demux_mkv: implement audio skipping/trimming

This mechanism was introduced for Opus, and allows correct skipping of
"preroll" data, as well as discarding trailing audio if the file's
length isn't a multiple of the audio frame size.

Not sure how to handle seeking. I don't understand the purpose of the
SeekPreRoll element.

This was tested with correctness_trimming_nobeeps.opus, remuxed to mka
with mkvmerge v7.2.0. It seems to be correct, although the reported file
duration is incorrect (maybe a mkvmerge issue).

TOOLS/lib/Parse/Matroska/Definitions.pm

@@ -232,6 +232,7 @@ sub define_matroska {
                     elem('Block',           'a1', 'binary'),
                     elem('BlockDuration',   '9b', 'uint'),
                     elem('ReferenceBlock*', 'fb', 'sint'),
+                    elem('DiscardPadding',  '75A2', 'sint'),
                 }),
                 elem('SimpleBlock*', 'a3', 'binary'),
             }),
@@ -256,6 +257,8 @@ sub define_matroska {
                     elem('CodecPrivate',         '63a2', 'binary'),
                     elem('CodecName',          '258688', 'str'),
                     elem('CodecDecodeAll',         'aa', 'uint'),
+                    elem('CodecDelay',           '56AA', 'uint'),
+                    elem('SeekPreRoll',          '56BB', 'uint'),
                     elem('Video', 'e0', {
                         elem('FlagInterlaced',  '9a', 'uint'),
                         elem('PixelWidth',      'b0', 'uint'),

demux/demux_mkv.c

@@ -112,6 +112,7 @@ typedef struct mkv_track {
     float a_osfreq;
 
     double default_duration;
+    double codec_delay;
 
     int default_track;
 
@@ -182,6 +183,7 @@ typedef struct mkv_demuxer {
 
     uint64_t skip_to_timecode;
     int v_skip_to_keyframe, a_skip_to_keyframe;
+    int a_skip_preroll;
     int subtitle_preroll;
 } mkv_demuxer_t;
 
@@ -626,6 +628,9 @@ static void parse_trackentry(struct demuxer *demuxer,
     if (entry->n_content_encodings)
         parse_trackencodings(demuxer, track, &entry->content_encodings);
 
+    if (entry->n_codec_delay)
+        track->codec_delay = entry->codec_delay / 1e9;
+
     mkv_d->tracks[mkv_d->num_tracks++] = track;
 }
 
@@ -1750,6 +1755,7 @@ static int demux_mkv_open(demuxer_t *demuxer, enum demux_check check)
     mkv_d->tc_scale = 1000000;
     mkv_d->segment_start = stream_tell(s);
     mkv_d->segment_end = end_pos;
+    mkv_d->a_skip_preroll = 1;
 
     if (demuxer->params && demuxer->params->matroska_was_valid)
         *demuxer->params->matroska_was_valid = true;
@@ -2258,7 +2264,7 @@ static bool mkv_parse_packet(mkv_track_t *track, bstr *raw, bstr *out)
 }
 
 struct block_info {
-    uint64_t duration;
+    uint64_t duration, discardpadding;
     bool simple, keyframe;
     uint64_t timecode;
     mkv_track_t *track;
@@ -2358,7 +2364,7 @@ static int handle_block(demuxer_t *demuxer, struct block_info *block_info)
         return 0;
     }
 
-    current_pts = tc / 1e9;
+    current_pts = tc / 1e9 - track->codec_delay;
 
     if (track->type == MATROSKA_TRACK_AUDIO) {
         if (mkv_d->a_skip_to_keyframe)
@@ -2418,6 +2424,13 @@ static int handle_block(demuxer_t *demuxer, struct block_info *block_info)
                         MPSWAP(double, dp->pts, dp->dts);
                     if (p == 0)
                         dp->duration = block_duration / 1e9;
+                    if (stream->type == STREAM_AUDIO) {
+                        unsigned int srate = track->a_sfreq;
+                        demux_packet_set_padding(dp,
+                            mkv_d->a_skip_preroll ? track->codec_delay * srate : 0,
+                            block_info->discardpadding / 1e9 * srate);
+                        mkv_d->a_skip_preroll = 0;
+                    }
                     demux_add_packet(stream, dp);
                     p++;
                 }
@@ -2456,6 +2469,12 @@ static int read_block_group(demuxer_t *demuxer, int64_t end,
             block->duration *= mkv_d->tc_scale;
             break;
 
+        case MATROSKA_ID_DISCARDPADDING:
+            block->discardpadding = ebml_read_uint(s);
+            if (block->discardpadding == EBML_UINT_INVALID)
+                goto error;
+            break;
+
         case MATROSKA_ID_BLOCK:
             if (read_block(demuxer, end, block) < 0)
                 goto error;
@@ -2745,6 +2764,7 @@ static void demux_mkv_seek(demuxer_t *demuxer, double rel_seek_secs, int flags)
 
         mkv_d->v_skip_to_keyframe = st_active[STREAM_VIDEO];
         mkv_d->a_skip_to_keyframe = st_active[STREAM_AUDIO];
+        mkv_d->a_skip_preroll = mkv_d->a_skip_to_keyframe;
 
         if (flags & SEEK_FORWARD) {
             mkv_d->skip_to_timecode = target_timecode;
@@ -2781,6 +2801,7 @@ static void demux_mkv_seek(demuxer_t *demuxer, double rel_seek_secs, int flags)
 
         mkv_d->v_skip_to_keyframe = st_active[STREAM_VIDEO];
         mkv_d->a_skip_to_keyframe = st_active[STREAM_AUDIO];
+        mkv_d->a_skip_preroll = mkv_d->a_skip_to_keyframe;
 
         if (index) {
             stream_seek(s, index->filepos);

demux/packet.c

@@ -20,6 +20,7 @@
 #include <assert.h>
 
 #include <libavcodec/avcodec.h>
+#include <libavutil/intreadwrite.h>
 
 #include "common/av_common.h"
 #include "common/common.h"
@@ -113,3 +114,18 @@ struct demux_packet *demux_copy_packet(struct demux_packet *dp)
     new->duration = dp->duration;
     return new;
 }
+
+int demux_packet_set_padding(struct demux_packet *dp, int start, int end)
+{
+    if (!start  && !end)
+        return 0;
+    if (!dp->avpacket)
+        return -1;
+    uint8_t *p = av_packet_new_side_data(dp->avpacket, AV_PKT_DATA_SKIP_SAMPLES, 10);
+    if (!p)
+        return -1;
+
+    AV_WL32(p + 0, start);
+    AV_WL32(p + 4, end);
+    return 0;
+}

demux/packet.h

@@ -44,4 +44,6 @@ void demux_packet_shorten(struct demux_packet *dp, size_t len);
 void free_demux_packet(struct demux_packet *dp);
 struct demux_packet *demux_copy_packet(struct demux_packet *dp);
 
+int demux_packet_set_padding(struct demux_packet *dp, int start, int end);
+
 #endif /* MPLAYER_DEMUX_PACKET_H */