A product team can spend a full quarter narrowing the panel shortlist and still leave one line of the spec nearly blank: the display interface. That single line decides where video decode lives, how wide the temple FPC runs, which cable the end customer plugs in, and whether one optical engine can serve both a tethered viewer and a standalone build. Across recent teardowns of shipped XR hardware, the interface map is strikingly consistent — and it is not the one monitor-era defaults assume.
Key Takeaways
- Shipped glasses run a two-layer interface map: multi-lane MIPI DSI inside the frame, USB-C DisplayPort Alt Mode outside it. Teardowns of current VR headsets and AR viewers document no full-size HDMI or DisplayPort ports on the eyewear itself.
- MIPI DSI is the internal panel standard: Snapdragon XR2-class platforms drive 4K display feeds at up to 120 Hz over multi-lane MIPI, and shipping designs multiplex those lanes — Pico Neo 3 routes them through a 3 GHz, 5-differential-pair switch shared between displays and cameras.
- LVDS has disappeared from current eyewear teardowns. It remains a monitor-era and industrial-panel link; in a new glasses BOM it should read as a legacy flag, not a default.
- Tethered viewers behave as external monitors: Xreal Air, Rokid Max, and RayNeo Air-class glasses receive video and audio from phones, consoles, and host stations over USB-C carrying DisplayPort signals.
- The video lanes are only half the interface spec: driver ICs and regulators sit on I²C buses running to 3.4 MHz, light engines store panel calibration in onboard SPI NOR flash, and binocular designs add a dedicated inter-eye sync FPC.
Two Layers, Not One Cable
In smart glasses, the display interface is two decisions stacked on top of each other. The inside decision connects the panel to the system: in standalone headsets such as Meta Quest 2, Pico Neo 3, and Pico 4, display data leaves the application processor over multi-lane MIPI and reaches the panels through multi-pin FPC connectors and board-to-board sockets. The same USB-C 3.0 port that charges these headsets also carries 5 Gbps-class data, Power Delivery, and wired PC VR streaming — but the panels themselves never speak USB.
The outside decision is how the glasses meet a host. Tethered AR viewing glasses — the Xreal Air, Rokid Max, RayNeo Air, and Viture class — present themselves to a phone, console, or workstation as an external monitor, receiving video and audio over USB-C DisplayPort Alternate Mode. The split is clean: wherever decode silicon lives, the link on the far side of it is MIPI; wherever the outside world plugs in, the connector is USB-C.
The Four Interface Families, Compared
Most interface confusion comes from treating these standards as competing alternatives. They are not — they occupy different positions in the link, and the table below reflects where each one actually ships in eyewear-class hardware today.
| Interface | Role in the link | Signal character | Eyewear reality |
|---|---|---|---|
| MIPI DSI | Panel link inside the frame | Scalable differential lanes, low swing, short FPC reach | Ships in every standalone headset and SoC-based glasses teardown |
| LVDS | Monitor-era and industrial panel link | Fixed differential pairs, higher swing, thicker harness | Not documented in current XR eyewear teardowns |
| HDMI | External A/V between consumer devices | Single transition-minimized link, connector bulk | No shipped glasses carry a native port; lives on hosts and docks |
| DP over USB-C Alt Mode | External video into tethered glasses | DisplayPort lanes plus USB data and power on one cable | The external standard shipped by AR viewing glasses |
Two clarifications the table cannot carry. First, DisplayPort and USB-C are not rivals here: Alt Mode is DisplayPort lanes tunneled through a USB-C connector, which is how one reversible cable ends up carrying video, audio, USB data, and charging at once. Second, HDMI is not absent from glasses ecosystems — it is absent from glasses themselves; it lives on the source devices, adapter docks, and driver boards that feed them.
Sizing the Link to Your Panel
Bandwidth follows pixels multiplied by refresh rate, and the shipped product landscape brackets the working range. At the lightweight end, waveguide information glasses run monochrome micro-LED at 640×350 or 480×640 per eye at 60 Hz. Birdbath viewers center on 0.49–0.70-inch micro-OLED at 1080p per eye with 60–120 Hz refresh. Snapdragon AR1-class platforms cap dual binocular output at 1280×1280 per eye at 60 fps. VR headsets sit higher: Pico 4 drives 2160×2160 per eye at 72 or 90 Hz, 1.03-inch micro-OLED panels reach 2560×1440 per eye at 120 Hz, and Vision Pro-class silicon pushes past 4000 PPI per eye. Each step up multiplies the lane requirement on the inside link — or moves the decode burden to the host on the outside link.
The practical rule: fix where decode lives before arguing connectors. Decode on the glasses means the panel interface is MIPI DSI, and you are sizing lanes against the SoC's display capability, the way XR2-class chips drive 4K feeds at up to 120 Hz. Decode on the host means the glasses receive DP Alt Mode over USB-C, and the internal question shrinks to a short bridge from the receiver to the panel's native link.
The Channels Around the Video
A panel interface spec that stops at video lanes is half a spec. The control and configuration channels decide how the display actually behaves in the field.
1. Control buses
Driver ICs, power regulators, and LED drivers are configured over I²C — running at up to 3.4 MHz in shipping glasses designs, with 400 kHz fast-mode devices alongside them. The bus map, device addresses, and voltage rails belong in the spec next to the video lanes.
2. Calibration storage
In shipping micro-LED light engines, an onboard SPI NOR flash holds panel calibration and drive configuration, loaded instantly at power-up across a board-to-board connector. That is what lets a calibrated engine stay calibrated through module-level handling — and it is a spec line, not an afterthought.
3. Binocular synchronization
Dual-panel designs need frame-level sync between eyes. One shipping design routes a 0.1 mm wide FPC across the front frame to carry a wired clock between left and right temples, keeping both micro-LED engines locked without wireless timing. However it is implemented, the sync architecture changes the FPC stack and should be specified with it.
A Decision Framework You Can Write Into the Spec
Choose MIPI DSI when the design carries its own SoC — and write down the lane count, the D-PHY or C-PHY version, and the connector family. Choose USB-C DisplayPort Alt Mode when the product is a tethered viewer — and write down the DP version, pin assignment, and which host devices you will certify against. Accept LVDS only when reusing an existing industrial panel whose ecosystem you cannot change; in a new eyewear design it is a legacy flag. Put full-size HDMI and DisplayPort connectors on host devices and docks, never on the glasses.
This is also where module-level customization earns its keep. In custom viewer and glasses programs, the panel, the bridge or driver board, and the temple FPC are specified together — a driver board that accepts USB-C DisplayPort or HDMI input and drives the panel over its native link is a standard variant, not a special project. Getting the interface definition right early is what keeps one optical engine reusable across tethered and standalone builds.
Not sure which interface combination fits your panel and host? Share your target resolution and host device — we will map out what is technically feasible, no strings attached.
FAQ
Q1: What display interface do smart glasses use?
Shipped glasses use a two-layer map: inside the frame, panels are driven over multi-lane MIPI DSI from the SoC; outside, tethered glasses receive video over USB-C DisplayPort Alt Mode. Standalone headsets add a USB-C 3.0 port for data, charging, and PC VR streaming, but the panels themselves speak MIPI.
Q2: Is LVDS still used in AR and VR glasses?
Current XR eyewear teardowns document MIPI panel links and no LVDS interconnects. LVDS remains common in legacy monitors and industrial panels, where thicker differential pairs and lower lane efficiency are acceptable. For a new glasses design, treat an LVDS quote as a legacy flag rather than a default.
Q3: Can VISGLASS help specify the interface and driver board for a custom glasses project?
Yes. Interface definition is part of every custom program: MIPI lane count, external input through USB-C DisplayPort or HDMI driver boards, control bus mapping, and calibration storage are specified together with the optics. Share your target resolution and host device, and we will map out the feasible options.
