Guide Video Duplicate Cleaner

📁 'Folder Scan' Tab

This is the main tab for finding similar video files in selected folders.

📍 Step 1: Selecting Folders

The "Select folder for scanning" button – adds a folder to the list. You can add several folders – the program will scan all videos inside them.

Managing the folder list (the "📋 Manage folder list" button) – allows you to remove selected folders from the list (multi‑selection).

Clear the entire list – the "🗑️ Clear scanning list" button.

📝 Important: The program automatically checks folder nesting. If you add a folder that is already inside another selected folder, a warning will appear – there will be no duplicate scanning.

🚫 Folder Ignore List

If you do not want to scan certain folders (e.g., system folders or backup folders):

  1. Click "🚫 Add folder to ignore list".
  2. Select the folder to be excluded from scanning.
  3. Manage the list via "📋 Manage ignore list" (add, remove, clear).

For more details about the ignore list, see also the "File Management" section.

⚙️ Step 2: Parameter Configuration

The main scanning parameters are configured using three elements on the panel (from left to right):

  1. Algorithm selection – a drop-down list with pHash and CNN+Faiss.
  2. "Scan Settings" button – opens a dialog that contains all parameters for the selected algorithm:
    • General similarity threshold (0.1–1.0).
    • Crop black borders – automatically removes dark edges (increases accuracy but may slow down processing).
    • For pHash: number of frames, neighbourhood size, adaptive neighbourhood, parallel processing.
    • For CNN+Faiss: number of frames, model selection, neighbourhood size, adaptive neighbourhood, parallel processing (with a load warning).
  3. "Pre-filter" button (⚙️) – opens a dialog for advanced file and pair filtering to speed up scanning. See the dedicated "Extended Pre-filtering" section for details.

The indicator to the right of the buttons shows whether filters are active (yellow ⚠️) or not (🔍).

📌 About Parallel Processing

In the scan settings dialog, there is a checkbox "Use multiple CPU cores" (separate for pHash and for CNN). It enables parallel frame processing to speed up analysis.

  • Enable parallel processing when:
    • ✅ Many videos (>20 medium-sized files)
    • ✅ Videos on SSD (handles parallel reads well)
    • ✅ Time is critical (need quick results)
    • ✅ Powerful computer (good cooling)
  • Disable parallel processing when:
    • ✅ Few videos (fewer than 20 medium-sized files)
    • ✅ Videos on HDD (noisy drive)
    • ✅ Working at night (don't want noise)
    • ✅ Laptop on battery (power saving)
  • Important: For pHash, parallel processing provides a significant speed boost (up to 2-3 times on large folders). For CNN+Faiss, the improvement is less noticeable (around 15-20%), since CNN already heavily utilizes the CPU.
  • Warning: ⚠️ Parallel processing uses all available CPU cores and may significantly load your system. Before enabling this feature, please close resource-intensive applications (browsers with many tabs, video editors, games) to avoid system freezes.
  • Tip: On small volumes (up to 20 videos), the speed improvement will be negligible — you can disable this feature to save resources.

For more details on algorithm settings and their parameters, see the "Algorithm Settings" section.

🚀 Step 3: Starting the Scan

Click "🚀 Start Optimized Scan". The program will show progress (stages: metadata collection, feature extraction, comparison), elapsed time, and estimated remaining time. You can pause or stop it.
For more details on process control, see the "Scan Control" section.

📊 Step 4: Analysing the Results

After the scan is complete, you will see a list of video pairs sorted by similarity descending:

  • Video pairs with similarity percentage and colour coding (green – better, red – worse) for characteristics (size, duration, resolution, FPS, number of frames).
  • Preview of each video (a random frame from the middle, so preview alone is not sufficient to judge similarity).
  • Action buttons:
    • 🎬 Synchronous View – opens a side‑by‑side view of the two videos with seeking.
    • 🔍 Compare – launches a detailed comparison of a sample of frames from both videos with frame‑by‑frame results.
  • "🗑️ DELETE" checkboxes – mark files for later deletion/moving.
  • Context menu (right‑click on the file panel) – Open file location.

Navigation through results (pagination and search) is described in the "Working with Results" section.

🗑️ Deletion Management

The top panel shows the number of marked files and their total size. The Clear marks button clears all marks. The 🗑️ Delete/move marked (X) button launches a bulk deletion with action selection (trash, permanent delete, move to folder).

💡 Tip: Before deleting, use the synchronous view or the "Compare" button to make sure the files are indeed duplicates or unwanted.