feat: mp4 chapter: track or udta

libmov/include/mov-format.h

@@ -29,6 +29,7 @@
 #define MOV_OBJECT_VP9      0xB1 // VP9 Video
 #define MOV_OBJECT_FLAC     0xC1 // nonstandard from FFMPEG
 #define MOV_OBJECT_VP8      0xC2 // nonstandard
+#define MOV_OBJECT_CHAPTER	0xC3 // chapter https://developer.apple.com/documentation/quicktime-file-format/base_media_information_header_atom
 #define MOV_OBJECT_H266		0xFC // ITU-T Recommendation H.266
 #define MOV_OBJECT_G711a	0xFD // ITU G.711 alaw
 #define MOV_OBJECT_G711u	0xFE // ITU G.711 ulaw

libmov/include/mov-udta.h

@@ -40,6 +40,13 @@ struct mov_udta_meta_t
 
 int mov_udta_meta_write(const struct mov_udta_meta_t* meta, void* data, int bytes);
 
+struct mov_udta_chapter_t
+{
+	uint64_t timestamp; // 1s = 10 * 1000 * 1000
+	const char* title; // optional
+};
+int mov_udta_chapter_write(const struct mov_udta_chapter_t* chapters, int count, void* data, int bytes);
+
 #ifdef __cplusplus
 }
 #endif

libmov/source/mov-internal.h

@@ -177,6 +177,8 @@ struct mov_track_t
 
 	struct mov_elst_t* elst;
 	size_t elst_count;
+
+	uint32_t chpl_track; // MOV_CHAPTER track
 
 	struct mov_sample_t* samples;
 	uint32_t sample_count;
@@ -254,6 +256,7 @@ int mov_read_text(struct mov_t* mov, const struct mov_box_t* box);
 int mov_read_smdm(struct mov_t* mov, const struct mov_box_t* box);
 int mov_read_coll(struct mov_t* mov, const struct mov_box_t* box);
 int mov_read_udta(struct mov_t* mov, const struct mov_box_t* box);
+int mov_read_chpl(struct mov_t* mov, const struct mov_box_t* box);
 
 size_t mov_write_ftyp(const struct mov_t* mov);
 size_t mov_write_mvhd(const struct mov_t* mov);
@@ -263,6 +266,7 @@ size_t mov_write_hdlr(const struct mov_t* mov);
 size_t mov_write_vmhd(const struct mov_t* mov);
 size_t mov_write_smhd(const struct mov_t* mov);
 size_t mov_write_nmhd(const struct mov_t* mov);
+size_t mov_write_gmhd(const struct mov_t* mov);
 size_t mov_write_sthd(const struct mov_t* mov);
 size_t mov_write_dinf(const struct mov_t* mov);
 size_t mov_write_dref(const struct mov_t* mov);
@@ -291,6 +295,7 @@ size_t mov_write_mehd(const struct mov_t* mov);
 size_t mov_write_sidx(const struct mov_t* mov, uint64_t offset);
 size_t mov_write_mfhd(const struct mov_t* mov, uint32_t fragment);
 size_t mov_write_edts(const struct mov_t* mov);
+size_t mov_write_tref(const struct mov_t* mov);
 size_t mov_write_stbl(const struct mov_t* mov);
 size_t mov_write_minf(const struct mov_t* mov);
 size_t mov_write_mdia(const struct mov_t* mov);

libmov/source/mov-minf.c

@@ -76,6 +76,25 @@ int mov_read_text(struct mov_t* mov, const struct mov_box_t* box)
 	return 0;
 }
 
+int mov_read_chpl(struct mov_t* mov, const struct mov_box_t* box)
+{
+	int i, count, len;
+	mov_buffer_r8(&mov->io); /* version */
+	mov_buffer_r24(&mov->io); /* flags */
+	mov_buffer_skip(&mov->io, 4); /* unknown*/
+
+	count = mov_buffer_r8(&mov->io); /* count */
+	for (i = 0; i < count; i++)
+	{
+		mov_buffer_r64(&mov->io); /* timestamp */
+		len = mov_buffer_r8(&mov->io); /* title size */
+		mov_buffer_skip(&mov->io, len); /* title */
+	}
+
+	(void)box;
+	return 0;
+}
+
 size_t mov_write_vmhd(const struct mov_t* mov)
 {
 	mov_buffer_w32(&mov->io, 20); /* size (always 0x14) */
@@ -103,6 +122,36 @@ size_t mov_write_nmhd(const struct mov_t* mov)
 	return 12;
 }
 
+size_t mov_write_gmhd(const struct mov_t* mov)
+{
+	mov_buffer_w32(&mov->io, 8 + 24 + 44); /* size */
+	mov_buffer_write(&mov->io, "gmhd", 4);
+
+	mov_buffer_w32(&mov->io, 24); /* size */
+	mov_buffer_write(&mov->io, "gmin", 4);
+	mov_buffer_w32(&mov->io, 0); /* version & flags */
+	mov_buffer_w16(&mov->io, 0x40); /* graphics mode */
+	mov_buffer_w16(&mov->io, 0x8000); /* opcolor red*/
+	mov_buffer_w16(&mov->io, 0x8000); /* opcolor green*/
+	mov_buffer_w16(&mov->io, 0x8000); /* opcolor blue*/
+	mov_buffer_w16(&mov->io, 0); /* balance */
+	mov_buffer_w16(&mov->io, 0); /* reserved */
+
+	mov_buffer_w32(&mov->io, 44);   /* size */
+	mov_buffer_write(&mov->io, "text", 4);
+	mov_buffer_w32(&mov->io, 0x00010000); /* u */
+	mov_buffer_w32(&mov->io, 0);
+	mov_buffer_w32(&mov->io, 0);
+	mov_buffer_w32(&mov->io, 0); /* v */
+	mov_buffer_w32(&mov->io, 0x00010000);
+	mov_buffer_w32(&mov->io, 0);
+	mov_buffer_w32(&mov->io, 0); /* w */
+	mov_buffer_w32(&mov->io, 0);
+	mov_buffer_w32(&mov->io, 0x40000000);
+
+	return 8 + 24 + 44;
+}
+
 /*
 ISO/IEC 14496-12:2015(E) 12.6.2 Subtitle media header (p185)
 	aligned(8) class SubtitleMediaHeaderBox extends FullBox ('sthd', version = 0, flags = 0){

libmov/source/mov-reader.c

@@ -208,6 +208,7 @@ static struct mov_parse_t s_mov_parse_table[] = {
 	{ MOV_TAG('a', 'v', '1', 'C'), MOV_NULL, mov_read_extra }, // av1-isobmff
 	{ MOV_TAG('a', 'v', 'c', 'C'), MOV_NULL, mov_read_extra }, // ISO/IEC 14496-15:2010(E) avcC
 	{ MOV_TAG('b', 't', 'r', 't'), MOV_NULL, mov_read_btrt }, // ISO/IEC 14496-15:2010(E) 5.3.4.1.1 Definition
+	{ MOV_TAG('c', 'h', 'p', 'l'), MOV_STBL, mov_read_chpl }, // chapter title
 	{ MOV_TAG('c', 'o', '6', '4'), MOV_STBL, mov_read_stco },
     { MOV_TAG('C', 'o', 'L', 'L'), MOV_STBL, mov_read_coll },
 	{ MOV_TAG('c', 't', 't', 's'), MOV_STBL, mov_read_ctts },

libmov/source/mov-stsd.c

@@ -196,6 +196,8 @@ static int mov_read_hint_sample_entry(struct mov_t* mov, struct mov_sample_entry
 	struct mov_box_t box;
 	mov_read_sample_entry(mov, &box, &entry->data_reference_index);
 	mov_buffer_skip(&mov->io, box.size - 16);
+	entry->object_type_indication = mov_tag_to_object(box.type);
+	entry->stream_type = MP4_STREAM_VISUAL;
 	mov->track->tag = box.type;
 	return mov_buffer_error(&mov->io);
 }
@@ -205,6 +207,8 @@ static int mov_read_meta_sample_entry(struct mov_t* mov, struct mov_sample_entry
 	struct mov_box_t box;
 	mov_read_sample_entry(mov, &box, &entry->data_reference_index);
 	mov_buffer_skip(&mov->io, box.size - 16);
+	entry->object_type_indication = mov_tag_to_object(box.type);
+	entry->stream_type = MP4_STREAM_VISUAL;
 	mov->track->tag = box.type;
 	return mov_buffer_error(&mov->io);
 }
@@ -226,7 +230,7 @@ static int mov_read_text_sample_entry(struct mov_t* mov, struct mov_sample_entry
 	mov_read_sample_entry(mov, &box, &entry->data_reference_index);
 	if (MOV_TEXT == box.type)
 	{
-		// https://developer.apple.com/library/archive/documentation/QuickTime/QTFF/QTFFChap3/qtff3.html#//apple_ref/doc/uid/TP40000939-CH205-69835
+		// https://developer.apple.com/documentation/quicktime-file-format/text_sample_description
 		//mov_buffer_r32(&mov->io); /* display flags */
 		//mov_buffer_r32(&mov->io); /* text justification */
 		//mov_buffer_r16(&mov->io); /* background color: 48-bit RGB color */
@@ -249,6 +253,8 @@ static int mov_read_text_sample_entry(struct mov_t* mov, struct mov_sample_entry
 		mov_buffer_skip(&mov->io, box.size - 16);
 	}
 
+	entry->object_type_indication = mov_tag_to_object(box.type);
+	entry->stream_type = MP4_STREAM_VISUAL;
 	mov->track->tag = box.type;
 	return mov_buffer_error(&mov->io);
 }
@@ -381,6 +387,18 @@ int mov_read_stsd(struct mov_t* mov, const struct mov_box_t* box)
 //	return size;
 //}
 
+static size_t mov_write_btrt(const struct mov_t* mov, const struct mov_sample_entry_t* entry)
+{
+	mov_buffer_w32(&mov->io, 20); /* size */
+	mov_buffer_write(&mov->io, "btrt", 4);
+	mov_buffer_w32(&mov->io, entry->u.bitrate.bufferSizeDB);
+	mov_buffer_w32(&mov->io, 0x00000014);
+	mov_buffer_w32(&mov->io, 0x00000014);
+	//mov_buffer_w32(&mov->io, entry->u.bitrate.maxBitrate);
+	//mov_buffer_w32(&mov->io, entry->u.bitrate.avgBitrate);
+	return 20;
+}
+
 static size_t mov_write_video(const struct mov_t* mov, const struct mov_sample_entry_t* entry)
 {
 	size_t size;
@@ -437,6 +455,7 @@ static size_t mov_write_video(const struct mov_t* mov, const struct mov_sample_e
 	else if (MOV_OBJECT_VP8 == entry->object_type_indication || MOV_OBJECT_VP9 == entry->object_type_indication)
 		size += mov_write_vpcc(mov);
 
+	//size += mov_write_btrt(mov, entry);
 	mov_write_size(mov, offset, size); /* update size */
 	return size;
 }
@@ -479,6 +498,7 @@ static size_t mov_write_audio(const struct mov_t* mov, const struct mov_sample_e
 	else if(MOV_OBJECT_OPUS == entry->object_type_indication)
 		size += mov_write_dops(mov);
 
+	//size += mov_write_btrt(mov, entry);
 	mov_write_size(mov, offset, size); /* update size */
 	return size;
 }
@@ -488,7 +508,7 @@ static int mov_write_subtitle(const struct mov_t* mov, const struct mov_sample_e
 	int size;
 	uint64_t offset;
 
-	size = 8 /* Box */ + 8 /* SampleEntry */ + entry->extra_data_size;
+	size = 8 /* Box */ + 8 /* SampleEntry */;
 
 	offset = mov_buffer_tell(&mov->io);
 	mov_buffer_w32(&mov->io, 0); /* size */
@@ -506,6 +526,8 @@ static int mov_write_subtitle(const struct mov_t* mov, const struct mov_sample_e
 	{
 		mov_buffer_write(&mov->io, entry->extra_data, entry->extra_data_size);
 		size += entry->extra_data_size;
+
+		size += mov_write_btrt(mov, entry);
 	}
 
 	mov_write_size(mov, offset, size); /* update size */

libmov/source/mov-stts.c

@@ -152,7 +152,7 @@ uint32_t mov_build_stts(struct mov_track_t* track)
     {
 		assert(track->samples[i + 1].dts >= track->samples[i].dts || i + 1 == track->sample_count);
         delta = (uint32_t)(i + 1 < track->sample_count && track->samples[i + 1].dts > track->samples[i].dts ? track->samples[i + 1].dts - track->samples[i].dts : 1);
-        if (NULL != sample && delta == sample->samples_per_chunk)
+        if (NULL != sample && (delta == sample->samples_per_chunk || i + 1 == track->sample_count) )
         {
             track->samples[i].first_chunk = 0;
             assert(sample->first_chunk > 0);

libmov/source/mov-tag.c

@@ -26,6 +26,7 @@ static struct mov_object_tag s_tags[] = {
 	{ MOV_OBJECT_TEXT,		MOV_TAG('t', 'x', '3', 'g') },
 	{ MOV_OBJECT_TEXT,		MOV_TAG('t', 'e', 'x', 't') },
 	{ MOV_OBJECT_TEXT,		MOV_TAG('c', '6', '0', '8') },
+	{ MOV_OBJECT_CHAPTER,	MOV_TAG('t', 'e', 'x', 't') },
     { MOV_OBJECT_OPUS,		MOV_OPUS },
     { MOV_OBJECT_VP8,		MOV_VP8 },
     { MOV_OBJECT_VP9,		MOV_VP9 },

libmov/source/mov-track.c

@@ -176,6 +176,28 @@ int mov_add_video(struct mov_track_t* track, const struct mov_mvhd_t* mvhd, uint
 
 int mov_add_subtitle(struct mov_track_t* track, const struct mov_mvhd_t* mvhd, uint32_t timescale, uint8_t object, const void* extra_data, size_t extra_data_size)
 {
+    static const uint8_t chapter_extra_data[] = {
+        // TextSampleEntry
+        0x00, 0x00, 0x00, 0x01, // displayFlags
+        0x00, 0x00,             // horizontal + vertical justification
+        0x00, 0x00, 0x00, 0x00, // bgColourRed/Green/Blue/Alpha
+        // BoxRecord
+        0x00, 0x00, 0x00, 0x00, // defTextBoxTop/Left
+        0x00, 0x00, 0x00, 0x00, // defTextBoxBottom/Right
+        // StyleRecord
+        0x00, 0x00, 0x00, 0x00, // startChar + endChar
+        0x00, 0x01,             // fontID
+        0x00, 0x00,             // fontStyleFlags + fontSize
+        0x00, 0x00, 0x00, 0x00, // fgColourRed/Green/Blue/Alpha
+        // FontTableBox
+        0x00, 0x00, 0x00, 0x0D, // box size
+        'f', 't', 'a', 'b',     // box atom name
+        0x00, 0x01,             // entry count
+        // FontRecord
+        0x00, 0x01,             // font ID
+        0x00,                   // font name length
+    };
+
     struct mov_sample_entry_t* subtitle;
 
     subtitle = &track->stsd.entries[0];
@@ -185,12 +207,12 @@ int mov_add_subtitle(struct mov_track_t* track, const struct mov_mvhd_t* mvhd, u
 
     assert(0 != mov_object_to_tag(object));
     track->tag = mov_object_to_tag(object);
-    track->handler_type = MOV_SBTL;
+    track->handler_type = track->tag == MOV_TAG('t', 'e', 'x', 't') ? MOV_TEXT : MOV_SBTL;
     track->handler_descr = "SubtitleHandler";
     track->stsd.entry_count = 1;
     track->offset = 0;
 
-    track->tkhd.flags = MOV_TKHD_FLAG_TRACK_ENABLE | MOV_TKHD_FLAG_TRACK_IN_MOVIE;
+    track->tkhd.flags = (track->tag == MOV_TAG('t', 'e', 'x', 't') ? 0 : MOV_TKHD_FLAG_TRACK_ENABLE) | MOV_TKHD_FLAG_TRACK_IN_MOVIE;
     track->tkhd.track_ID = mvhd->next_track_ID;
     track->tkhd.creation_time = mvhd->creation_time;
     track->tkhd.modification_time = mvhd->modification_time;
@@ -205,6 +227,12 @@ int mov_add_subtitle(struct mov_track_t* track, const struct mov_mvhd_t* mvhd, u
     track->mdhd.language = 0x55c4;
     track->mdhd.duration = 0; // placeholder
 
+    if (object == MOV_OBJECT_CHAPTER && 0 == extra_data_size)
+    {
+        extra_data = chapter_extra_data;
+        extra_data_size = sizeof(chapter_extra_data);
+    }
+
 	subtitle->extra_data = malloc(extra_data_size + 1);
     if (NULL == subtitle->extra_data)
         return -ENOMEM;
@@ -268,6 +296,10 @@ size_t mov_write_minf(const struct mov_t* mov)
     {
         size += mov_write_smhd(mov);
     }
+    else if (MOV_TEXT == track->handler_type)
+    {
+        size += mov_write_gmhd(mov);
+    }
     else if (MOV_SUBT == track->handler_type || MOV_SBTL == track->handler_type)
     {
         size += mov_write_nmhd(mov);
@@ -312,8 +344,9 @@ size_t mov_write_trak(const struct mov_t* mov)
     mov_buffer_write(&mov->io, "trak", 4);
 
     size += mov_write_tkhd(mov);
-    //size += mov_write_tref(mov);
     size += mov_write_edts(mov);
+    if(mov->track->chpl_track != 0)
+        size += mov_write_tref(mov);
     size += mov_write_mdia(mov);
 
     mov_write_size(mov, offset, size); /* update size */
@@ -338,3 +371,15 @@ size_t mov_write_edts(const struct mov_t* mov)
     mov_write_size(mov, offset, size); /* update size */
     return size;
 }
+
+size_t mov_write_tref(const struct mov_t* mov)
+{
+    mov_buffer_w32(&mov->io, 20); /* size */
+    mov_buffer_write(&mov->io, "tref", 4);
+
+    mov_buffer_w32(&mov->io, 12); /* size */
+    mov_buffer_write(&mov->io, "chap", 4);
+    mov_buffer_w32(&mov->io, mov->track->chpl_track);
+
+    return 20;
+}