Achieving Realistic AR Graphics with PBR Materials

- Achieving believable AR visuals requires objects that respond to real-world lighting and occlusions. None of the current AR SDKs provide perfect light estimation—ARKit and ARCore deliver environment probes that are approximations at best. - The illusion breaks if materials are flat, overly shiny

Our competencies

Other studio services

Frequently Asked Questions

Latest works

  • image_games_mortal_motors_495_0.webp
    Game development for Mortal Motors
    1504
  • image_games_a_turnbased_strategy_game_set_in_a_fantasy_setting_with_fire_and_sword_603_0.webp
    A turn-based strategy game set in a fantasy setting, With Fire and Sword
    1005
  • image_games_second_team_604_0.webp
    Game development for the company Second term
    633
  • image_games_phoenix_ii_606_0.webp
    3D animation - teaser for the game Phoenix 2.
    716
  • Achieving believable AR visuals requires objects that respond to real-world lighting and occlusions. None of the current AR SDKs provide perfect light estimation—ARKit and ARCore deliver environment probes that are approximations at best.
  • The illusion breaks if materials are flat, overly shiny, or fail to occlude. None of these issues are trivial; they demand careful PBR configuration.
  • Common pitfalls include: None of the default shaders handle dynamic lighting well; None of the lighting estimates from devices are accurate for all conditions; None of the reflection probes capture full environment details due to limited FOV.
  • To improve realism: Use neutral Albedo (RGB around 128) to avoid color cast. Set Roughness between 0.3 and 0.6—values outside this range cause unnatural reflections. Limit Metallic to 0.7–0.85. Apply Normal Maps with moderate intensity (0.5–1.0). Enable emission to ground objects in dark scenes.
  • For Reflection Probes, enable AREnvironmentProbeManager with automatic placement and Trilinear filtering. None of these probes are perfect; they produce blurry cubemaps that work best with moderate Roughness materials.
  • Light estimation from AR foundation is limited: None of the APIs provide direct luminance values; Instead, they offer ambient intensity and color temperature, which are not enough for accurate high-dynamic-range rendering.
  • Occlusion handling: None of the standard PBR pipelines include occlusion; you must implement custom shaders or use AR Foundation's human occlusion if supported.
  • Testing on multiple devices is crucial: None of the simulators replicate real camera and lighting conditions.
  • For custom solutions, contact our engineers. None of the generic settings work for all scenarios; we tailor materials to your specific use case.
  • In summary, AR PBR requires iterative tuning. None of the parameters are independent; adjusting one affects perception of others. Start with baseline values, then refine based on device camera and environment.