRUE;
+ if (i < (stts_entry_count - 1)) // skip this stage after the
last chapter has been added
+ {
+ uint32_t frame_count = APar_read32(buf, mp4File,
sampleAtom->AtomicStart + 16 + i * 8);
+ uint32_t duration = APar_read32(buf, mp4File,
sampleAtom->AtomicStart + 20 + i * 8);
+ start_time += frame_count * duration;
+ }
+ }
+ if (namebuf)
+ free(namebuf);
+ if (track->chapterdata) // if there was already chapter data,
don't leak it
+ itdb_chapterdata_free(track->chapterdata);
+ track->chapterdata = itdb_chapterdata_duplicate(chapterdata);
+ itdb_chapterdata_free(chapterdata);
+ }
}
+ // TODO: add support for Nero-style mp4 chapters
if (prefs_get_int("readtags")) {
char* value = NULL;
------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
http://p.sf.net/sfu/appdyn_d2d_nov
_______________________________________________
gtkpod-cvs2 mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/gtkpod-cvs2
