DAEMLABS
When working with video delivery systems, one problem quickly becomes apparent: encoding a video is easy, but understanding how encoding decisions affect quality is much harder. Resolution changes, bitrate constraints, compression artifacts, and visual degradation can all impact the viewing experience. To better understand this process, we built VidSage — a video transcoding and quality analysis platform that combines FFmpeg, OpenCV, and automated reporting into a single workflow.
This is the story of how we built a lightweight video engineering tool that automates transcoding, extracts media metadata, evaluates visual quality, and generates actionable reports for video analysis.
Modern video platforms process enormous amounts of content every day.
Before a video reaches viewers, it often passes through multiple stages:
While transcoding pipelines are well understood, evaluating the resulting output remains a challenge.
Questions such as:
often require manual inspection.
We wanted a system that could automate these checks and provide objective insights into video quality.
Video analysis involves several independent tasks that need to work together seamlessly.
Our requirements included:
Rather than building a monolithic application, we divided the workflow into modular components.
Input Video
↓
Transcoding
↓
Metadata Extraction
↓
Quality Analysis
↓
Visualization
↓
Report Generation
Each stage operates independently while contributing data to the final analysis report.
The first component of VidSage handles video transcoding.
Using FFmpeg, the system automatically generates multiple resolution variants from a single source file.
The workflow looks like:
Original Video
↓
FFmpeg
↓
┌─────────────┐
│ 480p │
├─────────────┤
│ 720p │
├─────────────┤
│ 1080p │
└─────────────┘
This mirrors the workflow used by adaptive bitrate streaming systems.
Instead of storing only one version of a video, streaming platforms typically create multiple renditions optimized for different network conditions and device capabilities.
VidSage automates this process and prepares the outputs for further inspection.
A video file contains much more information than just image frames.
Every stream carries metadata such as:
Understanding these properties is critical when diagnosing playback or quality issues.
To extract this information, VidSage uses ffprobe.
The metadata extraction stage gathers stream information and converts it into a structured format suitable for reporting.
Example output includes:
{
"codec": "h264",
"resolution": "1280x720",
"bitrate": "3200 kbps",
"duration": "95 seconds"
}
This gives engineers immediate visibility into how a file has been encoded without requiring manual inspection.
Resolution and bitrate alone do not tell the entire story.
Two videos with identical specifications can look dramatically different depending on encoding settings, source quality, and compression artifacts.
We wanted a lightweight metric capable of detecting visual degradation automatically.
The challenge was finding something computationally inexpensive while still providing meaningful insight.
Our solution was blur detection using Laplacian variance.
The intuition is simple:
Sharp images contain strong edges and high-frequency detail.
Blurry images contain fewer sharp transitions and lower edge variance.
For sampled video frames, the system calculates:
Frame
↓
Grayscale Conversion
↓
Laplacian Filter
↓
Variance Calculation
↓
Blur Score
A higher variance indicates a sharper frame.
A lower variance suggests increased blur.
By repeating this process across the entire video, VidSage produces a timeline of visual clarity.
Individual blur scores are useful, but trends are even more valuable.
VidSage generates frame-by-frame quality visualizations showing how sharpness changes throughout a video.
The analysis pipeline creates plots similar to:
Blur Score
▲
│
│ /\ /\
│ / \ / \
│____/____\___/____\____
└──────────────────────►
Time
These visualizations help identify:
Instead of manually reviewing hours of footage, engineers can immediately locate problematic regions.
A major goal of the project was reducing manual analysis effort.
Once transcoding and analysis are complete, VidSage automatically generates a Markdown report containing:
The reporting pipeline looks like:
Analysis Results
↓
Structured Data
↓
Markdown Generator
↓
report.md
This transforms raw technical information into a format that can be shared with engineers, QA teams, and stakeholders.
Video quality issues become exponentially harder to diagnose at scale.
Streaming platforms may process thousands of videos per day, making manual validation impossible.
Automated analysis helps answer critical questions:
By automating these checks, teams can focus on fixing issues rather than searching for them.
Several design choices proved particularly effective.
FFmpeg provides a battle-tested foundation for generating production-grade video outputs.
Each component performs a single responsibility, making the system easy to maintain and extend.
The blur detection approach is computationally lightweight while still providing meaningful quality insights.
Generating human-readable summaries significantly reduced analysis time.
Blur detection is useful, but modern metrics such as SSIM, PSNR, and VMAF would provide deeper quality evaluation.
Future versions could generate HLS and DASH outputs to better simulate production streaming environments.
The current implementation focuses on offline analysis. Live stream monitoring would expand its usefulness significantly.
Recent research has shown that neural networks can predict perceived video quality more accurately than traditional metrics.
A lightweight video engineering platform capable of transcoding media files, extracting stream metadata, measuring visual sharpness, generating quality visualizations, and producing automated reports.
By combining FFmpeg, OpenCV, and Python-based analytics, VidSage simulates key components of modern video delivery infrastructure while providing valuable insight into how encoding decisions impact visual quality.
The project demonstrates how automated quality assurance can simplify video workflows, reduce manual inspection effort, and provide engineers with actionable data before content reaches viewers.
Need a custom media processing or video analytics solution? We build scalable systems for transcoding, streaming infrastructure, computer vision, content moderation, and AI-powered media intelligence.