def test_box_header_force_box_extended_size():
bs = pack("uintbe:32, bytes:4, uintbe:64", 1, b"abcd", 32)
box_header = BoxHeader()
box_header.type = b"abcd"
box_header.box_ext_size = 32
assert box_header.type == b"abcd"
assert box_header.box_size == 32
assert box_header.box_ext_size == 32
assert box_header.header_size == 16
assert box_header.content_size == 16
assert bytes(box_header) == bs.bytes
def test_mett_box():
bs = pack("uintbe:32, bytes:4, "
"uintbe:8, uintbe:8, uintbe:8, uintbe:8, uintbe:8, uintbe:8, "
"uintbe:16, "
"bytes:1, bytes:11",
28, b"mett",
0x0, 0x0, 0x0, 0x0, 0x0, 0x0,
1,
b'\0', b'image/heif\0')
box_header = BoxHeader()
mett = bx_def.METT(box_header)
mett.header.type = b"mett"
mett.data_reference_index = 1
mett.content_encoding = b'\0'
mett.mime_format = b'image/heif\0'
mett.refresh_box_size()
box = mett
assert box.header.type == b"mett"
assert box.header.box_size == 28
assert box.data_reference_index == 1
assert box.content_encoding == b'\0'
assert box.mime_format == b'image/heif\0'
assert len(box.boxes) == 0
assert bytes(next(Parser.parse(bs))) == bs.bytes
assert bytes(box) == bs.bytes
def test_sbtt_box():
bs = pack("uintbe:32, bytes:4, "
"uintbe:8, uintbe:8, uintbe:8, uintbe:8, uintbe:8, uintbe:8, "
"uintbe:16, "
"bytes:1, bytes:11",
28, b"sbtt",
0x0, 0x0, 0x0, 0x0, 0x0, 0x0,
1,
b'\0', b'text/plain\0')
box_header = BoxHeader()
stxt = bx_def.STXT(box_header)
stxt.header.type = b"sbtt"
stxt.data_reference_index = 1
stxt.content_encoding = b'\0'
stxt.mime_format = b'text/plain\0'
stxt.refresh_box_size()
box = stxt
assert box.header.type == b"sbtt"
assert box.header.box_size == 28
assert box.data_reference_index == 1
assert box.content_encoding == b'\0'
assert box.mime_format == b'text/plain\0'
assert len(box.boxes) == 0
assert bytes(next(Parser.parse(bs))) == bs.bytes
assert bytes(box) == bs.bytes
def test_ftyp_box_w_no_type():
bs = pack("uintbe:32, bytes:4, "
"bytes:4, uintbe:32, bytes:8",
24, b"ftyp",
b"bzna", 10, b"mp42mp41")
box_header = BoxHeader()
ftyp = bx_def.FTYP(box_header)
ftyp.header.type = b"ftyp"
ftyp.major_brand = (1652190817,) # b"bzna"
ftyp.minor_version = (10,)
ftyp.compatible_brands = ([1836069938, # b"mp42"
1836069937],) # b"mp41"
ftyp.refresh_box_size()
box = ftyp
assert box.header.type == b"ftyp"
assert box.header.box_size == 24
assert box.major_brand == 1652190817 # b"bzna"
assert box.minor_version == 10
assert box.compatible_brands == [1836069938, # b"mp42"
1836069937] # b"mp41"
assert bytes(next(Parser.parse(bs))) == bs.bytes
assert bytes(box) == bs.bytes
def test_sample_entry_box():
bs = pack("uintbe:32, bytes:4, "
"uintbe:8, uintbe:8, uintbe:8, uintbe:8, uintbe:8, uintbe:8, "
"uintbe:16",
16, b"____",
0x0, 0x0, 0x0, 0x0, 0x0, 0x0,
1)
box_header = BoxHeader()
sample_entry_box = bx_def.SampleEntryBox(box_header)
sample_entry_box.header.type = b"____"
sample_entry_box.data_reference_index = 1
sample_entry_box.refresh_box_size()
box = sample_entry_box
assert box.header.type == b"____"
assert box.header.box_size == 16
assert box.data_reference_index == 1
assert len(box.boxes) == 0
assert bytes(box) == bs.bytes
def tune_video(filename, im_width, im_height):
bstr = ConstBitStream(filename=filename)
boxes = [box for box in Parser.parse(bstr)]
# skip 'free' box
boxes = boxes[:1] + boxes[2:]
ftyp = boxes[0]
mdat = boxes[1]
moov = boxes[2]
# pop 'udta' box
moov.boxes.pop()
trak = moov.boxes[-1]
# remove edts
mdia = trak.pop()
trak.pop()
trak.append(mdia)
for box in boxes:
box.load(bstr)
box.refresh_box_size()
ftyp.major_brand = 1769172845 # b"isom"
ftyp.minor_version = 0
ftyp.compatible_brands = [1769172845] # b"isom"
ftyp.refresh_box_size()
# moov.trak.mdia.minf.stbl
stbl = moov.boxes[-1].boxes[-1].boxes[-1].boxes[-1]
# moov.trak.mdia.minf.stbl.stsd.hvc1
hvc1 = stbl.boxes[0].boxes[0]
clap = bx_def.CLAP(BoxHeader())
clap.header.type = b"clap"
clap.clean_aperture_width_n = im_width
clap.clean_aperture_width_d = 1
clap.clean_aperture_height_n = im_height
clap.clean_aperture_height_d = 1
clap.horiz_off_n = im_width - 512
clap.horiz_off_d = 2
clap.vert_off_n = im_height - 512
clap.vert_off_d = 2
# insert clap before pasp
pasp = hvc1.pop()
hvc1.append(clap)
hvc1.append(pasp)
stco = stbl.boxes[-1]
stco.entries[
0].chunk_offset = ftyp.header.box_size + mdat.header.header_size
moov.refresh_box_size()
return b''.join(bytes(box) for box in boxes)
def test_avc1_box():
bs = pack("uintbe:32, bytes:4, "
"uintbe:8, uintbe:8, uintbe:8, uintbe:8, uintbe:8, uintbe:8, "
"uintbe:16, "
"uintbe:16, uintbe:16, uintbe:32, uintbe:32, uintbe:32, "
"uintbe:16, uintbe:16, uintbe:16, uintbe:16, uintbe:16, uintbe:16, "
"uintbe:32, "
"uintbe:16, bytes:32, uintbe:16, "
"intbe:16",
86, b"avc1",
0x0, 0x0, 0x0, 0x0, 0x0, 0x0,
1,
0x0, 0x0, 0x0, 0x0, 0x0,
512, 512, 72, 0, 72, 0,
0x0,
1, b'\0' * 32, 24,
-1)
box_header = BoxHeader()
avc1 = bx_def.AVC1(box_header)
avc1.header.type = b"avc1"
avc1.data_reference_index = 1
avc1.width = 512
avc1.height = 512
avc1.horizresolution = [72, 0]
avc1.vertresolution = [72, 0]
avc1.frame_count = 1
avc1.compressorname = b'\0' * 32
avc1.depth = 24
avc1.refresh_box_size()
box = avc1
assert box.header.type == b"avc1"
assert box.header.box_size == 86
assert box.data_reference_index == 1
assert box.width == 512
assert box.height == 512
assert box.horizresolution == [72, 0]
assert box.vertresolution == [72, 0]
assert box.frame_count == 1
assert box.compressorname == b'\0' * 32
assert box.depth == 24
assert len(box.boxes) == 0
assert bytes(next(Parser.parse(bs))) == bs.bytes
assert bytes(box) == bs.bytes
def test_visual_sample_entry_box():
bs = pack("uintbe:32, bytes:4, "
"uintbe:8, uintbe:8, uintbe:8, uintbe:8, uintbe:8, uintbe:8, "
"uintbe:16, "
"uintbe:16, uintbe:16, uintbe:32, uintbe:32, uintbe:32, "
"uintbe:16, uintbe:16, uintbe:16, uintbe:16, uintbe:16, uintbe:16, "
"uintbe:32, "
"uintbe:16, bytes:32, uintbe:16, "
"intbe:16",
86, b"____",
0x0, 0x0, 0x0, 0x0, 0x0, 0x0,
1,
0x0, 0x0, 0x0, 0x0, 0x0,
512, 512, 72, 0, 72, 0,
0x0,
1, b'\0' * 32, 24,
-1)
box_header = BoxHeader()
visual_sample_entry_box = bx_def.VisualSampleEntryBox(box_header)
visual_sample_entry_box.header.type = b"____"
visual_sample_entry_box.data_reference_index = 1
visual_sample_entry_box.width = 512
visual_sample_entry_box.height = 512
visual_sample_entry_box.horizresolution = [72, 0]
visual_sample_entry_box.vertresolution = [72, 0]
visual_sample_entry_box.frame_count = 1
visual_sample_entry_box.compressorname = b'\0' * 32
visual_sample_entry_box.depth = 24
visual_sample_entry_box.refresh_box_size()
box = visual_sample_entry_box
assert box.header.type == b"____"
assert box.header.box_size == 86
assert box.data_reference_index == 1
assert box.width == 512
assert box.height == 512
assert box.horizresolution == [72, 0]
assert box.vertresolution == [72, 0]
assert box.frame_count == 1
assert box.compressorname == b'\0' * 32
assert box.depth == 24
assert len(box.boxes) == 0
assert bytes(box) == bs.bytes
def test_box_header_extended_user_type_w_drop():
bs = pack("uintbe:32, bytes:4, uintbe:64, bytes:16", 1, b"uuid",
MAX_UINT_32 + 1, b":benzina\x00\x00\x00\x00\x00\x00\x00\x00")
box_header = BoxHeader()
box_header.type = b"uuid:benzina\x00\x00\x00\x00\x00\x00\x00\x00"
box_header.box_size = MAX_UINT_32 + 1
box_header.type = b"abcd"
box_header.box_size = 100
box_header.type = b"uuid:benzina\x00\x00\x00\x00\x00\x00\x00\x00"
box_header.box_size = MAX_UINT_32 + 1
assert box_header.type == b"uuid:benzina\x00\x00\x00\x00\x00\x00\x00\x00"
assert box_header.box_size == MAX_UINT_32 + 1
assert box_header.header_size == 32
assert box_header.content_size == MAX_UINT_32 + 1 - 32
assert bytes(box_header) == bs.bytes
def test_box_header():
bs = pack("uintbe:32, bytes:4", 100, b"abcd")
box_header = BoxHeader()
box_header.type = b"abcd"
box_header.box_size = 100
assert box_header.type == b"abcd"
assert box_header.box_size == 100
assert box_header.header_size == 8
assert box_header.content_size == 92
assert bytes(box_header) == bs.bytes
box_header = BoxHeader()
box_header.box_size = 100
box_header.type = b"abcd"
assert box_header.type == b"abcd"
assert box_header.box_size == 100
assert box_header.header_size == 8
assert box_header.content_size == 92
assert bytes(box_header) == bs.bytes
def test_box_header_w_drop():
bs = pack("uintbe:32, bytes:4", 100, b"abcd")
box_header = BoxHeader()
box_header.type = b"uuid:benzina\x00\x00\x00\x00\x00\x00\x00\x00"
box_header.box_size = MAX_UINT_32 + 1
box_header.type = b"abcd"
box_header.box_size = 100
assert box_header.type == b"abcd"
assert box_header.box_size == 100
assert box_header.header_size == 8
assert box_header.content_size == 92
assert bytes(box_header) == bs.bytes
def make_text_trak(creation_time, modification_time, name,
samples_sizes, samples_offsets):
trak = make_trak(creation_time, modification_time, samples_sizes, samples_offsets)
# MOOV.TRAK.MDIA
mdia = trak.boxes[-1]
# MOOV.TRAK.MDIA.HDLR
hdlr = mdia.boxes[1]
hdlr.handler_type = (b"text",)
hdlr.name = (name,)
# MOOV.TRAK.MDIA.MINF
minf = mdia.boxes[-1]
# MOOV.TRAK.MDIA.MINF.NMHD
nmhd = bx_def.NMHD(FullBoxHeader())
minf.boxes[0] = nmhd
nmhd.header.type = b"nmhd"
nmhd.header.version = (0,)
nmhd.header.flags = (b"\x00\x00\x00",)
# MOOV.TRAK.MDIA.MINF.STBL
stbl = minf.boxes[-1]
# MOOV.TRAK.MDIA.MINF.STBL.STSD
stsd = stbl.boxes[0]
# MOOV.TRAK.MDIA.MINF.STBL.STSD.STXT
stxt = bx_def.STXT(BoxHeader())
stxt.header.type = b"stxt"
stxt.data_reference_index = (1,)
stxt.content_encoding = (b'\0',)
stxt.mime_format = (b'text/plain\0',)
stsd.append(stxt)
return trak
def test_pasp_box():
bs = pack("uintbe:32, bytes:4, uintbe:32, uintbe:32",
16, b"pasp", 150, 157)
box_header = BoxHeader()
pasp = bx_def.PASP(box_header)
pasp.header.type = b"pasp"
pasp.h_spacing = 150
pasp.v_spacing = 157
pasp.refresh_box_size()
box = pasp
assert box.header.type == b"pasp"
assert box.header.box_size == 16
assert box.h_spacing == 150
assert box.v_spacing == 157
assert bytes(next(Parser.parse(bs))) == bs.bytes
assert bytes(box) == bs.bytes
def test_clap_box():
bs = pack("uintbe:32, bytes:4, "
"uintbe:32, uintbe:32, uintbe:32, uintbe:32, "
"intbe:32, uintbe:32, intbe:32, uintbe:32",
40, b"clap",
500, 1, 333, 1,
-12, 2, -179, 2)
box_header = BoxHeader()
clap = bx_def.CLAP(box_header)
clap.header.type = b"clap"
clap.clean_aperture_width_n = 500
clap.clean_aperture_width_d = 1
clap.clean_aperture_height_n = 333
clap.clean_aperture_height_d = 1
clap.horiz_off_n = -12
clap.horiz_off_d = 2
clap.vert_off_n = -179
clap.vert_off_d = 2
clap.refresh_box_size()
box = clap
assert box.header.type == b"clap"
assert box.header.box_size == 40
assert box.clean_aperture_width_n == 500
assert box.clean_aperture_width_d == 1
assert box.clean_aperture_height_n == 333
assert box.clean_aperture_height_d == 1
assert box.horiz_off_n == -12
assert box.horiz_off_d == 2
assert box.vert_off_n == -179
assert box.vert_off_d == 2
assert bytes(next(Parser.parse(bs))) == bs.bytes
assert bytes(box) == bs.bytes
def make_vide_trak(creation_time, modification_time, name,
samples_sizes, samples_offsets):
trak = make_trak(creation_time, modification_time, samples_sizes, samples_offsets)
# MOOV.TRAK.MDIA
mdia = trak.boxes[-1]
# MOOV.TRAK.MDIA.HDLR
hdlr = mdia.boxes[1]
hdlr.handler_type = (b"vide",)
hdlr.name = (name,)
# MOOV.TRAK.MDIA.MINF
minf = mdia.boxes[-1]
# MOOV.TRAK.MDIA.MINF.VMHD
vmhd = bx_def.VMHD(FullBoxHeader())
minf.boxes[0] = vmhd
vmhd.header.type = b"vmhd"
vmhd.header.version = (0,)
# flag is 1
vmhd.header.flags = (b"\x00\x00\x01",)
vmhd.graphicsmode = (0,)
vmhd.opcolor = ([0, 0, 0],)
# MOOV.TRAK.MDIA.MINF
stbl = minf.boxes[-1]
# MOOV.TRAK.MDIA.MINF.STBL.STSD
stsd = stbl.boxes[0]
# MOOV.TRAK.MDIA.MINF.STBL.STSD.AVC1
avc1 = bx_def.AVC1(BoxHeader())
avc1.header.type = b"avc1"
avc1.data_reference_index = (1,)
avc1.width = (-1,)
avc1.height = (-1,)
avc1.horizresolution = ([72, 0],)
avc1.vertresolution = ([72, 0],)
avc1.frame_count = (1,)
avc1.compressorname = (b'\0' * 32,)
avc1.depth = (24,)
# TODO: implement MOOV.TRAK.MDIA.MINF.STBL.STSD.AVC1.AVCC
avcC = bx_def.UnknownBox(BoxHeader())
avcC.header.type = b"avcC"
avcC.payload = b'\x01d\x10\x16\xff\xe1\x00\x1bgd\x10\x16\xac\xb8\x10\x02' \
b'\r\xff\x80K\x00N\xb6\xa5\x00\x00\x03\x00\x01\x00\x00\x03' \
b'\x00\x02\x04\x01\x00\x07h\xee\x01\x9cL\x84\xc0'
avc1.append(avcC)
# MOOV.TRAK.MDIA.MINF.STBL.STSD.AVC1.PASP
pasp = bx_def.PASP(BoxHeader())
pasp.header.type = b"pasp"
pasp.h_spacing = (1,)
pasp.v_spacing = (1,)
avc1.append(pasp)
stsd.append(avc1)
return trak
def test_hvcc_box():
bs = pack("uintbe:32, bytes:4, "
"uintbe:8, "
"int:2, int:1, int:5, uintbe:32, uintbe:48, uintbe:8, "
"bits:4, int:12, bits:6, int:2, bits:6, int:2, bits:5, int:3, bits:5, int:3, "
"uintbe:16, uint:2, uint:3, uint:1, uint:2, uint:8, "
"uint:1, bits:1, uint:6, uintbe:16, "
"uint:16, bytes:3",
39, b"hvcc",
1,
0, 0, 3, 1879048192, 193514046488576, 90,
'0b1111', 0, '0b111111', 0, '0b111111', 1, '0b11111', 0, '0b11111', 0,
0, 0, 1, 0, 3, 1,
1, '0b0', 32, 1,
3, b"321")
box_header = BoxHeader()
hvcc = bx_def.HVCC(box_header)
hvcc.header.type = b"hvcc"
hvcc.header.box_size = 39
hvcc.configuration_version = 1
hvcc.general_profile_space = 0
hvcc.general_tier_flag = 0
hvcc.general_profile_idc = 3
hvcc.general_profile_compatibility_flags = 1879048192
hvcc.general_constraint_indicator_flags = 193514046488576
hvcc.general_level_idc = 90
hvcc.min_spatial_segmentation_idc = 0
hvcc.parallelism_type = 0
hvcc.chroma_format = 1
hvcc.bit_depth_luma_minus_8 = 0
hvcc.bit_depth_chroma_minus_8 = 0
hvcc.avg_frame_rate = 0
hvcc.constant_frame_rate = 0
hvcc.num_temporal_layers = 1
hvcc.temporal_id_nested = 0
hvcc.length_size_minus_one = 3
array = hvcc.append_and_return()
array.array_completeness = 1
array.nal_unit_type = 32
nalu = array.append_and_return()
nalu.nal_unit_length = 3
nalu.nal_unit = (b"321", "bytes:3")
box = hvcc
assert box.header.type == b"hvcc"
assert box.header.box_size == 39
assert box.configuration_version == 1
assert box.general_profile_space == 0
assert box.general_tier_flag == 0
assert box.general_profile_idc == 3
assert box.general_profile_compatibility_flags == 1879048192
assert box.general_constraint_indicator_flags == 193514046488576
assert box.general_level_idc == 90
assert box.min_spatial_segmentation_idc == 0
assert box.parallelism_type == 0
assert box.chroma_format == 1
assert box.bit_depth_luma_minus_8 == 0
assert box.bit_depth_chroma_minus_8 == 0
assert box.avg_frame_rate == 0
assert box.constant_frame_rate == 0
assert box.num_temporal_layers == 1
assert box.temporal_id_nested == 0
assert box.length_size_minus_one == 3
assert box.num_of_arrays == 1
assert len(box.arrays) == 1
array = box.arrays[0]
assert array.array_completeness == 1
assert array.nal_unit_type == 32
assert array.num_nalus == 1
assert len(array.nalus) == 1
nalu = array.nalus[0]
assert nalu.nal_unit_length == 3
assert nalu.nal_unit == b"321"
parsed_box = next(Parser.parse(bs))
parsed_box.load(bs)
assert bytes(parsed_box) == bs.bytes
assert bytes(box) == bs.bytes
def make_trak(creation_time, modification_time, samples_sizes, samples_offsets):
# MOOV.TRAK
trak = bx_def.TRAK(BoxHeader())
trak.header.type = b"trak"
# MOOV.TRAK.TKHD
tkhd = bx_def.TKHD(FullBoxHeader())
tkhd.header.type = b"tkhd"
tkhd.header.version = (1,)
tkhd.header.flags = (b"\x00\x00\x00",)
tkhd.creation_time = (creation_time,)
tkhd.modification_time = (modification_time,)
tkhd.track_id = (-1,)
tkhd.duration = (20 * len(samples_sizes),)
tkhd.layer = (0,)
tkhd.alternate_group = (0,)
tkhd.volume = ([0, 0],)
# TODO: validate matrix (and check if those are 16.16 floats)
tkhd.matrix = ([65536, 0, 0, 0, 65536, 0, 0, 0, 1073741824],)
# TODO: make sure that this is the canvas size
tkhd.width = ([-1, 0],)
tkhd.height = ([-1, 0],)
trak.append(tkhd)
# # MOOV.TRAK.EDTS
# edts = bx_def.EDTS(BoxHeader())
# edts.header.type = b"edts"
#
# # MOOV.TRAK.EDTS.ELST
# elst = bx_def.ELST(FullBoxHeader())
#
# elst.header.type = b"elst"
# elst.header.version = (1,)
# elst.header.flags = (b"\x00\x00\x00",)
#
# entry = elst.append_and_return()
# entry.segment_duration = (60,)
# entry.media_time = (0,)
# entry.media_rate_integer = (1,)
# entry.media_rate_fraction = (0,)
#
# edts.append(elst)
#
# trak.append(edts)
# MOOV.TRAK.MDIA
mdia = bx_def.MDIA(BoxHeader())
mdia.header.type = b"mdia"
# MOOV.TRAK.MDIA.MDHD
mdhd = bx_def.MDHD(FullBoxHeader())
mdhd.header.type = b"mdhd"
mdhd.header.version = (1,)
mdhd.header.flags = (b"\x00\x00\x00",)
mdhd.creation_time = (creation_time,)
mdhd.modification_time = (modification_time,)
mdhd.timescale = (20,)
mdhd.duration = (20 * len(samples_sizes),)
# TODO: check the language code
mdhd.language = ([21, 14, 4],)
mdhd.pre_defined = (0,)
mdia.append(mdhd)
# MOOV.TRAK.MDIA.HDLR
hdlr = bx_def.HDLR(FullBoxHeader())
hdlr.header.type = b"hdlr"
hdlr.header.version = (0,)
hdlr.header.flags = (b"\x00\x00\x00",)
hdlr.pre_defined = (0,)
hdlr.handler_type = (b"____",)
hdlr.name = (b"\0",)
mdia.append(hdlr)
# MOOV.TRAK.MDIA.MINF
minf = bx_def.MINF(BoxHeader())
minf.header.type = b"minf"
# MOOV.TRAK.MDIA.MINF._MHD (placeholder)
_mhd = bx_def.UnknownBox(BoxHeader())
_mhd.header.type = b"_mhd"
minf.append(_mhd)
# MOOV.TRAK.MDIA.MINF.DINF
dinf = bx_def.DINF(BoxHeader())
dinf.header.type = b"dinf"
# MOOV.TRAK.MDIA.MINF.DINF.DREF
dref = bx_def.DREF(FullBoxHeader())
dref.header.type = b"dref"
dref.header.version = (0,)
dref.header.flags = (b"\x00\x00\x00",)
# MOOV.TRAK.MDIA.MINF.DINF.DREF.URL_
url_ = bx_def.URL_(FullBoxHeader())
url_.header.type = b"url "
url_.header.version = (0,)
# TODO: validate that this flags means that the data is in the same file
url_.header.flags = (b"\x00\x00\x01",)
dref.append(url_)
dinf.append(dref)
minf.append(dinf)
# MOOV.TRAK.MDIA.MINF.STBL
stbl = bx_def.STBL(BoxHeader())
stbl.header.type = b"stbl"
# MOOV.TRAK.MDIA.MINF.STBL.STSD
stsd = bx_def.STSD(FullBoxHeader())
stsd.header.type = b"stsd"
stsd.header.version = (0,)
stsd.header.flags = (b"\x00\x00\x00",)
stbl.append(stsd)
# MOOV.TRAK.MDIA.MINF.STBL.STTS
stts = bx_def.STTS(FullBoxHeader())
stts.header.type = b"stts"
stts.header.version = (0,)
stts.header.flags = (b"\x00\x00\x00",)
entry = stts.append_and_return()
# imges count
entry.sample_count = (len(samples_sizes),)
# 1 img / sec
entry.sample_delta = (20,)
stbl.append(stts)
# MOOV.TRAK.MDIA.MINF.STBL.STSZ
stsz = bx_def.STSZ(FullBoxHeader())
stsz.header.type = b"stsz"
stsz.header.version = (0,)
stsz.header.flags = (b"\x00\x00\x00",)
stsz.sample_size = (0,)
for samples_size in samples_sizes:
sample = stsz.append_and_return()
sample.entry_size = (samples_size,)
stbl.append(stsz)
# MOOV.TRAK.MDIA.MINF.STBL.STSC
stsc = bx_def.STSC(FullBoxHeader())
stsc.header.type = b"stsc"
stsc.header.version = (0,)
stsc.header.flags = (b"\x00\x00\x00",)
entry = stsc.append_and_return()
entry.first_chunk = (1,)
entry.samples_per_chunk = (1,)
entry.sample_description_index = (1,)
stbl.append(stsc)
# MOOV.TRAK.MDIA.MINF.STBL.STCO
co = bx_def.STCO(FullBoxHeader())
co.header.type = b"stco"
co.header.version = (0,)
co.header.flags = (b"\x00\x00\x00",)
for offset in gen_sample_offsets(samples_sizes, samples_offsets):
# If the offset gets bigger than 2^32-1, use the 64 bits implementation
if offset > MAX_UINT_32:
# MOOV.TRAK.MDIA.MINF.STBL.CO64
co = bx_def.CO64(FullBoxHeader())
co.header.type = b"co64"
co.header.version = (0,)
co.header.flags = (b"\x00\x00\x00",)
break
entry = co.append_and_return()
entry.chunk_offset = (offset,)
# If the offset gets bigger than 2^32-1, add again using co64
if isinstance(co, bx_def.CO64):
for offset in gen_sample_offsets(samples_sizes, samples_offsets):
entry = co.append_and_return()
entry.chunk_offset = (offset,)
stbl.append(co)
minf.append(stbl)
mdia.append(minf)
trak.append(mdia)
return trak
def test_mp4_dataset():
creation_time = to_mp4_time(datetime(2019, 9, 15, 0, 0, 0))
modification_time = to_mp4_time(datetime(2019, 9, 16, 0, 0, 0))
# FTYP
ftyp = bx_def.FTYP(BoxHeader())
ftyp.header.type = b"ftyp"
ftyp.major_brand = 1769172845 # b"isom"
ftyp.minor_version = 0
ftyp.compatible_brands = [1652190817, # b"bzna"
1769172845] # b"isom"
ftyp.refresh_box_size()
assert ftyp.header.type == b"ftyp"
assert ftyp.header.box_size == 24
assert ftyp.major_brand == 1769172845 # b"isom"
assert ftyp.minor_version == 0
assert ftyp.compatible_brands == [1652190817, # b"bzna"
1769172845] # b"isom"
assert bytes(ftyp) == pack("uintbe:32, bytes:4, bytes:4, uintbe:32, "
"bytes:8",
24, b"ftyp", b"isom", 0,
b"bznaisom")
# MDAT
mdat = bx_def.MDAT(BoxHeader())
mdat.header.type = b"mdat"
data = []
with open("tests/data/small_vid_mdat_im0", "rb") as f:
data.append(f.read())
with open("tests/data/small_vid_mdat_im1", "rb") as f:
data.append(f.read())
with open("tests/data/small_vid_mdat_im2", "rb") as f:
data.append(f.read())
data.append(b"/path/image_1_name.JPEG")
data.append(b"/path/image_2_name.JPEG")
data.append(b"/path/image_3_name.JPEG")
data.append((0).to_bytes(8, byteorder="little"))
data.append((1).to_bytes(8, byteorder="little"))
data.append((0).to_bytes(8, byteorder="little"))
mdat.data = b''.join(data)
mdat.refresh_box_size()
assert mdat.header.type == b"mdat"
assert mdat.header.box_size == 8 + sum(len(entry) for entry in data)
# MOOV
moov = bx_def.MOOV(BoxHeader())
moov.header.type = b"moov"
# MOOV.MVHD
mvhd = make_mvhd(creation_time, modification_time, 3)
# == total number of tracks
mvhd.next_track_id = 5
assert mvhd.next_track_id == 5
moov.append(mvhd)
# MOOV.TRAK
offset = ftyp.header.box_size + mdat.header.header_size
sizes = [198297, 127477, 192476]
trak = make_vide_trak(creation_time, modification_time, b"VideoHandler\0",
sizes, offset)
# MOOV.TRAK.TKHD
tkhd = trak.boxes[0]
# "\x00\x00\x01" trak is enabled
# "\x00\x00\x02" trak is used in the presentation
# "\x00\x00\x04" trak is used in the preview
# "\x00\x00\x08" trak size in not in pixel but in aspect ratio
tkhd.header.flags = b"\x00\x00\x03"
tkhd.track_id = 1
# TODO: make sure that this is the canvas size
tkhd.width = [512, 0]
tkhd.height = [512, 0]
tkhd.refresh_box_size()
assert tkhd.header.type == b"tkhd"
assert tkhd.header.box_size == 104
assert tkhd.header.flags == b"\x00\x00\x03"
assert tkhd.track_id == 1
assert tkhd.width == 512
assert tkhd.height == 512
assert bytes(tkhd) == \
pack("uintbe:32, bytes:4, uintbe:8, bits:24, "
"uintbe:64, uintbe:64, uintbe:32, "
"bits:32, "
"uintbe:64, "
"bits:32, bits:32, "
"uintbe:16, uintbe:16, uintbe:8, uintbe:8, "
"bits:16, "
"uintbe:32, uintbe:32, uintbe:32, uintbe:32, uintbe:32, uintbe:32, uintbe:32, uintbe:32, uintbe:32, "
"uintbe:16, uintbe:16, uintbe:16, uintbe:16",
104, b"tkhd", 1, b"\x00\x00\x03",
creation_time, modification_time, 1,
b"\x00" * 4,
60,
b"\x00" * 4, b"\x00" * 4,
0, 0, 0, 0,
b"\x00" * 2,
65536, 0, 0, 0, 65536, 0, 0, 0, 1073741824,
512, 0, 512, 0)
# MOOV.TRAK.MDIA.MINF.STBL.STSD.AVC1
avc1 = trak.boxes[-1].boxes[-1].boxes[-1].boxes[0].boxes[0]
avc1.width = 512
avc1.height = 512
assert avc1.header.type == b"avc1"
assert avc1.width == 512
assert avc1.height == 512
moov.append(trak)
# MOOV.TRAK
trak = make_meta_trak(creation_time, modification_time, b"bzna_inputs\0",
sizes, offset)
# MOOV.TRAK.TKHD
tkhd = trak.boxes[0]
# "\x00\x00\x01" trak is enabled
# "\x00\x00\x02" trak is used in the presentation
# "\x00\x00\x04" trak is used in the preview
# "\x00\x00\x08" trak size in not in pixel but in aspect ratio
tkhd.header.flags = b"\x00\x00\x00"
tkhd.track_id = 2
# TODO: make sure that this is the canvas size
tkhd.width = [0, 0]
tkhd.height = [0, 0]
tkhd.refresh_box_size()
assert tkhd.header.type == b"tkhd"
assert tkhd.header.box_size == 104
assert tkhd.header.flags == b"\x00\x00\x00"
assert tkhd.track_id == 2
assert tkhd.width == 0
assert tkhd.height == 0
assert bytes(tkhd) == \
pack("uintbe:32, bytes:4, uintbe:8, bits:24, "
"uintbe:64, uintbe:64, uintbe:32, "
"bits:32, "
"uintbe:64, "
"bits:32, bits:32, "
"uintbe:16, uintbe:16, uintbe:8, uintbe:8, "
"bits:16, "
"uintbe:32, uintbe:32, uintbe:32, uintbe:32, uintbe:32, uintbe:32, uintbe:32, uintbe:32, uintbe:32, "
"uintbe:16, uintbe:16, uintbe:16, uintbe:16",
104, b"tkhd", 1, b"\x00\x00\x00",
creation_time, modification_time, 2,
b"\x00" * 4,
60,
b"\x00" * 4, b"\x00" * 4,
0, 0, 0, 0,
b"\x00" * 2,
65536, 0, 0, 0, 65536, 0, 0, 0, 1073741824,
0, 0, 0, 0)
# MOOV.TRAK.MDIA.MDHD
mdhd = trak.boxes[-1].boxes[0]
mdhd.timescale = 20
mdhd.duration = 60
mdhd.refresh_box_size()
assert mdhd.header.type == b"mdhd"
assert mdhd.header.box_size == 44
assert mdhd.timescale == 20
assert mdhd.duration == 60
assert bytes(mdhd) == \
pack("uintbe:32, bytes:4, uintbe:8, bits:24, "
"uintbe:64, uintbe:64, uintbe:32, uintbe:64, "
"bits:1, uint:5, uint:5, uint:5, "
"bits:16",
44, b"mdhd", 1, b"\x00\x00\x00",
creation_time, modification_time, 20, 60,
0x1, 21, 14, 4,
b"\x00" * 2)
# MOOV.TRAK.MDIA.MINF.STBL.STSD.METT
mett = trak.boxes[-1].boxes[-1].boxes[-1].boxes[0].boxes[0]
mett.mime_format = b'video/h264\0'
mett.refresh_box_size()
assert mett.header.type == b"mett"
assert mett.header.box_size == 28
assert mett.mime_format == b'video/h264\0'
moov.append(trak)
# MOOV.TRAK
offset += sum(sizes)
sizes = [23, 23, 23]
trak = make_text_trak(creation_time, modification_time, b"bzna_fnames\0",
sizes, offset)
# MOOV.TRAK.TKHD
tkhd = trak.boxes[0]
# "\x00\x00\x01" trak is enabled
# "\x00\x00\x02" trak is used in the presentation
# "\x00\x00\x04" trak is used in the preview
# "\x00\x00\x08" trak size in not in pixel but in aspect ratio
tkhd.header.flags = b"\x00\x00\x00"
tkhd.track_id = 3
# TODO: make sure that this is the canvas size
tkhd.width = [0, 0]
tkhd.height = [0, 0]
tkhd.refresh_box_size()
assert tkhd.header.type == b"tkhd"
assert tkhd.header.box_size == 104
assert tkhd.header.flags == b"\x00\x00\x00"
assert tkhd.track_id == 3
assert tkhd.width == 0
assert tkhd.height == 0
assert bytes(tkhd) == \
pack("uintbe:32, bytes:4, uintbe:8, bits:24, "
"uintbe:64, uintbe:64, uintbe:32, "
"bits:32, "
"uintbe:64, "
"bits:32, bits:32, "
"uintbe:16, uintbe:16, uintbe:8, uintbe:8, "
"bits:16, "
"uintbe:32, uintbe:32, uintbe:32, uintbe:32, uintbe:32, uintbe:32, uintbe:32, uintbe:32, uintbe:32, "
"uintbe:16, uintbe:16, uintbe:16, uintbe:16",
104, b"tkhd", 1, b"\x00\x00\x00",
creation_time, modification_time, 3,
b"\x00" * 4,
60,
b"\x00" * 4, b"\x00" * 4,
0, 0, 0, 0,
b"\x00" * 2,
65536, 0, 0, 0, 65536, 0, 0, 0, 1073741824,
0, 0, 0, 0)
# MOOV.TRAK.MDIA.MDHD
mdhd = trak.boxes[-1].boxes[0]
mdhd.timescale = 20
mdhd.duration = 60
mdhd.refresh_box_size()
assert mdhd.header.type == b"mdhd"
assert mdhd.header.box_size == 44
assert mdhd.timescale == 20
assert mdhd.duration == 60
assert bytes(mdhd) == \
pack("uintbe:32, bytes:4, uintbe:8, bits:24, "
"uintbe:64, uintbe:64, uintbe:32, uintbe:64, "
"bits:1, uint:5, uint:5, uint:5, "
"bits:16",
44, b"mdhd", 1, b"\x00\x00\x00",
creation_time, modification_time, 20, 60,
0x1, 21, 14, 4,
b"\x00" * 2)
moov.append(trak)
# MOOV.TRAK
offset += sum(sizes)
sizes = [8, 8, 8]
trak = make_text_trak(creation_time, modification_time, b"bzna_targets\0",
sizes, offset)
# MOOV.TRAK.TKHD
tkhd = trak.boxes[0]
# "\x00\x00\x01" trak is enabled
# "\x00\x00\x02" trak is used in the presentation
# "\x00\x00\x04" trak is used in the preview
# "\x00\x00\x08" trak size in not in pixel but in aspect ratio
tkhd.header.flags = b"\x00\x00\x00"
tkhd.track_id = 4
# TODO: make sure that this is the canvas size
tkhd.width = [0, 0]
tkhd.height = [0, 0]
tkhd.refresh_box_size()
assert tkhd.header.type == b"tkhd"
assert tkhd.header.box_size == 104
assert tkhd.header.flags == b"\x00\x00\x00"
assert tkhd.track_id == 4
assert tkhd.width == 0
assert tkhd.height == 0
assert bytes(tkhd) == \
pack("uintbe:32, bytes:4, uintbe:8, bits:24, "
"uintbe:64, uintbe:64, uintbe:32, "
"bits:32, "
"uintbe:64, "
"bits:32, bits:32, "
"uintbe:16, uintbe:16, uintbe:8, uintbe:8, "
"bits:16, "
"uintbe:32, uintbe:32, uintbe:32, uintbe:32, uintbe:32, uintbe:32, uintbe:32, uintbe:32, uintbe:32, "
"uintbe:16, uintbe:16, uintbe:16, uintbe:16",
104, b"tkhd", 1, b"\x00\x00\x00",
creation_time, modification_time, 4,
b"\x00" * 4,
60,
b"\x00" * 4, b"\x00" * 4,
0, 0, 0, 0,
b"\x00" * 2,
65536, 0, 0, 0, 65536, 0, 0, 0, 1073741824,
0, 0, 0, 0)
# MOOV.TRAK.MDIA.MDHD
mdhd = trak.boxes[-1].boxes[0]
mdhd.timescale = 20
mdhd.duration = 60
mdhd.refresh_box_size()
assert mdhd.header.type == b"mdhd"
assert mdhd.header.box_size == 44
assert mdhd.timescale == 20
assert mdhd.duration == 60
assert bytes(mdhd) == \
pack("uintbe:32, bytes:4, uintbe:8, bits:24, "
"uintbe:64, uintbe:64, uintbe:32, uintbe:64, "
"bits:1, uint:5, uint:5, uint:5, "
"bits:16",
44, b"mdhd", 1, b"\x00\x00\x00",
creation_time, modification_time, 20, 60,
0x1, 21, 14, 4,
b"\x00" * 2)
moov.append(trak)
moov.refresh_box_size()
assert mvhd.next_track_id == len(moov.boxes)
assert moov.header.type == b"moov"
assert len(moov.boxes) == 5
mp4_bytes = b''.join([bytes(ftyp), bytes(mdat), bytes(moov)])
for box in moov.boxes:
if isinstance(box, bx_def.TRAK):
mdia = box.boxes[1]
else:
continue
# trak.mdia.tkhd.name == b"bzna_inputs\0"
if mdia.boxes[1].name == b"bzna_inputs\0":
pass
# trak.mdia.tkhd.name == b"bzna_names\0"
elif mdia.boxes[1].name == b"bzna_fnames\0":
# trak.mdia.minf.stbl.stsz
stsz = mdia.boxes[2].boxes[2].boxes[2]
# trak.mdia.minf.stbl.stco
stco = mdia.boxes[2].boxes[2].boxes[4]
for i, (sample, entry) in enumerate(zip(stsz.samples, stco.entries)):
sample_end = entry.chunk_offset + sample.entry_size
if i == 0:
assert mp4_bytes[entry.chunk_offset:sample_end] == \
b"/path/image_1_name.JPEG"
elif i == 1:
assert mp4_bytes[entry.chunk_offset:sample_end] == \
b"/path/image_2_name.JPEG"
elif i == 2:
assert mp4_bytes[entry.chunk_offset:sample_end] == \
b"/path/image_3_name.JPEG"
# trak.mdia.tkhd.name == b"bzna_targets\0"
elif mdia.boxes[1].name == b"bzna_targets\0":
# trak.mdia.minf.stbl.stsz
stsz = mdia.boxes[2].boxes[2].boxes[2]
# trak.mdia.minf.stbl.stco
stco = mdia.boxes[2].boxes[2].boxes[4]
for i, (sample, entry) in enumerate(zip(stsz.samples, stco.entries)):
sample_end = entry.chunk_offset + sample.entry_size
if i == 0:
assert mp4_bytes[entry.chunk_offset:sample_end] == \
(0).to_bytes(8, byteorder="little")
elif i == 1:
assert mp4_bytes[entry.chunk_offset:sample_end] == \
(1).to_bytes(8, byteorder="little")
elif i == 2:
assert mp4_bytes[entry.chunk_offset:sample_end] == \
(0).to_bytes(8, byteorder="little")
with open("tests/data/small_dataset.out.mp4", "rb") as f:
assert mp4_bytes == f.read()
def test_mp4_small_vid():
creation_time = to_mp4_time(datetime(2019, 9, 15, 0, 0, 0))
modification_time = to_mp4_time(datetime(2019, 9, 16, 0, 0, 0))
# FTYP
ftyp = bx_def.FTYP(BoxHeader())
ftyp.header.type = b"ftyp"
ftyp.major_brand = 1836069937 # b"mp41"
ftyp.minor_version = 0
ftyp.compatible_brands = [1836069937] # b'mp41'
ftyp.refresh_box_size()
assert ftyp.header.type == b"ftyp"
assert ftyp.header.box_size == 20
assert ftyp.major_brand == 1836069937 # b"mp41"
assert ftyp.minor_version == 0
assert ftyp.compatible_brands == [1836069937] # b'mp41'
assert bytes(ftyp) == pack("uintbe:32, bytes:4, "
"bytes:4, uintbe:32, bytes:4",
20, b"ftyp",
b"mp41", 0, b'mp41')
# MDAT
mdat = bx_def.MDAT(BoxHeader())
mdat.header.type = b"mdat"
data = []
with open("tests/data/small_vid_mdat_im0", "rb") as f:
data.append(f.read())
with open("tests/data/small_vid_mdat_im1", "rb") as f:
data.append(f.read())
with open("tests/data/small_vid_mdat_im2", "rb") as f:
data.append(f.read())
mdat.data = b''.join(data)
mdat.refresh_box_size()
assert mdat.header.type == b"mdat"
assert mdat.header.box_size == 518258
# MOOV
moov = bx_def.MOOV(BoxHeader())
moov.header.type = b"moov"
# MOOV.MVHD
mvhd = make_mvhd(creation_time, modification_time, 3)
# == total number of tracks
mvhd.next_track_id = 2
assert mvhd.next_track_id == 2
moov.append(mvhd)
# MOOV.TRAK
offset = ftyp.header.box_size + mdat.header.header_size
sizes = [198297, 127477, 192476]
trak = make_vide_trak(creation_time, modification_time, b"VideoHandler\0",
sizes, offset)
# MOOV.TRAK.TKHD
tkhd = trak.boxes[0]
# "\x00\x00\x01" trak is enabled
# "\x00\x00\x02" trak is used in the presentation
# "\x00\x00\x04" trak is used in the preview
# "\x00\x00\x08" trak size in not in pixel but in aspect ratio
tkhd.header.flags = b"\x00\x00\x03"
tkhd.track_id = 1
# TODO: make sure that this is the canvas size
tkhd.width = [512, 0]
tkhd.height = [512, 0]
tkhd.refresh_box_size()
assert tkhd.header.type == b"tkhd"
assert tkhd.header.box_size == 104
assert tkhd.header.flags == b"\x00\x00\x03"
assert tkhd.track_id == 1
assert tkhd.width == 512
assert tkhd.height == 512
assert bytes(tkhd) == \
pack("uintbe:32, bytes:4, uintbe:8, bits:24, "
"uintbe:64, uintbe:64, uintbe:32, "
"bits:32, "
"uintbe:64, "
"bits:32, bits:32, "
"uintbe:16, uintbe:16, uintbe:8, uintbe:8, "
"bits:16, "
"uintbe:32, uintbe:32, uintbe:32, uintbe:32, uintbe:32, uintbe:32, uintbe:32, uintbe:32, uintbe:32, "
"uintbe:16, uintbe:16, uintbe:16, uintbe:16",
104, b"tkhd", 1, b"\x00\x00\x03",
creation_time, modification_time, 1,
b"\x00" * 4,
60,
b"\x00" * 4, b"\x00" * 4,
0, 0, 0, 0,
b"\x00" * 2,
65536, 0, 0, 0, 65536, 0, 0, 0, 1073741824,
512, 0, 512, 0)
# MOOV.TRAK.MDIA.MDHD
mdhd = trak.boxes[-1].boxes[0]
mdhd.timescale = 20
mdhd.duration = 60
mdhd.refresh_box_size()
assert mdhd.header.type == b"mdhd"
assert mdhd.header.box_size == 44
assert mdhd.timescale == 20
assert mdhd.duration == 60
assert bytes(mdhd) == \
pack("uintbe:32, bytes:4, uintbe:8, bits:24, "
"uintbe:64, uintbe:64, uintbe:32, uintbe:64, "
"bits:1, uint:5, uint:5, uint:5, "
"bits:16",
44, b"mdhd", 1, b"\x00\x00\x00",
creation_time, modification_time, 20, 60,
0x1, 21, 14, 4,
b"\x00" * 2)
# MOOV.TRAK.MDIA.MINF.STBL.STSD
stsd = trak.boxes[-1].boxes[-1].boxes[-1].boxes[0]
assert stsd.header.type == b"stsd"
assert len(stsd.boxes) == 1
# MOOV.TRAK.MDIA.MINF.STBL.STSD.AVC1
avc1 = stsd.boxes[0]
avc1.width = 512
avc1.height = 512
assert avc1.header.type == b"avc1"
assert avc1.width == 512
assert avc1.height == 512
moov.append(trak)
moov.refresh_box_size()
assert mvhd.next_track_id == len(moov.boxes)
assert moov.header.type == b"moov"
assert len(moov.boxes) == 2
mp4_bytes = b''.join([bytes(ftyp), bytes(mdat), bytes(moov)])
with open("tests/data/small_vid.out.mp4", "rb") as f:
assert mp4_bytes == f.read()