
VMAF-E
Rethinking Video Quality for Live and On-Demand
VMAF-E, developed by MainConcept and part of the vScore video quality metrics suite, is an AI-powered breakthrough in video quality measurement. Building on Netflix’s widely-used VMAF metric, VMAF-E brings perceptual video analysis into live production, unlocking instant, actionable insights for content creators, broadcasters and streaming platforms.
Learn more:
VMAF-E is available in the MainConcept HEVC Encoder released with SDK 16.0.
Why VMAF-E?
Conventional video quality tools were designed for file-based workflows: slow, fragmented, and never intended for real-time feedback. VMAF-E changes that.
AI-Accelerated Speed
Up to 10x faster than traditional VMAF, delivering immediate scoring for both live and on-demand content.
Live-Ready Accuracy
Performance closely matches standard VMAF, but is optimized for dynamic, time-sensitive streams.
Zero-Latency Insight
Built into the encoding process, VMAF-E analyzes in coding order, not presentation order, so your results are truly real-time.
Trusted Results
Benchmarked against PSNR and VMAF-CUDA , VMAF-E’s scores consistently align within similar levels.
Engineered for maximum flexibility and efficiency
Ideal for Live Streaming, great for VOD too
Get summary scores at the session’s end, or per-frame results as encoding happens, perfect for monitoring or in-depth analysis.
Smarter quality control
Loop VMAF-E data directly into your workflow for real-time optimization, letting you adjust encoding on the fly.
CPU and Workflow-Friendly
Reduced processing requirements free up CPU resources for encoding, instead of after-the-fact analysis.
See a Tangible Impact for Live and VOD

Standard VMAF is too resource intensive for live use. Operators need to know if stream quality is holding up now, not later. With VMAF-E, instantly assess, optimize, and track key metrics frame-by-frame. No more waiting, no more cumbersome syncing.

VMAF for VOD often requires decoding source and output, perfect frame sync, and re-processing for every content edit or filter. With VMAF-E, direct, inline measurement keeps analysis perfectly aligned, eliminates redundant steps, and simplifies even the most customized post-production pipeline.
Technical Highlights
01
AI-Powered Estimation
Neural network delivers perceptual scoring, approximating VMAF to within ±2 points for robust decision-making.
02
Session and Frame Logging
Choose between instant one-shot results or granular per-frame records for deeper insights.
03
Industry-First for HEVC
VMAF-E is MainConcept’s first production-ready AI metric, a testament to our commitment to smarter, more accessible video analytics.
