Phantom Blade Zero PC Settings: What to Set, and What Each One Costs
The honest version of «best settings» for this game: the preset ladder is measured, ray tracing and path tracing are measured, and the twenty-odd sub-options below them are not — Game8 recommends Very High for every one of them, and no public source says what each costs on its own. So the values here are sourced, the reasoning is sourced, and the unknowns are labelled.
- 25
- settings covered
- 4
- groups
- 19
- items marked unverified
- 25
- entries
Set these first, in this order
- Resolution and upscaling decide more than any single quality option. On an RTX 5090 at 4K Ultra, measured runs go 103 fps at Low, 99 at Medium, 85 at High, 72 at Very High and 67 at Ultra — the step from High to Very High costs about 15 percent, while Ultra costs only another 6–7 percent on top of Very High.
- Ray tracing is the second big lever. On an RTX 5090 at 4K, enabling RT drops frame rate 20–22 percent; path tracing on top drops another 35–40 percent, roughly 50 percent combined. A separate RTX 5080 run measured RT at about 10 percent and path tracing at about twice the RT cost, with balanced upscaling halving the RT loss.
- Verify with the frame counter in one fixed spot: same scene, same preset, toggle one thing at a time. The official notes never publish frame numbers, so any claim about «this setting gave me +20» is the player's own measurement, not an official figure.
Upscaling and latency (settle these first)
Game8's PC table sets the upscaler to DLSS, the mode to Quality, frame generation Off and Reflex to Auto; the official patch notes add that Reflex on Auto now works with V-Sync enabled.
| Setting | Suggested value | What it moves | Why, and what is unverified |
|---|---|---|---|
| Super Resolution Method | DLSS | Frame rate, image clarity | Game8's stated reason is that DLSS gives the best mix of performance and clarity. Hardware caveat: multi-frame generation needs RTX 50-series with fifth-gen Tensor Cores, so 30-series cannot use it; AMD cards use a different upscaler. |
| DLSS Super Resolution Mode | Quality | Frame rate, clarity | Game8 recommends Quality. For scale: Balanced on an RTX 5080 measured +64%, +46% and +34% at 4K/2K/1080p — the higher the resolution the bigger the gain. A figure for Quality specifically is not published (unverified). |
| Dynamic Resolution Quality | 60 | Frame-rate stability | Game8's PC table lists 60 without explaining the value; its write-up does not cover this option separately (unverified: whether 60 is a frame-rate floor or a scaling threshold is not stated). |
| Super Resolution Output Scale | Auto | Frame rate, clarity | Game8 lists Auto with no elaboration. The manual range and the frame cost of each step are not public (unverified). |
| Frame Generation | Off (Game8's pick) | Frame rate, input latency | Game8 recommends Off. The size of the gain is measured: on an RTX 5080, 4K path tracing went from 41 to 194 fps (+373 percent claimed) and 2K Balanced from 102 to 429 (+320). Multi-frame generation needs RTX 50-series and generated frames add latency; the exact millisecond cost is not published (unverified). |
| NVIDIA Reflex | Enable Automatically | Input latency | Game8 recommends Auto. The November 4 patch specifically fixed this combination: Reflex on Auto previously did nothing with V-Sync on, and now works. No latency figure in milliseconds was published (unverified). |
| V-Sync | Depends on the display (no single value given) | Tearing, input latency | Neither source gives one recommended value (unverified). What is certain: after the patch Reflex works with V-Sync on, so «no tearing and low latency» is no longer a trade-off — the only conclusion the patch notes state. |
Ray tracing and the quality preset (the second lever)
Game8's PC table turns ray tracing off and sets the quality preset to Very High; the official specs give only one path-tracing target — 4K/60 on an RTX 5080, assuming upscaling is on.
| Setting | Suggested value | What it moves | Why, and what is unverified |
|---|---|---|---|
| Ray Tracing | Off (Game8's pick) | Frame rate (and VRAM) | Measured cost: 4K on an RTX 5090 loses 20–22 percent with RT on; a separate RTX 5080 run measured about 10 percent, halved with Balanced upscaling. The recommendation is a default, not a verdict for your specific card. |
| Path Tracing | Unverified (official 4K/60 target only for RTX 5080) | Frame rate (largest single drop) | Path tracing on top of RT adds another 35–40 percent drop on an RTX 5090 at 4K, roughly 50 percent combined; on the RTX 5080 run its cost was about twice RT's. Of the official four tiers only the path-tracing tier (RTX 5080, 4K/60) has a target; other cards have no official guidance. |
| Quality Preset | Very High | Frame rate, detail | Game8 recommends Very High, and the measurements support it: 4K on an RTX 5090 runs 85 fps at High, 72 at Very High (about −15 percent) and 67 at Ultra (only another −6–7). Very High is where the visible step ends and Ultra's return gets small. |
Detail options (Game8 says Very High across the board)
For these rows the only source is Game8's recommended value; no public data breaks down per-option frame cost. Whether to lower one is a test you run yourself in a fixed spot.
| Setting | Suggested value | What it moves | Why, and what is unverified |
|---|---|---|---|
| Shadow Quality | Very High | Frame rate, VRAM | Game8's value; per-option cost is not public (unverified). The general trade-off is shadow resolution and cascade count, but no source describes this game's tier differences. |
| Reflection Quality | Very High | Frame rate | Game8's value; per-option cost is not public (unverified). |
| Lighting Quality | Very High | Frame rate | Game8's value; per-option cost is not public (unverified). The measured preset step from Medium to High was attributed mainly to lighting, textures and model detail — useful context. |
| Effects Quality | Very High | Frame rate (more sensitive in heavy combat) | Game8's value; per-option cost is not public (unverified). |
| Material Quality | Very High | Frame rate, VRAM | Game8's value; per-option cost is not public (unverified). |
| Ambient Occlusion Quality | Very High | Frame rate | Game8's value; per-option cost is not public (unverified). |
| View Distance | Very High | Frame rate, CPU load | Game8's value; per-option cost is not public (unverified). View distance usually costs CPU time, but no public test covers it for this game. |
| High-Quality Hair | On | Frame rate, visuals | Game8 recommends On. Related: the November 4 patch fixed incorrect hair rendering on some AMD cards with certain drivers — if hair still looks wrong after updating, check the driver version first (the notes do not name versions; unverified). |
Display, feel and periphery
Most of these move comfort and controls rather than frame rate; the controller row comes from Game8's recommendation and the official post's talk of layout presets.
| Setting | Suggested value | What it moves | Why, and what is unverified |
|---|---|---|---|
| Motion Blur | Off | Clarity, motion sickness | Game8's reason: this is a high-motion action game, and turning motion blur off helps avoid motion sickness. Note this is a different control from the Motion Blur Quality slider. |
| Camera Shake | 50 | Comfort, motion sickness | Game8 lists 50 without a stated reason (unverified); like most comfort options it can be lowered to taste and does not affect performance. |
| Film Grain Intensity | 2 | Image texture, clarity | Game8 lists 2 (a low setting) with no stated reason (unverified). |
| Sharpness | Player preference | Clarity (no frame cost) | Game8 lists this as player preference with no number — there is no accepted optimum here. |
| Controller Layout (PS5 pad on PC) | Controller Preset 2 | Controls (no frame cost) | Game8's reason: Preset 2 moves dash from R1 to O and execution from O to R1, closer to the usual Souls-like layout and easier to reach. The official post mentions three layouts on PS5 while Game8 describes two default schemes — the counts differ (unverified); trust the in-game list. |
| NVIDIA driver version | 617.42 (official advice) | Stability, frame rate | The October 28 post: 621.01 showed unstable performance and 617.42 tested best; anything newer than 617.42 should be rolled back with a clean install until the next driver. A clean install wipes custom display settings, so note them first. |
| Controller connection (PC) | Prefer wired or 2.4 GHz; Bluetooth improved | Input stability | The November 4 patch reduced stutter with Bluetooth controllers, but the notes only say «reduced» (unverified whether input loss remains). If it still stutters, wired or a 2.4 GHz dongle is the direct next step. |
Related pages
Source: Recommended values come from Game8's «Recommended Settings» page for Phantom Blade Zero (updated October 31, 2026). Frame costs and preset deltas come from this site's PC benchmark write-up, which reports Wccftech's 15-GPU run (October 27) and a separate RTX 5080 sweep (October 30): 4K preset ladder 103/99/85/72/67 fps for Low to Ultra, RT at −20–22 percent on RTX 5090, path tracing a further −35–40 percent. Driver and Bluetooth advice come from S-GAME's October 28 post and the November 4 patch notes. Where no source states a reason or an optimum, the cell says unverified instead of guessing.