Repository navigation
Expand file tree
/
Copy pathAudioClipConverter.py
More file actions
220 lines (186 loc) · 6.38 KB
/
Copy pathAudioClipConverter.py
File metadata and controls
220 lines (186 loc) · 6.38 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
from __future__ import annotations
import ctypes
import os
import platform
from typing import TYPE_CHECKING, Dict, Union
from UnityPy.streams import EndianBinaryWriter
from ..helpers.ResourceReader import get_resource_data
try:
import numpy as np
except ImportError:
np = None
import struct
if TYPE_CHECKING:
from ..classes import AudioClip
# pyfmodex loads the dll/so/dylib on import so we have to adjust
# the environment vars before importing it
# This is done in import_pyfmodex()
# which will replace the global pyfmodex var
pyfmodex = None
NO_MODEX = False
def get_fmod_path(
system: Union["Windows", "Linux", "Darwin"], arch: ["x64", "x86", "arm", "arm64"]
) -> str:
if system == "Darwin":
# universal dylib
return "lib/FMOD/Darwin/libfmod.dylib"
if system == "Windows":
return f"lib/FMOD/Windows/{arch}/fmod.dll"
if system == "Linux":
if arch == "x64":
arch = "x86_64"
return f"lib/FMOD/Linux/{arch}/libfmod.so"
raise NotImplementedError(f"Unsupported system: {system}")
def import_pyfmodex():
global pyfmodex, NO_MODEX
if pyfmodex is not None or NO_MODEX:
return
# determine system - Windows, Darwin, Linux, Android
system = platform.system()
# determine architecture
arch = platform.architecture()[0]
machine = platform.machine()
if "arm" in machine:
arch = "arm"
elif "aarch64" in machine:
if system == "Linux":
arch = "arm64"
else:
arch = "arm"
elif arch == "32bit":
arch = "x86"
elif arch == "64bit":
arch = "x64"
fmod_rel_path = get_fmod_path(system, arch)
fmod_path = os.path.join(
os.path.dirname(os.path.dirname(os.path.realpath(__file__))), fmod_rel_path
)
os.environ["PYFMODEX_DLL_PATH"] = fmod_path
# build path and load library
# prepare the environment for pyfmodex
if system != "Windows":
# hotfix ctypes for pyfmodex for non windows systems
ctypes.windll = None
try:
import pyfmodex
except:
NO_MODEX = True
def extract_audioclip_samples(
audio: AudioClip, convert_pcm_float: bool = True
) -> Dict[str, bytes]:
"""extracts all the samples from an AudioClip
:param audio: AudioClip
:type audio: AudioClip
:return: {filename : sample(bytes)}
:rtype: dict
"""
if audio.m_AudioData:
audio_data = audio.m_AudioData
else:
resource = audio.m_Resource
audio_data = get_resource_data(
resource.m_Source,
audio.object_reader.assets_file,
resource.m_Offset,
resource.m_Size,
)
magic = memoryview(audio_data)[:8]
if magic[:4] == b"OggS":
return {f"{audio.m_Name}.ogg": audio_data}
elif magic[:4] == b"RIFF":
return {f"{audio.m_Name}.wav": audio_data}
elif magic[4:8] == b"ftyp":
return {f"{audio.m_Name}.m4a": audio_data}
return dump_samples(audio, audio_data, convert_pcm_float)
def dump_samples(
clip: AudioClip, audio_data: bytes, convert_pcm_float: bool = True
) -> Dict[str, bytes]:
if pyfmodex is None:
import_pyfmodex()
if not pyfmodex:
return {}
# init system
system = pyfmodex.System()
system.init(clip.m_Channels, pyfmodex.flags.INIT_FLAGS.NORMAL, None)
sound = system.create_sound(
bytes(audio_data),
pyfmodex.flags.MODE.OPENMEMORY,
exinfo=pyfmodex.structure_declarations.CREATESOUNDEXINFO(
length=len(audio_data),
numchannels=clip.m_Channels,
defaultfrequency=clip.m_Frequency,
),
)
# iterate over subsounds
samples = {}
for i in range(sound.num_subsounds):
if i > 0:
filename = "%s-%i.wav" % (clip.m_Name, i)
else:
filename = "%s.wav" % clip.m_Name
subsound = sound.get_subsound(i)
samples[filename] = subsound_to_wav(subsound, convert_pcm_float)
subsound.release()
sound.release()
system.release()
return samples
def subsound_to_wav(subsound, convert_pcm_float: bool = True) -> bytes:
# get sound settings
sound_format = subsound.format.format
sound_data_length = subsound.get_length(pyfmodex.enums.TIMEUNIT.PCMBYTES)
channels = subsound.format.channels
bits = subsound.format.bits
sample_rate = int(subsound.default_frequency)
if sound_format in [
pyfmodex.enums.SOUND_FORMAT.PCM8,
pyfmodex.enums.SOUND_FORMAT.PCM16,
pyfmodex.enums.SOUND_FORMAT.PCM24,
pyfmodex.enums.SOUND_FORMAT.PCM32,
]:
audio_format = 1
wav_data_length = sound_data_length
convert_pcm_float = False
elif sound_format == pyfmodex.enums.SOUND_FORMAT.PCMFLOAT:
if convert_pcm_float:
audio_format = 1
bits = 16
wav_data_length = sound_data_length // 2
else:
audio_format = 3
wav_data_length = sound_data_length
else:
raise NotImplementedError("Sound format " + sound_format + " is not supported.")
writer = EndianBinaryWriter(endian="<")
# riff chucnk
writer.write(b"RIFF")
# sizeof(FmtChunk) + sizeof(RiffChunk) + length
writer.write_int(wav_data_length + 36)
writer.write(b"WAVE")
# fmt chunck
writer.write(b"fmt ")
writer.write_int(16) # sizeof(FmtChunk) - sizeof(RiffChunk)
writer.write_short(audio_format)
writer.write_short(channels)
writer.write_int(sample_rate)
writer.write_int(sample_rate * channels * bits // 8)
writer.write_short(channels * bits // 8)
writer.write_short(bits)
# data chunk - sub chunk 2
writer.write(b"data") # sub chunk 2 id
writer.write_int(wav_data_length) # sub chunk 2 size
# sub chunk 2 data
lock = subsound.lock(0, sound_data_length)
for ptr, sound_data_length in lock:
ptr_data = ctypes.string_at(ptr, sound_data_length.value)
if convert_pcm_float:
if np is not None:
ptr_data = np.frombuffer(ptr_data, dtype=np.float32)
ptr_data = (ptr_data * 2**15).astype(np.int16).tobytes()
else:
ptr_data = struct.unpack("<%df" % (len(ptr_data) // 4), ptr_data)
ptr_data = struct.pack(
"<%dh" % len(ptr_data), *[int(f * 2**15) for f in ptr_data]
)
writer.write(ptr_data)
subsound.unlock(*lock)
return writer.save()