In a VR headset the display has become good enough to disappear into the experience, and input is now the part users complain about. A 0.49-inch Micro OLED at 1920x1080 with 3000 nits, 90Hz and a MIPI interface provides the image quality that makes presence possible, but presence depends just as much on how quickly and convincingly the system responds to the user. That response has two halves: sensing what the user did, and telling the user's hands what happened. Both have to be designed around the display's timing, not bolted on afterwards.
Key takeawaysThe 0.49-inch module provides 1920x1080 at 90Hz with 3000 nits and a MIPI interface in an ultra-compact form factor aimed at AR/VR, medical imaging and industrial applications. · A touch panel converts a physical contact into an electrical signal, and its report rate determines how finely the system can track a moving finger [1]. · Capacitive sensing detects the capacitance change caused by a conductive object approaching an electrode, which is why it works through a thin non-conductive cover [2]. · Refresh rate sets how often the image updates, and an input event can only be shown to the user at the next frame, so display timing is part of the perceived latency [3]. · Haptic feedback is most convincing when it is co-located with the visual event it accompanies, which usually means the hands rather than the headset.
VR hardware has settled into a division of labour: the headset carries the display and the tracking, and the controllers carry most of the interaction. That is not an accident. The hands are where haptics can be felt, and they are a comfortable place to put buttons and triggers, while the headset is a poor surface for reaching and touching.
The 0.49-inch Micro OLED at 1920x1080 with 3000 nits, 90Hz and a MIPI interface supplies the visual half of the experience, at a brightness that lets the optics deliver a convincing image.
This article looks at the sensing and feedback half: what touch on the headset is actually for, how sensing and display timing combine into perceived latency, and where haptics belong for the effect to feel real.
What touch on a headset can and cannot do
A touch panel converts contact into an electrical signal that the host can interpret [1], and on a headset the practical surfaces are the side arms, the brow and, on some designs, a small trackpad. None of them map naturally onto a virtual object, which limits what the input can mean. They are good for menus, mode changes and coarse adjustment, and poor for fine manipulation.
The sensor technology matters less than the surface. Capacitive sensing measures the change in capacitance when a conductive object such as a finger disturbs the field at an electrode [2], which works well through a plastic or glass cover and fails with gloves. Since a VR headset is often used with the hands occupied, the design should decide early whether glove compatibility is required.
There is also a hygienic argument for keeping the user's hands away from the lens area. Every touch near the optics risks transferring material to a surface that is magnified by the eyepiece, and cleaning a lens stack is a worse user experience than reaching for a controller.
Where the latency budget actually goes
Perceived latency is the time between a user action and the visual consequence. It includes sensor sampling, filtering, application processing, rendering and display update. The display contributes the final interval, because the new frame can only appear at the next refresh [3]. At 90Hz that interval is short, but it is not zero, and it adds to everything upstream.
This is why smoothing filters on a touch signal are a trade-off rather than a free improvement. Heavier filtering produces a cleaner signal and a more sluggish response, which the user experiences as input lag. In a headset, where the virtual image is already moving with the head, that lag is more noticeable than on a desk.
The practical discipline is to measure the whole chain rather than optimising one stage. A touch report rate faster than the display refresh buys nothing if the renderer is the bottleneck, and a fast renderer cannot compensate for a heavily filtered input signal.
It also helps to keep the timing budget visible as a table during development, with a measured value beside every stage rather than an estimate. Teams routinely discover that a stage they assumed was negligible, such as an operating system input queue, accounts for a noticeable share of the total. Once the numbers are written down, the cheapest improvement is usually obvious, and it is rarely the component with the most impressive specification. Where the display is the fixed part of the design, the same reasoning applies to how the whole interaction is arranged around the panel: reviewing the Learn more details for the module makes it easier to see which latency contributions belong to the display and which belong to everything upstream of it.
Designing input and feedback around the display
Do not judge input by the sensor specification alone. A touch controller with an excellent report rate can still feel unresponsive if the render path is slow, and a haptic pulse that arrives a frame late feels detached from the image. The measurement that matters is the interval between the user's action and the user's perception, taken on the assembled product.
Sources
Frequently asked questions
Is touch on the headset a replacement for VR controllers?
No. Headset-mounted sensing is suited to menus, mode changes and coarse adjustment, while fine manipulation and precise pointing are better served by tracked controllers. Designing for both is usually more effective than trying to do everything with one input.
How does display refresh rate affect perceived input latency?
A new frame can only appear at the next display update, so refresh rate sets the final part of the latency chain. Higher refresh rates shorten that interval, but they cannot compensate for slow sensing or rendering upstream.
Where should haptic actuators be placed?
Where the user can feel them, which in most systems means the controllers or hands. Feedback delivered to the head is hard to associate with an action performed by the hand, and the mismatch weakens the sense of presence.
Does a bright display change the input design?
It changes how visible contamination and surface defects become, so it argues for keeping hands and materials away from the optical window. Touch surfaces placed on the frame avoid that risk while still providing menu-level input.