import cv2 import numpy as np from PIL import Image, ImageDraw, ImageFont import os from tqdm import tqdm import time import multiprocessing as mp # 画像を作成するために使用されるASCII文字 ASCII_CHARS_BW = "@%#*+=-:. " ASCII_CHARS_COLOR = "@%#S&8BWM#oahkbdpqwmZO0QLCJUYXzcvunxrjft/|()1{}[]?-_+~<>i!lI;:,\"^`'. " # 各ピクセルをASCII文字に変換 def pixel_to_ascii(pixel_value, ascii_chars): return ascii_chars[int(pixel_value / 256 * len(ascii_chars))] # 画像をASCIIに変換 def image_to_ascii(image, num_cols, char_width, char_height, full_color): if full_color: resized_image = cv2.resize(image, (num_cols, int(num_cols * image.shape[0] / image.shape[1] * char_height / char_width))) ascii_image = [] for row in resized_image: ascii_row = "".join([pixel_to_ascii(pixel[0], ASCII_CHARS_COLOR) for pixel in row]) # Using the first channel for ASCII conversion ascii_image.append(ascii_row) return ascii_image, resized_image else: gray_image = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY) height, width = gray_image.shape aspect_ratio = height / float(width) num_rows = int(num_cols * aspect_ratio * (char_height / char_width)) resized_gray_image = cv2.resize(gray_image, (num_cols, num_rows)) ascii_image = [] for row in resized_gray_image: ascii_row = "".join([pixel_to_ascii(pixel, ASCII_CHARS_BW) for pixel in row]) ascii_image.append(ascii_row) return ascii_image, None # ビデオライターオブジェクトを作成 def create_video_writer(output_path, frame_size, fps): fourcc = cv2.VideoWriter_fourcc(*'mp4v') return cv2.VideoWriter(output_path, fourcc, fps, frame_size) def process_frame(frame, num_cols, width, height, font_path, font_size, full_color, bg_color=None, txt_color=None): font = ImageFont.truetype(font_path, font_size) char_width, char_height = font.getbbox('A')[2:4] ascii_frame, resized_image = image_to_ascii(frame, num_cols, char_width, char_height, full_color) if full_color: if bg_color is not None: img_pil = Image.new('RGB', (num_cols * char_width, len(ascii_frame) * char_height), color=(bg_color, bg_color, bg_color)) else: img_pil = Image.new('RGB', (num_cols * char_width, len(ascii_frame) * char_height)) draw = ImageDraw.Draw(img_pil) y_text = 0 for i, line in enumerate(ascii_frame): x_text = 0 for j, char in enumerate(line): color = tuple(resized_image[i, j][:3]) # Take RGB values from the resized image draw.text((x_text, y_text), char, font=font, fill=color) x_text += char_width y_text += char_height else: img_pil = Image.new('L', (num_cols * char_width, len(ascii_frame) * char_height), color=bg_color) draw = ImageDraw.Draw(img_pil) y_text = 0 for line in ascii_frame: draw.text((0, y_text), line, font=font, fill=txt_color) y_text += char_height img_np = np.array(img_pil) if full_color: img_bgr = cv2.cvtColor(img_np, cv2.COLOR_RGB2BGR) else: img_bgr = cv2.cvtColor(img_np, cv2.COLOR_GRAY2BGR) img_bgr = cv2.resize(img_bgr, (width, height)) return img_bgr def main(): # フェーズ1: Processing - ビデオファイルを読み込み video_path = '/video/path/is/here' cap = cv2.VideoCapture(video_path) # ビデオが正常に開かれたか確認 if not cap.isOpened(): print("エラー: ビデオを開くことができませんでした。") exit() # ビデオのプロパティを取得 fps = cap.get(cv2.CAP_PROP_FPS) width = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH)) height = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT)) total_frames = int(cap.get(cv2.CAP_PROP_FRAME_COUNT)) # ユーザーに背景色とテキスト色を入力してもらう background_color = input("背景色を入力してください (b/w/c): ").strip().lower() if background_color == "b" or background_color == "ｂ": text_color = "white" bg_color = 0 txt_color = 255 suffix = "（ASCII・Black）" full_color = False elif background_color == "w" or background_color == "ｗ": text_color = "black" bg_color = 255 txt_color = 0 suffix = "（ASCII・White）" full_color = False elif background_color == "c" or background_color == "ｃ": color_mode = input("カラーの背景色を選択してください (b/w): ").strip().lower() if color_mode == "b" or color_mode == "ｂ": text_color = "white" bg_color = 0 txt_color = None suffix = "（ASCII・FullColor・Black）" full_color = True elif color_mode == "w" or color_mode == "ｗ": text_color = "black" bg_color = 255 txt_color = None suffix = "（ASCII・FullColor・White）" full_color = True else: print("無効な入力です。b, wのいずれかを入力してください。") exit() else: print("無効な入力です。b, w, cのいずれかを入力してください。") exit() # ビデオライターを作成 output_path = video_path.rsplit('.', 1)[0] + suffix + '.mp4' out = create_video_writer(output_path, (width, height), fps) # フォント設定 font_path = "/usr/share/fonts/truetype/dejavu/DejaVuSansMono-Bold.ttf" font_size = 10 font = ImageFont.truetype(font_path, font_size) char_width, char_height = font.getbbox('A')[2:4] num_cols = width // char_width # スペースを考慮して列数を調整 # 使用するワーカー数を設定 (総CPU数 - 2) num_workers = mp.cpu_count() - 2 pool = mp.Pool(num_workers) # フェーズ2: Collecting - プログレスバー付きでフレームを読み取り処理 start_time = time.time() results = [] with tqdm(total=total_frames, desc="Processing", unit="frame") as pbar: for _ in range(total_frames): ret, frame = cap.read() if not ret: break if full_color: result = pool.apply_async(process_frame, args=(frame, num_cols, width, height, font_path, font_size, full_color, bg_color, txt_color)) else: result = pool.apply_async(process_frame, args=(frame, num_cols, width, height, font_path, font_size, full_color, bg_color, txt_color)) results.append(result) pbar.update(1) # フェーズ3: Writing - 処理済みフレームをビデオに書き込み for result in tqdm(results, desc="Writing", unit="frame"): out.write(result.get()) # ビデオキャプチャとライターオブジェクトを解放 cap.release() out.release() pool.close() pool.join() if __name__ == "__main__": main()