433 lines
17 KiB
Python
433 lines
17 KiB
Python
#!/usr/bin/env python3
|
|
|
|
import argparse
|
|
import subprocess
|
|
import sys
|
|
import os
|
|
import re
|
|
from collections import Counter
|
|
import shutil
|
|
import multiprocessing
|
|
import json
|
|
|
|
# ANSI color codes
|
|
COLOR_GREEN = "\033[92m"
|
|
COLOR_RED = "\033[91m"
|
|
COLOR_YELLOW = "\033[93m"
|
|
COLOR_RESET = "\033[0m"
|
|
|
|
def check_prerequisites():
|
|
"""Checks if required tools are available."""
|
|
print("--- Prerequisite Check ---")
|
|
all_found = True
|
|
for tool in ['ffmpeg', 'ffprobe']:
|
|
if not shutil.which(tool):
|
|
print(f"Error: '{tool}' command not found. Is it installed and in your PATH?")
|
|
all_found = False
|
|
if not all_found:
|
|
sys.exit(1)
|
|
print("All required tools found.")
|
|
|
|
def analyze_segment(task_args):
|
|
"""Function to be run by each worker process. Analyzes one video segment."""
|
|
seek_time, input_file, width, height = task_args
|
|
|
|
ffmpeg_args = [
|
|
'ffmpeg', '-hide_banner',
|
|
'-ss', str(seek_time),
|
|
'-i', input_file, '-t', '1', '-vf', 'cropdetect',
|
|
'-f', 'null', '-'
|
|
]
|
|
|
|
result = subprocess.run(ffmpeg_args, capture_output=True, text=True, encoding='utf-8')
|
|
|
|
if result.returncode != 0:
|
|
return [] # Return empty list on error
|
|
|
|
crop_detections = re.findall(r'crop=(\d+):(\d+):(\d+):(\d+)', result.stderr)
|
|
|
|
significant_crops = []
|
|
for w_str, h_str, x_str, y_str in crop_detections:
|
|
w, h, x, y = map(int, [w_str, h_str, x_str, y_str])
|
|
|
|
# Return the crop string along with the timestamp it was found at
|
|
significant_crops.append((f"crop={w}:{h}:{x}:{y}", seek_time))
|
|
|
|
return significant_crops
|
|
|
|
def get_frame_luma(input_file, seek_time):
|
|
"""Analyzes a single frame at a given timestamp to get its average luma."""
|
|
ffmpeg_args = [
|
|
'ffmpeg', '-hide_banner',
|
|
'-ss', str(seek_time),
|
|
'-i', input_file,
|
|
'-t', '1',
|
|
'-vf', 'signalstats',
|
|
'-f', 'null', '-'
|
|
]
|
|
result = subprocess.run(ffmpeg_args, capture_output=True, text=True, encoding='utf-8')
|
|
|
|
if result.returncode != 0:
|
|
return None # Error during analysis
|
|
|
|
# Find the average luma (YAVG) for the frame
|
|
match = re.search(r'YAVG:([0-9.]+)', result.stderr)
|
|
if match:
|
|
return float(match.group(1))
|
|
|
|
return None
|
|
|
|
def check_luma_for_group(task_args):
|
|
"""Worker function to check the luma for a single group."""
|
|
group_key, sample_ts, input_file, luma_threshold = task_args
|
|
luma = get_frame_luma(input_file, sample_ts)
|
|
is_bright = luma is not None and luma >= luma_threshold
|
|
return (group_key, is_bright)
|
|
|
|
KNOWN_ASPECT_RATIOS = [
|
|
{"name": "HDTV (16:9)", "ratio": 16/9},
|
|
{"name": "Widescreen (Scope)", "ratio": 2.39},
|
|
{"name": "Widescreen (Flat)", "ratio": 1.85},
|
|
{"name": "IMAX Digital (1.90:1)", "ratio": 1.90},
|
|
{"name": "Fullscreen (4:3)", "ratio": 4/3},
|
|
{"name": "IMAX 70mm (1.43:1)", "ratio": 1.43},
|
|
]
|
|
|
|
def snap_to_known_ar(w, h, x, y, video_w, video_h, tolerance=0.03):
|
|
"""Snaps a crop rectangle to the nearest standard aspect ratio if it's close enough."""
|
|
if h == 0: return f"crop={w}:{h}:{x}:{y}", None
|
|
detected_ratio = w / h
|
|
|
|
best_match = None
|
|
smallest_diff = float('inf')
|
|
|
|
for ar in KNOWN_ASPECT_RATIOS:
|
|
diff = abs(detected_ratio - ar['ratio'])
|
|
if diff < smallest_diff:
|
|
smallest_diff = diff
|
|
best_match = ar
|
|
|
|
# If the best match is not within the tolerance, return the original
|
|
if not best_match or (smallest_diff / best_match['ratio']) >= tolerance:
|
|
return f"crop={w}:{h}:{x}:{y}", None
|
|
|
|
# Match found, now snap the dimensions.
|
|
# Heuristic: if width is close to full video width, it's letterboxed.
|
|
if abs(w - video_w) < 16:
|
|
new_h = round(video_w / best_match['ratio'])
|
|
|
|
# Round height up to the nearest multiple of 8 for cleaner dimensions and less aggressive cropping.
|
|
if new_h % 8 != 0:
|
|
new_h = new_h + (8 - (new_h % 8))
|
|
|
|
new_y = round((video_h - new_h) / 2)
|
|
# Ensure y offset is an even number for compatibility.
|
|
if new_y % 2 != 0:
|
|
new_y -= 1
|
|
|
|
return f"crop={video_w}:{new_h}:0:{new_y}", best_match['name']
|
|
|
|
# Heuristic: if height is close to full video height, it's pillarboxed.
|
|
if abs(h - video_h) < 16:
|
|
new_w = round(video_h * best_match['ratio'])
|
|
|
|
# Round width up to the nearest multiple of 8.
|
|
if new_w % 8 != 0:
|
|
new_w = new_w + (8 - (new_w % 8))
|
|
|
|
new_x = round((video_w - new_w) / 2)
|
|
# Ensure x offset is an even number.
|
|
if new_x % 2 != 0:
|
|
new_x -= 1
|
|
|
|
return f"crop={new_w}:{video_h}:{new_x}:0", best_match['name']
|
|
|
|
# If not clearly letterboxed or pillarboxed, don't snap.
|
|
return f"crop={w}:{h}:{x}:{y}", None
|
|
|
|
def cluster_crop_values(crop_counts, tolerance=8):
|
|
"""Groups similar crop values into clusters based on the top-left corner."""
|
|
clusters = []
|
|
temp_counts = crop_counts.copy()
|
|
|
|
while temp_counts:
|
|
# Get the most frequent remaining crop as the new cluster center
|
|
center_str, _ = temp_counts.most_common(1)[0]
|
|
|
|
try:
|
|
_, values = center_str.split('=')
|
|
cw, ch, cx, cy = map(int, values.split(':'))
|
|
except (ValueError, IndexError):
|
|
del temp_counts[center_str] # Skip malformed strings
|
|
continue
|
|
|
|
cluster_total_count = 0
|
|
crops_to_remove = []
|
|
|
|
# Find all crops "close" to the center
|
|
for crop_str, count in temp_counts.items():
|
|
try:
|
|
_, values = crop_str.split('=')
|
|
w, h, x, y = map(int, values.split(':'))
|
|
if abs(x - cx) <= tolerance and abs(y - cy) <= tolerance:
|
|
cluster_total_count += count
|
|
crops_to_remove.append(crop_str)
|
|
except (ValueError, IndexError):
|
|
continue
|
|
|
|
if cluster_total_count > 0:
|
|
clusters.append({'center': center_str, 'count': cluster_total_count})
|
|
|
|
# Remove the clustered crops from the temporary counter
|
|
for crop_str in crops_to_remove:
|
|
del temp_counts[crop_str]
|
|
|
|
clusters.sort(key=lambda c: c['count'], reverse=True)
|
|
return clusters
|
|
|
|
def parse_crop_string(crop_str):
|
|
"""Parses a 'crop=w:h:x:y' string into a dictionary of integers."""
|
|
try:
|
|
_, values = crop_str.split('=')
|
|
w, h, x, y = map(int, values.split(':'))
|
|
return {'w': w, 'h': h, 'x': x, 'y': y}
|
|
except (ValueError, IndexError):
|
|
return None
|
|
|
|
def calculate_bounding_box(crop_keys):
|
|
"""Calculates a bounding box that contains all given crop rectangles."""
|
|
min_x = min_w = min_y = min_h = float('inf')
|
|
max_x = max_w = max_y = max_h = float('-inf')
|
|
|
|
for key in crop_keys:
|
|
parsed = parse_crop_string(key)
|
|
if not parsed:
|
|
continue
|
|
|
|
w, h, x, y = parsed['w'], parsed['h'], parsed['x'], parsed['y']
|
|
|
|
min_x = min(min_x, x)
|
|
min_y = min(min_y, y)
|
|
max_x = max(max_x, x + w)
|
|
max_y = max(max_y, y + h)
|
|
|
|
min_w = min(min_w, w)
|
|
min_h = min(min_h, h)
|
|
max_w = max(max_w, w)
|
|
max_h = max(max_h, h)
|
|
|
|
# Heuristic: if the bounding box is very close to the min/max, it means all crops were similar
|
|
if (max_x - min_x) <= 2 and (max_y - min_y) <= 2:
|
|
return None # Too uniform, don't create a bounding box
|
|
|
|
# Create a crop that spans the entire bounding box
|
|
bounding_crop = f"crop={max_x - min_x}:{max_y - min_y}:{min_x}:{min_y}"
|
|
|
|
return bounding_crop
|
|
|
|
def is_major_crop(crop_str, video_w, video_h, min_crop_size):
|
|
"""Checks if a crop is significant enough to be recommended by checking if any side is cropped by at least min_crop_size pixels."""
|
|
parsed = parse_crop_string(crop_str)
|
|
if not parsed:
|
|
return False
|
|
|
|
w, h, x, y = parsed['w'], parsed['h'], parsed['x'], parsed['y']
|
|
|
|
# Calculate how much is cropped from each side
|
|
crop_top = y
|
|
crop_bottom = video_h - (y + h)
|
|
crop_left = x
|
|
crop_right = video_w - (x + w)
|
|
|
|
# Return True if the largest crop on any single side meets the threshold
|
|
if max(crop_top, crop_bottom, crop_left, crop_right) >= min_crop_size:
|
|
return True
|
|
|
|
return False
|
|
|
|
def analyze_video(input_file, duration, width, height, num_workers, significant_crop_threshold, min_crop, debug=False):
|
|
"""Main analysis function for the video."""
|
|
print(f"\n--- Analyzing Video: {os.path.basename(input_file)} ---")
|
|
|
|
# Step 1: Analyze video in segments to detect crops
|
|
num_tasks = num_workers * 4
|
|
segment_duration = max(1, duration // num_tasks)
|
|
tasks = [(i * segment_duration, input_file, width, height) for i in range(num_tasks)]
|
|
|
|
print(f"Analyzing {len(tasks)} segments across {num_workers} worker(s)...")
|
|
|
|
crop_results = []
|
|
with multiprocessing.Pool(processes=num_workers) as pool:
|
|
total_tasks = len(tasks)
|
|
results_iterator = pool.imap_unordered(analyze_segment, tasks)
|
|
|
|
for i, result in enumerate(results_iterator, 1):
|
|
crop_results.append(result)
|
|
progress_message = f"Analyzing Segments: {i}/{total_tasks} completed..."
|
|
sys.stdout.write(f"\r{progress_message}")
|
|
sys.stdout.flush()
|
|
print()
|
|
|
|
all_crops_with_ts = [crop for sublist in crop_results for crop in sublist]
|
|
all_crop_strings = [item[0] for item in all_crops_with_ts]
|
|
if not all_crop_strings:
|
|
print(f"\n{COLOR_GREEN}Analysis complete. No black bars detected.{COLOR_RESET}")
|
|
return
|
|
|
|
crop_counts = Counter(all_crop_strings)
|
|
|
|
if debug:
|
|
print("\n--- Debug: Most Common Raw Detections ---")
|
|
for crop_str, count in crop_counts.most_common(10):
|
|
print(f" - {crop_str} (Count: {count})")
|
|
|
|
# Step 2: Cluster similar crop values
|
|
clusters = cluster_crop_values(crop_counts)
|
|
total_detections = sum(c['count'] for c in clusters)
|
|
|
|
if debug:
|
|
print("\n--- Debug: Detected Clusters ---")
|
|
for cluster in clusters:
|
|
percentage = (cluster['count'] / total_detections) * 100
|
|
print(f" - Center: {cluster['center']}, Count: {cluster['count']} ({percentage:.1f}%)")
|
|
|
|
# Step 3: Filter clusters that are below the significance threshold
|
|
significant_clusters = []
|
|
for cluster in clusters:
|
|
percentage = (cluster['count'] / total_detections) * 100
|
|
if percentage >= significant_crop_threshold:
|
|
significant_clusters.append(cluster)
|
|
|
|
# Step 4: Determine final recommendation based on significant clusters
|
|
print("\n--- Determining Final Crop Recommendation ---")
|
|
|
|
for cluster in significant_clusters:
|
|
parsed_crop = parse_crop_string(cluster['center'])
|
|
if parsed_crop:
|
|
_, ar_label = snap_to_known_ar(
|
|
parsed_crop['w'], parsed_crop['h'], parsed_crop['x'], parsed_crop['y'], width, height
|
|
)
|
|
cluster['ar_label'] = ar_label
|
|
else:
|
|
cluster['ar_label'] = None
|
|
|
|
if not significant_clusters:
|
|
print(f"{COLOR_RED}No single crop value meets the {significant_crop_threshold}% significance threshold.{COLOR_RESET}")
|
|
print("Recommendation: Do not crop. Try lowering the -sct threshold.")
|
|
|
|
elif len(significant_clusters) == 1:
|
|
dominant_cluster = significant_clusters[0]
|
|
parsed_crop = parse_crop_string(dominant_cluster['center'])
|
|
snapped_crop, ar_label = snap_to_known_ar(
|
|
parsed_crop['w'], parsed_crop['h'], parsed_crop['x'], parsed_crop['y'], width, height
|
|
)
|
|
|
|
print("A single dominant aspect ratio was found.")
|
|
if ar_label:
|
|
print(f"The detected crop snaps to the '{ar_label}' aspect ratio.")
|
|
|
|
# Check if the final crop is a no-op (i.e., matches source dimensions)
|
|
parsed_snapped = parse_crop_string(snapped_crop)
|
|
if parsed_snapped and parsed_snapped['w'] == width and parsed_snapped['h'] == height:
|
|
print(f"\n{COLOR_GREEN}The detected crop matches the source resolution. No crop is needed.{COLOR_RESET}")
|
|
else:
|
|
print(f"\n{COLOR_GREEN}Recommended crop filter: -vf {snapped_crop}{COLOR_RESET}")
|
|
|
|
else: # len > 1, mixed AR case
|
|
print(f"{COLOR_YELLOW}Mixed aspect ratios detected (e.g., IMAX scenes).{COLOR_RESET}")
|
|
print("Calculating a safe 'master' crop to contain all significant scenes.")
|
|
|
|
crop_keys = [c['center'] for c in significant_clusters]
|
|
bounding_box_crop = calculate_bounding_box(crop_keys)
|
|
|
|
if bounding_box_crop:
|
|
parsed_bb = parse_crop_string(bounding_box_crop)
|
|
snapped_crop, ar_label = snap_to_known_ar(
|
|
parsed_bb['w'], parsed_bb['h'], parsed_bb['x'], parsed_bb['y'], width, height
|
|
)
|
|
|
|
print("\n--- Detected Significant Ratios ---")
|
|
for cluster in significant_clusters:
|
|
percentage = (cluster['count'] / total_detections) * 100
|
|
label = f"'{cluster['ar_label']}'" if cluster['ar_label'] else "Custom AR"
|
|
print(f" - {label} ({cluster['center']}) was found in {percentage:.1f}% of samples.")
|
|
|
|
print(f"\n{COLOR_GREEN}Analysis complete.{COLOR_RESET}")
|
|
if ar_label:
|
|
print(f"The calculated master crop snaps to the '{ar_label}' aspect ratio.")
|
|
|
|
# Check if the final crop is a no-op
|
|
parsed_snapped = parse_crop_string(snapped_crop)
|
|
if parsed_snapped and parsed_snapped['w'] == width and parsed_snapped['h'] == height:
|
|
print(f"{COLOR_GREEN}The final calculated crop matches the source resolution. No crop is needed.{COLOR_RESET}")
|
|
else:
|
|
print(f"{COLOR_GREEN}Recommended safe crop filter: -vf {snapped_crop}{COLOR_RESET}")
|
|
else:
|
|
print(f"{COLOR_RED}Could not calculate a bounding box. Manual review is required.{COLOR_RESET}")
|
|
|
|
def main():
|
|
parser = argparse.ArgumentParser(
|
|
description="Analyzes a video file to detect black bars and recommend crop values. "
|
|
"Handles mixed aspect ratios by calculating a safe bounding box.",
|
|
formatter_class=argparse.RawTextHelpFormatter
|
|
)
|
|
parser.add_argument("input", help="Input video file")
|
|
parser.add_argument("-n", "--num_workers", type=int, default=max(1, multiprocessing.cpu_count() // 2), help="Number of worker threads. Defaults to half of available cores.")
|
|
parser.add_argument("-sct", "--significant_crop_threshold", type=float, default=5.0, help="Percentage a crop must be present to be considered 'significant'. Default is 5.0.")
|
|
parser.add_argument("-mc", "--min_crop", type=int, default=10, help="Minimum pixels to crop on any side for it to be considered a 'major' crop. Default is 10.")
|
|
parser.add_argument("--debug", action="store_true", help="Enable detailed debug logging.")
|
|
|
|
args = parser.parse_args()
|
|
|
|
input_file = args.input
|
|
num_workers = args.num_workers
|
|
significant_crop_threshold = args.significant_crop_threshold
|
|
min_crop = args.min_crop
|
|
|
|
# Validate input file
|
|
if not os.path.isfile(input_file):
|
|
print(f"{COLOR_RED}Error: Input file does not exist.{COLOR_RESET}")
|
|
sys.exit(1)
|
|
|
|
# Always probe the video file for metadata
|
|
print("--- Probing video file for metadata ---")
|
|
|
|
try:
|
|
probe_duration_args = [
|
|
'ffprobe', '-v', 'error', '-show_entries', 'format=duration', '-of', 'default=noprint_wrappers=1:nokey=1',
|
|
input_file
|
|
]
|
|
duration_str = subprocess.check_output(probe_duration_args, stderr=subprocess.STDOUT, text=True)
|
|
duration = int(float(duration_str))
|
|
print(f"Detected duration: {duration}s")
|
|
|
|
probe_res_args = [
|
|
'ffprobe', '-v', 'error', '-select_streams', 'v:0', '-show_entries', 'stream=width,height', '-of', 'csv=s=x:p=0',
|
|
input_file
|
|
]
|
|
resolution_str = subprocess.check_output(probe_res_args, stderr=subprocess.STDOUT, text=True)
|
|
width, height = map(int, resolution_str.strip().split('x'))
|
|
print(f"Detected resolution: {width}x{height}")
|
|
|
|
except Exception as e:
|
|
print(f"{COLOR_RED}Error probing video file: {e}{COLOR_RESET}")
|
|
sys.exit(1)
|
|
|
|
print(f"\n--- Video Analysis Parameters ---")
|
|
print(f"Input File: {os.path.basename(input_file)}")
|
|
print(f"Duration: {duration}s")
|
|
print(f"Resolution: {width}x{height}")
|
|
print(f"Number of Workers: {num_workers}")
|
|
print(f"Significance Threshold: {significant_crop_threshold}%")
|
|
print(f"Minimum Crop Size: {min_crop}px")
|
|
|
|
# Check for required tools
|
|
check_prerequisites()
|
|
|
|
# Analyze the video
|
|
analyze_video(input_file, duration, width, height, num_workers, significant_crop_threshold, min_crop, args.debug)
|
|
|
|
if __name__ == "__main__":
|
|
main()
|
|
|