Guides

Avowed PC Optimized Settings: Benchmarks + Comparisons

Areej Syed Follow on X Send an email February 15, 2025

Avowed is out in advanced access on PC and Xbox Series consoles. Despite recent concerns, the game is an excellent RPG, combining Obsidian’s masterful writing with addictive gameplay. Avowed is based on the Unreal Engine 5, and uses Lumen and Nanite to deliver next-gen visuals. The PC port features hardware ray tracing for even higher-quality lighting at a decent cost. Here’s our optimization guide for Avowed.

Windows/System Settings to Optimize

  • Enable Resizable BAR.
  • Turn on Game Mode.
  • Enable Hardware-accelerated GPU Scheduling (HAGS) and Windowed Optimizations.
  • Use the Windows “High Performance” power profile and set your GPU power management mode to the same.
  • Disable Memory Integrity. Windows Menu->VBS->Device Security.
  • Ensure you use the proper XMP/EXPO memory profile (if available).
  • Overclock your GPU if you’re narrowly missing the 60 FPS mark.
  • Here’s a guide with more detailed instructions.

Avowed: PC System Requirements

Minimum

  • OS: 64-bit Windows 10/11 64-bit
  • Processor: AMD Ryzen 5 2600/Intel i5-8400
  • Memory: 16 GB
  • Graphics: AMD RX 5700/Nvidia GTX 1070/Intel Arc A580
  • Storage: 75 GB

Recommended

  • OS: 64-bit Windows 10/11
  • Processor: AMD Ryzen 5 5600X/Intel i7-10700K
  • Memory: 16 GB RAM
  • Graphics: Graphics: AMD RX 6800 XT/Nvidia RTX 3080
  • Storage: 75 GB

Avowed: Graphics & Resolution Scaling

Avowed shows drastic resolution-based performance scaling: from 39 FPS at 4K to 69 FPS at 1440p and 96 FPS at 1080p at the highest quality preset. That’s a 2.46x improvement going from 4K to 1080p.

  • Because Hardware Lumen doesn’t trace rays on a pixel level, enabling ray tracing reduces resolution-framerate scaling.
  • Avowed uses the Nanite geometric engine. Consequently, the level of detail increases with resolution, contributing to the performance scaling.

Test Setup

  • CPU: Intel Core i9-12900K @ 5.3 GHz.
  • Cooler: Arctic Liquid Freezer III 420.
  • GPU: NVIDIA RTX 4090 FE.
  • Motherboard: MSI PRO Z790-P WIFI.
  • Memory: 16 GB x2 @ 6000 MT/s CL30.

The graphics presets show ample performance scaling as well. The lowest quality preset is 80% faster than the highest, while hardware ray tracing is 10-12% slower than the latter.

Graphics Presets

Ray Tracing: Hardware Lumen

Ray tracing enables hardware-accelerated Lumen global illumination and reflections. It is usually 12-14% slower than software Lumen.

  • Steeper framerate drops are witnessed near dense grass and vegetation as they increase BVH complexity and traversal costs.
  • Interestingly, it uses less VRAM than SW Lumen.
RT on vs. Off

Ray tracing enables Hardware Lumen:

Related Articles
  • Like traditional ray tracing algorithms, it replaces the distance fields with BVH which is rebuilt every frame.
  • Polygon intersection testing offers notable quality upgrades over distance fields.
  • However instead of pixels, it operates on probes, surface cache texels, and tiles.

Hardware Lumen works well with skinned meshes (moving objects), and produces more detailed reflections.

Ray tracing also uses Far Field traces which significantly improves distant lighting. It extends the global illumination and reflection coverage to 1 km from the camera.

Ray tracing considerably improves the lighting and reflection quality over Lumen:

  • Reflections are more detailed, retaining surface-level detail and distant geometry.
  • Diffuse lighting is more accurate, especially in between objects.
  • Nearby: Higher-quality sun occlusion, especially prominent with vegetation.
  • Distant: Thorough illumination of far-off geometry on account of Far Field traces.
RT Off vs. On

View Distance & Shadow Quality

View Distance sets the render distance and LOD of the scene geometry. However, its performance and quality impacts are negligible. The Nanite geometry engine adjusts the LOD scaling.

Nanite is among the core highlights of Unreal Engine 5. It has allowed for an unprecedented increase in geometric detail without unrealistic polygon counts or memory budgets. This is achieved by adopting cluster-based LOD scaling.

Cluster groups of varying detail are generated for every mesh + Different mesh parts are also rendered at different LODs

The viewing angle or viewport determines the LOD to ensure that the highest perceptible detail is rendered for each part of the mesh. This involves using high-resolution cluster groups for some in-focus and low-resolution groups for distant or partially visible objects.

The LOD changes with the viewport. The updated cluster groups are streamed in and out of storage in real time.

The screen resolution is also used to determine the LOD as the subtlest details are often lost on small or low-resolution displays. These minute triangles are culled to save resources without a (noticeable) loss in detail.

Nanite pre-calculates all the cluster LOD hierarchies beforehand, storing them in the memory. The GPU accesses this data using DMA (Direct Memory Access) to avoid pipeline stalls or pop-ups.

Shadow Quality

Avowed uses Virtual Shadow Maps at medium quality and above. It’s a substantial performance (12%) and memory drain (700 MB), but produces high-resolution contact hardening soft shadows using Shadow Map Ray Tracing.

  • Low produces low-quality, blocky shadows, and should be avoided.
  • High and above-quality options increase shadow map samples, producing larger penumbras and detailed slimmer shadows.

Shadow Map Ray Tracing involves intersection testing against the virtual shadow map instead of actual geometry.

  • Shadow rays are cast from the from the surface towards the light source.
  • Along them, numerous samples are projected and tested against the shadow maps to produce soft shadowing and contact hardening.
  • The shadow ray distribution is based on the light radius or angle.

Shading & Effects Quality

Shading Quality adjusts lighting shaders, but we couldn’t observe any tangible differences between the various quality options.

  • It subtly impacts shadow boundaries, making them more visible for finer and distant objects.
  • The performance remains unaffected.

Effects Quality adjusts the quality and density of special effects, including fire, smoke, embers, and other energy-based particles.

  • The performance impact is trivial, ranging from 2-3% at the higher quality options.
  • The high-quality option is the sweet spot.
Effects Quality

Foliage & Post Processing

Foliage Quality adjusts the density and variety of foliage, mainly grass, bushes, flowers, etc.

  • The performance impact ranges from 4-5% and is higher with ray tracing due to the higher BVH traversal times.
  • It’s best left at the highest quality option.
Foliage Quality

Post Processing enables various late-stage pipelines, including tone mapping, color grading, bloom, and blur effects.

  • The performance impact is trivial.
  • Most emissive effects are disabled at medium and replaced by low-quality bloom at low.
Post Processing Quality

Reflections & Global Illumination: Lumen

Reflection Quality adjusts the reflection resolution. The performance impact is about the same as screen space reflections, ranging from 4-8%.

  • Epic produces the highest quality (barring RT).
  • High reduces ambient lighting quality and surface detail.
  • Low disables all but the nearest reflections.
Reflection Quality

Lumen Explained: Is it Ray Tracing?

Lumen, by default, uses software ray tracing. This implementation includes Screen Tracing, Mesh Distance Fields, and Global Distance Fields, each used on different sections of the scene. The Final Gather is resolved by the Skylight, combining atmospheric and local lighting.

What are distance fields?

Screen Tracing is the first step in the Lumen pipeline.

  • It is conducted against objects in the depth buffer or screen space.
  • It is primarily used for object boundaries and crevices as a higher quality SSAO replacement.
  • Objects missed are served by the distance fields.

Mesh Distance Fields are 3D representations of an object (or set of combined objects).

  • Each point in an MDF stores the nearest distance to an object surface within the volume.
  • This is computed offline, and allows skipping the empty space in the MDF when ray marching.
  • Ray marching is an optimized form of ray tracing used to calculate diffuse lighting.
    • You march along a ray’s path in small steps.
    • At each step, the distance to the closest surface is calculated using an MDF.
  • Shading is applied if a surface is detected in the ray’s proximity.
  • The amount of shading applied depends on the distance to the object.
  • Upon intersection, shadow, diffuse, and reflection rays are cast outwards towards light sources or probes.
Mip-maps: High resolution MDFs for nearer objects, and scaled down variants for the rest

Global Distance Fields are abstract volumes obtained by combining all the MDFs in the scene.

  • The result is a bare-bones geometrical representation with minimal per-object detail.
  • The GDFs are used for large-scale or “global” lighting.
  • GDFs are cached and updated only when required.

Surface Cache forms Lumen’s backbone:

  • It stores the material and lighting data for various surface points, called cards.
  • Upon intersection (see SDFs), the lighting at a point is referenced from this cache.
  • It is calculated, cached, and updated gradually over frames.
Up to 12 cards per mesh/object

Indirect Lighting is calculated using light probes placed in the scene. The distribution is scant (1 per 4×4 tile). For each texel, data is interpolated from the four closest probes and from previous frames.

The Final Gather backs the software ray-tracing results.

  • It is based on the Screen Space Radiance Cache. Instead of texels, light probes are placed on pixels, and their results are interpolated spatially and temporally.
  • A separate, low-resolution World Radiance Cache is used for distant lighting.

Areas with detailed geometry use a denser probe grid, and ambient occlusion is added to the temporally sampled lighting for a refined result.

Light tracing is optimized by prioritizing sections with luminance in the last frame.

Lumen Global Illumination produces highly optimized diffuse lighting. Even so, it can be quite taxing at the epic quality option, reducing performance by over 15%.

  • The epic-quality option enables Detail Traces which uses individual distance fields for each object/mesh.
    • It is only used for nearby objects.
    • Works well with complex geometry, especially vegetation and multi-layered surfaces.
  • The high-quality option offers a good quality-performance ratio.
  • The low-quality option switches to Distance Field and Screen Space ambient occlusion, disabling color bleeding and skylight.
Global Illumination Quality

Upscaling & Frame Generation

Avowed features DLSS 3.5, FSR 3, and TSR upscalers. GeForce RTX users are advised to upgrade to DLSS 4 using the “DLSS Override” option in the NVIDIA app. Upscaling offers substantial performance improvements in Avowed, with the quality mode being 57%, balanced 78%, and performance 2x faster than native.

DLSS vs FSR (Performance)

FSR 3 looks nearly as good as DLSS in this game, even at the performance preset. However, while it excels in geometry reconstruction, it fails to retain as much RT lighting information as DLSS 4. And that’s without ray reconstruction.

Avowed is limited to DLSS-based frame generation which is only pairable with DLSS upscaling (no DLAA). It offers a ~50% performance uplift over base upscaling at 4K.

Avowed: VRAM Usage

Avowed uses up to ~11 GB of VRAM at 4K “Epic” without upscaling. It’s the first game to use lower graphics memory with hardware ray tracing than without. This is due to Lumen’s caching systems which increase the VRAM usage when using software ray tracing.

This is good news for 8 GB GPU owners as Avowed maxes out at 8.75 GB at 4K with upscaling set to quality mode. The VRAM usage should stay well below 8 GB at 1440p and 1080p.

Texture Quality

Avowed: CPU Bottlenecks

Like most Unreal Engine 5 titles, Avowed is completely GPU-bound with a busy deviation of <2% or lower across all scenarios.

1080p Epic

Avowed: Performance Summary

  • Ray Tracing: If you are running an RTX 3060 Ti or below, disable ray tracing for a 15% FPS bump.
  • Shadow Quality: Unreal 5’s Virtual Shadows can be quite taxing. If you’re short of your FPS target, we recommend switching to high.
  • Reflections: Lumen Reflections are quite optimized, but can strain older GPUs. Stick to high quality for a balanced experience.
  • Global Illumination: Lumen GI uses MDFs for detailed close-range lighting. However, it’s an FPS drain. Stick to high for the best performance.
  • Foliage Quality: Consider reducing this to high or medium if facing FPS drops near dense grass patches.

Best Graphics Settings for Avowed PC

Optimized SettingsHigh-endMidrangeLow-end
Resolution4K (3840×2160)1440p (2560×1400)1080p (1920×1080)
Target FPS75 FPS120 FPS60 FPS90 FPS+60 FPS60 FPS
Ray TracingOnOnOnOnOffOn
View DistanceEpicEpicEpicEpicEpicEpic
Shadow QualityEpicEpicEpicEpicHighEpic
Shading QualityEpicEpicEpicEpicEpicEpic
Texture QualityEpicEpicEpicEpicEpicEpic
Effects QualityEpicEpicEpicEpicEpicEpic
Post ProcessingEpicEpicEpicEpicEpicEpic
Foliage QualityEpicEpicEpicEpicEpicEpic
Reflection QualityEpicEpicEpicEpicEpicEpic
Global IlluminationEpicEpicEpicEpicHighEpic
Upscaling (DLSS/FSR)BalancedPerformBalancedPerformBalancedBalanced
Frame GenerationOffOnOffOnOffOn
Refer.High-end (4K)Mid-range (1440p)Low-end (1080p)
CPUCore i7-13700K|Ryzen 7 7700XCore i5-12600K|Ryzen 5 7600 Core i5-12400
AMD Ryzen 5 3600
GPUGeForce RTX 4090GeForce RTX 4070 SuperRTX 3060|RTX 4060
Memory32GB (dual-channel)16GB (dual-channel)Less than: 16GB (dual-channel)
Optimized SettingsRTX 4090RTX 4080 SRTX 4070 TiRTX 4070 SRTX 4070
Resolution4K4K4K1440p4K1440p4K1440p4K1440p
Target FPS75 FPS120 FPS60 FPS90 FPS65 FPS75 FPS60 FPS70 FPS60 FPS60 FPS
Ray TracingOnOnOnOnOnOnOnOnOnOn
View DistanceEpicEpicEpicEpicEpicEpicEpicEpicEpicEpic
Shadow QualityEpicEpicEpicEpicEpicEpicEpicEpicEpicEpic
Shading QualityEpicEpicEpicEpicEpicEpicEpicEpicEpicEpic
Texture QualityEpicEpicEpicEpicEpicEpicEpicEpicEpicEpic
Effects QualityEpicEpicEpicEpicEpicEpicEpicEpicEpicEpic
Post ProcessingEpicEpicEpicEpicEpicEpicEpicEpicEpicEpic
Foliage QualityEpicEpicEpicEpicEpicEpicEpicEpicEpicEpic
Reflection QualityEpicEpicEpicEpicEpicEpicEpicEpicEpicEpic
Global IlluminationEpicEpicEpicEpicEpicEpicEpicEpicEpicEpic
Upscaling (DLSS)BalancedPerformBalancedPerformBalancedBalancedBalancedBalancedPerformBalanced
Frame GenerationOffOnOffOnOnOffOnOffOnOff

Avowed: Best Settings for Low-end PC

Avowed runs extremely well on low-end builds, courtesy of well-implemented upscaling and frame generation. FSR 3-based frame generation can be enabled by adding the following lines of code to the “engine.ini” file located in C\Users\AppData\Local\Alabama\Saved\Config\Windows (or WinGDK):

[SystemSettings]

r.FidelityFX.FI.Enabled=1

r.FidelityFX.FSR3.Enabled=1

r.FidelityFX.FSR3.UseNativeDX12=1

Optimized SettingsRTX 3060RTX 3060 TiRTX 4060
Resolution1080p1080p1080p
Target FPS60 FPS75 FPS70 FPS
Ray TracingOnOnOn
View DistanceEpicEpicEpic
Shadow QualityEpicEpicEpic
Shading QualityEpicEpicEpic
Texture QualityEpicEpicEpic
Effects QualityEpicEpicEpic
Post ProcessingEpicEpicEpic
Foliage QualityEpicEpicEpic
Reflection QualityEpicEpicEpic
Global IlluminationEpicEpicEpic
Upscaling (DLSS/FSR)QualityQualityQuality
Frame GenerationOnOnOn

Areej Syed Follow on X Send an email February 15, 2025

Areej Syed

Processors, PC gaming, and the past. I have been writing about computer hardware for over seven years with more than 5000 published articles. Started off during engineering college and haven't stopped since. Find me at HardwareTimes and PC Opset.
Back to top button