The short answer

Start with a wired USB‑C connection or a clear 5GHz Wi‑Fi link, enable the PC’s hardware encoder (NVENC/QuickSync/AMF) in your streaming path, and lower encoder latency by reducing resolution/bitrate and buffers. Also set a high‑performance power plan and stop background tasks to remove micro‑stutters.

Time needed about two hours, spread over two sessions

Key takeaways
  • Measure latency first so you know which change actually helped.
  • Wired USB‑C or a clear 5GHz link reduces the largest, most variable delay.
  • Use the GPU or silicon hardware encoder instead of CPU encoding to cut encoding delay and CPU load.
  • Router placement, channel choice and a single QoS rule remove retransmits and jitter.
  • Windows power plan, background tasks and GPU scheduling commonly introduce micro‑stutters.
  • Test each headset and streaming app pair separately; buffering differs by device and app.
Before you start
  • A Windows PC with the streaming apps installed and administrator access to change drivers and power settings
  • Your smart glasses paired and able to receive streaming (wireless or wired) with the companion app or casting path working
  • A smartphone capable of high‑frame‑rate video (for the frame‑delay camera test)
  • Access to your Wi‑Fi router settings or the ability to place the PC and AP close together for wired testing

The steps

  1. Measure baseline latencyPlace the PC monitor and the glasses display side‑by‑side, run a rapid on‑screen counter or stopwatch on the PC, film both views with a high‑frame‑rate phone camera and record the frame difference across three short runs.You should see: Phone footage shows a consistent frame gap you can convert to milliseconds and an averaged baseline number.
  2. Switch to wired or clear 5GHzIf the headset supports USB‑C wired display, connect it and repeat the camera test. If not, move the AP and headset into line of sight, set a clear 5GHz channel and retest.You should see: The camera test shows a measurable reduction in frame gap when wired or on a clean 5GHz channel.
  3. Enable hardware encodingOpen the streaming app or encoder and switch the encoder to NVENC, QuickSync or AMF; choose a low‑latency or fast preset if available.You should see: The app reports the GPU encoder in use or CPU use drops, and the camera test shows fewer frames of extra delay.
  4. Lower resolution and bitrateReduce the stream resolution or lower the encoder bitrate while keeping FPS stable; run the camera test to observe changes.You should see: Latency improves on the camera test while visual quality remains acceptable for your use.
  5. Adjust Windows power and GPU optionsSet a high‑performance power plan, pause cloud sync and antivirus scans, and enable or test hardware‑accelerated GPU scheduling only if drivers support it; close background apps during tests.You should see: System CPU/GPU usage becomes steadier with fewer spikes and the camera test shows fewer jitter events.
  6. Tune router QoS and placementCreate a single QoS rule prioritising the PC or headset traffic (by IP or device), place the AP to reduce obstructions, choose a less congested channel, and retest with the camera method.You should see: Jitter and occasional spikes reduce and the camera test shows a more consistent latency number.
  7. Test device display or companion modesSwitch the headset between wired display mode and the vendor’s streaming app, check companion or developer buffer settings and measure each mode with the camera test.You should see: One mode shows a clear latency advantage; you can note the preferred mode and app for regular use.
  8. Log results and rollback if neededRecord each change and its camera test result. If a change increased latency, revert that change first and re-run the baseline test.You should see: You have a simple change log and can restore the last known good configuration quickly.

How to reduce lag when streaming PC apps to smart glasses

Start by switching to a wired USB‑C display or a clean 5GHz Wi‑Fi link, enable the PC’s hardware encoder in your streaming app, and stop background CPU/GPU activity. Those changes eliminate the most common sources of delay before you change apps or buy different hardware.

After those three moves, tune encoder buffers, resolution and bitrate to trade visual quality for lower latency.

Measure your baseline latency and reasonable targets

Measure current latency before you change anything. Use a high‑frame‑rate phone camera to film the PC monitor and the glasses display side‑by‑side while a rapid counter runs. Convert the frame gap to milliseconds and record an average across runs so you can compare after each tweak.

Also record any on‑screen latency readouts the headset or app provides, but rely on the camera test for a real‑world comparison.

  • Tools: ping to the headset or PC where supported, and a phone capable of high‑frame‑rate capture.
  • Run three short camera tests per configuration and average the result.

Why wired or 5GHz Wi‑Fi first

A wired USB‑C display mode removes radio interference and retransmits, giving the lowest and most consistent latency. If wired isn’t available, a dedicated 5GHz band on a clear channel comes next—only when the router and headset both have a strong signal and the channel is not crowded.

Avoid 2.4GHz for games or video streaming; it introduces variable delay from interference and contention.

  • Test wired USB‑C display mode if your headset supports it and compare results.
  • For Wi‑Fi, choose 5GHz (802.11ac/ax) and move the AP and headset into line of sight.

Router and Wi‑Fi tuning that actually reduces lag

Fix common wireless problems by moving the AP closer, picking a less crowded channel, and using a conservative channel width when neighbouring APs overlap. Add a single QoS rule that prioritises the PC or headset traffic. These changes reduce retransmits and jitter, which cause micro‑stutter.

Avoid enabling heavy inspection or scanning features on the router; those can add latency. Test any vendor 'gaming' mode and verify with your camera measurement.

  • Measure channel congestion with a Wi‑Fi analyser app before changing channel width.
  • If you must use 80 MHz, confirm neighbouring APs are quiet; a crowded 80 MHz is worse than a clean 40 MHz.
  • Use one simple QoS rule for the streaming device and verify improvement with the camera test.

Pick the right streaming app and enable hardware encoding

Choose a low‑latency streaming app that exposes encoder selection and switch from CPU (software) encode to the GPU or integrated hardware encoder (NVENC for NVIDIA, QuickSync for Intel, AMF for AMD). Hardware encoders perform the encode step faster and with less delay than software encoders.

Select a low‑latency or fast preset in the app, then lower resolution or bitrate if you still need to hit your latency target.

  • Apps to test: Moonlight + Sunshine, Parsec, Steam Link; for Vision Pro, test AirPlay/companion paths.
  • In OBS or similar, choose the GPU encoder in Output settings and confirm the host app uses it.

Windows and GPU settings that remove micro‑stutters

Use a performance power plan, stop background scans and backups, and close high‑CPU apps while testing. Enable hardware‑accelerated GPU scheduling only if your drivers support it and you verify it helps on your hardware; it can reduce scheduling delays but sometimes causes regressions.

Update GPU drivers from the vendor’s updater and disable frame‑capture overlays during latency tests to avoid extra encoding overhead.

  • Set a power plan that prevents aggressive CPU parking under load.
  • Temporarily pause cloud sync and antivirus scans while measuring latency.
  • If a driver or GPU scheduling option worsens latency, revert and retest.

Device-specific paths to cut latency (XREAL, Rokid, Vision Pro)

Headsets differ. If your XREAL supports USB‑C wired display, test that first. Rokid devices may offer a wired mode or vendor streaming app that reduces buffering. For Apple Vision Pro, test the AirPlay and companion PC workflows and check any developer buffer settings the headset exposes.

Try wired display mode first, then the vendor’s streaming app, then third‑party tools if needed; measure each mode with the camera test and pick the best pair.

Why this often fails in practice and what breaks next

People skip measuring and change several things at once, so they cannot identify the cause when latency shifts. A common pattern: switch to 5GHz without checking interference, then update drivers or enable multiple performance features that introduce new stalls.

If you see spikes rather than a steady improvement, roll back recent changes and isolate the single setting that introduced jitter.

  • Failure sign: lower average latency but frequent spikes — indicates jitter or retransmits.
  • Failure sign: steady higher latency after a driver or OS change — roll back the driver and retest.

Quick troubleshooting checklist and rollback rules

If latency worsens after a change, revert that change first. If you changed multiple items together, restore your baseline image or driver set and apply changes one at a time while re-running the camera test. Keep a short log of each change and its measured result.

Run short 30–60 second trials and repeat each configuration three times to avoid drawing conclusions from a single run.

  • Revert order: last change first; if that fails, return to the baseline configuration.
  • Keep three short camera measurements and average them for decisions.

Where to go next after the checklist

If you still have unacceptable lag after these steps, move to a wired path where possible, consider a better AP with stronger 5GHz performance, or use a lighter client and lower local resolution. Use the GlassesApps apps directory to compare streaming apps that users report as low‑latency for your device.

When changing hardware, compare how each device handles wired display mode and which streaming apps are recommended on its device page before buying.

  • Compare device pages for practical paths: XREAL, Rokid, Vision Pro.
  • Scope hardware changes after you have measured and documented current latency.

Where this usually goes wrong

  • Latency improves on average but shows frequent spikesWireless jitter or packet loss from interference or competing traffic.Move to wired if possible, reduce nearby interference, pick a cleaner 5GHz channel, or tighten QoS to stabilise packet timing.
  • Latency increases after a GPU driver or Windows updateA new driver or OS option introduced scheduling or encoder regressions.Roll back to the previous driver, disable the new scheduling option, and keep a stable driver for use while reporting the regression.
  • Encoding CPU usage is high and stutters persistThe stream is using CPU software encoding or a slow preset.Switch to the GPU hardware encoder (NVENC/QuickSync/AMF) and select a faster preset; if unavailable, reduce resolution/bitrate to relieve CPU load.
  • One streaming app has much worse latency than anotherApp‑level buffering, default presets, or lack of hardware encoder support.Use an app that supports hardware encoding and low‑latency presets, or run the game through a low‑latency capture pipeline configured for hardware encode.

The checklist

  • Measure baseline latency
  • Switch to wired or clear 5GHz
  • Enable hardware encoding
  • Lower resolution and bitrate
  • Adjust Windows power and GPU options
  • Tune router QoS and placement
  • Test device display or companion modes
  • Log results and rollback if needed

Frequently asked questions

Will wired USB‑C always fix latency completely?

Wired USB‑C usually gives the lowest and most consistent latency because it removes radio variability. It does not fix a slow encoder or a misconfigured streaming app; you still need to enable hardware encoding and use low buffer presets.

How much will enabling NVENC/QuickSync/AMF reduce latency?

Switching to a hardware encoder typically cuts encoding delay and lowers CPU load. Measure with the camera test before and after to see the real effect on your hardware and streaming app; results vary by GPU and encoder preset.

Is upgrading my router the first thing I should do?

No. First measure baseline, try wired or a clean 5GHz channel, and tune placement and a single QoS rule. Upgrade the router only if a clear 5GHz path and QoS still leave you with unacceptable latency or if your AP lacks 802.11ac/ax.

Can I expect the same latency for every app and headset?

No. Each headset and streaming app uses different buffering and compression decisions. Test each device/app pair with the camera method and tune per pair rather than assuming parity.

What’s the simplest rollback plan if something breaks?

Keep a short change log and a recorded baseline. Revert the last change first. If you changed drivers, reinstall the previous driver version; if a router firmware update caused the issue, restore the prior firmware if available.

Find apps that work on your glasses

Every app in the directory lists the glasses it runs on, how it works on each, and the official source that proves it.

Browse the app directory