Start with Frame Generation off and compare native resolution with FSR Quality in the same DOOM checkpoint. Then test generated output separately. This gives you a clearer way to choose settings after Update 5 without treating a larger FPS counter as proof of faster controls.
What Update 5 adds
DOOM: The Dark Ages Update 5, released on 16 November 2026, adds FSR Upscaling 4.1 and Frame Generation 4 on compatible GPUs. It also enables FSR frame generation on Intel GPUs that meet the required hardware and driver capabilities. The patch notes do not provide a complete per-card compatibility list or a measured FPS uplift.
RX 9000 owners are a useful audience for this comparison: AMD's current ML Frame Generation requires RX 9000-series hardware or newer and Windows 11. Upscaling and frame generation have separate requirements. An FSR option appearing on an older Radeon or Intel card does not establish that it is using the same ML frame generator.
Start with the updated game and a compatible driver for your exact GPU. Record the build, driver version, operating system and the settings actually offered. This guide proposes tests; it contains no HD OPTI benchmark or measured latency result.
FSR 4.1 on Radeon RX 6000: Support, Setup and Quality Testing
Native integration and native resolution mean different things
Native integration means the game includes the feature officially. Native resolution means rendering at the chosen output resolution. Update 5's official integration lets compatible systems test the new features without adding a third-party upscaler modification; it does not mean FSR Quality renders every pixel at native resolution.
FSR upscaling reconstructs the output from a lower-resolution input and has its own processing cost. Quality uses a standard 1.5× scale per dimension, Balanced 1.7× and Performance 2×. At 4K output, Quality starts from 2560 × 1440; at 1440p output, it starts from approximately 1707 × 960. Actual dimensions can be rounded by the game.
Use Quality as your first comparison because it retains more input pixels than Balanced or Performance. That is a starting recommendation, not a verdict on DOOM image quality. If Native AA is offered, label it separately: it uses full-resolution input and still applies reconstruction.
Build a 1440p or 4K comparison
Run the following sequence separately at each output resolution you use. Skip options unavailable on your system. Keep the graphics preset, sharpening, display mode, refresh rate, frame cap, V-Sync and VRR state fixed. Disable dynamic resolution for the comparison if the game offers it, so changing internal resolution does not obscure the test.
| Run | Configuration | Question |
|---|---|---|
| A | Native-resolution baseline; Frame Generation off | What picture and rendered performance are you starting from? |
| B | FSR Quality; Frame Generation off | Does reduced rendering resolution help at acceptable clarity? |
| C | FSR Balanced; Frame Generation off | Is the extra compromise worthwhile? |
| D | Preferred upscaling mode; Frame Generation on | Does generated output improve motion while response remains acceptable? |
- Save your existing settings. Choose a repeatable campaign checkpoint and a short route with both a steady camera pan and fast combat.
- Warm up that route, then capture the same duration for A, B and C. Repeat each condition at least three times and retain the individual results.
- Compare D with the same upscaling mode and generation off. If enabling it forces another setting to change, record that dependency.
- Return to the earlier configuration to check whether the difference repeats. Treat small changes within the spread between runs as inconclusive.
Slaughter Mode includes randomised encounters, so two runs can present different workloads. Use it to check your chosen profile under real combat pressure, but keep those observations separate from the controlled checkpoint comparison. Record the arena and round instead of combining every run into one average.
Read FPS, frame times and response separately
Frame Generation creates intermediate images between application-rendered frames. A higher displayed frame rate therefore does not establish a matching increase in rendered FPS or faster input response. AMD recommends starting from at least 60 FPS before interpolation for its general Frame Generation guidance; this is not a measured DOOM-specific threshold.
With generation off, record average FPS, 1% lows and the frame-time trace using the same capture method. With generation on, verify what each counter measures. AMD's SDK notes identify its PresentMon Frame Generation integration as DirectX 12 only. Do not assume a Vulkan capture separates rendered and generated frames; leave unsupported metrics unreported.
Inspect quick turns past nearby geometry, weapon edges, particles and the HUD. Compare the same movements with generation off and on, looking for lost detail, unstable edges or trails. These are proposed inspection points, not confirmed Update 5 defects.
Keep perceived mouse response in a separate notes column. Quantifying input-to-display latency requires an appropriate measurement setup; neither the average FPS nor a 1% low supplies that measurement. Keep generation only if its motion benefit suits you and the controls remain acceptable.
Bodycam PC Settings: Compare FSR, Frame Generation and Response
Compare older FSR only when you can identify it
The official patch notes do not establish that Update 5 offers an older-FSR selector. If your installation provides a supported choice, compare it at the same quality mode with generation off and record both implementations. If you cannot confirm which version is active, describe the settings you can verify and avoid calling it an FSR 3-versus-4.1 benchmark.
Earlier-build recordings can provide context, but a before-and-after patch comparison changes more than the upscaler. Update 5 also changes weapon behaviour and queued swap inputs. Differences in fight duration or how controls behave cannot be assigned to FSR alone.
| Observation | Next decision |
|---|---|
| Quality improves repeatable performance and preserves useful detail | Keep it as the baseline for the separate generation test. |
| Balanced loses distracting detail during turns | Return to Quality or the native-resolution baseline. |
| Lower input resolution barely changes FPS | Check whether the fixed frame cap masks headroom before diagnosing a hardware limit. |
| Displayed FPS rises but motion or controls become distracting | Return to generation off at the same upscaling mode. |
| Hitches persist despite a higher average | Compare the frame-time traces and retain the hitching as an unresolved issue. |
FAQ
Update 5 officially adds FSR Upscaling 4.1 and Frame Generation 4 on compatible GPUs. The comparison here uses that game integration and does not require a third-party upscaler modification. Confirm your GPU, driver and available game options first.
The patch notes confirm FSR frame generation on qualifying Intel GPUs but do not identify it as the RX 9000 ML path. AMD documents ML Frame Generation for RX 9000-series hardware or newer, with a separate analytical fallback. Do not equate the two from a menu label.
Start by comparing Quality against your native-resolution baseline with generation off. Quality retains more input pixels than Balanced at the same output resolution. Test Balanced next and keep it only if its repeatable performance benefit is worth the change in moving detail on your system.
Only if your installation exposes a supported choice and you can identify both implementations. The inspected patch notes do not confirm such a selector. Comparing different game builds also includes other patch changes, so it cannot isolate the upscaler by itself.




