Introduction

# From src/visual_sync.py:39-94
def extract_motion_energy(video_path: str, 
                          downsample: int = 4,
                          blur_size: int = 5,
                          center_crop: bool = True,
                          step: int = 3) -> Tuple[np.ndarray, float]:
    cap = cv2.VideoCapture(video_path)
    fps = cap.get(cv2.CAP_PROP_FPS)
    
    while True:
        ret, frame = cap.read()
        if not ret:
            break
            
        # Center crop and downsample for speed
        if center_crop:
            h, w = frame.shape[:2]
            frame = frame[int(h*0.25):int(h*0.75), int(w*0.25):int(w*0.75)]
        
        gray = cv2.cvtColor(small_frame, cv2.COLOR_BGR2GRAY)
        
        if prev_gray is not None:
            diff = cv2.absdiff(gray, prev_gray)
            energy = np.sum(thresh) / (thresh.shape[0] * thresh.shape[1] * 255)
            motion_energy.append(energy)

# From src/visual_sync.py:105-133
def correlate_motion_signals(sig1: np.ndarray, sig2: np.ndarray,
                             fps: float,
                             max_offset_sec: float = 20.0) -> Tuple[float, float]:
    # Normalize signals
    sig1_norm = (sig1 - np.mean(sig1)) / (np.std(sig1) + 1e-10)
    sig2_norm = (sig2 - np.mean(sig2)) / (np.std(sig2) + 1e-10)
    
    cc = correlate(sig1_norm, sig2_norm, mode='full')
    
    # Find peak within search window
    lag_frames = lag_idx - center
    offset_seconds = lag_frames / fps
    
    # Confidence scoring
    confidence = (peak - mean_cc) / (std_cc + 1e-10)
    return offset_seconds, confidence