Created
April 29, 2016 18:45
-
-
Save dmishin/56238588f712628fd9ec25c7498c51fb to your computer and use it in GitHub Desktop.
Unwrap imgage pixels into a 1-dimensional array along the Hilbert's curve, then convert it to sound using SOX
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
#!/usr/bin/env python | |
from __future__ import print_function, division | |
import os.path | |
from numpy import array, hstack, vstack, clip | |
from PIL import Image | |
import numpy as np | |
import subprocess # need the subprocess module | |
from tempfile import mkdtemp | |
import shutil | |
#SOX_EXEC = r"C:\Program Files\sox-14-4-2\sox.exe" | |
SOX_EXEC = r"sox" | |
def ilog2(x): | |
log = 0 | |
while x > 1: | |
log += 1 | |
x = x // 2 | |
return log | |
def hilbert_indices(N): | |
m = array([[0]], dtype=np.int) | |
for i in range(N): | |
d = 4**i | |
m1 = vstack( (hstack( (m.T, m.T[::-1, ::-1] + 3*d )), | |
hstack( (m+d, m+2*d) ) | |
) | |
) | |
m = m1 | |
return m | |
def saveSound(sound, ofile, rate): | |
if sound.dtype != np.uint8: | |
sound = sound.astype(np.uint8) | |
tempdir = mkdtemp() | |
try: | |
fname = os.path.join(tempdir, "audio.raw") | |
with open(fname, "wb") as rawfile: | |
sound.tofile( rawfile ) | |
print("saved temp file", fname) | |
args = [SOX_EXEC, | |
"-t", "RAW", | |
'-e','signed-integer', # input encode as signed ints | |
'-L', # input little endin | |
'-c', '1', # mono | |
'-b', '8', # input bits per sample | |
'-r', str(rate), | |
fname, | |
ofile] | |
print ("args:", " ".join(args)) | |
subprocess.check_call( args ) | |
finally: | |
shutil.rmtree(tempdir) | |
print ("Removed temp directory") | |
def hilbertImage2audio( filename, ofile, rate=44100 ): | |
#load mage | |
img = Image.open( filename ) | |
if img.mode != "L": | |
raise ValueError("Image must be grayscale") | |
arr = np.asarray(img) | |
del img | |
if len(arr.shape) == 3: | |
w,h,chans = arr.shape | |
else: | |
w,h = arr.shape | |
chans = 1 | |
if chans != 1: raise ValueError("Image must be grayscale") | |
if w!=h: raise ValueError("Image not square") | |
if (w-1)&w != 0: raise ValueError("Image sides must be powers of 2") | |
arr = arr.astype(np.uint8) | |
arr -= 127 | |
n= ilog2(w) | |
indices = hilbert_indices( n ) | |
sound = np.zeros( (w*h,), dtype = np.uint8 ) | |
sound[ indices ] = arr | |
del arr | |
saveSound(sound, ofile, rate) | |
if __name__=="__main__": | |
import sys | |
try: | |
imgfile = sys.argv[1] | |
audiofile = sys.argv[2] | |
except IndexError: | |
print("Usage: hilbert_image2audio.py IMAGE_FILE AUDIO_FILE") | |
exit(1) | |
hilbertImage2audio(imgfile, audiofile) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment