import subprocess import matplotlib.pyplot as plt import os import json import tkinter as tk from tkinter import filedialog import statistics # ===== CONFIGURATION ===== DROP_THRESHOLD_MULTIPLIER = 1.5 # Threshold multiplier for detecting dropped frames # TOLERANCE will be set automatically based on FPS # ========================== def select_file(): root = tk.Tk() root.withdraw() file_path = filedialog.askopenfilename( title="Select a video file", filetypes=[("Video files", "*.mp4 *.mkv *.avi *.mov *.flv *.wmv *.webm"), ("All files", "*.*")] ) return file_path def ffprobe_timestamps(video_file): result = subprocess.run( [ "ffprobe", "-v", "error", "-select_streams", "v:0", "-show_entries", "frame=pts_time", "-of", "csv=p=0", video_file ], stdout=subprocess.PIPE, stderr=subprocess.PIPE, text=True ) if result.returncode != 0: print("❌ ffprobe failed:", result.stderr) return [] return [float(line.strip()) for line in result.stdout.strip().splitlines() if line.strip()] def ffprobe_nominal(video_file): result = subprocess.run( [ "ffprobe", "-v", "error", "-select_streams", "v:0", "-show_entries", "stream=avg_frame_rate,r_frame_rate,nb_frames", "-show_entries", "format=duration", "-of", "json", video_file ], stdout=subprocess.PIPE, stderr=subprocess.PIPE, text=True ) if result.returncode != 0: print("❌ ffprobe (fps) failed:", result.stderr) return None, None, {} data = json.loads(result.stdout) stream = data["streams"][0] duration = float(data["format"]["duration"]) if "duration" in data["format"] else None def fps_from_fraction(frac): if frac and frac != "0/0": num, den = frac.split("/") return float(num) / float(den) return None candidates = { "avg_frame_rate": fps_from_fraction(stream.get("avg_frame_rate")), "r_frame_rate": fps_from_fraction(stream.get("r_frame_rate")), "nb_frames": int(stream["nb_frames"]) if stream.get("nb_frames") else None, "duration": duration, } chosen = None if candidates["avg_frame_rate"]: chosen = candidates["avg_frame_rate"] elif candidates["r_frame_rate"]: chosen = candidates["r_frame_rate"] elif candidates["nb_frames"] and candidates["duration"]: chosen = candidates["nb_frames"] / candidates["duration"] return chosen, (1.0 / chosen if chosen else None), candidates def auto_tolerance(fps): """Automatically set TOLERANCE based on FPS.""" if fps <= 30: return 1e-5 elif fps <= 60: return 5e-6 elif fps <= 120: return 1e-6 else: return 1e-7 def analyze(pts_list, expected_interval, TOLERANCE): if expected_interval is None: print("❌ Cannot analyze intervals without expected interval.") return None intervals = [pts_list[i+1] - pts_list[i] for i in range(len(pts_list) - 1)] drop_threshold = expected_interval * DROP_THRESHOLD_MULTIPLIER early_threshold = expected_interval * 0.5 paused = [(i, pts_list[i]) for i in range(len(intervals)) if abs(intervals[i]) < TOLERANCE] dropped = [(i, intervals[i]) for i in range(len(intervals)) if intervals[i] > drop_threshold] early = [(i, intervals[i]) for i in range(len(intervals)) if intervals[i] < early_threshold and intervals[i] > 0] out_of_order = [(i, pts_list[i], pts_list[i+1]) for i in range(len(pts_list) - 1) if pts_list[i+1] < pts_list[i]] valid_intervals = [iv for iv in intervals if iv > 0] observed_median = statistics.median(valid_intervals) if valid_intervals else None min_interval = min(valid_intervals) if valid_intervals else None max_interval = max(valid_intervals) if valid_intervals else None return { "intervals": intervals, "paused": paused, "dropped": dropped, "early": early, "out_of_order": out_of_order, "observed_median": observed_median, "min_interval": min_interval, "max_interval": max_interval } def plot_intervals(intervals, expected_interval, drop_threshold): plt.figure(figsize=(12, 5)) plt.plot(intervals, label='Frame Interval (s)', marker='o', linestyle='-') plt.axhline(expected_interval, color='green', linestyle='--', label='Expected Interval') plt.axhline(drop_threshold, color='red', linestyle='--', label='Drop Threshold') plt.axhline(expected_interval * 0.5, color='blue', linestyle='--', label='Early Threshold') plt.xlabel("Frame Number") plt.ylabel("Interval (s)") plt.title("Frame Interval Analysis") plt.legend() plt.grid(True) plt.tight_layout() plt.show() def main(): video_file = select_file() if not video_file or not os.path.isfile(video_file): print("❌ No valid file selected.") return print(f"[✓] Analyzing file: {video_file}") pts_list = ffprobe_timestamps(video_file) if not pts_list: print("❌ No timestamps extracted.") return fps, expected_interval, candidates = ffprobe_nominal(video_file) if not fps: print("❌ Could not determine nominal FPS.") return TOLERANCE = auto_tolerance(fps) print(f"[✓] Auto TOLERANCE set to {TOLERANCE} based on FPS {fps:.2f}") print("\n📊 ffprobe FPS candidates:") for k, v in candidates.items(): print(f" {k}: {v}") print(f"Chosen nominal FPS: {fps:.6f}, interval: {expected_interval:.9f} s") result = analyze(pts_list, expected_interval, TOLERANCE) if not result: return total_intervals = len(result['intervals']) paused_count = len(result['paused']) dropped_count = len(result['dropped']) early_count = len(result['early']) out_of_order_count = len(result['out_of_order']) print("\n📋 SUMMARY") print(f"Total frames: {len(pts_list)}") print(f"Expected interval (chosen): {expected_interval:.9f} s") print(f"Observed median interval: {result['observed_median']:.9f} s") print(f"Paused (duplicate PTS) frames: {paused_count} ({paused_count/total_intervals*100:.2f}%)") print(f"Dropped/irregular frames: {dropped_count} ({dropped_count/total_intervals*100:.2f}%)") print(f"Early (fast) frames: {early_count} ({early_count/total_intervals*100:.2f}%)") print(f"Out-of-order PTS frames: {out_of_order_count} ({out_of_order_count/total_intervals*100:.2f}%)") print(f"Min interval: {result['min_interval']:.9f} s, Max interval: {result['max_interval']:.9f} s") if result["early"]: print("\n⚡ Early frames (first 10):") for idx, gap in result["early"][:10]: print(f" Frame {idx} → {idx+1} gap: {gap:.9f} s") if len(result["early"]) > 10: print(" ...") plot_intervals(result["intervals"], expected_interval, expected_interval * DROP_THRESHOLD_MULTIPLIER) # Export JSON export = { "file": video_file, "fps_candidates": candidates, "fps_chosen": fps, "expected_interval": expected_interval, "observed_median_interval": result["observed_median"], "paused_frames": {"count": paused_count, "percentage": paused_count/total_intervals*100, "frames": result["paused"]}, "dropped_frames": {"count": dropped_count, "percentage": dropped_count/total_intervals*100, "frames": result["dropped"]}, "early_frames": {"count": early_count, "percentage": early_count/total_intervals*100, "frames": result["early"]}, "out_of_order_frames": {"count": out_of_order_count, "percentage": out_of_order_count/total_intervals*100, "frames": result["out_of_order"]}, "min_interval": result["min_interval"], "max_interval": result["max_interval"], "total_intervals": total_intervals } out_json = os.path.splitext(video_file)[0] + "_frame_analysis.json" with open(out_json, "w") as f: json.dump(export, f, indent=2) print(f"\n[✓] Exported analysis to {out_json}") if __name__ == "__main__": main()