Class: FormatParser::MOOVParser::Decoder

Inherits:
Object
  • Object
show all
Includes:
IOUtils
Defined in:
lib/parsers/moov_parser/decoder.rb

Overview

Handles decoding of MOV/MPEG4 atoms/boxes in a stream. Will recursively read atoms and parse their data fields if applicable. Also contains a few utility functions for finding atoms in a list etc. To know more about Atoms: developer.apple.com/library/archive/documentation/QuickTime/QTFF/QTFFChap2/qtff2.html

Defined Under Namespace

Classes: Atom

Constant Summary collapse

KNOWN_BRANCH_ATOM_TYPES =

Atoms (boxes) that are known to only contain children, no data fields. Avoid including udta or udta.meta here since we do not have methods for dealing with them yet.

%w(moov mdia trak clip edts minf dinf stbl)
KNOWN_BRANCH_AND_LEAF_ATOM_TYPES =

Mark that udta may contain both

[]
MAX_ATOMS_AT_LEVEL =

Limit how many atoms we scan in sequence, to prevent derailments

128

Instance Method Summary collapse

Methods included from IOUtils

#safe_read, #safe_skip

Instance Method Details

#extract_atom_stream(io, max_read, current_branch = []) ⇒ Object

Recursive descent parser - will drill down to atoms which we know are permitted to have leaf/branch atoms within itself, and will attempt to recover the data fields for leaf atoms



247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
 
# File 'lib/parsers/moov_parser/decoder.rb', line 247

def extract_atom_stream(io, max_read, current_branch = [])
  initial_pos = io.pos
  atoms = []
  MAX_ATOMS_AT_LEVEL.times do
    atom_pos = io.pos

    break if atom_pos - initial_pos >= max_read

    size_and_type = io.read(4 + 4)
    break if size_and_type.to_s.bytesize < 8

    atom_size, atom_type = size_and_type.unpack('Na4')

    # If atom_size is specified to be 1, it is larger than what fits into the
    # 4 bytes and we need to read it right after the atom type
    atom_size = read_64bit_uint(io) if atom_size == 1
    atom_header_size = io.pos - atom_pos
    atom_size_sans_header = atom_size - atom_header_size

    children, fields = if KNOWN_BRANCH_AND_LEAF_ATOM_TYPES.include?(atom_type)
      parse_atom_children_and_data_fields(io, atom_size_sans_header, atom_type, current_branch)
    elsif KNOWN_BRANCH_ATOM_TYPES.include?(atom_type)
      [extract_atom_stream(io, atom_size_sans_header, current_branch + [atom_type]), nil]
    else # Assume leaf atom
      [nil, parse_atom_fields_per_type(io, atom_size_sans_header, atom_type)]
    end

    atoms << Atom.new(atom_pos, atom_size, atom_type, current_branch + [atom_type], children, fields)

    io.seek(atom_pos + atom_size)
  end
  atoms
end

#find_atoms_by_path(atoms, atom_types) ⇒ Object 



51
52
53
54
55
56
57
58
59
60
61
62
63
 
# File 'lib/parsers/moov_parser/decoder.rb', line 51

def find_atoms_by_path(atoms, atom_types)
  type_to_find = atom_types.shift
  requisites = atoms.select { |e| e.atom_type == type_to_find }

  # Return if we found our match
  return requisites if atom_types.empty?

  # Return nil if we didn't find the match at this nesting level
  return unless requisites

  # ...otherwise drill further down
  find_atoms_by_path(requisites.flat_map(&:children).compact || [], atom_types)
end

#find_first_atom_by_path(atoms, *atom_types) ⇒ Object

Finds the first atom in the given Array of Atom structs that matches the type, drilling down if a list of atom names is given



37
38
39
40
41
42
43
44
45
46
47
48
49
 
# File 'lib/parsers/moov_parser/decoder.rb', line 37

def find_first_atom_by_path(atoms, *atom_types)
  type_to_find = atom_types.shift
  requisite = atoms.find { |e| e.atom_type == type_to_find }

  # Return if we found our match
  return requisite if atom_types.empty?

  # Return nil if we didn't find the match at this nesting level
  return unless requisite

  # ...otherwise drill further down
  find_first_atom_by_path(requisite.children || [], *atom_types)
end

#find_video_trak_atom(atoms) ⇒ Object

A file can have multiple tracks. To identify the type it is necessary to check the fields omponent_subtype in hdlr atom under the trak atom More details in developer.apple.com/library/archive/documentation/QuickTime/QTFF/QTFFChap2/qtff2.html#//apple_ref/doc/uid/TP40000939-CH204-DontLinkElementID_147



68
69
70
71
72
73
74
75
76
77
 
# File 'lib/parsers/moov_parser/decoder.rb', line 68

def find_video_trak_atom(atoms)
  trak_atoms = find_atoms_by_path(atoms, ['moov', 'trak'])

  return if trak_atoms.empty?

  trak_atoms.find do |trak_atom|
    hdlr_atom = find_first_atom_by_path([trak_atom], 'trak', 'mdia', 'hdlr')
    hdlr_atom.atom_fields[:component_type] == 'mhlr' && hdlr_atom.atom_fields[:component_subtype] == 'vide'
  end
end

#parse_atom_children_and_data_fields(io, atom_size_sans_header, atom_type, current_branch) ⇒ Object 



239
240
241
242
 
# File 'lib/parsers/moov_parser/decoder.rb', line 239

def parse_atom_children_and_data_fields(io, atom_size_sans_header, atom_type, current_branch)
  parse_atom_fields_per_type(io, atom_size_sans_header, atom_type)
  extract_atom_stream(io, atom_size_sans_header, current_branch + [atom_type])
end

#parse_atom_fields_per_type(io, atom_size, atom_type) ⇒ Object 



231
232
233
234
235
236
237
 
# File 'lib/parsers/moov_parser/decoder.rb', line 231

def parse_atom_fields_per_type(io, atom_size, atom_type)
  if respond_to?("parse_#{atom_type}_atom", true)
    send("parse_#{atom_type}_atom", io, atom_size)
  else
    nil # We can't look inside this leaf atom
  end
end

#parse_dref_atom(io, _) ⇒ Object 



163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
 
# File 'lib/parsers/moov_parser/decoder.rb', line 163

def parse_dref_atom(io, _)
  dict = {
    version: read_byte_value(io),
    flags: read_bytes(io, 3),
    num_entries: read_32bit_uint(io),
  }
  num_entries = dict[:num_entries]
  entries = (1..num_entries).map do
    entry = {
      size: read_32bit_uint(io),
      type: read_bytes(io, 4),
      version: read_bytes(io, 1),
      flags: read_bytes(io, 3),
    }
    entry[:data] = read_bytes(io, entry[:size] - 12)
    entry
  end
  dict[:entries] = entries
  dict
end

#parse_elst_atom(io, _) ⇒ Object 



184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
 
# File 'lib/parsers/moov_parser/decoder.rb', line 184

def parse_elst_atom(io, _)
  dict = {
    version: read_byte_value(io),
    flags: read_bytes(io, 3),
    num_entries: read_32bit_uint(io),
  }
  is_v1 = dict[:version] == 1 # Usual is 0, version 1 has 64bit durations
  num_entries = dict[:num_entries]
  entries = (1..num_entries).map do
    {
      track_duration: is_v1 ? read_64bit_uint(io) : read_32bit_uint(io),
      media_time: is_v1 ? read_64bit_uint(io) : read_32bit_uint(io),
      media_rate: read_32bit_uint(io),
    }
  end
  dict[:entries] = entries
  dict
end

#parse_ftyp_atom(io, atom_size) ⇒ Object 



79
80
81
82
83
84
85
86
87
88
89
90
 
# File 'lib/parsers/moov_parser/decoder.rb', line 79

def parse_ftyp_atom(io, atom_size)
  # Subtract 8 for the atom_size+atom_type,
  # and 8 once more for the major_brand and minor_version. The remaining
  # numbr of bytes is reserved for the compatible brands, 4 bytes per
  # brand.
  num_brands = (atom_size - 8 - 8) / 4
  {
    major_brand: read_bytes(io, 4),
    minor_version: read_binary_coded_decimal(io),
    compatible_brands: (1..num_brands).map { read_bytes(io, 4) },
  }
end

#parse_hdlr_atom(io, atom_size) ⇒ Object 



203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
 
# File 'lib/parsers/moov_parser/decoder.rb', line 203

def parse_hdlr_atom(io, atom_size)
  sub_io = StringIO.new(io.read(atom_size - 8))
  version = read_byte_value(sub_io)
  base_fields = {
    version: version,
    flags: read_bytes(sub_io, 3),
    component_type: read_bytes(sub_io, 4),
    component_subtype: read_bytes(sub_io, 4),
    component_manufacturer: read_bytes(sub_io, 4),
  }
  if version == 1
    version1_fields = {
      component_flags: read_bytes(sub_io, 4),
      component_flags_mask: read_bytes(sub_io, 4),
      component_name: sub_io.read,
    }
    base_fields.merge(version1_fields)
  else
    base_fields
  end
end

#parse_mdhd_atom(io, _) ⇒ Object 



114
115
116
117
118
119
120
121
122
123
124
125
126
127
 
# File 'lib/parsers/moov_parser/decoder.rb', line 114

def parse_mdhd_atom(io, _)
  version = read_byte_value(io)
  is_v1 = version == 1
  {
    version: version,
    flags: read_bytes(io, 3),
    ctime: is_v1 ? read_64bit_uint(io) : read_32bit_uint(io),
    mtime: is_v1 ? read_64bit_uint(io) : read_32bit_uint(io),
    tscale: read_32bit_uint(io),
    duration: is_v1 ? read_64bit_uint(io) : read_32bit_uint(io),
    language: read_32bit_uint(io),
    quality: read_32bit_uint(io),
  }
end

#parse_meta_atom(io, atom_size) ⇒ Object 



225
226
227
228
229
 
# File 'lib/parsers/moov_parser/decoder.rb', line 225

def parse_meta_atom(io, atom_size)
  return if atom_size == 0 # this atom can be empty

  parse_hdlr_atom(io, atom_size)
end

#parse_mvhd_atom(io, _) ⇒ Object 



140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
 
# File 'lib/parsers/moov_parser/decoder.rb', line 140

def parse_mvhd_atom(io, _)
  version = read_byte_value(io)
  is_v1 = version == 1
  {
    version: version,
    flags: read_bytes(io, 3),
    ctime: is_v1 ? read_64bit_uint(io) : read_32bit_uint(io),
    mtime: is_v1 ? read_64bit_uint(io) : read_32bit_uint(io),
    tscale: read_32bit_uint(io),
    duration: is_v1 ? read_64bit_uint(io) : read_32bit_uint(io),
    preferred_rate: read_32bit_uint(io),
    reserved: read_bytes(io, 10),
    matrix_structure: (1..9).map { read_32bit_fixed_point(io) },
    preview_time: read_32bit_uint(io),
    preview_duration: read_32bit_uint(io),
    poster_time: read_32bit_uint(io),
    selection_time: read_32bit_uint(io),
    selection_duration: read_32bit_uint(io),
    current_time: read_32bit_uint(io),
    next_trak_id: read_32bit_uint(io),
  }
end

#parse_tkhd_atom(io, _) ⇒ Object 



92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
 
# File 'lib/parsers/moov_parser/decoder.rb', line 92

def parse_tkhd_atom(io, _)
  version = read_byte_value(io)
  is_v1 = version == 1
  {
    version: version,
    flags: read_chars(io, 3),
    ctime: is_v1 ? read_64bit_uint(io) : read_32bit_uint(io),
    mtime: is_v1 ? read_64bit_uint(io) : read_32bit_uint(io),
    trak_id: read_32bit_uint(io),
    reserved_1: read_chars(io, 4),
    duration: is_v1 ? read_64bit_uint(io) : read_32bit_uint(io),
    reserved_2: read_chars(io, 8),
    layer: read_16bit_uint(io),
    alternate_group: read_16bit_uint(io),
    volume: read_16bit_uint(io),
    reserved_3: read_chars(io, 2),
    matrix_structure: (1..9).map { read_32bit_fixed_point(io) },
    track_width: read_32bit_fixed_point(io),
    track_height: read_32bit_fixed_point(io),
  }
end

#parse_vmhd_atom(io, _) ⇒ Object 



129
130
131
132
133
134
135
136
137
138
 
# File 'lib/parsers/moov_parser/decoder.rb', line 129

def parse_vmhd_atom(io, _)
  {
    version: read_byte_value(io),
    flags: read_bytes(io, 3),
    graphics_mode: read_bytes(io, 2),
    opcolor_r: read_32bit_uint(io),
    opcolor_g: read_32bit_uint(io),
    opcolor_b: read_32bit_uint(io),
  }
end

#read_16bit_fixed_point(io) ⇒ Object 



281
282
283
 
# File 'lib/parsers/moov_parser/decoder.rb', line 281

def read_16bit_fixed_point(io)
  _whole, _fraction = safe_read(io, 2).unpack('CC')
end

#read_16bit_uint(io) ⇒ Object 



301
302
303
 
# File 'lib/parsers/moov_parser/decoder.rb', line 301

def read_16bit_uint(io)
  safe_read(io, 2).unpack('n').first
end

#read_32bit_fixed_point(io) ⇒ Object 



285
286
287
 
# File 'lib/parsers/moov_parser/decoder.rb', line 285

def read_32bit_fixed_point(io)
  _whole, _fraction = safe_read(io, 4).unpack('nn')
end

#read_32bit_uint(io) ⇒ Object 



305
306
307
 
# File 'lib/parsers/moov_parser/decoder.rb', line 305

def read_32bit_uint(io)
  safe_read(io, 4).unpack('N').first
end

#read_64bit_uint(io) ⇒ Object 



309
310
311
 
# File 'lib/parsers/moov_parser/decoder.rb', line 309

def read_64bit_uint(io)
  safe_read(io, 8).unpack('Q>').first
end

#read_binary_coded_decimal(io) ⇒ Object 



313
314
315
316
 
# File 'lib/parsers/moov_parser/decoder.rb', line 313

def read_binary_coded_decimal(io)
  bcd_string = safe_read(io, 4)
  [bcd_string].pack('H*').unpack('C*')
end

#read_byte_value(io) ⇒ Object 



293
294
295
 
# File 'lib/parsers/moov_parser/decoder.rb', line 293

def read_byte_value(io)
  safe_read(io, 1).unpack('C').first
end

#read_bytes(io, n) ⇒ Object 



297
298
299
 
# File 'lib/parsers/moov_parser/decoder.rb', line 297

def read_bytes(io, n)
  safe_read(io, n)
end

#read_chars(io, n) ⇒ Object 



289
290
291
 
# File 'lib/parsers/moov_parser/decoder.rb', line 289

def read_chars(io, n)
  safe_read(io, n)
end