movenc: Allow interleaving samples when writing fragmented files

This is incompatible with the omit_tfhd_offset flag (writing
position independent fragments with interleaving requires the
default_base_moof flag).

This makes the moof atoms slightly bigger, but can be better for
playback (improving locality of sample data in the mdat).

Signed-off-by: Martin Storsjö <martin@martin.st>

libavformat/movenc.c

@@ -73,6 +73,7 @@ static const AVOption options[] = {
     { "brand",    "Override major brand", offsetof(MOVMuxContext, major_brand),   AV_OPT_TYPE_STRING, {.str = NULL}, .flags = AV_OPT_FLAG_ENCODING_PARAM },
     { "use_editlist", "use edit list", offsetof(MOVMuxContext, use_editlist), AV_OPT_TYPE_INT, {.i64 = -1}, -1, 1, AV_OPT_FLAG_ENCODING_PARAM},
     { "fragment_index", "Fragment number of the next fragment", offsetof(MOVMuxContext, fragments), AV_OPT_TYPE_INT, {.i64 = 1}, 1, INT_MAX, AV_OPT_FLAG_ENCODING_PARAM},
+    { "frag_interleave", "Interleave samples within fragments (max number of consecutive samples, lower is tighter interleaving, but with more overhead)", offsetof(MOVMuxContext, frag_interleave), AV_OPT_TYPE_INT, {.i64 = 0}, 0, INT_MAX, AV_OPT_FLAG_ENCODING_PARAM },
     { NULL },
 };
 
@@ -3118,6 +3119,32 @@ static void mov_parse_vc1_frame(AVPacket *pkt, MOVTrack *trk)
     }
 }
 
+static int mov_flush_fragment_interleaving(AVFormatContext *s, MOVTrack *track)
+{
+    MOVMuxContext *mov = s->priv_data;
+    int ret, buf_size;
+    uint8_t *buf;
+    int i, offset;
+
+    if (!track->mdat_buf)
+        return 0;
+    if (!mov->mdat_buf) {
+        if ((ret = avio_open_dyn_buf(&mov->mdat_buf)) < 0)
+            return ret;
+    }
+    buf_size = avio_close_dyn_buf(track->mdat_buf, &buf);
+    track->mdat_buf = NULL;
+
+    offset = avio_tell(mov->mdat_buf);
+    avio_write(mov->mdat_buf, buf, buf_size);
+    av_free(buf);
+
+    for (i = track->entries_flushed; i < track->entry; i++)
+        track->cluster[i].pos += offset;
+    track->entries_flushed = track->entry;
+    return 0;
+}
+
 static int mov_flush_fragment(AVFormatContext *s)
 {
     MOVMuxContext *mov = s->priv_data;
@@ -3180,14 +3207,28 @@ static int mov_flush_fragment(AVFormatContext *s)
         return 0;
     }
 
+    if (mov->frag_interleave) {
+        for (i = 0; i < mov->nb_streams; i++) {
+            MOVTrack *track = &mov->tracks[i];
+            int ret;
+            if ((ret = mov_flush_fragment_interleaving(s, track)) < 0)
+                return ret;
+        }
+
+        if (!mov->mdat_buf)
+            return 0;
+        mdat_size = avio_tell(mov->mdat_buf);
+    }
+
     for (i = 0; i < mov->nb_streams; i++) {
         MOVTrack *track = &mov->tracks[i];
-        if (mov->flags & FF_MOV_FLAG_SEPARATE_MOOF)
+        if (mov->flags & FF_MOV_FLAG_SEPARATE_MOOF || mov->frag_interleave)
             track->data_offset = 0;
         else
             track->data_offset = mdat_size;
-        if (!track->mdat_buf)
+        if (!track->entry)
             continue;
+        if (track->mdat_buf)
             mdat_size += avio_tell(track->mdat_buf);
         if (first_track < 0)
             first_track = i;
@@ -3227,10 +3268,18 @@ static int mov_flush_fragment(AVFormatContext *s)
         if (track->entry)
             track->frag_start += duration;
         track->entry = 0;
+        track->entries_flushed = 0;
+        if (!mov->frag_interleave) {
             if (!track->mdat_buf)
                 continue;
             buf_size = avio_close_dyn_buf(track->mdat_buf, &buf);
             track->mdat_buf = NULL;
+        } else {
+            if (!mov->mdat_buf)
+                continue;
+            buf_size = avio_close_dyn_buf(mov->mdat_buf, &buf);
+            mov->mdat_buf = NULL;
+        }
 
         avio_write(s->pb, buf, buf_size);
         av_free(buf);
@@ -3269,6 +3318,13 @@ int ff_mov_write_packet(AVFormatContext *s, AVPacket *pkt)
     if (mov->flags & FF_MOV_FLAG_FRAGMENT) {
         int ret;
         if (mov->moov_written || mov->flags & FF_MOV_FLAG_EMPTY_MOOV) {
+            if (mov->frag_interleave && mov->fragments > 0) {
+                if (trk->entry - trk->entries_flushed >= mov->frag_interleave) {
+                    if ((ret = mov_flush_fragment_interleaving(s, trk)) < 0)
+                        return ret;
+                }
+            }
+
             if (!trk->mdat_buf) {
                 if ((ret = avio_open_dyn_buf(&trk->mdat_buf)) < 0)
                     return ret;
@@ -3741,6 +3797,21 @@ static int mov_write_header(AVFormatContext *s)
     if (!mov->use_editlist && s->avoid_negative_ts == AVFMT_AVOID_NEG_TS_AUTO)
         s->avoid_negative_ts = AVFMT_AVOID_NEG_TS_MAKE_ZERO;
 
+    /* Clear the omit_tfhd_offset flag if default_base_moof is set;
+     * if the latter is set that's enough and omit_tfhd_offset doesn't
+     * add anything extra on top of that. */
+    if (mov->flags & FF_MOV_FLAG_OMIT_TFHD_OFFSET &&
+        mov->flags & FF_MOV_FLAG_DEFAULT_BASE_MOOF)
+        mov->flags &= ~FF_MOV_FLAG_OMIT_TFHD_OFFSET;
+
+    if (mov->frag_interleave &&
+        mov->flags & (FF_MOV_FLAG_OMIT_TFHD_OFFSET | FF_MOV_FLAG_SEPARATE_MOOF)) {
+        av_log(s, AV_LOG_ERROR,
+               "Sample interleaving in fragments is mutually exclusive with "
+               "omit_tfhd_offset and separate_moof\n");
+        return AVERROR(EINVAL);
+    }
+
     /* Non-seekable output is ok if using fragmentation. If ism_lookahead
      * is enabled, we don't support non-seekable output at all. */
     if (!s->pb->seekable &&

libavformat/movenc.h

@@ -123,6 +123,7 @@ typedef struct MOVTrack {
     int64_t     data_offset;
     int64_t     frag_start;
     int         frag_discont;
+    int         entries_flushed;
 
     int         nb_frag_info;
     MOVFragmentInfo *frag_info;
@@ -172,6 +173,7 @@ typedef struct MOVMuxContext {
     AVFormatContext *fc;
 
     int use_editlist;
+    int frag_interleave;
 } MOVMuxContext;
 
 #define FF_MOV_FLAG_RTP_HINT              (1 <<  0)