Also add initial support for v2 metadata (--v2). Currently supported options:

* stereo_mode = {none | top-bottom | left-right} * projection = {none | equirectangular }

Author: nbirkbeck

spatialmedia/__main__.py

@@ -48,6 +48,12 @@ def main():
  help=
  "injects spatial media metadata into the first file specified (.mp4 or "
  ".mov) and saves the result to the second file specified")
+ parser.add_argument(
+ "-2",
+ "--v2",
+ action="store_true",
+ help=
+ "Uses v2 of the video metadata spec")
  video_group = parser.add_argument_group("Spherical Video")
  video_group.add_argument("-s",
  "--stereo",
@@ -91,10 +97,13 @@ def main():
  console("Injecting metadata requires both an input file and output file.")
  return
 
- metadata = metadata_utils.Metadata()
- metadata.video = metadata_utils.generate_spherical_xml(args.projection,
- args.stereo_mode,
- args.crop)
+ metadata = metadata_utils.Metadata(args.projection, args.stereo_mode)
+ if not args.v2:
+ metadata.projection = None
+ metadata.stereo_mode = None
+ metadata.video = metadata_utils.generate_spherical_xml(args.projection,
+ args.stereo_mode,
+ args.crop)
 
  if args.spatial_audio:
  parsed_metadata = metadata_utils.parse_metadata(args.file[0], console)
@@ -110,7 +119,7 @@ def main():
  "spatial audio format." % (parsed_metadata.num_audio_channels))
  return
 
- if metadata.video:
+ if metadata.video or metadata.projection or metadata.stereo_mode:
  metadata_utils.inject_metadata(args.file[0], args.file[1], metadata,
  console)
  else:

spatialmedia/metadata_utils.py

@@ -95,7 +95,9 @@
 ]
 
 class Metadata(object):
- def __init__(self):
+ def __init__(self, projection=None, stereo_mode=None):
+ self.projection = None if (not projection or projection == "none") else projection
+ self.stereo_mode = None if (not stereo_mode or stereo_mode == "none") else stereo_mode
  self.video = None
  self.audio = None
 
@@ -152,7 +154,7 @@ def spherical_uuid(metadata):
  return uuid_leaf
 
 
-def mpeg4_add_spherical(mpeg4_file, in_fh, metadata):
+def mpeg4_add_spherical_xml_v1(mpeg4_file, in_fh, metadata):
  """Adds a spherical uuid box to an mpeg4 file for all video tracks.
 
  Args:
@@ -184,6 +186,72 @@ def mpeg4_add_spherical(mpeg4_file, in_fh, metadata):
  mpeg4_file.resize()
  return True
 
+def mpeg4_add_spherical_v2(mpeg4_file, in_fh, projection, stereo_mode):
+ for element in mpeg4_file.moov_box.contents:
+ if element.name == mpeg.constants.TAG_TRAK:
+ for sub_element in element.contents:
+ if sub_element.name != mpeg.constants.TAG_MDIA:
+ continue
+ for mdia_sub_element in sub_element.contents:
+ if mdia_sub_element.name != mpeg.constants.TAG_HDLR:
+ continue
+ position = mdia_sub_element.content_start() + 8
+ in_fh.seek(position)
+ if in_fh.read(4) == mpeg.constants.TAG_VIDE:
+ ret = inject_spatial_video_v2_atoms(
+ in_fh, sub_element, projection, stereo_mode)
+ mpeg4_file.resize()
+ return ret
+
+
+def inject_spatial_video_v2_atoms(in_fh, video_media_atom, projection, stereo_mode):
+ """Adds spherical v2 boxes to an mpeg4 file for all video tracks.
+
+ Args:
+ mpeg4_file: mpeg4, Mpeg4 file structure to add metadata.
+ in_fh: file handle, Source for uncached file contents.
+ metadata: string, xml metadata to inject into spherical tag.
+ """
+ for atom in video_media_atom.contents:
+ if atom.name != mpeg.constants.TAG_MINF:
+ continue
+ for element in atom.contents:
+ if element.name != mpeg.constants.TAG_STBL:
+ continue
+ for sub_element in element.contents:
+ if sub_element.name != mpeg.constants.TAG_STSD:
+ continue
+ for sample_description in sub_element.contents:
+ if sample_description.name in\
+ mpeg.constants.VIDEO_SAMPLE_DESCRIPTIONS:
+ in_fh.seek(sample_description.position +
+ sample_description.header_size + 16)
+ # Should remove any existing boxes...
+ if stereo_mode:
+ st3d_atom = mpeg.sv3d.ST3DBox.create()
+ st3d_atom.name = mpeg.constants.TAG_ST3D
+ st3d_atom.set_stereo_mode_from_string(stereo_mode)
+
+ sample_description.remove(st3d_atom.name)
+ sample_description.add(st3d_atom)
+
+ if projection:
+ proj_atom = mpeg.container.Container(header_size=8)
+ proj_atom.name = mpeg.constants.TAG_PROJ
+
+ proj_atom.add(mpeg.sv3d.PRHDBox.create())
+ proj_atom.add(mpeg.sv3d.EQUIBox.create())
+
+ sv3d_atom = mpeg.container.Container(header_size=8)
+ sv3d_atom.name = mpeg.constants.TAG_SV3D
+
+ sv3d_atom.add(proj_atom)
+
+ sample_description.remove(sv3d_atom.name)
+ sample_description.add(sv3d_atom)
+
+ return True
+
 
 def mpeg4_add_spatial_audio(mpeg4_file, in_fh, audio_metadata, console):
  """Adds spatial audio metadata to the first audio track of the input
@@ -355,6 +423,21 @@ def parse_spherical_mpeg4(mpeg4_file, fh, console):
  if sa3d_elem.name == mpeg.constants.TAG_SA3D:
  sa3d_elem.print_box(console)
  metadata.audio = sa3d_elem
+
+ for sv3d_container_elem in stsd_elem.contents:
+ if sv3d_container_elem.name not in \
+ mpeg.constants.VIDEO_SAMPLE_DESCRIPTIONS:
+ continue
+ for sub_elem in sv3d_container_elem.contents:
+ if sub_elem.name == mpeg.constants.TAG_SV3D:
+ console("\t\tSV3D {")
+ sub_elem.print_box(console)
+ console("\t\t}")
+ elif sub_elem.name == mpeg.constants.TAG_ST3D:
+ console("\t\tST3D {")
+ sub_elem.print_box(console)
+ console("\t\t} ")
+
  return metadata
 
 def parse_mpeg4(input_file, console):
@@ -378,9 +461,13 @@ def inject_mpeg4(input_file, output_file, metadata, console):
  if mpeg4_file is None:
  console("Error file could not be opened.")
 
- if not mpeg4_add_spherical(mpeg4_file, in_fh, metadata.video):
+ if metadata.video and not mpeg4_add_spherical_xml_v1(mpeg4_file, in_fh, metadata.video):
  console("Error failed to insert spherical data")
 
+ if ((metadata.projection or metadata.stereo_mode)
+ and not mpeg4_add_spherical_v2(mpeg4_file, in_fh, metadata.projection, metadata.stereo_mode)):
+ console("Error failed to insert spherical data v2")
+
  if metadata.audio:
  if not mpeg4_add_audio_metadata(
  mpeg4_file, in_fh, metadata.audio, console):

spatialmedia/mpeg/constants.py

@@ -29,8 +29,14 @@
 TAG_FTYP = b"ftyp"
 TAG_ESDS = b"esds"
 TAG_SOUN = b"soun"
+TAG_VIDE = b"vide"
 TAG_SA3D = b"SA3D"
 
+TAG_PRHD = b"prhd"
+TAG_EQUI = b"equi"
+TAG_SVHD = b"svhd"
+TAG_ST3D = b"st3d"
+
 # Container types.
 TAG_MOOV = b"moov"
 TAG_UDTA = b"udta"
@@ -43,6 +49,10 @@
 TAG_UUID = b"uuid"
 TAG_WAVE = b"wave"
 
+TAG_SV3D = b"sv3d"
+TAG_PROJ = b"proj"
+
+
 # Sound sample descriptions.
 TAG_NONE = b"NONE"
 TAG_RAW_ = b"raw "
@@ -58,6 +68,14 @@
 TAG_MP4A = b"mp4a"
 TAG_OPUS = b"Opus"
 
+# Video sample descriptions.
+TAG_AVC1 = b"avc1"
+TAG_VP09 = b"vp09"
+TAG_AV01 = b"av01"
+TAV_HEV1 = b"hev1"
+TAG_DVH1 = b"dvh1"
+TAG_APCN = b"apcn"
+
 SOUND_SAMPLE_DESCRIPTIONS = frozenset([
  TAG_NONE,
  TAG_RAW_,
@@ -74,6 +92,16 @@
  TAG_OPUS,
  ])
 
+VIDEO_SAMPLE_DESCRIPTIONS = frozenset([
+ TAG_NONE,
+ TAG_AVC1,
+ TAG_VP09,
+ TAG_AV01,
+ TAV_HEV1,
+ TAG_DVH1,
+ TAG_APCN,
+ ])
+
 CONTAINERS_LIST = frozenset([
  TAG_MDIA,
  TAG_MINF,
@@ -83,4 +111,6 @@
  TAG_TRAK,
  TAG_UDTA,
  TAG_WAVE,
- ]).union(SOUND_SAMPLE_DESCRIPTIONS)
+ TAG_SV3D,
+ TAG_PROJ
+ ]).union(SOUND_SAMPLE_DESCRIPTIONS).union(VIDEO_SAMPLE_DESCRIPTIONS)

spatialmedia/mpeg/container.py

@@ -25,6 +25,7 @@
 from spatialmedia.mpeg import box
 from spatialmedia.mpeg import constants
 from spatialmedia.mpeg import sa3d
+from spatialmedia.mpeg import sv3d
 
 def load(fh, position, end):
  if position is None:
@@ -34,14 +35,15 @@ def load(fh, position, end):
  header_size = 8
  size = struct.unpack(">I", fh.read(4))[0]
  name = fh.read(4)
-
  is_box = name not in constants.CONTAINERS_LIST
  # Handle the mp4a decompressor setting (wave -> mp4a).
  if name == constants.TAG_MP4A and size == 12:
  is_box = True
  if is_box:
  if name == constants.TAG_SA3D:
  return sa3d.load(fh, position, end)
+ if sv3d.is_supported_box_name(name):
+ return sv3d.load(fh, position, end)
  return box.load(fh, position, end)
 
  if size == 1:
@@ -74,6 +76,17 @@ def load(fh, position, end):
  else:
  print("Unsupported sample description version:",
  sample_description_version)
+ if name in constants.VIDEO_SAMPLE_DESCRIPTIONS:
+ current_pos = fh.tell()
+ fh.seek(current_pos + 8)
+ sample_description_version = struct.unpack(">h", fh.read(2))[0]
+ fh.seek(current_pos)
+
+ if sample_description_version == 0:
+ padding = 78
+ else:
+ print("Unsupported video sample description version:",
+ sample_description_version)
 
  new_box = Container()
  new_box.name = name
@@ -106,10 +119,10 @@ def load_multiple(fh, position=None, end=None):
 class Container(box.Box):
  """MPEG4 container box contents / behaviour."""
 
- def __init__(self, padding=0):
+ def __init__(self, padding=0, header_size=0):
  self.name = ""
  self.position = 0
- self.header_size = 0
+ self.header_size = header_size
  self.content_size = 0
  self.contents = list()
  self.padding = padding
@@ -122,6 +135,10 @@ def resize(self):
  element.resize()
  self.content_size += element.size()
 
+ def print_box(self, console):
+ for child in self.contents:
+ child.print_box(console)
+
  def print_structure(self, indent=""):
  """Prints the box structure and recurses on contents."""
  size1 = self.header_size