This is awesome! Here's a python version (not a real issue)

[ZachNagengast]

I see this is quite an old repo now but it helped me get to something that worked where ChatGPT failed 😅 Thanks for posting it

Here's a version in python:

import numpy as np
from PIL import Image
import soundfile as sf

def img2spect(img_data, minfreq, maxfreq, dur, Fs, invert):
    """
    Convert an image to audio such that the image will appear on a spectrogram of the audio.
    
    Parameters:
    - img_data: Image data as a numpy array
    - minfreq: Minimum frequency represented in the spectrogram
    - maxfreq: Maximum frequency represented in the spectrogram
    - dur: Duration of the output audio in seconds
    - Fs: Sample rate of the output audio
    - invert: If true, invert the image's colors
    
    Returns:
    - audio: A numpy array containing the audio samples encoding the image
    """
    if invert:
        img_data = 255 - img_data
    imd = np.flipud(img_data) / 255.0
    audio = np.zeros(int(dur * Fs))
    oldlin = np.linspace(0, dur, img_data.shape[1])
    newlin = np.linspace(0, dur, int(dur * Fs))
    np.random.seed(1)  # for reproducibility
    
    for row in range(img_data.shape[0]):
        amplitudes = np.interp(newlin, oldlin, imd[row, :], left=0, right=0)
        freq = minfreq * (maxfreq / minfreq) ** (row / img_data.shape[0])
        phase_offset = np.random.rand()
        signal = np.sin(2 * np.pi * freq * (newlin + phase_offset))
        audio += amplitudes * signal
    
    audio = audio / np.max(np.abs(audio))
    return audio

Example usage:

image_path = 'path_to_input_image.png'
output_audio_path = 'path_to_output_file.wav'
min_freq = 500  # in Hz
max_freq = 10000  # in Hz
duration = 5  # in seconds
sample_rate = 44100  # in Hz
invert_colors = False

# Load the image
img = Image.open(image_path).convert('L')
img_data = np.array(img)

# Generate the audio signal
audio_signal = img2spect(img_data, min_freq, max_freq, duration, sample_rate, invert_colors)

# Save the audio signal to a file
sf.write(output_audio_path, audio_signal, sample_rate)