Refresh rate: Difference between revisions
Xinreality (talk | contribs) No edit summary |
Xinreality (talk | contribs) No edit summary |
||
| (One intermediate revision by the same user not shown) | |||
| Line 52: | Line 52: | ||
While OLEDs have a clear advantage in response time, manufacturing complexities historically meant that high-end LCD panels could sometimes achieve higher refresh rates in a given product generation. For instance, the LCD-based [[Valve Index]] supports up to 144 Hz, which surpassed many contemporary OLED headsets.<ref name="steamoled" /> However, this technological gap is closing. Modern fast-switching LCDs have become highly effective for VR, while new [[Micro-OLED]] displays are pushing the boundaries of resolution and efficiency at high refresh rates. | While OLEDs have a clear advantage in response time, manufacturing complexities historically meant that high-end LCD panels could sometimes achieve higher refresh rates in a given product generation. For instance, the LCD-based [[Valve Index]] supports up to 144 Hz, which surpassed many contemporary OLED headsets.<ref name="steamoled" /> However, this technological gap is closing. Modern fast-switching LCDs have become highly effective for VR, while new [[Micro-OLED]] displays are pushing the boundaries of resolution and efficiency at high refresh rates. | ||
This convergence has led to market diversification rather than a single "winner." High-refresh-rate LCDs are common in the mainstream consumer market ( | This convergence has led to market diversification rather than a single "winner." High-refresh-rate LCDs are common in the mainstream consumer market (for example [[Meta Quest 2]], [[Meta Quest 3]], [[Valve Index]]), offering a strong balance of performance and cost. Meanwhile, high-end Micro-OLEDs are featured in premium, next-generation devices (for example [[Apple Vision Pro]], Bigscreen Beyond) where ultimate contrast, color, and form factor are prioritized over cost.<ref name="panoxmicro" /> | ||
Other trade-offs influence the choice as well. OLEDs provide perfect blacks and infinite contrast, which greatly enhances immersion in dark environments. LCDs, on the other hand, can often achieve higher peak brightness and may use a full [[RGB]] subpixel layout that can reduce the [[screen-door effect]] compared to the [[PenTile]] subpixel arrangement often found in OLED panels.<ref name="steamoled" /> | Other trade-offs influence the choice as well. OLEDs provide perfect blacks and infinite contrast, which greatly enhances immersion in dark environments. LCDs, on the other hand, can often achieve higher peak brightness and may use a full [[RGB]] subpixel layout that can reduce the [[screen-door effect]] compared to the [[PenTile]] subpixel arrangement often found in OLED panels.<ref name="steamoled" /> | ||
| Line 82: | Line 82: | ||
==Importance in VR and AR== | ==Importance in VR and AR== | ||
In VR and AR, refresh rate is a critical factor affecting user comfort and the sense of [[presence]] | In VR and AR, refresh rate is a critical factor affecting user comfort and the sense of [[presence]]: the psychological feeling of "being there" in the virtual environment. A higher refresh rate results in lower latency between frames, leading to smoother motion and reduced visual artifacts.<ref name="kommando" /> VR and AR devices typically require high refresh rates to maintain a comfortable and immersive experience. | ||
===Visual Comfort and Presence=== | ===Visual Comfort and Presence=== | ||
| Line 111: | Line 111: | ||
* '''72 Hz:''' Acceptable for OLED displays with low persistence | * '''72 Hz:''' Acceptable for OLED displays with low persistence | ||
* '''90 Hz:''' Industry minimum standard | * '''90 Hz:''' Industry minimum standard | ||
* '''120 Hz:''' Optimal | * '''120 Hz:''' Optimal threshold, reduces nausea incidence by approximately half compared to 60 Hz<ref name="antaeus" /> | ||
In one controlled study, a VR forklift training simulator achieved only a 40% completion rate at 60 fps (average Simulator Sickness Questionnaire score: 54), while optimization to 90 Hz increased completion to 95% (average SSQ score: 8).<ref name="antaeus" /> | In one controlled study, a VR forklift training simulator achieved only a 40% completion rate at 60 fps (average Simulator Sickness Questionnaire score: 54), while optimization to 90 Hz increased completion to 95% (average SSQ score: 8).<ref name="antaeus" /> | ||
| Line 119: | Line 119: | ||
Higher refresh rates reduce motion sickness through several interconnected mechanisms: | Higher refresh rates reduce motion sickness through several interconnected mechanisms: | ||
'''Visual-Vestibular Conflict Reduction:''' VR displays visual motion without corresponding [[vestibular system]] (inner ear) signals. The brain detects this mismatch and triggers a "poison response" | '''Visual-Vestibular Conflict Reduction:''' VR displays visual motion without corresponding [[vestibular system]] (inner ear) signals. The brain detects this mismatch and triggers a "poison response", nausea evolved to expel neurotoxins causing sensory confusion. Higher refresh rates minimize this conflict by reducing temporal gaps between visual updates and actual head position.<ref name="weech" /> | ||
'''Reduced Prediction Error:''' The brain constantly predicts sensory input based on internal models. Large mismatches between predicted and actual input trigger discomfort. Higher refresh rates minimize temporal gaps that create prediction errors. [[EEG]] studies show motion sickness correlates with increased delta/theta/alpha band power (6-12 Hz) in brain activity.<ref name="nurnberger" /> | '''Reduced Prediction Error:''' The brain constantly predicts sensory input based on internal models. Large mismatches between predicted and actual input trigger discomfort. Higher refresh rates minimize temporal gaps that create prediction errors. [[EEG]] studies show motion sickness correlates with increased delta/theta/alpha band power (6-12 Hz) in brain activity.<ref name="nurnberger" /> | ||
| Line 127: | Line 127: | ||
==Motion-to-Photon Latency== | ==Motion-to-Photon Latency== | ||
[[Motion-to-Photon Latency]] (MTP) is the total time elapsed from when a user initiates a movement ( | [[Motion-to-Photon Latency]] (MTP) is the total time elapsed from when a user initiates a movement (for example turning their head) to the moment the corresponding change in the virtual world is fully illuminated on the display. It is arguably the single most important metric for user comfort in VR.<ref name="unitymtp" /> This latency is a cumulative result of several stages in the VR pipeline: | ||
# Sensor sampling (tracking head/controller position) | # Sensor sampling (tracking head/controller position) | ||
| Line 163: | Line 163: | ||
===The Motion Blur Problem=== | ===The Motion Blur Problem=== | ||
Most conventional [[LCD]] and [[OLED]] screens are '''"sample-and-hold"''' displays. This means that when a pixel is set to a specific color for a frame, it holds that color and remains continuously lit for the entire duration of the refresh cycle ( | Most conventional [[LCD]] and [[OLED]] screens are '''"sample-and-hold"''' displays. This means that when a pixel is set to a specific color for a frame, it holds that color and remains continuously lit for the entire duration of the refresh cycle (for example for the full 11.1 ms at 90 Hz).<ref name="googlevr" /> | ||
Traditional displays use sample-and-hold presentation where each frame remains visible for the entire refresh period. During head movement in VR, this creates motion blur because the brain receives the same static image even as the user's head position changes. The longer a frame persists, the less accurate it becomes relative to the current head position, causing a visible "smearing" effect.<ref name="uploadvr" /> | Traditional displays use sample-and-hold presentation where each frame remains visible for the entire refresh period. During head movement in VR, this creates motion blur because the brain receives the same static image even as the user's head position changes. The longer a frame persists, the less accurate it becomes relative to the current head position, causing a visible "smearing" effect.<ref name="uploadvr" /> | ||
| Line 237: | Line 237: | ||
===Asynchronous Spacewarp (ASW)=== | ===Asynchronous Spacewarp (ASW)=== | ||
[[Asynchronous spacewarp]] (ASW) is a more advanced Meta technology that extends reprojection to handle positional movement and object motion. It builds on ATW and uses a fast extrapolation algorithm that analyzes differences between previous frames to predict what a synthetic "in-between" frame should look like using motion vectors. It typically activates when an application's frame rate consistently drops to half the display's refresh rate ( | [[Asynchronous spacewarp]] (ASW) is a more advanced Meta technology that extends reprojection to handle positional movement and object motion. It builds on ATW and uses a fast extrapolation algorithm that analyzes differences between previous frames to predict what a synthetic "in-between" frame should look like using motion vectors. It typically activates when an application's frame rate consistently drops to half the display's refresh rate (for example 45 FPS on a 90 Hz display).<ref name="uploadvrrepro" /> | ||
When enabled, ASW automatically forces the application to run at half framerate (45 FPS on 90 Hz displays, 60 FPS on 120 Hz displays) and synthetically generates every alternate frame. ASW 2.0 incorporates depth buffer information to greatly reduce visual artifacts, requiring developer support to submit depth data.<ref name="metaasw" /> | When enabled, ASW automatically forces the application to run at half framerate (45 FPS on 90 Hz displays, 60 FPS on 120 Hz displays) and synthetically generates every alternate frame. ASW 2.0 incorporates depth buffer information to greatly reduce visual artifacts, requiring developer support to submit depth data.<ref name="metaasw" /> | ||
| Line 381: | Line 381: | ||
However, unlike gaming monitors which commonly support Variable Refresh Rate (VRR) technologies like [[G-Sync]] and [[FreeSync]], current VR headsets do not implement true variable refresh rate.<ref name="overclockers" /> | However, unlike gaming monitors which commonly support Variable Refresh Rate (VRR) technologies like [[G-Sync]] and [[FreeSync]], current VR headsets do not implement true variable refresh rate.<ref name="overclockers" /> | ||
The fundamental challenge is that lowering refresh rate in VR increases "pose age" | The fundamental challenge is that lowering refresh rate in VR increases "pose age", how old the tracking data is, making head movement feel less smooth and responsive. When refresh rate drops, the reprojection systems that compensate for head movement also run at lower frequency, potentially causing discomfort. Instead of VRR, VR platforms rely on [[asynchronous reprojection]] techniques to handle framerate variations while maintaining consistent display refresh.<ref name="overclockers" /> | ||
The [[Apple Vision Pro]] represents the closest implementation to VRR with its adaptive refresh system that switches between 90 Hz, 96 Hz, 100 Hz, and 120 Hz (M5 model) based on content requirements, though this differs from gaming-style VRR that continuously varies refresh within a range. | The [[Apple Vision Pro]] represents the closest implementation to VRR with its adaptive refresh system that switches between 90 Hz, 96 Hz, 100 Hz, and 120 Hz (M5 model) based on content requirements, though this differs from gaming-style VRR that continuously varies refresh within a range. | ||