Apple-2-SW/ii-vision
1
0
Fork 0
mirror of https://github.com/KrisKennaway/ii-vision.git synced 2025-01-02 05:30:05 +00:00
Code Issues Projects Releases Wiki Activity

Extract audio channel from movie file and emit 5-bit audio opcodes

Browse Source 
at 14KHz.
This commit is contained in:
kris 2019-03-05 20:47:34 +00:00
parent 6e2c83c1e5
commit 2f12407d3c
1 changed files with 68 additions and 15 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

83
audio.py
View File

@ -1,34 +1,87 @@
import numpy as np import numpy as np
import audioread
import librosa import librosa
import soundfile as sf
import opcodes import opcodes
import video import video
TICKS = {
4: opcodes.Tick4,
6: opcodes.Tick6,
8: opcodes.Tick8,
10: opcodes.Tick10,
12: opcodes.Tick12,
14: opcodes.Tick14,
16: opcodes.Tick16,
18: opcodes.Tick18,
20: opcodes.Tick20,
22: opcodes.Tick22,
24: opcodes.Tick24,
26: opcodes.Tick26,
28: opcodes.Tick28,
30: opcodes.Tick30,
32: opcodes.Tick32,
34: opcodes.Tick34,
36: opcodes.Tick36,
38: opcodes.Tick38,
40: opcodes.Tick40,
42: opcodes.Tick42,
44: opcodes.Tick44,
46: opcodes.Tick46,
48: opcodes.Tick48,
50: opcodes.Tick50,
52: opcodes.Tick52,
54: opcodes.Tick54,
56: opcodes.Tick56,
58: opcodes.Tick58,
60: opcodes.Tick60,
62: opcodes.Tick62,
64: opcodes.Tick64,
66: opcodes.Tick66,
}
class Audio: class Audio:
def encode_audio(self, audio): def encode_audio(self, audio):
for a in audio: for a in audio:
yield opcodes.Tick(a) a = max(-30, min(a * 2, 32)) + 34
yield opcodes.Tick(50-a) yield TICKS[a]()
def main(): def main():
filename = librosa.util.example_audio_file() filename = "Computer Chronicles - 06x05 - The Apple II.mp4"
data, samplerate = sf.read(filename, dtype='float32')
data = data.T
a = librosa.resample(data, samplerate, 20000).flatten()
a = librosa.util.normalize(a)
a = (a * 10 + 25).astype(np.int)
s = video.Video(frame_rate=None) s = video.Video(frame_rate=None)
au = Audio() au = Audio()
with open("out.bin", "wb") as out: with audioread.audio_open(filename) as f:
out.write(bytes(s.emit_stream(au.encode_audio(a)))) with open("out.bin", "wb") as out:
out.write(bytes(s.done())) for buf in f.read_data(128 * 1024):
print(f.channels, f.samplerate, f.duration)
data = np.frombuffer(buf, dtype='int16').astype(
'float32').reshape((f.channels, -1), order='F')
a = librosa.core.to_mono(data)
a = librosa.resample(a, f.samplerate, 14000).flatten()
# Normalize to 95%ile
# norm = max(
# abs(np.percentile(a, 5, axis=0)),
# abs(np.percentile(a, 95, axis=0))
# )
# print(min(a),max(a))
# print(norm)
# XXX how to estimate normalization without reading whole file?
norm = 12000
a /= norm # librosa.util.normalize(a)
a = (a * 32).astype(np.int)
out.write(bytes(s.emit_stream(au.encode_audio(a))))
out.write(bytes(s.done()))
if __name__ == "__main__": if __name__ == "__main__":
main() main()