Machine-Learning Video Format Selection for Display-Matched Quality

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing content sharing platforms inefficiently utilize computational resources to calculate quality scores for video formats, leading to suboptimal format selection that wastes bandwidth without improving perceptual quality, especially in large-scale systems with millions of video uploads.

Innovation Solution

Training a machine learning model using historical videos to predict quality scores for transcoded versions at various resolutions and display configurations, allowing efficient optimization of format selection without extensive computational overhead.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If computational resources are used to calculate quality scores for all video formats, then format selection accuracy is improved, but computational resource consumption increases

Engineering Contradiction:

Improvequality score accuracyVSAvoidcomputational resource consumption

Core Design Contradiction:

Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The system pre-calculates and stores quality scores for multiple video formats in advance during video processing. These pre-computed quality scores are then used during playback to determine the optimal format without requiring real-time computational analysis, thus reducing computational resource consumption while maintaining accuracy.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

Instead of calculating quality scores from scratch for each video format, the system uses pre-computed quality score data structures that replicate the necessary quality information. This allows the system to access quality metrics without performing the full computational analysis again, reducing energy consumption while preserving measurement precision.

Inventive Principle:

Principle #26Copying

2Reliability

If multiple video formats are served to maximize quality options, then viewer quality experience is improved, but bandwidth consumption increases

Engineering Contradiction:

Improvequality experienceVSAvoidbandwidth consumption

Core Design Contradiction:

ReliabilityVSLoss of energy

Solution Approach 1:

The system dynamically adjusts video format parameters (resolution, codec, bitrate) based on predicted quality scores and viewer conditions. By changing parameters adaptively rather than serving all possible formats, the system optimizes quality experience while reducing unnecessary bandwidth consumption for formats that would not improve perceived quality.

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

Instead of providing all possible video formats (excessive action), the system selectively provides only the necessary formats based on quality predictions and viewer capabilities (partial action). This reduces bandwidth consumption by excluding formats that would not contribute to improved quality experience.

Inventive Principle:

Principle #16Partial or excessive action

3Measurement precision

If traditional quality scoring methods are used for all videos, then quality assessment is improved, but processing time increases

Engineering Contradiction:

Improvequality assessment accuracyVSAvoidprocessing time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

Quality scores are calculated in advance during the video processing pipeline before the video is made available for playback. This preliminary computation allows the video to be ready for immediate playback without requiring time-consuming quality analysis at the moment of viewing, thus reducing processing time while maintaining assessment accuracy.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system extracts and stores only the essential quality metric data from comprehensive quality analyses, separating the necessary quality information from the full computational process. This extraction allows rapid retrieval of quality scores during playback without repeating the entire quality assessment computation, reducing processing time while preserving measurement precision.

Inventive Principle:

Principle #2Taking out (Extraction)

Data Source

PatentUS20250337969A1Optimal format selection for video players based on predicted visual quality using machine learning

Publication Date: 2025.10.30 GOOGLE LLC

AI summary

A system and methods are disclosed for optimal format selection for video players based on visual quality. The method includes determining, based on sampled frames of a video, a plurality of quality scores for the video, wherein each quality score is associated with a corresponding parameter combination that includes a corresponding video format, a corresponding transcoding configuration, and a corresponding display resolution. The method further includes identifying, among the plurality of quality scores, a first quality score that is associated with a first parameter combination, which includes a first display resolution matching a display resolution of a client device, and causing a first video format to be selected for the client device using the first parameter combination that includes the first display resolution matching the display resolution of the client device.