Eval release pending
CoreVQ Suite
CoreVQ Studio and the bundled corevq command line provide local interactive analysis, automation, and structured evidence.
Video quality measurement and engineering
CoreVQ
Measure video quality, compare encodes, and diagnose streaming performance. Use the local Suite, commission engineering Services, or plan a future SDK integration.
CoreVQ offerings
From direct analysis to product integration
Choose local analysis, a paid engineering engagement, or an integration assessment for your product. Start with the outcome you need.
Paid engineering engagement
CoreVQ Services
Commission CoreVQ customization, workflow or product integration, or in-house OTT video-quality analysis under defined deliverables and acceptance criteria.
Explore Services →Planned integration product
CoreVQ SDK
A planned native SDK for embedding quality analysis in media products, automated tests, and OEM systems. Evaluation and licensing begin after a suitable package is qualified.
Explore SDK integration →Engineering outcomes
Use quality evidence where decisions are made
CoreVQ supports direct investigation today and a clear path toward customer-specific or embedded deployment.
Measure and compareAssess video quality, compare encoded renditions with references, and diagnose stream startup and playback behavior.
Customize and qualifyIntegrate additional AI-based quality scores, specialized formats, test cases, and customer-specific analysis workflows.
Optimize OTT qualityEvaluate encoding ladders for QoE, bitrate, compute, delivery cost, and the required analysis scale.
Embed measurementBuild automated tests, streaming monitors, media platforms, appliances, and OEM integrations.
Versioned human-opinion verification
Perceptual score evidence
1,310held-out videos scored
- KoNViD-1k720 videos
- 0.6633SROCC
- Microsoft Video Call MOS590 videos
- 0.8160SROCC
- Combined1,310 videos
- 0.7301SROCC
SROCC measures rank agreement with human opinion. These results support relative quality comparison on the evaluated test partitions; the score does not predict numeric MOS.
Scope, result states, sources, and hashesInspectable technology
Technology behind CoreVQ
One native engine connects video measurements, score composition, and execution evidence across Studio and the command line.
- Measurement
- Native metric values, normalized quality, confidence, and score contribution.
- Execution
- Analyzed ranges, frame coverage, completion state, warnings, and partial results.
- Consistency
- Studio and the bundled CLI use the same CoreVQEngine implementation.
- Retention
- CSV and JSON preserve results and provenance for review and automation.
CoreVQ's scoring, workflow, and evidence layers build on established metrics and third-party components. The published technology material explains those boundaries and integration paths.