Supprimer gravure_upic/python/README.md

Reviewed-on: #3

README.md

@@ -1,14 +1 @@
-# git repo for art num
+git repo for Vivien Perrot
-*the wiki will be updated with more information and usefull snipet. fell free to contribute*
-
-test port ssh 
-
-## how to 
-1. clone this repo to you own computer 
-    ` git clone https://git.erg.school/P039/art_num_2024.git`
-2. check before eatch courses for update 
-    ` git pull ` 
-
-## content 
-
- 

arduino/README.md

@@ -1 +0,0 @@
-git repo for art num

bash/README.md

@@ -1 +0,0 @@
-git repo for art num

gravure_upic/README.md

@@ -0,0 +1,4 @@
+
+# Un système UPIC pour un orchestre de gravures
+
+C'est un projet prenant pour source de création des gravures faites avec la technique d'eau forte sur plaque de zinc. Il inclut la transcription visuelle dans le champ sonore fréquentiel dont le mouvement du dessin fixe va altérer le comportement sonore dans le temps.

gravure_upic/python/circular_to_plane/circular_to_plain.py

@@ -0,0 +1,26 @@
+import cv2
+import math
+
+ring = cv2.imread('../circular.png')
+
+size = ring.shape[0]
+outer_radius = size // 2
+inner_radius = 0
+
+unwrapped = cv2.warpPolar(
+    ring,
+    (size, int(size * math.pi)),
+    (outer_radius, outer_radius),
+    outer_radius,
+    flags = 0
+)
+
+unwrapped = cv2.rotate(unwrapped, cv2.ROTATE_90_COUNTERCLOCKWISE)
+
+unwrapped = unwrapped[inner_radius:, :]
+
+cv2.imshow("original", ring)
+cv2.imshow("Unwrapped", unwrapped)
+cv2.waitKey(0)
+cv2.destroyAllWindows()
+cv2.imwrite("output.png", unwrapped)

gravure_upic/python/img_to_freq/img_to_freq.py

@@ -0,0 +1,101 @@
+from tqdm import tqdm
+from scipy.ndimage import uniform_filter1d
+from scipy.io.wavfile import write
+import math, cv2, sys, getopt, wave
+import numpy as np
+
+def unwrap_img(input_file):
+    ring_img = cv2.imread(input_file)
+
+    size = ring_img.shape[0]
+    outer_radius = size // 2
+    inner_radius = 0
+
+    unwrapped_img = cv2.warpPolar(
+        ring_img,
+        (size, int(size * math.pi)),
+        (outer_radius, outer_radius),
+        outer_radius,
+        flags=0
+    )
+
+    rotated_unwrapped = cv2.rotate(unwrapped_img, cv2.ROTATE_90_COUNTERCLOCKWISE)
+    cropped_unwrapped = rotated_unwrapped[inner_radius:, :]
+
+    inverted_unwrapped = cv2.bitwise_not(cropped_unwrapped)
+
+    return inverted_unwrapped
+
+
+def image_to_audio(input_img, out_wav, duration_seconds, sample_rate, vertical_res, amp_threshold):
+    max_freq = 10000
+    min_freq = 50
+
+    duration_seconds = float(duration_seconds)
+
+    # Downsample image vertically to reduce number of frequencies
+    input_img = input_img[::vertical_res]  # Use one row every 20 pixels
+    height, width, _ = input_img.shape
+
+    num_samples = int(sample_rate * duration_seconds)
+    freqs = np.logspace(np.log10(min_freq), np.log10(max_freq), height)[::-1]
+
+    brightness = np.mean(input_img / 255.0, axis=2)
+    amplitudes = np.where(brightness >= 0.1, brightness, 0)
+
+    samples = np.zeros(num_samples, dtype=np.float32)
+    chunk_size = 10000
+
+    for start in tqdm(range(0, num_samples, chunk_size)):
+        end = min(start + chunk_size, num_samples)
+        t = np.linspace(start / sample_rate, end / sample_rate, end - start)
+
+        pixel_xs = (t * width / duration_seconds).astype(int)
+        pixel_xs = np.clip(pixel_xs, 0, width - 1)
+
+        amp_per_sample = amplitudes[:, pixel_xs]
+        amp_per_sample[amp_per_sample < amp_threshold] = 0
+        amp_per_sample = uniform_filter1d(amp_per_sample, size=7, axis=1)  # Smooth
+        
+        phases = 2 * np.pi * freqs[:, None] * t[None, :]
+        active = np.count_nonzero(amp_per_sample, axis=0)
+        chunk = np.sum(amp_per_sample * np.sin(phases), axis=0)
+
+        # Normalize chunk by number of active oscillators to avoid "snow"
+        chunk = np.where(active > 0, chunk / active, 0)
+        samples[start:end] = chunk
+
+    # Final normalization
+    samples /= np.max(np.abs(samples) + 1e-8)
+
+    # Convert to int16 for wav
+    wav_samples = (samples * 32767).astype(np.int16)
+    write(out_wav, sample_rate, wav_samples)
+    print(f"Audio saved to {out_wav}")
+
+
+if __name__ == '__main__':
+    input_file = ''
+    output_file = ''
+    duration = 10
+
+    try:
+        opts, args = getopt.getopt(sys.argv[1:], "hi:o:d:yRes:thresh:")
+    except getopt.GetoptError:
+        print('error: img_to_freq.py -i <input_picture> -o <output_sound> -d <audio_duration>')
+
+    for opt, arg in opts:
+        if opt == '-h':
+            print('error: img_to_freq.py -i <input_picture> -o <output_sound> -d <audio_duration>')
+            sys.exit(2)
+        elif opt == '-i':
+            input_file = arg
+        elif opt == '-o':
+            output_file = arg
+        elif opt == '-d':
+            duration = arg
+
+    unwrapped_img = unwrap_img(input_file)
+    cv2.imwrite('output.png', unwrapped_img)
+
+    image_to_audio(unwrapped_img, output_file, duration, 44100, 20, 0.1)

python/README.md

@@ -1 +0,0 @@
-git repo for art num

references/README.md

@@ -1 +0,0 @@
-git repo for art num

references/python_arti_ref_mara.md

@@ -1,31 +0,0 @@
-# artistic ref : usage de python
-
-## [Computational Poems : Les deux, Nick Montfort](https://nickm.com/2/les_deux.html) 
-
-  - US digital artist / chercheur
-  - générateur de poème online dynamique (javascript)
-  - poème multilangue (fr, esp, cn) => dispositif de traduction (js)
-    
-## [The Great Netfix, *Ritasdatter & Gansing*](http://netflix.lnd4.net/)
-*a video store after the end of the world*
-
-- notion de de-clouding : proposition speculative de redistribution de la “cloub-base” contemporaine
-- activité de **scraping** de film Netflix via VPN (utilitaire de deplacement immatérielle de la localisation du client) & enregistrement **VHS**
-- dispositif rspi (WLAN) - tape recorder VHS
-
-## [Videogrep, *Sam Lavigne* (2014)](https://antiboredom.github.io/videogrep/)
-
-- python script that searches through dialog on videos and combine then in a flesh video
-- e.g : condense toute les itération d’une expression d’une video originale
-
-- visibilisation de normalisation d’usage de stratégie marketing (element de langage) dans contexte politique -partisant-
-  
-- commande ligne tool / python module en libre acces sur archive github du project
-  ```videogrep -- input path/to/vid.mp4 --search 'search phrase'```
-
-## [Unerasable Characters, *Winnie Soon*](https://calls.ars.electronica.art/2023/prix/winners/7149/)
-Prix Ars Electronica, 2023
-
-- scraping data censurées/suprimées from Weibo (chinese social media == twitter)
-- dispersion des ideogram dans matrice lumineuse physique
-- concatenation de l’ensemble des caractère par machine learning (Tensor Flow) pour republication sur source (Weibo) et production d’une édition physique

upython/README.md

@@ -1 +0,0 @@
-git repo for art num