SoftMaskPro – Unity UI Rendering Performance Study

Overview

This project is an empirical performance analysis of soft masking in Unity UI under controlled load conditions.

Rather than modifying a UI tool superficially, this repository builds a reproducible profiling harness to measure the real CPU and GPU cost of soft masking in modern Unity (URP).

The focus is deterministic bottleneck isolation, overdraw amplification analysis, and frame-time stability under UI stress.

---

Objective

Soft masking in Unity UI introduces additional fragment processing and can significantly amplify overdraw in layered interface scenarios.

This project investigates:

• How soft masks affect fragment cost
• Fill-rate saturation under layered transparency
• Canvas rebuild CPU spikes
• Early-Z limitations in UI rendering
• Dirty mask optimization effects
• Frame-time variance under stress

The goal is to quantify these effects using profiler-validated measurements rather than assumptions.

---

Test Environment

• Unity 6
• Universal Render Pipeline (URP)
• Desktop profiling
• Mock HMD configuration (for XR-aligned testing conditions)
• MSAA configurable (disabled/enabled for comparison)

All tests are reproducible and isolated from unrelated systems.

---

Experimental Design

The project includes controlled scenes designed to isolate specific variables:

1. Baseline UI Scene

Minimal UI elements without soft masking to establish baseline CPU and GPU frame times.

2. SoftMask Stress Scene

Multiple layered masked elements to simulate worst-case overdraw amplification.

Measured:

• GPU frame time
• Draw calls
• SetPass calls
• Overdraw heatmap behavior

3. Dynamic Layout Stress

Frequent UI invalidation to measure:

• Canvas rebuild spikes
• Main thread CPU impact
• Layout recalculation cost

4. Dirty Mask Optimization Validation

Testing selective mask invalidation to reduce unnecessary redraw cost.

---

Measurement Methodology

All measurements were captured using:

• Unity Profiler (CPU + GPU timeline)
• Frame Debugger
• Overdraw visualization mode
• Deterministic test toggles

Each experiment isolates one variable at a time to avoid cross-system noise.

Frame time comparisons are reported in milliseconds, not FPS, to preserve accuracy.

---

Key Observations

• Soft masking increases fragment cost under heavy overlap scenarios.
• Overdraw scales non-linearly with stacked transparency.
• Canvas rebuild spikes dominate CPU cost when layout invalidation is frequent.
• UI rendering bypasses many early-Z optimizations, amplifying fill-rate pressure.
• Optimization via controlled invalidation reduces CPU spikes but does not eliminate fragment cost.

---

Architectural Notes

The project is structured as a controlled Unity harness:

SoftMaskPro-Performance-Study/
├── Assets/
├── Packages/
├── ProjectSettings/
├── README.md
└── LICENSE

All unnecessary upstream repository artifacts were removed to maintain clarity and authorship integrity.

---

License Notice

This project contains modified components derived from SoftMaskForUGUI (MIT License).

Original Author: Coffee

All performance experiments, optimization validation, profiling harness design, and analysis are authored by:

James De Raja Senior Real-Time Performance Engineer

---

Why This Project Exists

Most UI performance discussions rely on anecdotal experience.

This repository demonstrates:

• Measured engineering
• Controlled experimentation
• Deterministic bottleneck isolation
• Systems-level thinking in rendering pipelines

This project complements the XR Performance Stress Lab by focusing specifically on UI fill-rate and masking behavior.

---

Author

James De Raja Senior Real-Time Performance Engineer Unity Rendering | Frame Pacing | XR Optimization | CPU/GPU Bottleneck Analysis Website Portfolio