VQTDL
No source file? No problem.
Introducing our brand new no-reference video quality algorithm allowing you to evaluate image and video quality in a way that closely matches human perception. Get accurate video quality evaluations and data you can trust.
Trusted by Startups, SMEs and Enterprise businesses worldwide
How does it work
Efficient and accurate video evaluation
As a no-reference algorithm, VQTDL is able to accurately assess video quality without needing the original file. This flexibility makes it ideal for any video application where access to a source file may be difficult or downright impossible.
- MOS-based quality score prediction for better precision
- No-reference evaluation—no original video file needed
- Based on deep learning and built using the latest AI technologies
- Adjustable for any scenarios and network conditions
- Requires half the computing power of standard metrics
Discover how Zoom leveraged our competitive analysis
We helped Zoom compare their solution against competitors and assess performance under different network conditions—turning data into actionable insights.
Why choose VQTDL
Built for precision and real-world use cases
VQTDL excels in providing reliable, consistent, and accurate data on video performance across all video apps, use cases, and industries. Its many advantages make it a clear choice for precise video quality assessment over other, more popular metrics.
- No source file required for quality assessments
- Adapts effectively to UI changes and processes all image resolutions
- Optimized for real-world scenarios like group calls and streaming
- Delivers consistent insights across all video apps
- Demonstrates better stability and accuracy in group call scenarios
- Consumes significantly less RAM than other alternatives
Comparison
One metric to rule them all
See how VQTDL measures up against human perception and other popular video quality metrics with our comprehensive side-by-side comparison. Compare capabilities and you be the judge.
| Subjective Scoring | BRISQUE | VMAF | VQTDL | |
|---|---|---|---|---|
| Works Without Reference Video | Yes (requires representative samples) | Yes | No | Yes |
| MOS-based Scoring | True MOS from human users | No | Approximate MOS via model fusion | Predicts MOS |
| AI / Deep Learning | Human judgment | Traditional ML | CNN + handcrafted features | Transformer + MoE |
| Accuracy (PLCC*) | Highest with proper testing | ~80% | ~90% | >96% across network conditions |
| Stability with UI Changes | Humans adjust intuitively | Degrades with UI overlays | Impacted by layout changes | 33% lower deviation vs BRISQUE |
Coverage
Accurate video data across all apps
Get precise, reliable quality and performance insights across all video applications—from real-time communications and video conferencing to streaming and video-on-demand. Receive accurate data every time—no reference required.
- Real-time communications
- Video conferencing
- Streaming services
- Video-on-demand
- Short-form media
Deliver the quality your users deserve!
Join world-class companies in paving the way for quality audio and video experiences.
