diff mbox series

[FFmpeg-devel] avformat/mxf: support MCA audio information

Message ID 20211105160643.33567-1-marc-antoine.arnaud@luminvent.com
State New
Headers show
Series [FFmpeg-devel] avformat/mxf: support MCA audio information | expand

Checks

Context Check Description
andriy/make_x86 success Make finished
andriy/make_fate_x86 success Make fate finished
andriy/make_ppc success Make finished
andriy/make_fate_ppc success Make fate finished

Commit Message

Marc-Antoine ARNAUD Nov. 5, 2021, 4:06 p.m. UTC 
---
 libavformat/mxf.h    |   1 +
 libavformat/mxfdec.c | 280 ++++++++++++++++++++++++++++++++++++++++++-
 2 files changed, 275 insertions(+), 6 deletions(-)

Comments

Marton Balint Nov. 7, 2021, 11:32 a.m. UTC | #1 
On Fri, 5 Nov 2021, Marc-Antoine Arnaud wrote:

> ---
> libavformat/mxf.h    |   1 +
> libavformat/mxfdec.c | 280 ++++++++++++++++++++++++++++++++++++++++++-
> 2 files changed, 275 insertions(+), 6 deletions(-)
>

Could you mention in the commit description which 
standards/recommendations cover MCA in mxf?

[...]

> @@ -2328,7 +2480,9 @@ static int mxf_parse_structural_metadata(MXFContext *mxf)
>         AVStream *st;
>         FFStream *sti;
>         AVTimecode tc;
> +        enum AVAudioServiceType *ast;

This declaration can be pushed down to where it is used.

[...]

> +                // Soundfield group
> +                if (IS_KLV_KEY(mca_sub_descriptor->mca_label_dictionary_id, mxf_soundfield_group)) {
> +                    MXFSoundfieldGroupUL* group_ptr = (MXFSoundfieldGroupUL*)&mxf_soundfield_groups[0];

const MXFSoundfieldGroupUL *group_ptr = mxf_soundfield_groups;

> +
> +                    while (group_ptr->uid[0]) {
> +                        if (IS_KLV_KEY(group_ptr->uid, mca_sub_descriptor->mca_label_dictionary_id)) {
> +                            st->codecpar->channel_layout = group_ptr->id;
> +                            break;
> +                        }
> +                        group_ptr++;
> +                    }
> +                }
> +
> +                // Audio channel
> +                if (IS_KLV_KEY(mca_sub_descriptor->mca_label_dictionary_id, mxf_audio_channel)) {
> +                    MXFChannelOrderingUL* channel_ordering_ptr = (MXFChannelOrderingUL*)&mxf_channel_ordering[0];

const MXFChannelOrderingUL *channel_ordering_ptr = mxf_channel_ordering;

> +
> +                    while (channel_ordering_ptr->uid[0]) {
> +                        if (IS_KLV_KEY(channel_ordering_ptr->uid, mca_sub_descriptor->mca_label_dictionary_id)) {

You should check if current_channel < desciptor->channels here, and if 
not, then warn the user and break out of the loop. Otherwise 
current_channel can grow out of array limits.

It should also be checked that channel_ordering_ptr->index < 
descriptor->channels, and if not, then similarly, warn the user and break 
out.

Maybe a hard failure (returning AVERROR_INVALIDDATA) is not necessary, to 
allow some slightly invalid metadata?

> +                            source_track->channel_ordering[current_channel] = channel_ordering_ptr->index;
> +
> +                            if(channel_ordering_ptr->service_type != AV_AUDIO_SERVICE_TYPE_NB) {
> +                                uint8_t* side_data = av_stream_new_side_data(st, AV_PKT_DATA_AUDIO_SERVICE_TYPE, sizeof(*ast));
> +                                if (!side_data)
> +                                    goto fail_and_free;
> +                                ast = (enum AVAudioServiceType*)side_data;
> +                                *ast = channel_ordering_ptr->service_type;
> +                            }
> +
> +                            current_channel += 1;
> +                            break;
> +                        }
> +                        channel_ordering_ptr++;
> +                    }
> +                }
> +

[...]

> +            source_track->require_reordering = 0;
> +            for (j = 0; j < descriptor->channels; ++j) {
> +                if (source_track->channel_ordering[j] != j) {
> +                    source_track->require_reordering = 1;
> +                    break;
> +                }
> +            }

If require_reordering == 1 here but either codec is not PCM or 
skip_audio_reordering is set then channel_layout should be reset to 0 
(unknown). Maybe you should also set require_reordering to 0 in either of 
these cases, so further along you no longer have to check codec and 
skip_audio_reordering, only require_reordering.

> +
> +            if (source_track->require_reordering && is_pcm(st->codecpar->codec_id)) {
> +                current_channel = 0;
> +                av_log(mxf->fc, AV_LOG_DEBUG, "MCA Audio mapping (");
> +                for(j = 0; j < descriptor->channels; ++j) {
> +                    for(int k = 0; k < descriptor->channels; ++k) {
> +                        if(source_track->channel_ordering[k] == current_channel) {
> +                            av_log(mxf->fc, AV_LOG_DEBUG, "%d -> %d", source_track->channel_ordering[k], k);
> +                            if (current_channel != descriptor->channels - 1)
> +                                av_log(mxf->fc, AV_LOG_DEBUG, ", ");
> +                            current_channel += 1;
> +                        }
> +                    }
> +                }
> +                av_log(mxf->fc, AV_LOG_DEBUG, ")\n");
> +            }

[...]

> @@ -3920,6 +4185,9 @@ static const AVOption options[] = {
>     { "eia608_extract", "extract eia 608 captions from s436m track",
>       offsetof(MXFContext, eia608_extract), AV_OPT_TYPE_BOOL, {.i64 = 0}, 0, 1,
>       AV_OPT_FLAG_DECODING_PARAM },
> +    { "skip_audio_reordering", "skip audio reordering based on Multi-Channel Audio labelling",
> +      offsetof(MXFContext, skip_audio_reordering), AV_OPT_TYPE_BOOL, {.i64 = 0}, 0, 1,
> +      AV_OPT_FLAG_DECODING_PARAM },

Docs update (doc/demuxers.texi) and libavformat version micro bump is 
missing for the new option.

Thanks,
Marton
Tomas Härdin Nov. 9, 2021, 1:44 p.m. UTC | #2 
sön 2021-11-07 klockan 12:32 +0100 skrev Marton Balint:
> 
> > +
> > +                    while (channel_ordering_ptr->uid[0]) {
> > +                        if (IS_KLV_KEY(channel_ordering_ptr->uid,
> > mca_sub_descriptor->mca_label_dictionary_id)) {
> 
> You should check if current_channel < desciptor->channels here, and
> if 
> not, then warn the user and break out of the loop. Otherwise 
> current_channel can grow out of array limits.
> 
> It should also be checked that channel_ordering_ptr->index < 
> descriptor->channels, and if not, then similarly, warn the user and
> break 
> out.
> 
> Maybe a hard failure (returning AVERROR_INVALIDDATA) is not
> necessary, to 
> allow some slightly invalid metadata?

We should be as strict as we can get away with. Else we encourage
laxness in other implementations.

/Tomas
diff mbox series

Patch

diff --git a/libavformat/mxf.h b/libavformat/mxf.h
index fe9c52732c..cddbcb13c9 100644
--- a/libavformat/mxf.h
+++ b/libavformat/mxf.h
@@ -50,6 +50,7 @@  enum MXFMetadataSetType {
     TaggedValue,
     TapeDescriptor,
     AVCSubDescriptor,
+    MCASubDescriptor,
 };
 
 enum MXFFrameLayout {
diff --git a/libavformat/mxfdec.c b/libavformat/mxfdec.c
index af9d33f796..deee54991a 100644
--- a/libavformat/mxfdec.c
+++ b/libavformat/mxfdec.c
@@ -51,11 +51,14 @@ 
 #include "libavutil/mastering_display_metadata.h"
 #include "libavutil/mathematics.h"
 #include "libavcodec/bytestream.h"
+#include "libavcodec/internal.h"
+#include "libavutil/channel_layout.h"
 #include "libavutil/intreadwrite.h"
 #include "libavutil/parseutils.h"
 #include "libavutil/timecode.h"
 #include "libavutil/opt.h"
 #include "avformat.h"
+#include "avlanguage.h"
 #include "internal.h"
 #include "mxf.h"
 
@@ -177,6 +180,8 @@  typedef struct {
     int body_sid;
     MXFWrappingScheme wrapping;
     int edit_units_per_packet; /* how many edit units to read at a time (PCM, ClipWrapped) */
+    int require_reordering;
+    int channel_ordering[FF_SANE_NB_CHANNELS];
 } MXFTrack;
 
 typedef struct MXFDescriptor {
@@ -205,6 +210,8 @@  typedef struct MXFDescriptor {
     unsigned int vert_subsampling;
     UID *file_descriptors_refs;
     int file_descriptors_count;
+    UID *sub_descriptors_refs;
+    int sub_descriptors_count;
     int linked_track_id;
     uint8_t *extradata;
     int extradata_size;
@@ -217,6 +224,15 @@  typedef struct MXFDescriptor {
     size_t coll_size;
 } MXFDescriptor;
 
+typedef struct MXFMCASubDescriptor {
+    MXFMetadataSet meta;
+    UID uid;
+    UID mca_link_id;
+    UID mca_group_link_id;
+    UID mca_label_dictionary_id;
+    char *language;
+} MXFMCASubDescriptor;
+
 typedef struct MXFIndexTableSegment {
     MXFMetadataSet meta;
     int edit_unit_byte_count;
@@ -290,6 +306,7 @@  typedef struct MXFContext {
     int nb_index_tables;
     MXFIndexTable *index_tables;
     int eia608_extract;
+    int skip_audio_reordering;
 } MXFContext;
 
 /* NOTE: klv_offset is not set (-1) for local keys */
@@ -311,6 +328,9 @@  static const uint8_t mxf_system_item_key_cp[]              = { 0x06,0x0e,0x2b,0x
 static const uint8_t mxf_system_item_key_gc[]              = { 0x06,0x0e,0x2b,0x34,0x02,0x53,0x01,0x01,0x0d,0x01,0x03,0x01,0x14 };
 static const uint8_t mxf_klv_key[]                         = { 0x06,0x0e,0x2b,0x34 };
 static const uint8_t mxf_apple_coll_prefix[]               = { 0x06,0x0e,0x2b,0x34,0x01,0x01,0x01,0x0e,0x0e,0x20,0x04,0x01,0x05,0x03,0x01 };
+static const uint8_t mxf_audio_channel[]                   = { 0x06,0x0e,0x2b,0x34,0x04,0x01,0x01,0x0d,0x03,0x02,0x01 };
+static const uint8_t mxf_soundfield_group[]                = { 0x06,0x0e,0x2b,0x34,0x04,0x01,0x01,0x0d,0x03,0x02,0x02 };
+
 /* complete keys to match */
 static const uint8_t mxf_crypto_source_container_ul[]      = { 0x06,0x0e,0x2b,0x34,0x01,0x01,0x01,0x09,0x06,0x01,0x01,0x02,0x02,0x00,0x00,0x00 };
 static const uint8_t mxf_encrypted_triplet_key[]           = { 0x06,0x0e,0x2b,0x34,0x02,0x04,0x01,0x07,0x0d,0x01,0x03,0x01,0x02,0x7e,0x01,0x00 };
@@ -323,6 +343,15 @@  static const uint8_t mxf_indirect_value_utf16be[]          = { 0x42,0x01,0x10,0x
 static const uint8_t mxf_apple_coll_max_cll[]              = { 0x06,0x0e,0x2b,0x34,0x01,0x01,0x01,0x0e,0x0e,0x20,0x04,0x01,0x05,0x03,0x01,0x01 };
 static const uint8_t mxf_apple_coll_max_fall[]             = { 0x06,0x0e,0x2b,0x34,0x01,0x01,0x01,0x0e,0x0e,0x20,0x04,0x01,0x05,0x03,0x01,0x02 };
 
+static const uint8_t mxf_mca_label_dictionary_id[]         = { 0x06,0x0e,0x2b,0x34,0x01,0x01,0x01,0x0e,0x01,0x03,0x07,0x01,0x01,0x00,0x00,0x00 };
+static const uint8_t mxf_mca_tag_symbol[]                  = { 0x06,0x0e,0x2b,0x34,0x01,0x01,0x01,0x0e,0x01,0x03,0x07,0x01,0x02,0x00,0x00,0x00 };
+static const uint8_t mxf_mca_tag_name[]                    = { 0x06,0x0e,0x2b,0x34,0x01,0x01,0x01,0x0e,0x01,0x03,0x07,0x01,0x03,0x00,0x00,0x00 };
+static const uint8_t mxf_mca_link_id[]                     = { 0x06,0x0e,0x2b,0x34,0x01,0x01,0x01,0x0e,0x01,0x03,0x07,0x01,0x05,0x00,0x00,0x00 };
+static const uint8_t mxf_soundfield_group_link_id[]        = { 0x06,0x0e,0x2b,0x34,0x01,0x01,0x01,0x0e,0x01,0x03,0x07,0x01,0x06,0x00,0x00,0x00 };
+static const uint8_t mxf_mca_rfc5646_spoken_language[]     = { 0x06,0x0e,0x2b,0x34,0x01,0x01,0x01,0x0d,0x03,0x01,0x01,0x02,0x03,0x15,0x00,0x00 };
+
+static const uint8_t mxf_sub_descriptor[]                  = { 0x06,0x0e,0x2b,0x34,0x01,0x01,0x01,0x09,0x06,0x01,0x01,0x04,0x06,0x10,0x00,0x00 };
+
 static const uint8_t mxf_mastering_display_prefix[13]      = { FF_MXF_MasteringDisplay_PREFIX };
 static const uint8_t mxf_mastering_display_uls[4][16] = {
     FF_MXF_MasteringDisplayPrimaries,
@@ -343,6 +372,11 @@  static void mxf_free_metadataset(MXFMetadataSet **ctx, int freectx)
         av_freep(&((MXFDescriptor *)*ctx)->mastering);
         av_freep(&((MXFDescriptor *)*ctx)->coll);
         av_freep(&((MXFDescriptor *)*ctx)->file_descriptors_refs);
+        av_freep(&((MXFDescriptor *)*ctx)->sub_descriptors_refs);
+        break;
+    case MCASubDescriptor:
+        if (((MXFMCASubDescriptor *)*ctx)->language)
+            av_freep(&((MXFMCASubDescriptor *)*ctx)->language);
         break;
     case Sequence:
         av_freep(&((MXFSequence *)*ctx)->structural_components_refs);
@@ -906,6 +940,30 @@  static int mxf_read_strong_ref_array(AVIOContext *pb, UID **refs, int *count)
     return 0;
 }
 
+static inline int mxf_read_us_ascii_string(AVIOContext *pb, int size, char** str)
+{
+    int ret;
+    size_t buf_size;
+
+    if (size < 0 || size > INT_MAX - 1)
+        return AVERROR(EINVAL);
+
+    buf_size = size + 1;
+    av_free(*str);
+    *str = av_malloc(buf_size);
+    if (!*str)
+        return AVERROR(ENOMEM);
+
+    ret = avio_get_str(pb, size, *str, buf_size);
+
+    if (ret < 0) {
+        av_freep(str);
+        return ret;
+    }
+
+    return ret;
+}
+
 static inline int mxf_read_utf16_string(AVIOContext *pb, int size, char** str, int be)
 {
     int ret;
@@ -1360,11 +1418,34 @@  static int mxf_read_generic_descriptor(void *arg, AVIOContext *pb, int tag, int
                 descriptor->coll->MaxFALL = avio_rb16(pb);
             }
         }
+
+        if (IS_KLV_KEY(uid, mxf_sub_descriptor))
+            mxf_read_strong_ref_array(pb, &descriptor->sub_descriptors_refs, &descriptor->sub_descriptors_count);
+
         break;
     }
     return 0;
 }
 
+static int mxf_read_mca_sub_descriptor(void *arg, AVIOContext *pb, int tag, int size, UID uid, int64_t klv_offset)
+{
+    MXFMCASubDescriptor *mca_sub_descriptor = arg;
+
+    if (IS_KLV_KEY(uid, mxf_mca_label_dictionary_id))
+        avio_read(pb, mca_sub_descriptor->mca_label_dictionary_id, 16);
+
+    if (IS_KLV_KEY(uid, mxf_mca_link_id))
+        avio_read(pb, mca_sub_descriptor->mca_link_id, 16);
+
+    if (IS_KLV_KEY(uid, mxf_soundfield_group_link_id))
+        avio_read(pb, mca_sub_descriptor->mca_group_link_id, 16);
+
+    if (IS_KLV_KEY(uid, mxf_mca_rfc5646_spoken_language))
+        return mxf_read_us_ascii_string(pb, size, &mca_sub_descriptor->language);
+
+    return 0;
+}
+
 static int mxf_read_indirect_value(void *arg, AVIOContext *pb, int size)
 {
     MXFTaggedValue *tagged_value = arg;
@@ -1494,6 +1575,56 @@  static const MXFCodecUL mxf_data_essence_container_uls[] = {
     { { 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00 },  0, AV_CODEC_ID_NONE },
 };
 
+typedef struct MXFSoundfieldGroupUL {
+    UID uid;
+    int64_t id;
+} MXFSoundfieldGroupUL;
+
+static const MXFSoundfieldGroupUL mxf_soundfield_groups[] = {
+    { { 0x06,0x0e,0x2b,0x34,0x04,0x01,0x01,0x0d,0x03,0x02,0x02,0x01,0x00,0x00,0x00,0x00 }, AV_CH_LAYOUT_5POINT1 }, // 5.1
+    { { 0x06,0x0e,0x2b,0x34,0x04,0x01,0x01,0x0d,0x03,0x02,0x02,0x02,0x00,0x00,0x00,0x00 }, AV_CH_LAYOUT_7POINT1 }, // 7.1 DS
+    { { 0x06,0x0e,0x2b,0x34,0x04,0x01,0x01,0x0d,0x03,0x02,0x02,0x03,0x00,0x00,0x00,0x00 }, AV_CH_LAYOUT_7POINT1_WIDE }, // 7.1 SDS
+    { { 0x06,0x0e,0x2b,0x34,0x04,0x01,0x01,0x0d,0x03,0x02,0x02,0x04,0x00,0x00,0x00,0x00 }, AV_CH_LAYOUT_6POINT1 }, // 6.1
+    { { 0x06,0x0e,0x2b,0x34,0x04,0x01,0x01,0x0d,0x03,0x02,0x02,0x05,0x00,0x00,0x00,0x00 }, AV_CH_LAYOUT_MONO }, // 1.0 Monoral
+    { { 0x06,0x0e,0x2b,0x34,0x04,0x01,0x01,0x0d,0x03,0x02,0x02,0x20,0x01,0x00,0x00,0x00 }, AV_CH_LAYOUT_STEREO }, // Standard stereo
+    { { 0x06,0x0e,0x2b,0x34,0x04,0x01,0x01,0x0d,0x03,0x02,0x02,0x20,0x02,0x00,0x00,0x00 }, AV_CH_LAYOUT_STEREO }, // Dual mono
+    { { 0x06,0x0e,0x2b,0x34,0x04,0x01,0x01,0x0d,0x03,0x02,0x02,0x20,0x04,0x00,0x00,0x00 }, AV_CH_LAYOUT_SURROUND }, // 3.0
+    { { 0x06,0x0e,0x2b,0x34,0x04,0x01,0x01,0x0d,0x03,0x02,0x02,0x20,0x05,0x00,0x00,0x00 }, AV_CH_LAYOUT_4POINT0 }, // 4.0
+    { { 0x06,0x0e,0x2b,0x34,0x04,0x01,0x01,0x0d,0x03,0x02,0x02,0x20,0x06,0x00,0x00,0x00 }, AV_CH_LAYOUT_5POINT0 }, // 5.0
+    { { 0x06,0x0e,0x2b,0x34,0x04,0x01,0x01,0x0d,0x03,0x02,0x02,0x20,0x07,0x00,0x00,0x00 }, AV_CH_LAYOUT_6POINT0 }, // 6.0
+    { { 0x06,0x0e,0x2b,0x34,0x04,0x01,0x01,0x0d,0x03,0x02,0x02,0x20,0x08,0x00,0x00,0x00 }, AV_CH_LAYOUT_7POINT0 }, // 7.0
+    { { 0x06,0x0e,0x2b,0x34,0x04,0x01,0x01,0x0d,0x03,0x02,0x02,0x20,0x09,0x00,0x00,0x00 }, AV_CH_LAYOUT_STEREO }, // LrRt
+    { { 0x06,0x0e,0x2b,0x34,0x04,0x01,0x01,0x0d,0x03,0x02,0x02,0x20,0x0a,0x00,0x00,0x00 }, AV_CH_LAYOUT_5POINT1 }, // 5.1 EX
+    { { 0x06,0x0e,0x2b,0x34,0x04,0x01,0x01,0x0d,0x03,0x02,0x02,0x20,0x0b,0x00,0x00,0x00 }, AV_CH_LAYOUT_MONO }, // Hearing accessibility
+    { { 0x06,0x0e,0x2b,0x34,0x04,0x01,0x01,0x0d,0x03,0x02,0x02,0x20,0x0c,0x00,0x00,0x00 }, AV_CH_LAYOUT_MONO }, // Visual accessibility
+    { { 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00 },  0 },
+};
+
+typedef struct MXFChannelOrderingUL {
+    UID uid;
+    int index;
+    enum AVAudioServiceType service_type;
+} MXFChannelOrderingUL;
+
+static const MXFChannelOrderingUL mxf_channel_ordering[] = {
+    { { 0x06,0x0e,0x2b,0x34,0x04,0x01,0x01,0x0d,0x03,0x02,0x01,0x01,0x00,0x00,0x00,0x00 }, 0, AV_AUDIO_SERVICE_TYPE_NB }, // Left audio channel
+    { { 0x06,0x0e,0x2b,0x34,0x04,0x01,0x01,0x0d,0x03,0x02,0x01,0x02,0x00,0x00,0x00,0x00 }, 1, AV_AUDIO_SERVICE_TYPE_NB }, // Right audio channel
+    { { 0x06,0x0e,0x2b,0x34,0x04,0x01,0x01,0x0d,0x03,0x02,0x01,0x03,0x00,0x00,0x00,0x00 }, 2, AV_AUDIO_SERVICE_TYPE_NB }, // Center audio channel
+    { { 0x06,0x0e,0x2b,0x34,0x04,0x01,0x01,0x0d,0x03,0x02,0x01,0x04,0x00,0x00,0x00,0x00 }, 5, AV_AUDIO_SERVICE_TYPE_NB }, // LFE audio channel
+    { { 0x06,0x0e,0x2b,0x34,0x04,0x01,0x01,0x0d,0x03,0x02,0x01,0x05,0x00,0x00,0x00,0x00 }, 3, AV_AUDIO_SERVICE_TYPE_NB }, // Left surround audio channel
+    { { 0x06,0x0e,0x2b,0x34,0x04,0x01,0x01,0x0d,0x03,0x02,0x01,0x06,0x00,0x00,0x00,0x00 }, 4, AV_AUDIO_SERVICE_TYPE_NB }, // Right surround audio channel
+    { { 0x06,0x0e,0x2b,0x34,0x04,0x01,0x01,0x0d,0x03,0x02,0x01,0x07,0x00,0x00,0x00,0x00 }, 3, AV_AUDIO_SERVICE_TYPE_NB }, // Left side surround audio channel
+    { { 0x06,0x0e,0x2b,0x34,0x04,0x01,0x01,0x0d,0x03,0x02,0x01,0x08,0x00,0x00,0x00,0x00 }, 4, AV_AUDIO_SERVICE_TYPE_NB }, // Right side surround audio channel
+    { { 0x06,0x0e,0x2b,0x34,0x04,0x01,0x01,0x0d,0x03,0x02,0x01,0x09,0x00,0x00,0x00,0x00 }, 6, AV_AUDIO_SERVICE_TYPE_NB }, // Left rear surround audio channel
+    { { 0x06,0x0e,0x2b,0x34,0x04,0x01,0x01,0x0d,0x03,0x02,0x01,0x0a,0x00,0x00,0x00,0x00 }, 7, AV_AUDIO_SERVICE_TYPE_NB }, // Right rear surround audio channel
+    { { 0x06,0x0e,0x2b,0x34,0x04,0x01,0x01,0x0d,0x03,0x02,0x01,0x0b,0x00,0x00,0x00,0x00 }, 6, AV_AUDIO_SERVICE_TYPE_NB }, // Left center audio channel
+    { { 0x06,0x0e,0x2b,0x34,0x04,0x01,0x01,0x0d,0x03,0x02,0x01,0x0c,0x00,0x00,0x00,0x00 }, 7, AV_AUDIO_SERVICE_TYPE_NB }, // Right center audio channel
+    { { 0x06,0x0e,0x2b,0x34,0x04,0x01,0x01,0x0d,0x03,0x02,0x01,0x0d,0x00,0x00,0x00,0x00 }, 6, AV_AUDIO_SERVICE_TYPE_NB }, // Center surround audio channel
+    { { 0x06,0x0e,0x2b,0x34,0x04,0x01,0x01,0x0d,0x03,0x02,0x01,0x0e,0x00,0x00,0x00,0x00 }, 0, AV_AUDIO_SERVICE_TYPE_VISUALLY_IMPAIRED }, // Hearing impaired audio channel
+    { { 0x06,0x0e,0x2b,0x34,0x04,0x01,0x01,0x0d,0x03,0x02,0x01,0x0f,0x00,0x00,0x00,0x00 }, 0, AV_AUDIO_SERVICE_TYPE_HEARING_IMPAIRED }, // Visually impaired narrative audio channel
+    { { 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00 }, 0, AV_AUDIO_SERVICE_TYPE_NB },
+};
+
 static MXFWrappingScheme mxf_get_wrapping_kind(UID *essence_container_ul)
 {
     int val;
@@ -2292,6 +2423,27 @@  static enum AVColorRange mxf_get_color_range(MXFContext *mxf, MXFDescriptor *des
     return AVCOL_RANGE_UNSPECIFIED;
 }
 
+static int is_pcm(enum AVCodecID codec_id)
+{
+    /* we only care about "normal" PCM codecs until we get samples */
+    return codec_id >= AV_CODEC_ID_PCM_S16LE && codec_id < AV_CODEC_ID_PCM_S24DAUD;
+}
+
+static int set_language(AVFormatContext *s, const char *rfc5646, AVDictionary **met)
+{
+    // language abbr should contain at least 2 chars
+    if (rfc5646 && strlen(rfc5646) > 1) {
+        char primary_tag[4] =
+            {rfc5646[0], rfc5646[1], rfc5646[2] != '-' ? rfc5646[2] : '\0', '\0'};
+
+        const char *iso6392       = ff_convert_lang_to(primary_tag,
+                                                       AV_LANG_ISO639_2_BIBL);
+        if (iso6392)
+            return(av_dict_set(met, "language", iso6392, 0));
+    }
+    return 0;
+}
+
 static int mxf_parse_structural_metadata(MXFContext *mxf)
 {
     MXFPackage *material_package = NULL;
@@ -2328,7 +2480,9 @@  static int mxf_parse_structural_metadata(MXFContext *mxf)
         AVStream *st;
         FFStream *sti;
         AVTimecode tc;
+        enum AVAudioServiceType *ast;
         int flags;
+        int current_channel;
 
         if (!(material_track = mxf_resolve_strong_ref(mxf, &material_package->tracks_refs[i], Track))) {
             av_log(mxf->fc, AV_LOG_ERROR, "could not resolve material track strong ref\n");
@@ -2688,6 +2842,93 @@  static int mxf_parse_structural_metadata(MXFContext *mxf)
                 sti->need_parsing = AVSTREAM_PARSE_FULL;
             }
             st->codecpar->bits_per_coded_sample = av_get_bits_per_sample(st->codecpar->codec_id);
+
+            current_channel = 0;
+
+            if (descriptor->channels >= FF_SANE_NB_CHANNELS) {
+                av_log(mxf->fc, AV_LOG_ERROR, "max number of channels %d reached\n", FF_SANE_NB_CHANNELS);
+                return AVERROR_INVALIDDATA;
+            }
+
+            for (j = 0; j < descriptor->channels; ++j) {
+                source_track->channel_ordering[j] = j;
+            }
+
+            for (j = 0; j < descriptor->sub_descriptors_count; j++) {
+                MXFMCASubDescriptor *mca_sub_descriptor = mxf_resolve_strong_ref(mxf, &descriptor->sub_descriptors_refs[j], MCASubDescriptor);
+                if (mca_sub_descriptor == NULL) {
+                    continue;
+                }
+
+                // Soundfield group
+                if (IS_KLV_KEY(mca_sub_descriptor->mca_label_dictionary_id, mxf_soundfield_group)) {
+                    MXFSoundfieldGroupUL* group_ptr = (MXFSoundfieldGroupUL*)&mxf_soundfield_groups[0];
+
+                    while (group_ptr->uid[0]) {
+                        if (IS_KLV_KEY(group_ptr->uid, mca_sub_descriptor->mca_label_dictionary_id)) {
+                            st->codecpar->channel_layout = group_ptr->id;
+                            break;
+                        }
+                        group_ptr++;
+                    }
+                }
+
+                // Audio channel
+                if (IS_KLV_KEY(mca_sub_descriptor->mca_label_dictionary_id, mxf_audio_channel)) {
+                    MXFChannelOrderingUL* channel_ordering_ptr = (MXFChannelOrderingUL*)&mxf_channel_ordering[0];
+
+                    while (channel_ordering_ptr->uid[0]) {
+                        if (IS_KLV_KEY(channel_ordering_ptr->uid, mca_sub_descriptor->mca_label_dictionary_id)) {
+                            source_track->channel_ordering[current_channel] = channel_ordering_ptr->index;
+
+                            if(channel_ordering_ptr->service_type != AV_AUDIO_SERVICE_TYPE_NB) {
+                                uint8_t* side_data = av_stream_new_side_data(st, AV_PKT_DATA_AUDIO_SERVICE_TYPE, sizeof(*ast));
+                                if (!side_data)
+                                    goto fail_and_free;
+                                ast = (enum AVAudioServiceType*)side_data;
+                                *ast = channel_ordering_ptr->service_type;
+                            }
+
+                            current_channel += 1;
+                            break;
+                        }
+                        channel_ordering_ptr++;
+                    }
+                }
+
+                // set language from MCA spoken language information
+                if (mca_sub_descriptor->language) {
+                    ret = set_language(mxf->fc, mca_sub_descriptor->language, &st->metadata);
+                    if (ret < 0) {
+                        return ret;
+                    }
+                }
+            }
+
+            // check if the mapping is not required
+            source_track->require_reordering = 0;
+            for (j = 0; j < descriptor->channels; ++j) {
+                if (source_track->channel_ordering[j] != j) {
+                    source_track->require_reordering = 1;
+                    break;
+                }
+            }
+
+            if (source_track->require_reordering && is_pcm(st->codecpar->codec_id)) {
+                current_channel = 0;
+                av_log(mxf->fc, AV_LOG_DEBUG, "MCA Audio mapping (");
+                for(j = 0; j < descriptor->channels; ++j) {
+                    for(int k = 0; k < descriptor->channels; ++k) {
+                        if(source_track->channel_ordering[k] == current_channel) {
+                            av_log(mxf->fc, AV_LOG_DEBUG, "%d -> %d", source_track->channel_ordering[k], k);
+                            if (current_channel != descriptor->channels - 1)
+                                av_log(mxf->fc, AV_LOG_DEBUG, ", ");
+                            current_channel += 1;
+                        }
+                    }
+                }
+                av_log(mxf->fc, AV_LOG_DEBUG, ")\n");
+            }
         } else if (st->codecpar->codec_type == AVMEDIA_TYPE_DATA) {
             enum AVMediaType type;
             container_ul = mxf_get_codec_ul(mxf_data_essence_container_uls, essence_container_ul);
@@ -2888,6 +3129,8 @@  static const MXFMetadataReadTableEntry mxf_metadata_read_table[] = {
     { { 0x06,0x0e,0x2b,0x34,0x02,0x53,0x01,0x01,0x0d,0x01,0x01,0x01,0x01,0x01,0x5c,0x00 }, mxf_read_generic_descriptor, sizeof(MXFDescriptor), Descriptor }, /* VANC/VBI - SMPTE 436M */
     { { 0x06,0x0e,0x2b,0x34,0x02,0x53,0x01,0x01,0x0d,0x01,0x01,0x01,0x01,0x01,0x5e,0x00 }, mxf_read_generic_descriptor, sizeof(MXFDescriptor), Descriptor }, /* MPEG2AudioDescriptor */
     { { 0x06,0x0e,0x2b,0x34,0x02,0x53,0x01,0x01,0x0d,0x01,0x01,0x01,0x01,0x01,0x64,0x00 }, mxf_read_generic_descriptor, sizeof(MXFDescriptor), Descriptor }, /* DC Timed Text Descriptor */
+    { { 0x06,0x0e,0x2b,0x34,0x02,0x53,0x01,0x01,0x0d,0x01,0x01,0x01,0x01,0x01,0x6c,0x00 }, mxf_read_mca_sub_descriptor, sizeof(MXFMCASubDescriptor), MCASubDescriptor }, /* Soundfield Group Label Subdescriptor */
+    { { 0x06,0x0e,0x2b,0x34,0x02,0x53,0x01,0x01,0x0d,0x01,0x01,0x01,0x01,0x01,0x6b,0x00 }, mxf_read_mca_sub_descriptor, sizeof(MXFMCASubDescriptor), MCASubDescriptor }, /* Audio Channel Label Subdescriptor */
     { { 0x06,0x0e,0x2b,0x34,0x02,0x53,0x01,0x01,0x0d,0x01,0x01,0x01,0x01,0x01,0x3A,0x00 }, mxf_read_track, sizeof(MXFTrack), Track }, /* Static Track */
     { { 0x06,0x0e,0x2b,0x34,0x02,0x53,0x01,0x01,0x0d,0x01,0x01,0x01,0x01,0x01,0x3B,0x00 }, mxf_read_track, sizeof(MXFTrack), Track }, /* Generic Track */
     { { 0x06,0x0e,0x2b,0x34,0x02,0x53,0x01,0x01,0x0d,0x01,0x01,0x01,0x01,0x01,0x14,0x00 }, mxf_read_timecode_component, sizeof(MXFTimecodeComponent), TimecodeComponent },
@@ -3187,12 +3430,6 @@  static void mxf_compute_essence_containers(AVFormatContext *s)
     }
 }
 
-static int is_pcm(enum AVCodecID codec_id)
-{
-    /* we only care about "normal" PCM codecs until we get samples */
-    return codec_id >= AV_CODEC_ID_PCM_S16LE && codec_id < AV_CODEC_ID_PCM_S24DAUD;
-}
-
 static MXFIndexTable *mxf_find_index_table(MXFContext *mxf, int index_sid)
 {
     int i;
@@ -3619,6 +3856,25 @@  static int mxf_set_pts(MXFContext *mxf, AVStream *st, AVPacket *pkt)
     return 0;
 }
 
+static int mxf_audio_remapping(int* channel_ordering, uint8_t* data, int size, int sample_size, int channels)
+{
+    int sample_offset = channels * sample_size;
+    int number_of_samples = size / sample_offset;
+    uint8_t tmp[FF_SANE_NB_CHANNELS * 4];
+    uint8_t* data_ptr = data;
+
+    for (int sample = 0; sample < number_of_samples; ++sample) {
+        memcpy(tmp, data_ptr, sample_offset);
+
+        for (int channel = 0; channel < channels; ++channel) {
+            memcpy(&data_ptr[sample_size * channel_ordering[channel]], &tmp[sample_size * channel],  sample_size);
+        }
+
+        data_ptr += sample_offset;
+    }
+    return 0;
+}
+
 static int mxf_read_packet(AVFormatContext *s, AVPacket *pkt)
 {
     KLVPacket klv;
@@ -3733,6 +3989,15 @@  static int mxf_read_packet(AVFormatContext *s, AVPacket *pkt)
                 return ret;
             }
 
+            // for audio, process audio remapping if MCA label requires it 
+            if (st->codecpar->codec_type == AVMEDIA_TYPE_AUDIO && track->require_reordering && !mxf->skip_audio_reordering) {
+                int byte_per_sample = st->codecpar->bits_per_coded_sample / 8;
+                ret = mxf_audio_remapping(track->channel_ordering, pkt->data, pkt->size, byte_per_sample, st->codecpar->channels);
+                if (ret < 0) {
+                    return ret;
+                }
+            }
+
             /* seek for truncated packets */
             avio_seek(s->pb, klv.next_klv, SEEK_SET);
 
@@ -3920,6 +4185,9 @@  static const AVOption options[] = {
     { "eia608_extract", "extract eia 608 captions from s436m track",
       offsetof(MXFContext, eia608_extract), AV_OPT_TYPE_BOOL, {.i64 = 0}, 0, 1,
       AV_OPT_FLAG_DECODING_PARAM },
+    { "skip_audio_reordering", "skip audio reordering based on Multi-Channel Audio labelling",
+      offsetof(MXFContext, skip_audio_reordering), AV_OPT_TYPE_BOOL, {.i64 = 0}, 0, 1,
+      AV_OPT_FLAG_DECODING_PARAM },
     { NULL },
 };