MMT Live - Real-Time Media Monitoring Tool¶

📦 Source Code: View on GitHub

Overview¶

MMT Live is a real-time version of the Media Monitoring Tool that can detect video fragments in live streams or during live playback. It's designed for monitoring broadcasts in real-time, detecting advertisements as they air, and triggering immediate actions when specific content is detected.

Key Features¶

Real-Time Detection: Identify fragments as video plays
Live Stream Support: Monitor RTSP, HTTP, and file streams
Instant Notifications: Immediate alerts when content detected
Continuous Monitoring: 24/7 operation capability
Ad Detection: Real-time commercial identification
Low Latency: Near-instantaneous detection
Live Preview: Monitor stream while processing

Differences from Standard MMT¶

Feature	MMT	MMT Live
Processing	Post-recording	Real-time
Input	Files only	Files + Streams
Detection	Batch	Continuous
Results	After completion	Immediate
Use Case	Analysis	Monitoring

User Interface¶

Main Components¶

Live Media Player: Shows current stream/playback
Fragment Library: Pre-loaded detection targets
Detection Log: Real-time detection events
Status Indicators: Stream health and processing state
Settings Panel: Live adjustment of parameters

How to Use¶

Setup Workflow¶

Prepare Fragment Library:
Load commercials/clips to detect
Generate fingerprints in advance
Organize by priority/category
Configure Input Source:
File: Select video for monitoring
Stream: Enter RTSP/HTTP URL
Device: Select capture device
Set Detection Parameters:
Sensitivity threshold
Minimum match duration
Alert preferences
Start Monitoring:
Click "Start" to begin
Video plays while analyzing
Detections appear immediately

Real-Time Operation¶

Continuous Processing: Analyzes video as it plays
Rolling Buffer: Maintains recent video history
Instant Matching: Compares against fragment library
Event Logging: Records all detections with timestamps

Use Cases¶

1. Broadcast Compliance¶

Ensure ads play as scheduled
Verify content restrictions
Monitor competitor advertising
Track program segments

2. Live Stream Monitoring¶

Detect copyrighted content
Monitor multiple channels
Track brand appearances
Quality assurance

3. Automated Actions¶

Trigger recording on detection
Send notifications/alerts
Switch streams automatically
Generate real-time reports

4. Advertisement Tracking¶

Count commercial airings
Verify ad placement
Monitor ad frequency
Competitive analysis

Configuration¶

Input Sources¶

File Playback:

Simulates live monitoring
Useful for testing
Supports all video formats

RTSP Streams:

rtsp://camera.example.com:554/stream
rtsp://username:password@server/path

HTTP Streams:

http://server.com/stream.m3u8
http://server.com/live.mjpeg

Detection Settings¶

Buffer Size: Video history (5-60 seconds)
Check Interval: How often to analyze (1-5 seconds)
Confidence Threshold: Match quality (70-95%)
Fragment Priority: Which fragments to check first

Performance Optimization¶

System Requirements¶

CPU: Multi-core recommended
RAM: 8-16GB for smooth operation
Network: Stable connection for streams
Storage: Fast SSD for fragment library

Optimization Tips¶

Fragment Library:
Keep under 100 active fragments
Pre-generate all fingerprints
Remove unused fragments
Stream Quality:
Use consistent bitrate
Avoid very high resolutions
Ensure stable connection
Processing:
Adjust check interval based on CPU
Use appropriate buffer size
Enable GPU acceleration if available

Advanced Features¶

Multi-Stream Monitoring¶

Monitor multiple streams simultaneously
Separate detection threads per stream
Consolidated reporting
Resource management

Custom Actions¶

Configure actions for detections:

Email notifications
HTTP webhooks
File logging
Database recording
Stream recording triggers

Detection Zones¶

Define time windows for detection
Schedule different fragment sets
Ignore certain time periods
Priority scheduling

Troubleshooting¶

No Detections¶

Verify fragments are loaded
Check stream is playing
Confirm fingerprints generated
Adjust sensitivity lower

High CPU Usage¶

Reduce check frequency
Lower stream resolution
Decrease buffer size
Limit active fragments

Stream Issues¶

Check network connectivity
Verify stream URL
Test in media player first
Monitor bandwidth usage

Delayed Detections¶

Increase processing priority
Reduce buffer size
Check system resources
Optimize fragment count

Best Practices¶

Preparation¶

Test fragments in standard MMT first
Optimize fragment quality and length
Build comprehensive fragment library
Document expected detections

Operation¶

Monitor system resources
Regular fragment library updates
Periodic detection accuracy checks
Maintain detection logs

Maintenance¶

Clean old detection logs
Update fragment fingerprints
Review false positives/negatives
Optimize based on results

Integration Options¶

API Integration¶

REST API for detection events
WebSocket for real-time updates
Database logging options
Third-party integrations

Automation¶

Scheduled monitoring
Automatic report generation
Alert escalation
Stream switching

Comparison with Alternatives¶

vs Standard MMT:

Real-time vs post-processing
Continuous vs batch operation
Immediate vs delayed results

vs Manual Monitoring:

Automated vs human observation
24/7 vs limited hours
Consistent vs variable accuracy

MMT: Post-recording analysis version
vfp_search: Command-line fragment search
DVS: Duplicate video detection