What is visual noise in Percy?
Visual noise refers to visual differences that do not represent meaningful UI changes. These include shifts caused by dynamic content, animations, timestamps, ads, or minor rendering variations. Noise increases review time and reduces review efficiency.
Percy offers multiple features to help you reduce visual noise and focus on meaningful changes, including:
How can Percy AI help reduce noise?
Percy AI automatically classifies visual changes as meaningful or noisy based on learned review behavior.
Enable Percy AI to reduce manual effort and speed up reviews. This helps you:
* Focus reviews on real UI changes
* Automatically de-prioritize trivial diffs
* Improve your Review Efficiency Score over time
How does Diff Sensitivity affect noise?
Diff Sensitivity controls how strict Percy is when detecting visual differences.
* Lower sensitivity detects even small changes and may increase noise
* Higher sensitivity ignores minor pixel differences and reduces noise
Adjust Diff Sensitivity when you see frequent low-value diffs that do not impact the UI experience.
When should you use Ignore Regions?
Use Ignore Regions to exclude specific areas of the page from visual comparison.
Ignore Regions work best for:
* Dynamic content (timestamps, counters, rotating banners)
* Third-party widgets or ads
* User-specific or environment-specific data
By ignoring unstable regions, Percy focuses comparisons on stable, meaningful UI elements.
Should I use multiple noise-reduction techniques together?
Yes. The best results come from combining:
* Percy AI for intelligent diff classification
* Diff Sensitivity tuning for global control
* Ignore Regions for targeted stabilization
Together, these approaches reduce review noise and improve overall build health.
How do I know if noise is reduced?
You can track improvements through:
* A higher Review Efficiency Score
* Fewer snapshots marked as noisy
* Reduced manual review time in the Percy Insights Dashboard
