- Add check in loadPersistedConvertConfig to default FrameRate to "Source" if empty
- Add check after loading persisted config to ensure FrameRate is "Source" if not set
- Prevents unwanted frame rate conversions from persisted config overriding safe defaults
This ensures that frame rate always defaults to "Source" and users won't
accidentally convert all their videos to 23.976fps or another frame rate
if they had previously saved a config with a specific frame rate set.
- Add upscaleFrameRate and upscaleMotionInterpolation fields to appState
- Add Frame Rate section to upscale UI with dropdown and motion interpolation checkbox
- Pass frame rate settings through upscale job config
- Implement frame rate conversion in executeUpscaleJob using minterpolate or fps filter
- Frame rate section appears after resolution selection in upscale settings
Frame rate control is now available in both convert and upscale modules,
allowing users to standardize content from different regions with optional
motion interpolation for smooth conversion.
- Add UseMotionInterpolation field to convertConfig struct
- Implement minterpolate filter for smooth frame rate changes when enabled
- Add UI checkbox in advanced settings for motion interpolation option
- Use minterpolate with high-quality settings (mci mode, aobmc, bidir ME, vsbmc)
- Falls back to simple fps filter when motion interpolation is disabled
- Fix pre-existing statusBar function calls (renamed to moduleFooter)
Motion interpolation provides smooth frame rate conversion (e.g., 24fps→60fps)
using motion-compensated interpolation instead of simple frame duplication.
This is useful for standardizing content from different regions.
Removed confirmation dialog popups when converting files with
chapters to DVD format. Instead, show a non-intrusive inline
warning label that appears/disappears based on format selection.
Warning label:
- Shows only when file has chapters AND DVD format is selected
- Displays inline below format selector in both simple and advanced modes
- No user action required - just informational
- Text: "Chapters will be lost - DVD format doesn't support embedded chapters. Use MKV/MP4 to preserve chapters."
When converting a file with chapters to DVD/MPEG format, show
a confirmation dialog warning the user that chapters will be lost.
MPEG format does not support embedded chapters - they require
full DVD authoring with IFO files. Users are warned and given
the option to cancel or continue.
Warning appears for both 'Convert Now' and 'Add to Queue' buttons.
Added -show_chapters flag to ffprobe command to retrieve chapter
information. Parse chapters from JSON output and set HasChapters
field when chapters are present.
Files with chapters will now correctly show 'Chapters: Yes' in
the file information display.
Added explicit stream mapping to preserve all streams:
- Map video, audio, and subtitle streams (subtitles optional)
- Added -map_chapters to preserve chapter information
- Added -map_metadata to preserve all file metadata
- Copy subtitle streams without re-encoding
Applies to both conversions with and without cover art.
Works for all output formats that support these features.
Removed -target ntsc-dvd and -target pal-dvd preset flags which
strip metadata including chapters. Instead, manually specify all
DVD parameters (bitrate, maxrate, bufsize, format) to maintain
full control and preserve chapter metadata.
Chapters now retained in both new 'dvd' format and legacy DVD formats.
When 'Clear Completed' empties the queue, return to the previous
module instead of staying in an empty queue view. If jobs remain
after clearing, stay in queue view and refresh.
Changed 'Clear All' behavior to return to the last active module
instead of always going to main menu. Falls back to main menu
if no previous module is tracked or if coming from queue itself.
Added clearer format descriptions:
- "Fast Merge (No Re-encoding)" instead of "MKV (Copy streams)"
- "Lossless MKV (Best Quality)" - new option with slow preset, CRF 18, FLAC audio
- "High Quality MP4 (H.264/H.265)" instead of technical codec names
- "DVD Format" with conditional region/aspect selectors
- "Blu-ray Format" instead of "Blu-ray (H.264)"
DVD Format improvements:
- When "DVD Format" is selected, shows Region (NTSC/PAL) and Aspect (16:9/4:3) options
- Options hidden for other formats
- Stored in state and passed to merge job config
- Updated execution to use DVD region/aspect settings
Maintains backward compatibility with legacy DVD format codes.
WMV Encoder Fix:
- WMV files now use wmv2 encoder (ffmpeg compatible) instead of wmv3
- Audio uses wmav2 for WMV files
- High quality bitrate (2000k) for WMV video
- Fallback handling for unsupported source codecs
UI Simplification:
- Changed "High Quality (source format/codecs)" to "Match Source Format"
- Simplified hint text to just "Unchecked = Use Conversion Settings"
- More concise and less confusing labels
Snippet Quality Improvements:
- High Quality mode now detects and uses source codecs (WMV stays WMV)
- Uses conversion panel's encoder preset (e.g., 'slow') instead of hardcoded 'ultrafast'
- Uses conversion panel's CRF setting for quality control
- Outputs to source file extension in High Quality mode
- Updated UI label to "High Quality (source format/codecs)"
Workflow Streamlining:
- Removed popup dialog when loading multiple videos
- Showing convert view is sufficient feedback
- Failed files logged instead of shown in dialog
UI Fixes:
- Status label no longer wraps to new line on action bar
- Set text truncation to keep status on single line
Changed snippet "Default Format" mode from stream copy to re-encoding with
high quality settings (libx264, CRF 17, ultrafast preset). Stream copy
cannot provide precise durations as it can only cut at keyframe boundaries.
Both snippet modes now output MP4 and re-encode. The difference is quality:
- High Quality mode: CRF 17, ultrafast preset
- Conversion Settings mode: Uses configured output settings
Updated UI labels to reflect "Snippet Quality" instead of output format.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Reverts the problematic -accurate_seek and -to flags that caused wildly incorrect durations (9:40 instead of 10s). Returns to the standard, reliable FFmpeg pattern for stream copy:
ffmpeg -ss START -i input -t DURATION -c copy output
This places -ss before -i for fast keyframe seeking and uses -t for duration (not -to which is an absolute timestamp causing incorrect extraction). Should now correctly extract the configured snippet length centered on video midpoint.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Improves snippet timing accuracy for Default Format mode by:
- Adding -accurate_seek flag for precise keyframe seeking
- Changing from -t (duration) to -to (end time) for better accuracy
- Adding -avoid_negative_ts make_zero to fix timestamp issues with problematic containers like WMV
This should resolve issues where snippets were 1:20 or 0:21 instead of the configured length (e.g., 10s). Stream copy still uses keyframe-level precision but should now be much closer to target duration.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Changes multi-video drag-and-drop behavior to load videos into memory instead of automatically adding them to conversion queue. This allows users to:
- Adjust conversion settings before queuing
- Generate snippets instead of converting
- Navigate through videos before deciding to convert
Creates new loadMultipleVideos() function that loads all videos into loadedVideos array and shows informative dialog. Users can now use Convert or Snippet buttons to manually process videos as needed.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Changes checkbox label from "Use Source Format (stream copy)" to "Snippet to Default Format (preserves source quality)". Unchecked state is now "Snippet to Output Format (uses conversion settings)". This clarifies that default format preserves the source file's quality, bitrate, codec, and container without any conversion artifacts.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Updates snippet conversion mode to use configured video/audio codecs, presets, CRF, and bitrates from the Convert tab instead of hardcoded h264/AAC. Output extension now matches selected format (e.g., .mkv, .webm, .mp4). This allows true comparison between source format snippets and conversion preview snippets using user's exact conversion settings.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Implements configurable snippet output mode with two options:
1. Source Format (default): Uses stream copy to preserve exact video/audio quality with source container format. Duration uses keyframe-level precision (may not be frame-perfect).
2. Conversion Format: Re-encodes to h264/AAC MP4 with frame-perfect duration control.
Adds checkbox control in snippet settings UI. Users can now compare source format snippets for merge testing and conversion format snippets for output preview.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Changes snippet generation to always output .mp4 files instead of preserving source extension. This fixes duration accuracy issues caused by container/codec mismatch (e.g., h264 video in .wmv container). MP4 is the proper container for h264-encoded video and ensures FFmpeg respects the -t duration parameter correctly.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Updates application version constant, documentation, and completion tracking to reflect dev18 release. Build output now correctly shows v0.1.0-dev18.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Enhances screenshot module with comprehensive technical metadata display including audio bitrate, adds 8px padding between thumbnails for professional contact sheets. Implements new Player module for video playback access. Adds complete Filters and Upscale modules with traditional FFmpeg scaling methods (Lanczos, Bicubic, Spline, Bilinear) and resolution presets (720p-8K). Introduces configurable snippet length (5-60s, default 20s) with batch generation capability for all loaded videos. Fixes snippet duration precision by re-encoding instead of stream copy to ensure frame-accurate cutting at configured length.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Changed thumbnail module to match convert module behavior with two
action buttons:
GENERATE NOW (High Importance):
- Adds job to queue and starts it immediately
- Runs right away if queue is idle
- Queues for later if jobs are running
- Shows "Thumbnail generation started!" message
Add to Queue (Medium Importance):
- Adds job to queue without starting
- Allows queuing multiple jobs
- Shows "Thumbnail job added to queue!" message
Implementation:
- Refactored job creation into createThumbJob() helper function
- Both buttons use same job creation logic
- Generate Now auto-starts queue if not running
- Follows same pattern as convert module
Benefits:
- Immediate generation when queue is idle
- Queue multiple jobs without starting
- Consistent UX with convert module
- Clear user feedback on action taken
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Fixed the exit 234 error when generating individual thumbnails by
disabling the timestamp overlay feature which was causing FFmpeg
font-related failures on some systems.
Changes:
- ShowTimestamp: false (was true)
- ShowMetadata: only true for contact sheets (was always true)
The timestamp overlay was causing issues because:
1. DejaVu Sans Mono font might not be available on all systems
2. FFmpeg exits with code 234 when drawtext filter fails
3. Individual thumbnails don't need timestamp overlays anyway
Contact sheets still get metadata headers, which is the main use case
for showing video information on thumbnails.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Fixed Thumbnail Count Issue:
- Changed frame selection from hardcoded 30fps to timestamp-based
- Now uses gte(t,timestamp) filter for accurate frame selection
- This fixes the issue where 5x8 grid only generated 34 instead of 40 thumbnails
Improved Contact Sheet Display:
- Reduced thumbnail width from 320px to 200px for better window fit
- Changed background color from black to app theme (#0B0F1A)
- Contact sheets now match the VideoTools dark blue theme
Added Viewing Capability:
- New "View Results" button in thumbnail module
- Contact sheet mode: Shows image in full-screen dialog (900x700)
- Individual mode: Opens thumbnail folder in file manager
- Button checks if output exists before showing
- Provides user-friendly messages when no results found
Benefits:
- Correct number of thumbnails generated for any grid size
- Contact sheets fit better in display window
- Visual consistency with app theme
- Easy access to view generated results within the app
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Added full job queue integration for thumbnail generation:
Job Queue Integration:
- Implemented executeThumbJob() to handle thumbnail generation in queue
- Changed "Generate Thumbnails" to "Add to Queue" button
- Added "View Queue" button to thumbnail module
- Removed direct generation code in favor of queue system
Progress Tracking:
- Jobs now show in queue with progress bar
- Contact sheet mode: shows grid dimensions in description
- Individual mode: shows count and width in description
- Job title: "Thumbnails: {filename}"
Benefits:
- Real-time progress tracking via queue progress bar
- Can queue multiple thumbnail jobs
- Access queue from thumbnail screen
- Consistent with other modules (convert, merge, snippet)
- Background processing without blocking UI
The thumbnail module now uses the same job queue system as other
modules, providing progress tracking and background processing.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Fixed the total thumbnails label to update in real-time when adjusting
columns or rows sliders in contact sheet mode.
Changes:
- Created totalLabel before sliders so both callbacks can access it
- Both column and row slider OnChanged callbacks now update the total
- Total recalculates as: columns × rows on each slider change
The total now updates immediately as you adjust the grid dimensions,
providing instant feedback on how many thumbnails will be generated.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Fixed Issues:
- Exit 234 error: Added font parameter to drawtext filter for individual
thumbnails (was missing, causing FFmpeg to fail)
- Output directory: Changed from temp to video's directory, creating a
folder named "{video}_thumbnails" next to the source file
New Features:
- Added video preview window to thumbnail module (640x360)
- Split layout: preview on left (55%), settings on right (45%)
- Preview uses same buildVideoPane as other modules for consistency
The thumbnail module now has a proper preview window for reviewing
the loaded video before generating thumbnails, and outputs are saved
in a logical location next to the source file.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Changed snippet extraction to use stream copy instead of re-encoding:
- Removed all convert config and encoding logic
- Now uses `-c copy` to copy all streams without re-encoding
- Uses same file extension as source for container compatibility
- Much faster extraction with no quality loss
- Updated job description to indicate "source settings"
This makes snippet generation instant instead of requiring full re-encode.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Implemented metadata header rendering on contact sheets showing:
- Filename and file size
- Video resolution and duration
Uses FFmpeg pad and drawtext filters to create an 80px header area
with white text on black background.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
New Features:
- Thumbnail extraction package with FFmpeg integration
- Individual thumbnails or contact sheet generation
- Configurable thumbnail count (3-50 thumbnails)
- Adjustable thumbnail width (160-640 pixels)
- Contact sheet mode with customizable grid (2-10 columns/rows)
- Timestamp overlay on thumbnails
- Auto-open generated thumbnails folder
Technical Implementation:
- internal/thumbnail package with generator
- FFmpeg-based frame extraction
- Video duration and dimension detection
- Aspect ratio preservation
- JPEG quality control
- PNG lossless option support
UI Features:
- Thumbnail module in main menu (Orange tile)
- Load video via file picker
- Real-time configuration sliders
- Contact sheet toggle with grid controls
- Generate button with progress feedback
- Success dialog with folder open option
Integration:
- Added to module routing system
- State management for thumb module
- Proper Fyne threading with DoFromGoroutine
- Cross-platform folder opening support
Module is fully functional and ready for testing.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Features added:
- Auto-detection in Inspect module: runs QuickAnalyze automatically when video is loaded
- Interlacing results display in Inspect metadata panel
- Deinterlace preview generation: side-by-side comparison button in Convert view
- Analyze button integration in Simple menu deinterlacing section
- Auto-apply deinterlacing settings when recommended
The Inspect module now automatically analyzes videos for interlacing when loaded via:
- Load button
- Drag-and-drop to main menu tile
- Drag-and-drop within Inspect view
Results appear directly in the metadata panel with full detection details.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Major features in this release:
1. Fixed merge job progress reporting
- Progress counter was jumping to 100% immediately due to incorrect
time unit conversion (microseconds vs milliseconds)
- Now shows accurate real-time progress throughout merge operations
2. Hardware encoder benchmarking system
- Automatic test video generation (30s 1080p test pattern)
- Detects available hardware encoders (NVENC, QSV, AMF, VideoToolbox)
- Tests all available encoders with multiple presets
- Measures FPS performance and ranks results
- Provides optimal encoder recommendation for user's hardware
- Real-time progress tracking with live results display
3. Benchmark history tracking
- Stores up to 10 most recent benchmark runs
- Browse past benchmark results with detailed comparisons
- View all encoder/preset combinations tested in each run
- Compare performance across different presets and encoders
- Apply recommendations from any past benchmark
- Persistent storage in ~/.config/VideoTools/benchmark.json
UI improvements:
- "Run Benchmark" button in main menu
- "View Results" button to browse benchmark history
- Live progress view showing current test and results
- Comprehensive results view with all encoder data
- Fixed merge module file list to use full vertical space
Bug fixes:
- Fixed merge progress calculation (microseconds issue)
- Fixed Fyne threading errors in benchmark UI updates
- Fixed progress bar percentage display (0-100 range)
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Extended the benchmark system to maintain a complete history of all
benchmark runs (up to last 10) with full results for each encoder/preset
combination tested.
Features:
- Stores complete benchmark run data including all test results
- History browser UI to view past benchmark runs
- Click any run to see detailed results for all encoders tested
- Compare performance across different presets and encoders
- Apply recommendations from past benchmarks
- Automatic history limit (keeps last 10 runs)
UI Changes:
- Renamed "Benchmark" button to "Run Benchmark"
- Added "View Results" button to main menu
- New benchmark history view showing all past runs
- Each run displays timestamp, recommended encoder, and test count
- Clicking a run shows full results with all encoder/preset combinations
Data Structure:
- benchmarkRun: stores single test run with all results
- benchmarkConfig: maintains array of benchmark runs
- Saves to ~/.config/VideoTools/benchmark.json
This allows users to review past benchmark results and make informed
decisions about which encoder settings to use by comparing FPS across
all available options on their hardware.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Implemented a full benchmark system that automatically detects available
hardware encoders, tests them with different presets, measures FPS
performance, and recommends optimal settings for the user's system.
Features:
- Automatic test video generation (30s 1080p test pattern)
- Hardware encoder detection (NVENC, QSV, AMF, VideoToolbox)
- Comprehensive encoder testing across multiple presets
- Real-time progress UI with live results
- Performance scoring based on FPS metrics
- Top 10 results display with recommendation
- Config persistence for benchmark results
- One-click apply to use recommended settings
UI Components:
- Benchmark button in main menu header
- Progress view showing current test and results
- Final results view with ranked encoders
- Apply/Close actions for recommendation
Integration:
- Added to main menu between "Benchmark" and "Logs" buttons
- Saves results to ~/.config/VideoTools/benchmark.json
- Comprehensive debug logging for troubleshooting
This allows users to optimize their encoding settings based on their
specific hardware capabilities rather than guessing which encoder
will work best.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
The issue was that FFmpeg's out_time_ms field is actually in microseconds
(not milliseconds despite the name). We were dividing by 1,000 when we
should have been dividing by 1,000,000 to convert to seconds.
This caused the progress calculation to be off by 1000x, making it
immediately jump to 100% even though the job was just starting.
Also added comprehensive debug logging to track progress samples and
identify calculation issues in the future.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Issue: File list only used half the vertical space, wasting screen real estate.
Changed left panel from VBox to Border layout:
- Top: "Clips to Merge" label and Add/Clear buttons (fixed size)
- Center: File list scroll area (expands to fill remaining space)
The border layout gives the scroll area priority to expand vertically,
maximizing the visible file list area. This is especially important
when merging many clips.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
User reports progress jumps to 100% within 10 seconds but merge continues for 45s total.
Added comprehensive debug logging to track:
- Individual clip durations as they're summed
- Total expected duration for the merge
- Exact moment when progress hits 100% with actual vs expected times
- Only update progress when it changes by ≥0.1% (reduces callback spam)
This will help diagnose whether:
- Clip durations are being calculated incorrectly
- FFmpeg's out_time_ms doesn't match expected total duration
- Concat demuxer reports different output duration than sum of inputs
Logging appears in logs/videotools.log with CatFFMPEG category.
To view: tail -f logs/videotools.log | grep FFMPEG
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Issues fixed:
- Missing file extensions caused FFmpeg errors (user's job 234 failure)
- Limited codec options (only copy or H.265)
- Manual codec mode selector was redundant
Changes:
1. Auto file extension handling:
- Automatically adds/corrects extension based on selected format
- .mkv for MKV/Blu-ray formats
- .mpg for DVD formats
- .mp4 for MP4 formats
- Validates and fixes extension in addMergeToQueue
2. Expanded format options:
- MKV (Copy streams) - stream copy, no re-encoding
- MKV (H.264) - re-encode with H.264, CRF 23
- MKV (H.265) - re-encode with H.265, CRF 28
- MP4 (H.264) - H.264 + AAC audio, web-optimized
- MP4 (H.265) - H.265 + AAC audio, web-optimized
- DVD NTSC/PAL (16:9 and 4:3)
- Blu-ray (H.264)
3. Removed redundant codec mode selector:
- Format dropdown now explicitly includes codec choice
- Cleaner, more intuitive UI
- Backward compatible with old queue jobs
Extension is auto-updated when:
- User selects a different format (updates existing path extension)
- User adds merge to queue (validates/fixes before encoding)
- Prevents errors from missing or wrong file extensions
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
The -progress flag was being added AFTER the output path in the FFmpeg command,
causing FFmpeg to not recognize it and therefore not output progress information.
Moved -progress pipe:1 -nostats to appear BEFORE the output path.
Now merge jobs will correctly report progress as they encode:
- Progress starts at 0%
- Updates based on out_time_ms from FFmpeg progress output
- Calculates percentage based on total duration of all clips
- Shows accurate real-time progress in queue view and stats bar
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
The Merge module's ui.NewDroppable wrappers weren't receiving drop events
because the window-level handleDrop function was intercepting them first.
Added merge module handling to handleDrop function:
- Accepts individual video files and adds them sequentially to merge clips
- Accepts multiple files at once and processes all in order
- Accepts folders and recursively finds all video files
- Probes each video to get duration and metadata
- Sets chapter names defaulting to filename
- Auto-sets output path to "merged.mkv" once 2+ clips are added
- Refreshes UI after each clip is added
Now drag-and-drop works consistently across all modules (Convert, Compare, Inspect, Merge).
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Implements video transformation features:
- Horizontal flip (mirror effect) using hflip filter
- Vertical flip (upside down) using vflip filter
- Rotation support: 90°, 180°, 270° clockwise using transpose filters
UI additions in Advanced mode:
- New "VIDEO TRANSFORMATIONS" section
- Two checkboxes for flip controls with descriptive labels
- Dropdown selector for rotation angles
- Hint text explaining transformation purpose
Filter implementation:
- Applied after aspect ratio conversion, before frame rate conversion
- Works in both queue-based and direct conversion paths
- Uses FFmpeg standard filters: hflip, vflip, transpose
Addresses user request to add flip/rotation capabilities inspired by Jake's script using -vf hflip.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
This commit includes three critical bug fixes and Windows build improvements:
**Bug Fixes:**
1. **Queue Conversion Progress Tracking** (main.go:1471-1534)
- Enhanced executeConvertJob() to parse FPS, speed, and ETA from FFmpeg output
- Queue jobs now show detailed progress metrics matching direct conversions
- Stats stored in job.Config for display in the conversion stats bar
2. **AMD AMF Hardware Acceleration** (main.go)
- Added "amf" to hardware acceleration options
- Support for h264_amf, hevc_amf, and av1_amf encoders
- Added AMF-specific error detection in FFmpeg output parsing
3. **DVD Format Resolution Forcing** (main.go:1080-1103, 4504-4517)
- Removed automatic resolution forcing when DVD format is selected
- Removed -target parameter usage which was forcing 720×480/720×576
- Resolution now defaults to "Source" unless explicitly changed
- DVD compliance maintained through manual bitrate/GOP/codec parameters
**Windows Build Improvements:**
- Updated build.bat to enable CGO (required for Fyne/OpenGL)
- Added automatic GCC/MinGW-w64 detection and installation
- Automated setup via winget for one-command Windows builds
- Improved error messages with fallback manual instructions
**Documentation:**
- Added comprehensive Windows setup guides
- Created platform.go for future platform-specific code
- Updated .gitignore for Windows build artifacts
All changes tested and working. Ready for production use.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Features:
- FPS counter in conversion status showing real-time encoding speed
- Job queue now displays FPS, encoding speed (e.g., "1.2x"), and ETA for running conversions
- Copy Comparison button exports side-by-side metadata comparison report
- Auto-compare checkbox in Convert module - automatically loads Compare view after conversion
- Convert Now properly adds job to queue and displays in Job Queue with live stats
- Module badge colors in job queue now match main menu tile colors
- Fixed fullscreen compare window sizing (reduced player dimensions to prevent overflow)
Bug Fixes:
- Fixed queue state management - only one job runs at a time (prevents multiple jobs showing "running")
- Fixed Compare module slot assignment - single video drops now fill empty slot instead of overwriting
- Fixed job queue scroll rubber banding (no longer jumps back to top)
- Enhanced crop detection validation for WMV/AVI formats with dimension clamping and bounds checking
Documentation:
- VT_Player integration notes with API requirements for keyframing and trim features
- LosslessCut feature analysis for Trim module inspiration
- Video metadata guide covering MP4/MKV custom fields and NFO generation
- Trim module design specification
- Compare fullscreen mode documentation
- Updated VIDEO_PLAYER_FORK.md to mark fork as completed
Technical Changes:
- Added state tracking for FPS, speed, and ETA (main.go:197-199)
- Enhanced queue processJobs() to check for running jobs before starting new ones
- Improved Compare module drag-and-drop logic with smart slot assignment (both code paths)
- Added deferred scroll position restoration to prevent UI jumping
- Job queue Config map now carries conversion stats for display
🤖 Generated with Claude Code (https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Fixed thumbnails not displaying:
- Added preview frame generation to Compare module
- Thumbnails now load asynchronously when videos are loaded
- Uses capturePreviewFrames() just like Convert module
- Thumbnails appear after brief generation delay
Added Clear All button:
- Positioned to the right of instructions text
- Clears both File 1 and File 2 slots
- Refreshes view to show empty state
- Low importance styling (not highlighted)
Layout improvements:
- Instructions row now uses Border layout
- Clear All button aligned to the right
- Clean, accessible button placement
Both videos now show thumbnails (240x135) automatically
when loaded, providing visual confirmation of loaded content.
Fixed issue where dragging single videos would overwrite existing data:
Smart slot filling logic:
- Single video dropped: Fills first empty slot (File 1 then File 2)
- If both slots full: Shows dialog asking user to Clear first
- Multiple videos dropped: Fills both slots (replaces existing)
Behavior changes:
1. Drag first video → goes to slot 1
2. Drag second video → goes to slot 2
3. Drag third video → shows "Both Slots Full" message
4. Drag 2+ videos together → replaces both slots
User experience improvements:
- No more accidental overwrites when loading one at a time
- Clear feedback when slots are full
- Can now build comparison by dragging videos individually
- Or drag both at once to start fresh
Main menu drag-and-drop to Compare tile:
- Already working correctly
- Loads both videos sequentially then shows module
- No changes needed to that path
This makes the Compare workflow much more intuitive and prevents
losing loaded video data when adding the second video.
Fixed critical bug where loading second video would overwrite first:
- Changed parallel goroutines to sequential loading
- Load file 1, then file 2, then refresh UI once
- Prevents race condition from multiple showCompareView() calls
- Both files now display correctly side by side
Added action buttons for each file:
- Copy Metadata button: Copies formatted metadata to clipboard
- Clear button: Removes video from slot and refreshes display
- Buttons arranged horizontally: Load | Copy | Clear
- Low importance styling for secondary actions
Changes to drag-and-drop handlers:
- Within Compare module: sequential loading, single refresh
- From main menu: already sequential, no changes needed
- Both paths now work correctly
This fixes the "second file overwrites first" issue and adds
the requested metadata copy and clear functionality.
Fixed window resizing issues for better cross-platform behavior:
Convert module video pane:
- Reduced video pane minimum from 460x260 to 320x180
- Removed rigid MinSize on outer container (commented out)
- Removed rigid MinSize on image element
- Set stage minimum to 200x113 (reasonable 16:9 minimum)
- Video pane now scales down allowing smaller windows
Compare module:
- Reduced thumbnail minimum from 320x180 to 240x135
- Reduced metadata scroll minimum from 300x200 to 250x150
- More compact layout allows better window resizing
Benefits:
- Window can now shrink to fit smaller screens
- Better behavior on tiling window managers
- More flexible for cross-platform (Windows, macOS, Linux)
- Content scales intelligently instead of forcing window size
Note: Double-click titlebar maximize is a Fyne framework limitation.
Maximize via window controls or OS shortcuts (F11, Super+Up) works.
Added window centering to improve initial presentation:
- Call w.CenterOnScreen() after setting window size
- Window now opens centered rather than at OS default position
- Maintains existing resizing and maximization support
The window is already maximizable via SetFixedSize(false).
Users can maximize using OS window controls (double-click
titlebar, maximize button, or OS shortcuts like F11/Super+Up).
Prevents long filenames from manipulating window size:
- Truncate filenames longer than 35 characters
- Smart truncation preserves file extension
- Format: "long-filename-na...mp4" instead of wrapping
- Falls back to simple truncation for very long extensions
- Removed text wrapping from labels (no longer needed)
Examples:
- "my-very-long-video-filename.mp4" → "my-very-long-video-fi....mp4"
- "short.mp4" → "short.mp4" (unchanged)
- "filename.mkv" → kept as-is if under 35 chars
This ensures the Compare module labels stay compact and
predictable regardless of filename length.
Resolved UI framing issues where metadata was crushed and not
taking available vertical space:
Layout improvements:
- Used container.NewBorder to make metadata areas expand properly
- Set minimum sizes for scroll containers (300x200)
- Removed outer VScroll - individual metadata areas now scroll
- Grid columns now properly fill available vertical space
- Instructions fixed at top, metadata expands to fill remaining space
Text wrapping fixes:
- Added fyne.TextWrapBreak to file labels
- Prevents long filenames from stretching the window horizontally
- Labels now wrap to multiple lines as needed
Architecture changes:
- Separated file headers (label + button) from content
- Each column uses Border layout: header at top, metadata fills center
- Metadata scroll containers have explicit minimum sizes
- Two-column grid properly distributes horizontal space
The layout now feels more modern with better space utilization
and smooth scrolling within the metadata panels.
Major improvements to Compare module user experience:
- Auto-populate metadata when files are loaded (no Compare button needed)
- Show video thumbnails for both files (320x180)
- Support drag-and-drop onto Compare tile from main menu
- Load up to 2 videos when dropped on Compare tile
- Show dialog if more than 2 videos dropped
- Files loaded via drag show immediately with metadata
Changes to handleModuleDrop:
- Added special handling for Compare module
- Loads videos into compareFile1 and compareFile2 state
- Shows module with files already populated
Changes to buildCompareView:
- Added thumbnail display with dark background placeholders
- Created helper functions: formatMetadata(), loadThumbnail(), updateFile1(), updateFile2()
- Initialize view with any preloaded files
- Removed manual Compare button - metadata shows automatically
- Button handlers now call update functions to refresh display
- Cleaner, more intuitive workflow
This addresses the user feedback that dragging videos onto Compare
didn't load the module, and adds the requested thumbnail previews.
The Compare module now has colored bars at the top and bottom matching
its pink visual identity from the main menu. This creates visual
consistency with the Convert module and strengthens the app's
overall design language.
Changes:
- Added top bar with back button using ui.TintedBar
- Added bottom bar with module color
- Restructured layout to use container.NewBorder
- Made content area scrollable
The colored bars use the module's color (#FF44AA pink) as defined
in modulesList and retrieved via moduleColor().
- Implement drag-and-drop file loading in Compare module
- Accepts up to 2 video files
- Shows dialog if more than 2 videos dropped
- Automatically loads first two videos
- Integrated into global window drop handler
- Enhance metadata display with organized sections
- FILE INFO: path, file size, format
- VIDEO: codec, resolution, aspect ratio, frame rate, bitrate,
pixel format, color space, color range, field order, GOP size
- AUDIO: codec, bitrate, sample rate, channels
- OTHER: duration, SAR, chapters, metadata
- Both file panels now show identical detailed information
Critical Bug Fix:
- H.264 profile and level were being applied globally (-profile:v, -level:v)
- When cover art is present, this affected the PNG encoder stream
- PNG encoder doesn't support H.264 profiles, causing exit code 234
- Error: "Unable to parse option value 'main'" on PNG stream
Solution:
- Use stream-specific specifiers when cover art present
- Apply -profile✌️0 and -level✌️0 instead of -profile:v / -level:v
- This targets only the first video stream (main video)
- PNG cover art stream (1:v) is unaffected
- Fixed in both executeConvertJob() and startConvert()
UI Fix:
- Long output filenames were stretching the settings panel
- Added outputHint.Wrapping = fyne.TextWrapWord
- Filename now wraps properly instead of expanding horizontally
Tested with:
- Video with embedded cover art
- H.264 profile=main encoding
- Long filename conversion
This commit enhances the encoder preset selector with detailed information
about speed vs quality trade-offs for each preset option.
Preset Information:
- Ultrafast: ~10x faster than slow, ~30% larger files
- Superfast: ~7x faster than slow, ~20% larger files
- Very Fast: ~5x faster than slow, ~15% larger files
- Faster: ~3x faster than slow, ~10% larger files
- Fast: ~2x faster than slow, ~5% larger files
- Medium: Balanced baseline (default)
- Slow: ~2x slower than medium, ~5-10% smaller (recommended)
- Slower: ~3x slower than medium, ~10-15% smaller
- Very Slow: ~5x slower than medium, ~15-20% smaller
UI Enhancements:
- Dynamic hint label below encoder preset dropdown
- Updates automatically when preset changes
- Visual icons for different speed categories:
- ⚡ Ultrafast/Superfast/Very Fast (prioritize speed)
- ⏩ Faster/Fast (good balance)
- ⚖️ Medium (baseline)
- 🎯 Slow/Slower (recommended for quality)
- 🐌 Very Slow (maximum compression)
Implementation:
- updateEncoderPresetHint() function provides preset details
- Called on preset selection change
- Initialized with current preset on view load
- Positioned directly under preset dropdown for visibility
Benefits:
- Helps users understand encoding time implications
- Shows file size impact of each preset
- Recommends "slow" as best quality/size ratio
- Prevents confusion about preset differences
- Enables informed decisions about encoding settings
Technical:
- All presets already supported by FFmpeg
- No changes to command generation needed
- Works with all video codecs (H.264, H.265, VP9, etc.)
- Preset names match FFmpeg standards
This commit implements the frame rate conversion feature with intelligent
file size estimation and user guidance.
Frame Rate Options:
- Added all standard frame rates: 23.976, 24, 25, 29.97, 30, 50, 59.94, 60
- Maintained "Source" option to preserve original frame rate
- Replaced limited [24, 30, 60] with full broadcast standard options
- Supports both film (24 fps) and broadcast (25/29.97/30 fps) standards
Size Estimation:
- Calculates approximate file size reduction when downconverting
- Shows "Converting X → Y fps: ~Z% smaller file" hint
- Example: 60→30 fps shows "~50% smaller file"
- Dynamically updates hint when frame rate or video changes
- Only shows hint when conversion would reduce frame rate
User Warnings:
- Detects upscaling (target > source fps)
- Warns with ⚠ icon: "Upscaling from X to Y fps (may cause judder)"
- Prevents confusion about interpolation limitations
- No hint shown when target equals source
Implementation:
- updateFrameRateHint() function recalculates on changes
- Parses frame rate strings to float64 for comparison
- Calculates reduction percentage: (1 - target/source) * 100
- Updates automatically when video loaded or frame rate changed
- Positioned directly under frame rate dropdown for visibility
Technical:
- Uses FFmpeg fps filter (already implemented)
- Works in both direct convert and queue execution
- Integrated with existing frame rate handling
- No changes to FFmpeg command generation needed
Benefits:
- 40-50% file size reduction for 60→30 fps conversions
- Clear visual feedback before encoding
- Prevents accidental upscaling
- Helps users make informed compression decisions
This commit implements the highest priority dev13 feature: automatic
cropdetect with manual override capability.
Features:
- Added detectCrop() function that analyzes 10 seconds of video
- Samples from middle of video for stable detection
- Parses FFmpeg cropdetect output using regex
- Shows estimated file size reduction percentage (15-30% typical)
- User confirmation dialog before applying crop values
UI Changes:
- Added "Auto-Detect Black Bars" checkbox in Advanced mode
- Added "Detect Crop" button to trigger analysis
- Button shows "Detecting..." status during analysis
- Runs detection in background to avoid blocking UI
- Dialog shows before/after dimensions and savings estimate
Implementation:
- Added CropWidth, CropHeight, CropX, CropY to convertConfig
- Crop filter applied before scaling for best results
- Works in both direct convert and queue job execution
- Proper error handling for videos without black bars
- Defaults to center crop if X/Y offsets not specified
Technical Details:
- Uses FFmpeg cropdetect filter with threshold 24
- Analyzes last detected crop value (most stable)
- 30-second timeout for detection process
- Regex pattern: crop=(\d+):(\d+):(\d+):(\d+)
- Calculates pixel reduction for savings estimate
Benefits:
- 15-30% file size reduction with zero quality loss
- Automatic detection eliminates manual measurement
- Confirmation dialog prevents accidental crops
- Clear visual feedback during detection
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
This commit implements two new features:
1. Compare Module:
- New UI module for side-by-side video comparison
- Loads two video files and displays detailed metadata comparison
- Shows format, resolution, codecs, bitrates, frame rate, color info, etc.
- Accessible via GUI module button or CLI: videotools compare <file1> <file2>
- Added formatBitrate() helper function for consistent bitrate display
2. Target File Size Encoding Mode:
- New bitrate mode "Target Size" for convert module
- Allows users to specify desired output file size (e.g., "25MB", "100MB", "8MB")
- Automatically calculates required video bitrate based on:
* Target file size
* Video duration
* Audio bitrate
* Container overhead (3% reserved)
- Implemented ParseFileSize() to parse size strings (KB, MB, GB)
- Implemented CalculateBitrateForTargetSize() for bitrate calculation
- Works in both GUI convert view and job queue execution
Additional changes:
- Updated printUsage() to include compare command
- Added compare button to module grid with pink color
- Added compareFile1 and compareFile2 to appState
- Consistent "Target Size" naming throughout (UI and code)
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
