Video Quality Prediction via Audio Proxy Metrics

The patent introduces an intermediary system that uses audio quality metrics (which are easily measurable) as a proxy to estimate video quality. By treating audio quality measurement as an intermediary step, the system bridges the gap between simple network capability assessment and accurate video quality prediction, allowing operators to use straightforward audio metrics while still obtaining meaningful video quality estimates.

The patent transforms the approach by changing from direct video quality measurement to using audio quality parameters as substitutes. It systematically maps audio quality metrics (signal-to-noise ratio, bit error rate, etc.) to corresponding video quality predictions, allowing the system to use easily measurable audio parameters to infer video quality characteristics without directly measuring complex video stream metrics.

The patent creates a universal measurement framework that works across diverse video stream types by using audio quality metrics as a common denominator. The system designs audio-based measurement tools that can be applied universally to different video codecs, resolutions, and formats, providing a single unified approach that maintains measurement precision regardless of video stream diversity.

Instead of attempting to measure video quality directly across diverse formats, the patent inverts the approach by measuring audio quality (which has standardized metrics) and using those measurements to infer video quality. This reverse engineering approach bypasses the complexity of diverse video formats by working with the more standardized audio domain.